Game: #2077

Open
opened 2026-11-21 14:50:49 +00:00 by MayShowers · 0 comments

Longshot version

Nightly 680 9dd41326

Channel

Nightly

What kind of problem

Graphics (missing, black or wrong-coloured things, lighting, flicker)

What happened

Newest Nightly release broke Hyrule field the field appears segmented with seams between chunks? seams disappear when you get close to them but return when far away.

Where in the game

Hyrule Field

How often

Every time

Graphics card and driver version

RX 6750XT

CPU, RAM and operating system

Ryzen 7 5700X 32GB ddr4, win11

Changes from the defaults

I have the no fog mod installed but tested the bug with with it off.

longshot_log.txt

[hos host] longshot nightly-680-9dd41326 (source e287b3d8)
[hos host] run_id: bf8f392313b2578d0aeadbc8c186b6e4
[a64] unimplemented import ZN2nn2ro12ProtectRelroEPKvS2_S2_S2_S2 (lr 0x81237908)
[a64] unimplemented import _ZN2nn2vi36SetDisplay120HzRefreshRateCapabilityEPNS0_7DisplayEb (lr 0x8348bf64)
[hos host] crash window: watching this run (HOS_CRASH_WATCH=0: off)
[hos host] options: the settings file is D:\Fan games and Decomps\Longshot Ocarina of Time Remake\LongshotLauncher-data\user\longshot_options.ini
[hos host] oe: output resolution 2560x1440
[hos host] boot: widescreen: a 1.7778 screen is 16:9: the game's own picture
[hos host] boot: image 96 MB at 0x80000000, 9809 MB of memory for the game
[hos host] boot: 32694 MB of memory, 17408 MB of swap
[hos host] mods: the list (the order and the switches) is D:\Fan games and Decomps\Longshot Ocarina of Time Remake\LongshotLauncher-data\user\mods.json
[hos host] mods: Remove Fog: off (the Mods tab)
[hos host] mods: 1 mods from D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\mods: 0 RomFS entries
[hos host] vk: reserved core selection: 8 cores, 16 logical, best class=0, fast cores=8, selected=0xc000 (of the logical processors 0xffff)
[hos host] vk: core 0: mask=0x3, class=0, popcount=2
[hos host] vk: core 1: mask=0xc, class=0, popcount=2
[hos host] vk: core 2: mask=0x30, class=0, popcount=2
[hos host] vk: core 3: mask=0xc0, class=0, popcount=2
[hos host] vk: core 4: mask=0x300, class=0, popcount=2
[hos host] vk: core 5: mask=0xc00, class=0, popcount=2
[hos host] vk: core 6: mask=0x3000, class=0, popcount=2
[hos host] vk: core 7: mask=0xc000, class=0, popcount=2
[hos host] vk: reserved core mask (enhanced): 0xc000 (8 cores; a core the process may run on wholly, of 0xffff)
[hos MainThread] mem: the working set's minimum is 4096 MB of the PC's 32694 MB: Windows keeps that much of the game in RAM before it trims below it (HOS_WS_FLOOR_MB=0: Windows' own)
[hos MainThread] boot: _init_libc0 (0x85717e50)
[hos MainThread] boot: nnosInitialize (0x8510f720)
[hos MainThread] libc: memcpy, memmove, memset, memcmp, strlen and strcmp are the host's (HOS_GUEST_LIBC=1: the game's own)
[hos MainThread] boot: _init_libc1 (0x85717e60)
[hos MainThread] boot: nninitStartup (0x8262bb10)
[hos MainThread] heap: 6894 MB at 0x1000000000
[hos MainThread] mem: 0x1000000040 committed on first touch (nothing handed it out)
[hos MainThread] mem: 0x1000010000 committed on first touch (nothing handed it out)
[hos MainThread] boot: _init_libc2 (0x85717e80)
[hos MainThread] boot: main DT_INIT (0x812377e0)
[hos MainThread] boot: subsdk0 DT_INIT (0x84ec0040)
[hos MainThread] boot: sdk DT_INIT (0x84fa0060)
[hos MainThread] boot: nnMain (0x81222cd0)
[hos MainThread] mem: 0x1002400010 committed on first touch (nothing handed it out)
[hos sead::MainThread] mem: 0x10024109c0 committed on first touch (nothing handed it out)
[hos sead::MainThread] fs: mounted content: -> D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\romfs
[hos sead::MainThread] mem: 0x11aecffff2 committed on first touch (nothing handed it out)
[hos sead::MainThread] mem: 0x11aed00010 committed on first touch (nothing handed it out)
[hos sead::MainThread] mem: 0x11aed27998 committed on first touch (nothing handed it out)
[hos sead::MainThread] zstd: the game's whole-buffer decode on the host's libzstd 1.5.7 (HOS_NATIVE_ZSTD_FRAMES)
[hos sead::MainThread] mem: 0x1003f01a88 committed on first touch (nothing handed it out)
[hos sead::MainThread] mem: 0x11aedfff52 committed on first touch (nothing handed it out)
[hos sead::MainThread] oe: language en-US (language.txt)
[hos sead::MainThread] oe: GetCurrentGraphicsScale -> 1.00
[hos sead::MainThread] gs: BootGraphicsSystem(0, 0): no graphics server yet (headless)
[hos host] display: fullscreen (F11 for a window)
[hos sead::MainThread] nvn: native NVN layer: 429 procs
[hos sead::MainThread] nvn: native NVN device (flags 0): no Switch GPU emulation
[a64] unimplemented import _ZN2nn3aoc17CountAddOnContentEv (lr 0x83392a38)
[hos sead::MainThread] nvn: window origin upper left
[hos sead::MainThread] nvn: depth mode near = -w (GL)
[hos sead::MainThread] mem: 0x1004032000 committed on first touch (nothing handed it out)
[hos sead::MainThread] nvn: texture pool of 33068 descriptors
[hos sead::MainThread] vi: GetNativeWindow -> 0x1102400
[hos sead::MainThread] mem: 0x1004133000 committed on first touch (nothing handed it out)
[hos sead::MainThread] mem: 0x1006e37000 committed on first touch (nothing handed it out)
[hos sead::MainThread] nvn: window of 3 textures (3 active), 2560x1440 format 0x25
[hos sead::MainThread] mem: 0x1006f3b000 committed on first touch (nothing handed it out)
[hos sead::MainThread] mem: 0x10072b8020 committed on first touch (nothing handed it out)
[hos sead::MainThread] mem: 0x10073b9000 committed on first touch (nothing handed it out)
[hos sead::MainThread] mem: 0x10073ca000 committed on first touch (nothing handed it out)
[hos sead::MainThread] sleep: at most 4 of the game's threads yield through a short sleep at a time, the others wait on the timer (16 logical processors; HOS_SHORT_SLEEP_POLLERS=0: all yield)
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 1 made): 65537 lookups, 37293 missed (56.90 %)
[hos host] vk: Intel XeSS: no libxess.dll in D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\bin, D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326 (the readme says where Intel's file comes from)
[hos ParallelDecompThread] zstd: 1 stream(s) decoding on the host at once
[hos sead::MainThread] applet: render target pool 512 MB (the game's 305 MB)
[hos Resource Memory] mem: the game's mapped pages 360 MB (peak 360); heap 6894 MB; the console's total 9809 MB
[hos Resource Memory] mem: the game's mapped pages 512 MB (peak 512); heap 6894 MB; the console's total 9809 MB
[hos ParallelDecompThread] zstd: 2 stream(s) decoding on the host at once
[hos ARes Loading 1] texts: mals/usen.product.110.sarc.zs <- D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\texts/usen.product.110.sarc.zs (the console's wording made the PC's)
[hos host] hid: 55 controls from longshot_options.ini
[hos host] vk: the Output resolution's sizes: 31 (the screen's modes, with the common sizes of their shapes, and the game's own); shapes 5:4, 4:3, 16:10, 16:9
[hos host] vk: device 0: AMD Radeon RX 6750 XT, Vulkan 1.4, driver 0x800161, type 2, id 1002:73df
[hos host] vk: float atomic add on storage buffers: compare-exchange loop, by default only with NVIDIA's driver and Intel's Windows driver on Meteor Lake, Arrow Lake, Panther Lake and Battlemage (RX_VK_NATIVE_FADD=1: native where the GPU has it)
[hos host] vk: subgroups: default 64 lanes, 32-64, pinned to 32 on stages 0x20; RADV on GCN 1-4: no
[hos host] vk: shader caps: subgroup stages 0x3fff; denorm independence 0, fp32 denorm preserve 1 flush 1, signed zero fp16 1 fp32 1; int64 buffer atomics 1
[hos host] vk: mesh-only and task + mesh programs use the compute fallback (RX_VK_TASK_MESH_FALLBACK=0: native)
[hos host] vk: mesh shaders: Automatic, from the default: mesh-only and task + mesh draws on the compute path
[hos host] vk: robust buffer access on
[hos host] vk: vendor fixes for AMD proprietary driver (26.11.2 (AMD proprietary shader compiler)): out-of-bounds texel fetches 0 1, NaN tests by bits 1, gather offset 1
[hos host] vk: programs are loaded when the game makes them (create): the default, on the compute fallback too (RX_VK_LOAD=start: every listed program before the title) (RX_VK_LOAD=create or start says which)
[hos host] vk: the presenter thread waits for its presents to be shown: no
[hos host] vk: colour outputs NaN-guarded
[hos host] vk: positions invariant (not an NVIDIA card)
[hos host] vk: matrix multiplies in software (the cooperative matrices lack 16x8x16 f16)
[hos host] vk: software matrix multiplies looped (not RADV or RDNA 4 on AMD's proprietary driver, RX_VK_MMA_UNROLL=1 tries the unrolled form)
[hos host] vk: AMD Anti-Lag 2: available (VK_AMD_anti_lag)
[hos host] vk: AMD Radeon RX 6750 XT, Vulkan 1.4, driver 0x800161; mesh shaders used, host memory import yes, pipeline libraries yes
[hos host] vk: driver: AMD proprietary driver 26.11.2 (AMD proprietary shader compiler)
[hos host] vk: policy: facts 1002:73df, driver id 1, version 0x800161 (26.11.2 (AMD proprietary shader compiler)), type 2, Windows; VRAM 12272 MB, allowed at the start 11474 MB, RAM 32694 MB, resizable BAR 1; CPUs 16 (16 usable, 8 cores); mesh shaders 1, float add 0, cooperative matrices 1, 16-bit stage inputs 1, lanes 64 to 64 (32 on stages 0x20)
[hos host] vk: policy: driver fixes on driver version 0x800161: fetch_bounds=1 nan_bits=1 gather_offset=1 frag_shuffle_fallback=1 position_invariant=1 equal_depth_bias=0 mesh_no_optimizer=1 amd_retry_no_optimizer=1 amd_cloth_no_optimizer=1 fault_gate_serial=1 task_mesh_fallback=1(seen on 0x800161) mesh_mode=automatic fallback_phi_limit=0 intel_compile_gate=0 compute_builds_serial=0 pcache_fetch_lock=0 split_before_draw=0 astc_on_cpu=0 inline_deku_web_loads=0 networks_left_out=1 dma_buf_imports=0 nv_checkpoints=0 retired=after a wait idle_import_high=0% allowance_at_count_rule=1
[hos host] vk: policy: can and measured: mirrors=1 import_chunk=4 MB ub_snapshot=0 split_ub_flush=1 native_float_add=0 warp64=1 batch_submits=0 mma_unroll_try=0 global_table_loop=0 occ_copy=0 intel_side_by_side_auto=0 fsr4_tune=1 gpu_clock_peak=0 readback_late_newest=0 acquire_ahead=1 pcache_small_pc=0
[hos host] vk: policy: sizes by the allowance at the start (the sites read it later): textures 7158 MB, mirrors 2048, target pool 768, compaction from 512, shadows 1434, staging kept 256, spare captures no limit; blocks at the start 24, of 64 MB (targets 32); count line 97% room_in_blocks_rule=1 squeeze_damped=0 linux_hold_caps_budget=0 small_vram_textures=0 astc_bc_all=0 dlss_cnn=0
[hos host] vk: policy: auto settings and threads: auto_tier=4 auto_render_third=0 rtx=0 rdna4=0; workers 6, capture helpers 4, texture threads 8 (gather 6), program workers in play 4 threaded_auto_render=1 threaded_auto_recorder=1 drop_modules_small_ram=0 native_zstd_keep_core=0
[hos host] vk: policy: switches (* above: set by its switch; the other values are this PC's defaults): none of these is set
[hos host] report: gpu AMD Radeon RX 6750 XT, AMD proprietary driver 26.11.2 (AMD proprietary shader compiler)
[hos host] report: cpu AMD Ryzen 7 5700X 8-Core Processor (16 threads), 32 GB
[hos host] report: os Windows 11 (build 26300)
[hos host] report: renderer Vulkan
[hos host] vk: switches set: none
[hos host] vk: Auto (this PC) would be: 12272 MB of video memory: tier 4; the game's own upscaler (AMD Radeon RX 6750 XT, 12272 MB of video memory, dedicated, screen 3840x2160 at 165 Hz, 32694 MB of RAM, 16 threads)
[hos host] vk: surface caps images 2..16 extent 3840x2160..3840x2160 usage 0x9f alpha 0x1 transform 0x1
[hos host] discord: connected (application 1570559772111933578)
[hos host] discord: the boss heart containers: 8 of 8 found in the RomFS
[hos host] discord: the status: N680 9dd41326 / -
[hos host] discord: status set
[hos host] vk: swapchain 3840x2160, 4 images, fifo (V-Sync On)
[hos host] vk: guest memory imported in place (alignment 4096)
[hos host] vk: replay ready (window; push descriptors yes (32), extended dynamic state 3 yes, custom border colours yes)
[hos host] vk: driver shader cache at the start: 2968 MB (%USERPROFILE%\AppData\Local\AMD\VkCache 2968 MB)
[hos host] vk: VRAM mirror of buffer pools on
[hos host] vk: Threaded rendering: Auto: on, the render thread and the recorder thread (8 cores of 16 logical processors)
[hos host] vk: Compile shaders: Most (the default): the predicted pipelines are left out; a pipeline no list names is made when it is first drawn
[hos host] vk: handle tables as descriptor arrays: on
[hos host] vk: pipeline cache file read: not NVIDIA's own driver
[hos host] vk: acquire ahead: on by default: the next image asked for right after each present (RX_VK_ACQUIRE_AHEAD=0: off)
[hos host] vk: RX_VK_RECORDER=thread: vkCmd* packed on the replay thread and run in order by a recorder thread, handed over by part (RX_VK_RECORDER_PARTS): a part's packets at its end, and its end and submit made by the recorder; the replay waits for it at a reset, begin, destruction or direct recording, and for a part's submit where a wait, a fence or the queue needs it; ring 8192 KB, the recorder waits on the ring after 10 us
[hos host] vk: AMD FSR 4: D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\amd_fidelityfx_loader_vk.dll
[hos host] vk: ui: the game's own menu font, from its files
[hos host] vk: ui: Dear ImGui 1.92.9b on the window's images (format 44)
[hos host] vk: fsr3: AMD FSR 3.1 can run on this GPU
[hos host] vk: AMD FSR 4: upscalers 4.1.2, 3.1.6, 2.3.4; using 4.1.2
[hos host] menu: the upscalers asked in the background while the start-up screen shows (21 ms)
[hos host] vk: images from a pool of 32 MB blocks (memory type 0), 0 made at the start
[hos host] vk: presenter thread: each present made on a thread of its own once its copy is done
[hos host] vk: AMD Anti-Lag 2: on
[hos host] vk: the window's screen is 3840x2160 (16:9): 0 sizes added to the Output resolution's
[hos host] vk: pipeline cache D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\cache/longshot_vk_pipelines.bin (1973875 KB)
[hos host] vk: pipeline cache: 6 play files merged (968 KB)
[hos host] vk: ASTC is decoded on the GPU (Eden's decoder, Switch 2 tiling)
[hos host] vk: big ASTC textures are kept as BC3 (re-encoded on the GPU; ASTC as RGBA8 or RX_VK_ASTC_RGBA8=1: as RGBA8)
[hos host] vk: big block-linear textures are unswizzled on the GPU from their pools
[hos host] vk: texture pipelines made before the game in 1 ms (ASTC decoder yes, BC encoder yes, GPU unswizzle yes)
[hos host] vk: program cache D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\cache/longshot_programs.bin: 47798 program keys kept, 147312 KB
[hos host] vk: shader store %USERPROFILE%\AppData\Local\Longshot\shaders/: 50685 SPIR-V blobs, for every game folder
[hos host] vk: one build of a program goes to the driver at a time (AMD; RX_VK_DRIVER_FAULTS=0: as many as ask)
[hos host] vk: vertex format NVN 0x29 -> vk 97
[hos host] vk: vertex format NVN 0x22 -> vk 106
[hos host] vk: vertex format NVN 0x69 -> vk 65
[hos host] vk: vertex format NVN 0x25 -> vk 37
[hos host] vk: vertex format NVN 0x27 -> vk 41
[hos host] vk: vertex format NVN 0x12 -> vk 77
[hos host] vk: vertex format NVN 0x13 -> vk 78
[hos host] vk: vertex format NVN 0x11 -> vk 83
[hos host] vk: vertex format NVN 0x16 -> vk 103
[hos host] vk: vertex format NVN 0x2e -> vk 109
[hos host] vk: vertex format NVN 0x3 -> vk 13
[hos host] vk: the driver's shader cache has 51 of 51 noted pipelines asked (a sample, in 99 ms)
[hos host] vk: start-up: pipelines 0 of 4528, 0% of all
[hos host] vk: programs: load at create: 37641 notes of 23925 programs indexed, 4 type signatures, 3357 compute programs listed, in 14.5 ms (45798 kept program keys); 2 workers in play, 2 while the game loads; pipelines from: pipeline caches
[hos host] vk: precompile: load at create (RX_VK_LOAD=create): 37641 noted pipelines found, none built; 3357 compute programs named (3357 in the shipped list, 325 dispatched in play here); 2264 mesh fallback programs kept ready, 1589 of their noted pipelines with them (167 in no cache); in 680 ms. No other program is kept ready by a list: each is loaded when the game makes it (the notes' file walked 1 times this start, in 11072 ms: start_note_refs)
[hos host] vk: presenter thread: the copy's fence 7.33 ms, the present 0.44 ms a present
[hos host] vk: start-up: the shaders took 20.9 s
[hos host] vk: pipeline cache: what play makes goes into play files (RX_VK_PLAY_CACHE=0: into the one cache)
[hos host] rx: renderer: Vulkan (longshot_vk.dll)
[hos host] vk: the replay runs on a thread of its own, the render thread (RX_VK_ASYNC=0: on the submitting thread)
[hos host] vk: the replay thread runs on logical processors 0xc000, kept from the game's threads (its ideal processor there; free to go to the other fastest cores when another program takes it)
[hos host] vk: the replay thread may run on logical processors 0xffff
[hos host] rx: 123 texture and sampler calls made before the renderer started (33 texel writes), given to it
[hos sead::MainThread] nvn: QueueSubmitCommands #1: queue 00001000010FBF90 (its #1), 1 handles (lr 0x812c3cf4)
[hos sead::MainThread] vk: submit-time capture of written pages on
[hos sead::MainThread] mem: the write marks' walk: 256 marks a test, else 64, in 32-byte loads (AVX2) (HOS_COLLECT_VECTOR=0: word loads)
[hos sead::MainThread] vk: RX_VK_READBACK_EARLY left off: the render thread is on (RX_VK_ASYNC, Threaded rendering), which it was not measured with; reads of GPU results wait for their whole submission
[hos host] [debug] vk: 78 pools mapped
[hos host] vk: texture storage in the submit-time capture on
[hos host] vk: a 32 MB capture buffer made ahead in 7.8 ms, off the game's thread
[hos host] vk: a 64 MB capture buffer made ahead in 13.9 ms, off the game's thread
[hos host] vk: a 128 MB capture buffer made ahead in 24.9 ms, off the game's thread
[hos host] vk: sparse VRAM mirrors (RX_VK_MIRROR_SPARSE): on
[hos host] vk: occlusion queries on (RX_VK_OCC_QUERIES): the game's samples-passed reports get the GPU's counts; read from the pool
[hos host] vk: render target pool: up to 768 MB (RX_VK_TARGET_POOL_MB)
[hos host] vk: render targets' images from blocks of their own (32 MB, memory type 0)
[hos host] vk: texture 0x103d650000 32x32x32 x1 levels 1 NVN 0x26 -> vk 38 (1 textures, 0 MB)
[hos host] [debug] vk: a staging arena of 16 MB took 6.7 ms to make
[hos host] vk: staging arenas: 1 (16 MB in this one)
[hos host] vk: 1 pools imported
[hos host] vk: sparse mirrors' buffers are kept for the next mirror of their size, up to 4096 MB of address space (RX_VK_SPARSE_SPARE_MB=0: destroyed at once, as before)
[hos host] vk: pool 0x103d800000 (4 MB): mirror made at its first binding (1 so)
[hos host] vk: batched submits off (RX_VK_BATCH)
[a64] unimplemented import _ZN2nn5audio34GetAudioRendererVoiceDropParameterENS0_19AudioRendererHandleE (lr 0x81044370)
[a64] unimplemented import _ZN2nn5audio39IsDeviceOutputVolumeAutoTuneModeEnabledEv (lr 0x81043560)
[a64] unimplemented import ZN2nn5audio36AcquireAudioOutputDeviceNotificationEPNS_2os11SystemEventE (lr 0x80f72978)
[a64] unimplemented import ZN2nn5audio40SetDeviceOutputVolumeAutoTuneModeEnabledEb (lr 0x80f729e0)
[a64] unimplemented import ZN2nn5audio26GetAudioDeviceOutputVolumeEPKNS0_15AudioDeviceNameE (lr 0x80f72e00)
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 1 made): 222210 lookups, 73443 missed (33.05 %)
[hos GeneralFormatAsync
] vcores: 3363 takes, 472 waits (mean 1273 us, longest 14.8 ms, 0 over HOS_VCORE_LOG_MS), 184 kicks, 168 yields, 0 starved grants, 0 escapes in 22 s
[hos ARes Loading 0] prompts: switch icons, 16 buttons on these pad's controls
[hos ParallelDecompThread] zstd: the game's stream decode on the host's libzstd 1.5.7 (HOS_NATIVE_ZSTD=1, this GPU's default: the core given up while it decodes)
[hos sead::MainThread] audio: renderer opened
[hos host] audio: stereo for the game (the selected/default output has 2 channels)
[hos sead::MainThread] audio: device sink on mix 0, 6 channels (0 1 2 3 4 5)
[hos host] audio: output on the default device (3 devices), stereo
[a64] unimplemented import ZN2nn5prepo30GetOverSaveFrequencyLimitEventEv (lr 0x80e5ec00)
[hos sead::MainThread] fs: mounted save: -> D:\Fan games and Decomps\Longshot Ocarina of Time Remake\LongshotLauncher-data\user\save\5a3c9e1f0b7d24680000000000000001
[hos BootupBackgroundWorkThread] fs: mounted savt: -> D:\Fan games and Decomps\Longshot Ocarina of Time Remake\LongshotLauncher-data\user\temp
[hos GeneralFormatAsync
] vcores: GeneralFormatAsync
(priority 24) waited 131.4 ms for a core: given by sead::MainThread (priority 16), which held it 131.4 ms and was never asked; the cores it may use now: core 0: it;
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 68.7 ms for a core: given by GeneralFormatAsync
(priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 0: it;
[hos PlayReporter] vcores: PlayReporter (priority 29) waited 117.3 ms for a core: given by PreActorModelCompose (priority 26), which held it 0.0 ms and was never asked; the cores it may use now: core 0: it;
[a64] unimplemented import _ZN2nn5audio34SetAudioRendererVoiceDropParameterENS0_19AudioRendererHandleEf (lr 0x836f755c)
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 498.5 ms for a core: given by sead::MainThread (priority 16), which held it 580.6 ms and was never asked; the cores it may use now: core 0: it;
[hos host] vk: module identifiers: 757 known (longshot_module_ids.bin): a learned compute variant's pipeline is made at its program's create call without its code
[a64] unimplemented import _ZN2nn4nifm10InitializeEv (lr 0x8119aea8)
[a64] unimplemented import _ZN2nn2vi30QueryDisplayCurrentRefreshRateEPNS0_11RefreshRateEPbPKNS0_7DisplayE (lr 0x83b6eab8)
[a64] unimplemented import _ZN2nn3hid25SetGyroscopeZeroDriftModeERKNS0_19SixAxisSensorHandleERKNS0_22GyroscopeZeroDriftModeE (lr 0x834898d0)
[a64] unimplemented import _ZN2nn6socket10InitializeERKNS0_6ConfigE (lr 0x83c55b24)
[a64] unimplemented import _ZN2nn3ssl10InitializeEv (lr 0x83c55b30)
[a64] unimplemented import _ZN2nn6socket13IsInitializedEv (lr 0x824d92f8)
[a64] unimplemented import _ZN2nn5audio15AddBiquadFilterEPNS0_19AudioRendererConfigEPNS0_16BiquadFilterTypeEPNS0_10SubMixTypeE (lr 0x80dc7494)
[a64] unimplemented import _ZN2nn5audio26SetBiquadFilterInputOutputEPNS0_16BiquadFilterTypeEPKaS4_i (lr 0x80dc740c)
[a64] unimplemented import _ZN2nn5audio22SetBiquadFilterEnabledEPNS0_16BiquadFilterTypeEb (lr 0x80de9464)
[a64] unimplemented import _ZN2nn6socket7EventFdEmNS0_12EventFdFlagsE (lr 0x82213be4)
[a64] unimplemented import _ZN2nn6socket5CloseEi (lr 0x82213bf4)
[a64] unimplemented import _ZN2nn5audio24SetBiquadFilterParameterEPNS0_16BiquadFilterTypeERKNS0_22TBiquadFilterParameterIfEE (lr 0x836fcdb4)
[a64] unimplemented import _ZN2nn4nifm22GetCurrentProxySettingEPNS0_12ProxySettingE (lr 0x82378f6c)
[a64] unimplemented import _ZN2nn5audio8AddDelayEPNS0_19AudioRendererConfigEPNS0_9DelayTypeEPvmPNS0_10SubMixTypeENS_8TimeSpanEi (lr 0x80dd53b8)
[a64] unimplemented import _ZN2nn5audio19SetDelayInputOutputEPNS0_9DelayTypeEPKaS4_i (lr 0x80dd531c)
[a64] unimplemented import _ZN2nn5audio15SetDelayEnabledEPNS0_9DelayTypeEb (lr 0x80de9464)
[a64] unimplemented import _ZN2nn4nifm18IsNetworkAvailableEv (lr 0x82145cec)
[a64] unimplemented import _ZN2nn5audio14IsDelayEnabledEPKNS0_9DelayTypeE (lr 0x8119f138)
[a64] unimplemented import _ZN2nn5audio12GetDelayTimeEPKNS0_9DelayTypeE (lr 0x8119f164)
[a64] unimplemented import _ZN2nn5audio12SetDelayTimeEPNS0_9DelayTypeENS_8TimeSpanE (lr 0x8119f178)
[a64] unimplemented import _ZN2nn5audio15SetDelayDryGainEPNS0_9DelayTypeEf (lr 0x8119f184)
[a64] unimplemented import _ZN2nn5audio20SetDelayFeedbackGainEPNS0_9DelayTypeEf (lr 0x8119f190)
[a64] unimplemented import _ZN2nn5audio14SetDelayInGainEPNS0_9DelayTypeEf (lr 0x8119f19c)
[a64] unimplemented import _ZN2nn5audio21SetDelayChannelSpreadEPNS0_9DelayTypeEf (lr 0x8119f1a8)
[a64] unimplemented import _ZN2nn5audio21SetDelayLowPassAmountEPNS0_9DelayTypeEf (lr 0x8119f1b4)
[a64] unimplemented import _ZN2nn4nifm27SubmitNetworkRequestAndWaitEv (lr 0x8120873c)
[a64] unimplemented import _ZN2nn5audio9AddReverbEPNS0_19AudioRendererConfigEPNS0_10ReverbTypeEPvmPNS0_10SubMixTypeEi (lr 0x80de96ec)
[a64] unimplemented import _ZN2nn5audio20SetReverbInputOutputEPNS0_10ReverbTypeEPKaS4_i (lr 0x80de965c)
[a64] unimplemented import _ZN2nn5audio16SetReverbEnabledEPNS0_10ReverbTypeEb (lr 0x80de9464)
[a64] unimplemented import _ZN2nn5audio15IsReverbEnabledEPKNS0_10ReverbTypeE (lr 0x80de91a0)
[a64] unimplemented import _ZN2nn5audio18SetReverbEarlyModeEPNS0_10ReverbTypeENS1_9EarlyModeE (lr 0x80de91c4)
[a64] unimplemented import _ZN2nn5audio18SetReverbEarlyGainEPNS0_10ReverbTypeEf (lr 0x80de91d0)
[a64] unimplemented import _ZN2nn5audio21SetReverbPredelayTimeEPNS0_10ReverbTypeENS_8TimeSpanE (lr 0x80de91e4)
[a64] unimplemented import _ZN2nn5audio17SetReverbLateModeEPNS0_10ReverbTypeENS1_8LateModeE (lr 0x80de91f0)
[a64] unimplemented import ZN2nn5audio17SetReverbLateGainEPNS0_10ReverbTypeEf (lr 0x80de91fc)
[a64] unimplemented import ZN2nn5audio18SetReverbDecayTimeEPNS0_10ReverbTypeENS_8TimeSpanE (lr 0x80de9208)
[a64] unimplemented import ZN2nn5audio32SetReverbHighFrequencyDecayRatioEPNS0_10ReverbTypeEf (lr 0x80de9214)
[a64] unimplemented import ZN2nn5audio19SetReverbColorationEPNS0_10ReverbTypeEf (lr 0x80de9220)
[a64] unimplemented import ZN2nn5audio19SetReverbReverbGainEPNS0_10ReverbTypeEf (lr 0x80de922c)
[a64] unimplemented import ZN2nn5audio16SetReverbOutGainEPNS0_10ReverbTypeEf (lr 0x80de9238)
[a64] unimplemented import ZN2nn5audio16SetReverbDryGainEPNS0_10ReverbTypeEf (lr 0x80de9244)
[a64] unimplemented import ZN2nn5audio20GetReverbInputOutputEPKNS0_10ReverbTypeEPaS4_Pii (lr 0x80de9260)
[a64] unimplemented import ZN2nn5prepo27AddInternetConnectionStatusEPNS0_10PlayReportEPKcS4_S4 (lr 0x80e5bd60)
[a64] unimplemented import ZN2nn5prepo28RequestImmediateTransmissionEv (lr 0x80df26c0)
[hos] guest backtrace: lr 0x804fc1ec 0x83b6f194 0x83b6f080
[hos] guest backtrace: lr 0x8348c5e0 0x83b6f08c
[hos sead::MainThread] audio in: stream 0 opened: format 2, 1 channel, 48000 Hz, 4800 frames
[hos sead::MainThread] audio in: stream 1 opened (raw): format 2, 1 channel, 48000 Hz, 4800 frames
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 54.2 ms for a core: given by sead::MainThread (priority 16), which held it 122.0 ms and was never asked; the cores it may use now: core 0: it;
[hos sead::MainThread] vi: QueryDisplayHdrStatus -> SDR
[hos sead::MainThread] hid: the game reads npad 0 (No.1, lr 0x83489670)
[hos sead::MainThread] sdl: SDL3
[hos sead::MainThread] hid: game window not in front (foreground window: process 7404, class "MozillaWindowClass")
[hos sead::MainThread] hid: six-axis handles: 2 asked for, 3 reported
[hos ModuleSystemWorker5] tex: the game's texture heap is compacted at most every 1000 ms (HOS_TEX_COMPACT_MS; 0: at every frame, the game's own; off: never)
[hos ModuleSystemWorker4] nvn: the game reports counter type 16 (lr 0x8074b278)
[hos ModuleSystemWorker4] nvn: the game reports counter type 0 (lr 0x8074b290)
[hos sead::MainThread] nvn: window crop 0,0 2560x1440
[hos sead::MainThread] nvn: EventSignal (CPU) #1: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #1: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #1: event 0x100d6c8000 mode 0 value 0 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #2: event 0x100d6c8000 mode 0 value 0 (now 0) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #3: event 0x100d6c8000 mode 0 value 0 (now 0) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #2: queue 000010100D695020 (its #1), 1 handles (lr 0x80666be0)
[hos Porter] fs: mounted npln: -> D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\romfs
[hos sead::MainThread] nvn: CommandBufferWaitEvent #4: event 0x100d6c8000 mode 0 value 0 (now 0) (lr 0x806661f8)
[hos host] [debug] vk: 1299 pools mapped
[hos sead::MainThread] nvn: QueueSubmitCommands #3: queue 000010100D6BF020 (its #1), 1 handles (lr 0x80666be0)
[hos Porter] fiber 0x100ac54230: function 0x8504b440 arg 0x100ac541f8 stack 0x100acc1000+0x10000 flags 1
[hos Porter] fiber 0x100acd1090: function 0x8504b440 arg 0x100acd1058 stack 0x100acd4000+0x10000 flags 1
[hos nn.npln.Worker] fiber 0x100ac543d0: function 0x8504b440 arg 0x100ac54398 stack 0x100ace6000+0x10000 flags 1
[hos Porter] fiber 0x100acf7090: function 0x8504b440 arg 0x100acf7058 stack 0x100acfa000+0x10000 flags 1
[hos nn.npln.Worker] fiber 0x100ac3c5d0: function 0x8504b440 arg 0x100ac3c598 stack 0x100ad0a000+0x10000 flags 1
[hos host] vk: a 128 MB capture buffer made ahead in 19.0 ms, off the game's thread
[hos sead::MainThread] nvn: EventSignal (CPU) #2: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos Presentation Thread] nvn: QueueSubmitCommands #4: queue 000010100D5EB020 (its #1), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #5: queue 000010100D640020 (its #1), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #6: queue 000010100D6AA020 (its #1), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #7: queue 0000101004133020 (its #1), 1 handles (lr 0x80590358)
[hos host] vk: mirror syncs' pages from the captures by a compute pass (RX_VK_MSYNC_PASS)
[hos host] [debug] vk: 1699 pools mapped
[hos sead::MainThread] nvn: EventSignal (CPU) #3: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #5: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #2: event 0x100d6c8000 mode 0 value 1 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #6: event 0x100d6c8000 mode 0 value 1 (now 0) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #7: event 0x100d6c8000 mode 0 value 1 (now 0) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #8: queue 000010100D695020 (its #2), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #8: event 0x100d6c8000 mode 0 value 1 (now 0) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #9: queue 000010100D6BF020 (its #2), 1 handles (lr 0x80666be0)
[hos host] [debug] vk: 1819 pools mapped
[hos Presentation Thread] nvn: QueueSubmitCommands #10: queue 000010100D5EB020 (its #2), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #11: queue 000010100D640020 (its #2), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #12: queue 000010100D6AA020 (its #2), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #13: queue 0000101004133020 (its #2), 1 handles (lr 0x80590358)
[hos host] vk: 2 pools imported
[hos host] vk: written bindings over 1024 KB are written back only where the CPU reads or writes them, after 8 submissions whole (RX_VK_WB_MAX_KB)
[hos host] vk: texture 0x1001f0ffe00 2560x1440x1 x1 levels 1 NVN 0x3f -> vk 122 (2 textures, 14 MB)
[hos host] vk: texture 0x1004137000 2560x1440x1 x1 levels 1 NVN 0x25 -> vk 37 (3 textures, 28 MB)
[hos host] vk: upscaler: faded: while a temporal upscaler or frame generation draws, the depth draws of a skinned mesh that had colour and has had none for 3 frames are left out until its colour is back (RX_VK_DLSS_FADED=0 draws them)
[hos host] vk: descriptor slotgen ready: draws keep their descriptors through the registrations of slots they do not sample (menu: Descriptor reuse)
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 5, mip filter linear, LOD bias 0; made with 0, an image of one level: not the row's)
[hos host] vk: 3 pools imported
[hos host] vk: import 2.5 ms: commit+properties 0.0, buffer 0.0, allocate 2.5, bind+address 0.0
[hos host] vk: pool 0x102dd34020+0x40 imported in 2.6 ms as 0x102dc00000..0x102e000000 (4 imports, 7 ms in all)
[hos host] vk: 4 pools imported
[hos host] vk: write-back diff on (RX_VK_WB_DIFF=0: off)
[hos sead::MainThread] nvn: EventSignal (CPU) #4: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: present texture 0x1004137000 2560x1440 format 0x25: rendered
[hos host] vk: present: 2560x1440, format 37 (its image's 37); the Insert menu's HDR 0, HDR10 swapchain no: SDR
[hos host] wide ui: picture 0.0000: the HUD at 100% of its 16:9 size, its sides moved 0 and its top and bottom 0 canvas units; the menus at 100%
[hos host] vk: GPU hang watch: work that finishes nothing for 30 s ends the game (RX_VK_GPU_HANG_SEC)
[hos host] vk: acquire ahead: not now, the swapchain is FIFO (V-Sync On, or the window's only mode): each image asked for at its present
[hos Presentation Thread] vi: real-time step on: ratio 3.035, game's frame rate 20
[hos sead::MainThread] nvn: EventSignal (CPU) #5: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #9: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #3: event 0x100d6c8000 mode 0 value 2 (lr 0x80666598)
[hos host] vk: late pages' fix on (RX_VK_LATE_FIX=0: the wait, as before)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #10: event 0x100d6c8000 mode 0 value 2 (now 0) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #11: event 0x100d6c8000 mode 0 value 2 (now 0) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #14: queue 000010100D695020 (its #3), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #12: event 0x100d6c8000 mode 0 value 2 (now 0) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #15: queue 000010100D6BF020 (its #3), 1 handles (lr 0x80666be0)
[hos Presentation Thread] nvn: QueueSubmitCommands #16: queue 000010100D5EB020 (its #3), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #17: queue 000010100D640020 (its #3), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #18: queue 000010100D6AA020 (its #3), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #19: queue 0000101004133020 (its #3), 1 handles (lr 0x80590358)
[hos host] vk: submission 11 (queue 0x10100d5eb020) stopped at command buffer 0's wait for event 0x100d6c8000 == 1 (has 0)
[hos host] vk: submission 12 (queue 0x10100d640020) stopped at command buffer 0's wait for event 0x100d6c8000 == 1 (has 0)
[hos host] vk: submission 13 (queue 0x10100d6aa020) stopped at command buffer 0's wait for event 0x100d6c8000 == 1 (has 0)
[hos host] vk: submission 14 (queue 0x101004133020) stopped at command buffer 0's wait for event 0x1006e38004 == 1 (has 0)
[hos host] vk: texture 0x1005037000 2560x1440x1 x1 levels 1 NVN 0x25 -> vk 37 (4 textures, 42 MB)
[hos sead::MainThread] nvn: EventSignal (CPU) #6: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: the game sets present interval 1 -> 3
[hos sead::MainThread] nvn: the game sets present interval 3 -> 1
[hos host] vk: present texture 0x1005037000 2560x1440 format 0x25: rendered
[hos sead::MainThread] nvn: EventSignal (CPU) #7: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos host] [debug] vk: 1857 pools mapped
[hos sead::MainThread] nvn: CommandBufferWaitEvent #13: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #4: event 0x100d6c8000 mode 0 value 3 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #14: event 0x100d6c8000 mode 0 value 3 (now 1) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #15: event 0x100d6c8000 mode 0 value 3 (now 1) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #20: queue 000010100D695020 (its #4), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #16: event 0x100d6c8000 mode 0 value 3 (now 1) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #21: queue 000010100D6BF020 (its #4), 1 handles (lr 0x80666be0)
[hos Presentation Thread] nvn: QueueSubmitCommands #22: queue 000010100D5EB020 (its #4), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #23: queue 000010100D640020 (its #4), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #24: queue 000010100D6AA020 (its #4), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #25: queue 0000101004133020 (its #4), 1 handles (lr 0x80590358)
[hos host] vk: submission 18 (queue 0x10100d5eb020) stopped at command buffer 0's wait for event 0x100d6c8000 == 2 (has 1)
[hos host] vk: submission 19 (queue 0x10100d640020) stopped at command buffer 0's wait for event 0x100d6c8000 == 2 (has 1)
[hos host] vk: submission 20 (queue 0x10100d6aa020) stopped at command buffer 0's wait for event 0x100d6c8000 == 2 (has 1)
[hos host] vk: submission 21 (queue 0x101004133020) stopped at command buffer 0's wait for event 0x1006e38000 == 1 (has 0)
[hos host] vk: texture 0x1005f37000 2560x1440x1 x1 levels 1 NVN 0x25 -> vk 37 (5 textures, 56 MB)
[hos sead::MainThread] nvn: EventSignal (CPU) #8: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: present texture 0x1005f37000 2560x1440 format 0x25: rendered
[hos host] [debug] vk: 1858 pools mapped
[hos sead::MainThread] nvn: EventSignal (CPU) #9: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #17: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #5: event 0x100d6c8000 mode 0 value 4 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #18: event 0x100d6c8000 mode 0 value 4 (now 2) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #19: event 0x100d6c8000 mode 0 value 4 (now 2) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #26: queue 000010100D695020 (its #5), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #20: event 0x100d6c8000 mode 0 value 4 (now 2) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #27: queue 000010100D6BF020 (its #5), 1 handles (lr 0x80666be0)
[hos Presentation Thread] nvn: QueueSubmitCommands #28: queue 000010100D5EB020 (its #5), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #29: queue 000010100D640020 (its #5), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #30: queue 000010100D6AA020 (its #5), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #31: queue 0000101004133020 (its #5), 1 handles (lr 0x80590358)
[hos host] [debug] vk: 1862 pools mapped
[hos host] vk: submission 25 (queue 0x10100d5eb020) stopped at command buffer 0's wait for event 0x100d6c8000 == 3 (has 2)
[hos host] [debug] vk: 1927 pools mapped
[hos host] vk: submission 26 (queue 0x10100d640020) stopped at command buffer 0's wait for event 0x100d6c8000 == 3 (has 2)
[hos host] vk: submission 27 (queue 0x10100d6aa020) stopped at command buffer 0's wait for event 0x100d6c8000 == 3 (has 2)
[hos host] vk: submission 28 (queue 0x101004133020) stopped at command buffer 0's wait for event 0x1006e38004 == 1 (has 0)
[hos sead::MainThread] nvn: EventSignal (CPU) #10: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: present texture 0x1004137000 2560x1440 format 0x25: rendered
[hos sead::MainThread] nvn: EventSignal (CPU) #11: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #21: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #6: event 0x100d6c8000 mode 0 value 5 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #22: event 0x100d6c8000 mode 0 value 5 (now 3) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #23: event 0x100d6c8000 mode 0 value 5 (now 3) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #32: queue 000010100D695020 (its #6), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #24: event 0x100d6c8000 mode 0 value 5 (now 3) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #33: queue 000010100D6BF020 (its #6), 1 handles (lr 0x80666be0)
[hos Presentation Thread] nvn: QueueSubmitCommands #34: queue 000010100D5EB020 (its #6), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #35: queue 000010100D640020 (its #6), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #36: queue 000010100D6AA020 (its #6), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #37: queue 0000101004133020 (its #6), 1 handles (lr 0x80590358)
[hos SceneLowPriority] nfp: the game uses NFC tags: the Insert menu's NFC tab holds one of %USERPROFILE%\AppData\Local\Longshot\nfc's to the reader
[hos host] hid: SDL controller support ready (DualSense, DualShock 4, Switch Pro, Xbox and others)
[hos host] hid: controller "XInput Controller #1" connected (SDL); guid=0300fa67c82d00000631000014017801; mapping=available; mapping-name=""
[hos host] hid: guid=0300fa67c82d00000631000014017801; mapping-string[1/1]="0300fa67c82d00000631000014017801,
,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b10,leftshoulder:b4,leftstick:b8,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b9,righttrigger:a5,rightx:a3,righty:a4,start:b7,x:b2,y:b3,platform:Windows,"
[hos host] vibration: RenHDrum 0.5.0 loaded
[hos host] vibration: "XInput Controller #1": four motors (RenHDrum)
[hos sead::MainThread] nvn: EventSignal (CPU) #12: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #13: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #25: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #7: event 0x100d6c8000 mode 0 value 6 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #26: event 0x100d6c8000 mode 0 value 6 (now 4) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #27: event 0x100d6c8000 mode 0 value 6 (now 4) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #38: queue 000010100D695020 (its #7), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #28: event 0x100d6c8000 mode 0 value 6 (now 4) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #39: queue 000010100D6BF020 (its #7), 1 handles (lr 0x80666be0)
[hos Presentation Thread] nvn: QueueSubmitCommands #40: queue 000010100D5EB020 (its #7), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #41: queue 000010100D640020 (its #7), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #42: queue 000010100D6AA020 (its #7), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #43: queue 0000101004133020 (its #7), 1 handles (lr 0x80590358)
[hos sead::MainThread] nvn: EventSignal (CPU) #14: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos ModuleSystemWorker1] ocb: a new neighbour to preload is taken once it stays the choice 2000 ms (HOS_OCB_HOLD_MS; 0: the game's own at every frame)
[hos sead::MainThread] nvn: EventSignal (CPU) #15: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #29: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #8: event 0x100d6c8000 mode 0 value 7 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #30: event 0x100d6c8000 mode 0 value 7 (now 5) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #31: event 0x100d6c8000 mode 0 value 7 (now 5) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #44: queue 000010100D695020 (its #8), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #32: event 0x100d6c8000 mode 0 value 7 (now 5) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #45: queue 000010100D6BF020 (its #8), 1 handles (lr 0x80666be0)
[hos Presentation Thread] nvn: QueueSubmitCommands #46: queue 000010100D5EB020 (its #8), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #47: queue 000010100D640020 (its #8), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #48: queue 000010100D6AA020 (its #8), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #49: queue 0000101004133020 (its #8), 1 handles (lr 0x80590358)
[hos sead::MainThread] nvn: EventSignal (CPU) #16: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #17: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #33: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #9: event 0x100d6c8000 mode 0 value 8 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #34: event 0x100d6c8000 mode 0 value 8 (now 6) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #35: event 0x100d6c8000 mode 0 value 8 (now 6) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #50: queue 000010100D695020 (its #9), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #36: event 0x100d6c8000 mode 0 value 8 (now 6) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #51: queue 000010100D6BF020 (its #9), 1 handles (lr 0x80666be0)
[hos Presentation Thread] nvn: QueueSubmitCommands #52: queue 000010100D5EB020 (its #9), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #53: queue 000010100D640020 (its #9), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #54: queue 000010100D6AA020 (its #9), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #55: queue 0000101004133020 (its #9), 1 handles (lr 0x80590358)
[hos sead::MainThread] nvn: EventSignal (CPU) #18: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: texture budget: 7158 MB (driver allows 11474 MB, the card has 12272 MB, RAM 32694 MB)
[hos sead::MainThread] nvn: EventSignal (CPU) #19: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #37: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #10: event 0x100d6c8000 mode 0 value 9 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #38: event 0x100d6c8000 mode 0 value 9 (now 8) (lr 0x806661f8)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #39: event 0x100d6c8000 mode 0 value 9 (now 8) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #56: queue 000010100D695020 (its #10), 1 handles (lr 0x80666be0)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #40: event 0x100d6c8000 mode 0 value 9 (now 8) (lr 0x806661f8)
[hos sead::MainThread] nvn: QueueSubmitCommands #57: queue 000010100D6BF020 (its #10), 1 handles (lr 0x80666be0)
[hos Presentation Thread] nvn: QueueSubmitCommands #58: queue 000010100D5EB020 (its #10), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #59: queue 000010100D640020 (its #10), 1 handles (lr 0x805282f4)
[hos Presentation Thread] nvn: QueueSubmitCommands #60: queue 000010100D6AA020 (its #10), 1 handles (lr 0x805282f4)
[hos sead::MainThread] nvn: EventSignal (CPU) #20: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: RX_VK_RECORDER=thread: a command buffer or recorded call on a thread without the replay lock while the recorder was busy: made after the recorder caught up (counted in the stats line; logged once)
[hos sead::MainThread] nvn: EventSignal (CPU) #21: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #11: event 0x100d6c8000 mode 0 value 10 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #22: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #23: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #12: event 0x100d6c8000 mode 0 value 11 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #24: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #25: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #13: event 0x100d6c8000 mode 0 value 12 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #26: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #27: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #14: event 0x100d6c8000 mode 0 value 13 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #28: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: vertex format NVN 0x8 -> vk 74
[hos sead::MainThread] nvn: EventSignal (CPU) #29: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #15: event 0x100d6c8000 mode 0 value 14 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #30: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos ModuleSystemWorker2] barrier: the game's job barriers (f_800998a0, f_800738e0, f_8007e200) wait until their masks move instead of sleeping 20 us (HOS_OOT_BARRIER_WAIT)
[hos sead::MainThread] nvn: EventSignal (CPU) #31: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #16: event 0x100d6c8000 mode 0 value 15 (lr 0x80666598)
[hos ModuleSystemWorker1] vk: invalidates with nothing to wait for keep their virtual core: on
[hos sead::MainThread] nvn: EventSignal (CPU) #32: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #33: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #17: event 0x100d6c8000 mode 0 value 16 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #34: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #35: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #18: event 0x100d6c8000 mode 0 value 17 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #36: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #37: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #19: event 0x100d6c8000 mode 0 value 18 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #38: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: EventSignal (CPU) #39: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #20: event 0x100d6c8000 mode 0 value 19 (lr 0x80666598)
[hos sead::MainThread] nvn: EventSignal (CPU) #40: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #21: event 0x100d6c8000 mode 0 value 20 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #22: event 0x100d6c8000 mode 0 value 21 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #23: event 0x100d6c8000 mode 0 value 22 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #24: event 0x100d6c8000 mode 0 value 23 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #25: event 0x100d6c8000 mode 0 value 24 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #26: event 0x100d6c8000 mode 0 value 25 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #27: event 0x100d6c8000 mode 0 value 26 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #28: event 0x100d6c8000 mode 0 value 27 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #29: event 0x100d6c8000 mode 0 value 28 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #30: event 0x100d6c8000 mode 0 value 29 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #31: event 0x100d6c8000 mode 0 value 30 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #32: event 0x100d6c8000 mode 0 value 31 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #33: event 0x100d6c8000 mode 0 value 32 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #34: event 0x100d6c8000 mode 0 value 33 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #35: event 0x100d6c8000 mode 0 value 34 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #36: event 0x100d6c8000 mode 0 value 35 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #37: event 0x100d6c8000 mode 0 value 36 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #38: event 0x100d6c8000 mode 0 value 37 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #39: event 0x100d6c8000 mode 0 value 38 (lr 0x80666598)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #40: event 0x100d6c8000 mode 0 value 39 (lr 0x80666598)
[hos host] vk: vertex format NVN 0xd -> vk 16
[hos host] vk: vertex format NVN 0xf -> vk 20
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 260 dirty pages from their submits' captures, 1222 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 148 dispatches, 326 chunks; 0 syncs as copies (no table room)
[hos host] vk: read live, by pool:
[hos Resource Memory] mem: the game's mapped pages 951 MB (peak 951); heap 6894 MB; the console's total 9809 MB
[hos ARes Loading 1] prompts: the Talk ring's A B X Y at places 0 1 2 3, the ocarina hint as is
[hos ARes Loading 0] prompts: switch textures, 1373 runs for this pad's controls
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.03 over 1185 ranges, 702 MB), 122 pages written, 122 copied (0.5 MB)
[hos host] vk: a capture taken (ms): all 0.008; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (58 MB), 12 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: late pages left to their reads off: the pools have mirrors
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos ModuleSystemWorker5] prompts: the textures' pool at 0x7fc3035000 (+0xe86c000): a new face or new controls show at once
[hos sead::MainThread] prompts: the layouts at 0x7fe1a31000: the Talk ring and the ocarina hint follow at once
[hos host] vk: import 2.1 ms: commit+properties 0.0, buffer 0.0, allocate 2.0, bind+address 0.0
[hos host] vk: pool 0x107db39100+0x40 imported in 2.1 ms as 0x107d800000..0x107e000000 (5 imports, 9 ms in all)
[hos host] vk: 5 pools imported
[hos host] vk: import 3.4 ms: commit+properties 0.0, buffer 0.0, allocate 3.4, bind+address 0.0
[hos host] vk: pool 0x107cd70000+0x120 imported in 3.5 ms as 0x107cc00000..0x107d800000 (6 imports, 12 ms in all)
[hos host] vk: 6 pools imported
[hos host] vk: write-back diff: a shadow for pool 0x100d400000 (4 MB; 4 MB of shadows)
[a64] unimplemented import ZN2nn5movie10InitializeEv (lr 0x83b7d25c)
[a64] unimplemented import ZN2nn5movie13IsInitializedEv (lr 0x85467d74)
[hos host] vk: 3901 programs from the program cache
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 3.10, present 0.32, draw 0.08, descriptors 0.01, texture lookups 0.13, uploads 0.06, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.00, memory sync 0.48, queue submit 0.00, slot wait 0.21
[hos host] vk: a frame: mirrors 6 (15 MB): 0.16 MB synced in 3.4 copies, 0.00 MB written back
[hos host] vk: 120 presents, 4.7 fps; a frame: 1 draws (0 skipped), 0 dispatches, 2 passes, 1 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 3426 pipelines (661 ms), 0 programs (0 failed), 5 textures (56 MB)
[hos host] vk: host time in GPU resource waits: 45.03 ms a frame (fences 43.28, timelines 1.76, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 23.4 full barriers recorded, 13.9 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 0.0; 0x12: 1.0; 0x78: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 0.0, an op 1.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.5 ms a frame (the longest submission 2.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.5 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 108.2 ms for a core: given by GeneralFormatAsync
(priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 53.4 ms for a core: given by tera::FileResourceMgr (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos ModuleSystemWorker2] oe: startup logo finished
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos Resource Memory] mem: the game's mapped pages 1027 MB (peak 1027); heap 6894 MB; the console's total 9809 MB
[hos Resource Memory] mem: the game's mapped pages 1283 MB (peak 1283); heap 6894 MB; the console's total 9809 MB
[hos host] vk: late pages found at the replay (600 takes): 7 pages in 7 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: vertex format NVN 0x2c -> vk 95
[hos OptimizeChangeBanc] [debug] nvn: 1 copies of recordings whose control memory was reused let go (304 live)
[hos host] vk: vertex format NVN 0x26 -> vk 38
[hos MovieOperateThread] video: nn::codec::video on libavcodec
[hos MovieOperateThread] video: AV1 on dav1d 1.5.4 (a thread per logical core)
[hos MovieOperateThread] video: AV1 decoder ready on dav1d (1920x1080)
[hos AudioFrameworkThread] audio: voice 0: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s
[hos AudioFrameworkThread] audio: voice 1: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s
[hos AudioFrameworkThread] audio: voice 2: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s
[hos AudioFrameworkThread] audio: voice 3: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s
[hos AudioFrameworkThread] audio: voice 4: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s
[hos AudioFrameworkThread] audio: voice 5: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s
[hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x108964d000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2)
[hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x1089655000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2)
[hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x108965d000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2)
[hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x1089665000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2)
[hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x108966d000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2)
[hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x1089675000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2)
[hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x108964e000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2)
[hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x1089656000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2)
[hos AudioUserTaskThread] [debug] codec: Opus decoder 0: 48000 Hz, 1 channels
[hos AudioUserTaskThread] [debug] codec: Opus decoder 1: 48000 Hz, 1 channels
[hos AudioUserTaskThread] [debug] codec: Opus decoder 2: 48000 Hz, 1 channels
[hos AudioUserTaskThread] [debug] codec: Opus decoder 3: 48000 Hz, 1 channels
[hos AudioUserTaskThread] [debug] codec: Opus decoder 4: 48000 Hz, 1 channels
[hos AudioUserTaskThread] [debug] codec: Opus decoder 5: 48000 Hz, 1 channels
[hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe000, 49152): header 000000ae 07bf4b00 -> 182
[hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe0b6, 48970): header 00000094 036ed700 -> 156
[hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe152, 48814): header 000000a1 129a5e00 -> 169
[hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe1fb, 48645): header 000000a1 03898c00 -> 169
[hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe2a4, 48476): header 000000a1 66171600 -> 169
[hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe34d, 48307): header 000000a1 0322bb00 -> 169
[hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960)
[hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960)
[hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960)
[hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960)
[hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960)
[hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960)
[hos OptimizeChangeBanc] [debug] nvn: 4 copies of recordings whose control memory was reused let go (596 live)
[hos MovieOperateThread] video: the first buffer went out before the first picture was decoded (shown black)
[hos Presentation Thread] vi: a movie is playing: the frame rate goes to 240 for it (HOS_MOVIE_PACE=0 leaves it)
[hos host] vk: texture 0x10000000000 1920x1080x1 x1 levels 1 NVN 0x1 -> vk 9 (6 textures, 58 MB)
[hos host] vk: import 3.5 ms: commit+properties 0.0, buffer 0.0, allocate 3.4, bind+address 0.0
[hos host] vk: pool 0x10766a2000+0x1fa400 imported in 3.5 ms as 0x1075c00000..0x1077000000 (7 imports, 16 ms in all)
[hos host] vk: 7 pools imported
[hos host] vk: texture 0x1000021c200 960x540x1 x1 levels 1 NVN 0xd -> vk 16 (7 textures, 59 MB)
[hos host] vk: sampler 268 is nearest: min filter 4, mag filter 1, wrap 7/7/7, lod -1000..1000 bias 0, aniso 1, compare 0
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 4, mip filter linear, LOD bias 0; made with 0, an image of one level: not the row's)
[hos OptimizeChangeBanc] nvn: DrawMeshTasks: 1 so far (x1 0 x2 0x2 x3 0x400000 x4 0x83b70de0)
[hos OptimizeChangeBanc] nvn: DrawMeshTasks: 2 so far (x1 0 x2 0x2 x3 0x400000 x4 0x83b70de0)
[hos OptimizeChangeBanc] nvn: DrawMeshTasks: 3 so far (x1 0 x2 0x2 x3 0x400000 x4 0x83b70de0)
[hos host] vk: textures from a pool of their own, 32 MB blocks (RX_VK_COMPACT), 24 made at the start
[hos host] vk: texture 0x10755b4000 1920x1080x1 x1 levels 1 NVN 0x25 -> vk 37 (8 textures, 67 MB)
[hos host] vk: import 10.4 ms: commit+properties 0.2, buffer 0.0, allocate 10.2, bind+address 0.0
[hos host] vk: pool 0x10755b4000+0x870000 imported in 10.4 ms as 0x1072c00000..0x1077000000 (8 imports, 26 ms in all)
[hos host] vk: pool 0x1075c00000 remade larger (1 retired)
[hos host] vk: 7 pools imported
[hos host] vk: in play: compute e409733c45530d75 (e409733c45530d75 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program e409733c45530d75 not translated: a network left out on this card
[hos host] vk: rejected compute e409733c45530d75 (control 0x1075e24768) grid 900x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x1072e26320+0xfd2000; ssbo2 0x1072dea000+0x1200; ssbo3 0x1072deb200+0x20; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute 8306877b7dc8784b (8306877b7dc8784b 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program 8306877b7dc8784b not translated: a network left out on this card
[hos host] vk: rejected compute 8306877b7dc8784b (control 0x1075e260d8) grid 720x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x10745e1320+0xfd2000; ssbo1 0x1072e26320+0xfd2000; ssbo2 0x1072deb300+0x1200; ssbo3 0x1072dec500+0x20; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute de3c10dba4ff39d3 (de3c10dba4ff39d3 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program de3c10dba4ff39d3 not translated: a network left out on this card
[hos host] vk: rejected compute de3c10dba4ff39d3 (control 0x1075e27948) grid 97x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x1072e26320+0x7e9000; ssbo1 0x10745e1320+0xfd2000; ssbo2 0x1072dec600+0x1000; ssbo3 0x1072ded600+0x40; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute a41c78faa56ccf7d (a41c78faa56ccf7d 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program a41c78faa56ccf7d not translated: a network left out on this card
[hos host] vk: rejected compute a41c78faa56ccf7d (control 0x1075e28860) grid 120x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x107360f320+0x7e9000; ssbo1 0x1072e26320+0x7e9000; ssbo2 0x1072ded700+0x4800; ssbo3 0x1072df1f00+0x40; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute e959b9b1451491c0 (e959b9b1451491c0 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos OptimizeChangeBanc] [debug] nvn: 3 copies of recordings whose control memory was reused let go (2160 live)
[hos host] [trace] vk: program e959b9b1451491c0 not translated: a network left out on this card
[hos host] vk: rejected compute e959b9b1451491c0 (control 0x1075e2a100) grid 70x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x1072e26320+0x3f4800; ssbo1 0x107360f320+0x7e9000; ssbo2 0x1072df2000+0x4000; ssbo3 0x1072df6000+0x80; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute 5a4d7378db437fed (5a4d7378db437fed 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program 5a4d7378db437fed not translated: a network left out on this card
[hos host] vk: rejected compute 5a4d7378db437fed (control 0x1075e2b038) grid 60x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x107321ab20+0x3f4800; ssbo1 0x1072e26320+0x3f4800; ssbo2 0x1072df6100+0x12000; ssbo3 0x1072e08100+0x80; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute 8b73eaee02efc5e4 (8b73eaee02efc5e4 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program 8b73eaee02efc5e4 not translated: a network left out on this card
[hos host] vk: rejected compute 8b73eaee02efc5e4 (control 0x1075e2c8f0) grid 85x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x1073df8320+0x7e9000; ssbo1 0x1072e26320+0x3f4800; ssbo2 0x1072e1a300+0x4000; ssbo3 0x1072e1e300+0x40; ssbo4 0x107360f320+0x7e9000; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute 14140d4c1dbe754f (14140d4c1dbe754f 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program 14140d4c1dbe754f not translated: a network left out on this card
[hos host] vk: rejected compute 14140d4c1dbe754f (control 0x1075e2d970) grid 8100x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x107360f320+0xfd2000; ssbo1 0x1072e26320+0x7e9000; ssbo2 0x1072e22d00+0x4000; ssbo3 0x1072e23d00+0x40; ssbo4 0x10745e1320+0xfd2000; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute 4c2c129fe950b041 (4c2c129fe950b041 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program 4c2c129fe950b041 not translated: a network left out on this card
[hos host] vk: rejected compute 4c2c129fe950b041 (control 0x1075e2e9c8) grid 5400x1x1: img 0x1075e31000 1920x1080x1 fmt 0x3d; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x10745e1320+0xfd2000; ssbo1 0x10745e1320+0xfd2000; ssbo2 0x1072e25100+0x1200; ssbo3 0x1072e26300+0x20; ssbo4 0x10745e1320+0xfd2000; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100;
[hos host] vk: in play: compute 3f6f2b1d671badcd (3f6f2b1d671badcd 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants
[hos host] [trace] vk: program 3f6f2b1d671badcd not translated: a network left out on this card
[hos host] vk: rejected compute 3f6f2b1d671badcd (control 0x1072da4fe0) grid 80x180x1: img 0x1075e31000 1920x1080x1 fmt 0x3d; img 0x10000348400 2560x1440x1 fmt 0x38; tex 0x1075e31000 1920x1080x1 fmt 0x3d; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x1072ca0200 512x512x1 fmt 0x3d; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x10745e1320+0xfd2000; ssbo1 0x10745e1320+0xfd2000; ssbo2 0x1072da1f00+0x3000; ssbo3 0x1072e26300+0x20; ssbo4 0x10745e1320+0xfd2000; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100;
[hos host] vk: texture 0x10000348400 2560x1440x1 x1 levels 1 NVN 0x38 -> vk 37 (9 textures, 81 MB)
[hos host] vk: import 3.9 ms: commit+properties 0.0, buffer 0.0, allocate 3.9, bind+address 0.0
[hos host] vk: pool 0x100df5f400+0xf00000 imported in 3.9 ms as 0x100dc00000..0x100f000000 (9 imports, 30 ms in all)
[hos host] vk: movie upscaling uses the linear compute path
[hos Presentation Thread] vi: the movie is over: back to the game's frame rate cap 60
[hos host] vk: pool 0x1072c00000..0x1077000000: 6540 KB the CPU wrote since retired pool 0x1075c00000's last sync kept from guest memory, not its mirror
[hos host] vk: pool 0x1072c00000..0x1077000000: mirror made with the GPU's contents of 1 retired pool(s)
[hos host] vk: late pages' fix: stamps for 0x100d6c0000 (pool 0x100d400000; 1 blocks, 64 KB)
[hos host] vk: late pages' fix: stamps for 0x100d400000 (pool 0x100d400000; 2 blocks, 128 KB)
[hos host] vk: late pages' fix: stamps for 0x100d410000 (pool 0x100d400000; 3 blocks, 192 KB)
[hos host] vk: late pages' fix: stamps for 0x100d420000 (pool 0x100d400000; 4 blocks, 256 KB)
[hos host] vk: late pages' fix: stamps for 0x100d430000 (pool 0x100d400000; 5 blocks, 320 KB)
[hos host] vk: late pages' fix: stamps for 0x100d440000 (pool 0x100d400000; 6 blocks, 384 KB)
[hos host] vk: late pages' fix: stamps for 0x100d450000 (pool 0x100d400000; 7 blocks, 448 KB)
[hos host] vk: late pages' fix: stamps for 0x100d460000 (pool 0x100d400000; 8 blocks, 512 KB)
[hos host] vk: late pages' fix: stamps for 0x100d7f0000 (pool 0x100d400000; 64 blocks, 4096 KB)
[hos Presentation Thread] vi: a movie is playing: the frame rate goes to 240 for it (HOS_MOVIE_PACE=0 leaves it)
[hos host] vk: sparse mirrors: 7, 79 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 0; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 9 binds (1.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 0 kept as spares, 0 destroyed; 0 spares now, 0 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 136 dirty pages from their submits' captures, 3325 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 183 dispatches, 500 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 40 runs, 0.0 MB compared; 40 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 60 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: mirror of pool 0x1075c00000 (20 MB) given back: retired, 50 submissions since its sync
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 1265 ranges, 907 MB), 153 pages written, 153 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.019 (2 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (90 MB), 8 spare buffers, 0 texture pages copied (0 MB)
[hos OptimizeChangeBanc] [debug] nvn: 6 copies of recordings whose control memory was reused let go (2974 live)
[hos Resource Memory] mem: the game's mapped pages 1536 MB (peak 1536); heap 6894 MB; the console's total 9809 MB
[hos sead::MainThread] nvn: GPU waits in 5 s: 46 sync waits (570 ms), 0 polls timed out (0 ms), 1 queue finishes (46 ms)
[hos ParallelDecompThread] zstd: 3 stream(s) decoding on the host at once
[hos sead::MainThread] [debug] nvn: 3 copies of recordings whose control memory was reused let go (3512 live)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.10
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3602 dirty pages from their submits' captures, 32761 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2620 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 90 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 4427 programs from the program cache
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 3.54, present 0.05, draw 0.06, descriptors 0.01, texture lookups 1.18, uploads 0.14, pipelines 0.00, programs 0.00, commands 0.04, dispatch 2.10, memory sync 0.75, queue submit 0.00, slot wait 0.17
[hos host] vk: a frame: mirrors 8 (78 MB): 2.17 MB synced in 5.1 copies, 0.00 MB written back
[hos host] vk: 240 presents, 91.4 fps; a frame: 2 draws (9 skipped), 2 dispatches, 3 passes, 11 barriers, 6 submits, 0.0 uploads (0.2 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 5049 pipelines (1030 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 7.22 ms a frame (fences 3.16, timelines 4.06, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 9.3
[hos host] vk: a frame: 25.3 full barriers recorded, 26.3 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 10.1; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 10.1, an op 11.8, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 2.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.2 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.02 ms
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.03 over 1296 ranges, 948 MB), 161 pages written, 161 copied (0.6 MB)
[hos OptimizeChangeBanc] [debug] nvn: 3 copies of recordings whose control memory was reused let go (4005 live)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (84 MB), 8 spare buffers, 0 texture pages copied (0 MB)
[hos BaseProcExecutor0] vcores: 123740 takes, 27998 waits (mean 83 us, longest 498.5 ms, 7 over HOS_VCORE_LOG_MS), 13303 kicks, 12240 yields, 0 starved grants, 0 escapes in 5 s
[hos OptimizeChangeBanc] [debug] nvn: 3 copies of recordings whose control memory was reused let go (4541 live)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 120 pages in 120 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.11
[hos host] vk: vertex format NVN 0xe -> vk 17
[hos host] vk: sparse mirrors: 7, 78 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 11 binds (1.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 1 kept as spares, 0 destroyed; 1 spares now, 20 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2769 dirty pages from their submits' captures, 30218 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2414 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 95 in captures, 1 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos OptimizeChangeBanc] [debug] nvn: 4 copies of recordings whose control memory was reused let go (5437 live)
[hos OptimizeChangeBanc] [debug] nvn: 3 copies of recordings whose control memory was reused let go (5723 live)
[hos Resource Memory] mem: the game's mapped pages 1792 MB (peak 1792); heap 6894 MB; the console's total 9809 MB
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.03 over 1318 ranges, 1024 MB), 166 pages written, 166 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (130 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.16, present 0.05, draw 0.06, descriptors 0.01, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.04, dispatch 0.06, memory sync 0.68, queue submit 0.00, slot wait 0.17
[hos host] vk: a frame: mirrors 8 (78 MB): 1.68 MB synced in 5.0 copies, 0.00 MB written back
[hos host] vk: 360 presents, 164.9 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 6231 pipelines (1331 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 5.41 ms a frame (fences 2.78, timelines 2.64, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 27.9 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.6 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 0 KB; 0 made on demand (0 MB kept now); c1 stages: at most 0 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6532 dirty pages from their submits' captures, 27295 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2439 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 85 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: vertex format NVN 0x1 -> vk 9
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.09
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 1361 ranges, 1139 MB), 191 pages written, 191 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (104 MB), 8 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 147 pages in 147 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 7, 78 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 11 binds (1.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 1 kept as spares, 0 destroyed; 1 spares now, 20 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2984 dirty pages from their submits' captures, 29347 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2367 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 103 in captures, 1 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: 5240 programs from the program cache
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.25, present 0.05, draw 0.06, descriptors 0.01, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.04, dispatch 0.07, memory sync 0.70, queue submit 0.00, slot wait 0.19
[hos host] vk: a frame: mirrors 8 (78 MB): 1.73 MB synced in 5.1 copies, 0.00 MB written back
[hos host] vk: 480 presents, 165.0 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 7234 pipelines (1629 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 5.35 ms a frame (fences 2.84, timelines 2.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 27.4 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.1 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.11
[hos Presentation Thread] vi: 503 frames presented (100.5 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 358, 15.5-17.8 134, 17.8-22 2, 22-30 0, 30-40 4, over 40 4
[hos OptimizeChangeBanc] [debug] nvn: 1 copies of recordings whose control memory was reused let go (13140 live)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 1396 ranges, 1222 MB), 177 pages written, 177 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (112 MB), 10 spare buffers, 0 texture pages copied (0 MB)
[hos OptimizeChangeBanc] nvn: 50000 recordings made, 13368 live copies (1.3 MB), 7 texture copies; 2 command buffers' segments (0.1 MB)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 8529 dirty pages from their submits' captures, 23677 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2348 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 90 in captures, 1 only in the collection, 1 only by the replay itself
[hos host] vk: read live, by pool:
[hos Resource Memory] mem: the game's mapped pages 2048 MB (peak 2048); heap 6894 MB; the console's total 9809 MB
[hos host] vk: presenter thread: the copy's fence 3.45 ms, the present 0.77 ms a present
[hos host] vk: memory: working set 9754 MB (peak 9754), private 12033 MB, staging 16 MB in 1 arenas, 1 made so far (7 ms)
[hos host] vk: memory heaps: heap 0 (system): 12 blocks 592 MB, 37 allocations 540 MB; used 756 of 15579 MB; heap 1 (device): 83 blocks 1136 MB, 1472 allocations 306 MB; used 2616 of 11474 MB;
[hos host] vk: image blocks: 2 blocks 64 MB, 2 images 63 MB; 1 free ranges, the largest 0 MB
[hos host] vk: image blocks (render targets): 3 blocks 96 MB, 8 images 73 MB; 3 free ranges, the largest 17 MB
[hos host] vk: image blocks (textures, compacted): 24 blocks 768 MB, 1 images 7 MB; 24 free ranges, the largest 32 MB
[hos host] vk: render target images: 0 reused from the pool, 8 made new, 0 destroyed from it, 0 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 0 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 8 with the pool empty (0 of them with targets still in the dead queue), 0 with none in it a present yet, 0 with none of a kind they can use, 0 with all smaller (the largest 0% short on average), 0 with all bigger (over 1.5 times), 0 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 0 idle
[hos host] vk: VRAM: 2616 MB used of 11474 MB allowed (829 MB of it free in the blocks); textures 81 MB (budget 7158 MB), 9 textures
[hos host] vk: budget: VRAM 1136 of 11474 MB (2616 MB by the driver, 1480 MB unaccounted): blocks free 829, textures 81, mirrors 78, other images 63, other buffers 62, staging 16, shadows 4; RAM 703 MB (the process's private memory 12033 MB): capture spares 420, imports 124, captures 106, retired imports 20, capture table 19, other buffers 14, others 0
[hos host] vk: textures: decoded ASTC 0 MB, rendered 81 MB (1 drawn in one frame, kept: 128 KB), other 0 MB; used in the last 8 presents 81 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 0 MB of blocks hashed in 0 ms
[hos host] vk: 14 MB: 0x1004137000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 2 presents ago
[hos host] vk: 14 MB: 0x1005f37000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 3 presents ago
[hos host] vk: 14 MB: 0x1001f0ffe00 2560x1440x1 x1 levels 1 NVN 0x3f rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x10000348400 2560x1440x1 x1 levels 1 NVN 0x38 rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x1005037000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago
[hos host] vk: 7 MB: 0x10755b4000 1920x1080x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago
[hos host] vk: 1 MB: 0x10000000000 1920x1080x1 x1 levels 1 NVN 0x1 rendered, last used 1 presents ago
[hos host] vk: 0 MB: 0x1000021c200 960x540x1 x1 levels 1 NVN 0xd rendered, last used 1 presents ago
[hos Presentation Thread] vk: presents after the CPU's signal (600): 567 waited for it, 3.00 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.11
[hos host] vk: status: 600 frames, the slowest 275.6 ms; VRAM 2.6 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 0 program records loaded (0.0 ms), 0 new graphics pipelines, pipelines from the driver's cache 0, from the start's cache 0, compiled 0; 0 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 3114 queued, 3114 loaded (median 1.05 ms, 99th percentile 17.17 ms, worst 34.27 ms, 7049 ms and 171.6 MB read in all); 3836 programs (583 of them compute variants without their code, 0 graphics programs no note names), 6183 graphics pipelines (0 not in a cache), 365 compute pipelines (19 not in a cache); a first use ran 0 itself and waited for 0 (0.0 ms); 3 pipelines made at create taken by draws; 1295 create calls with nothing kept for their stages, 0 before the index; units 4409 (3114 ready, 0 loading), 8148 programs alive; 0 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 6422 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 757 known, 0 learned in this run; 365 compute pipelines made by them at a create call, 19 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.29, present 0.06, draw 0.06, descriptors 0.01, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.04, dispatch 0.07, memory sync 0.75, queue submit 0.00, slot wait 0.18
[hos host] vk: a frame: mirrors 8 (78 MB): 1.73 MB synced in 5.0 copies, 0.00 MB written back
[hos host] vk: 600 presents, 165.2 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 7869 pipelines (1827 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 5.26 ms a frame (fences 2.80, timelines 2.47, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 27.6 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 0.00 ms a frame): its jobs 6.01 ms a frame (submits 1.46, presents 3.81, others 0.74), idle -6.01; the game waited for room at 431 presents, 1.79 ms each (1.28 a frame; 16.09 jobs queued at a present); 43 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 1.91 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 580 captures at a CPU signal; the submits' waits read in 0.002 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.05 ms after its hand-over (265.75 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 7.98 + 0.12 (the window's copy is in the idle), after it 0.09 + 0.00; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 89.92 (0.13 ms)
[hos host] vk: frame ahead (600 presents): 1.00 parts a frame recorded in the signal's job, 17.8% of them took nothing that came with the signal; of the others, 0.82 a frame took pages of the signal's capture into a mirror (47.6 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 103 of 579 frames, its first three in 103
[hos host] vk: frame ahead, what came with the signal: 579 signals with a capture, each of 163.4 pages in 2.3 pools; between a frame's last submit and that signal 2.07 texture and sampler registrations, 2.78 releases and 0.00 other CPU signals a signal (295 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.09+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.04 over 1445 ranges, 1320 MB), 183 pages written, 183 copied (0.7 MB)
[hos host] vk: sparse mirrors: 7, 78 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 11 binds (1.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 1 kept as spares, 0 destroyed; 1 spares now, 20 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2750 dirty pages from their submits' captures, 30326 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2400 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 86 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: a capture taken (ms): all 0.007; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (108 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 143 pages in 143 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 7531 dirty pages from their submits' captures, 26285 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2428 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 74 in captures, 1 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1445 ranges, 1320 MB), 142 pages written, 142 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (96 MB), 9 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 0.94, present 0.05, draw 0.04, descriptors 0.00, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.06, memory sync 0.43, queue submit 0.00, slot wait 0.21
[hos host] vk: a frame: mirrors 8 (78 MB): 1.70 MB synced in 4.7 copies, 0.00 MB written back
[hos host] vk: 720 presents, 164.9 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 7869 pipelines (1827 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 5.46 ms a frame (fences 2.63, timelines 2.83, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 27.8 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.6 ms a frame (the longest submission 3.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 0 KB; 0 made on demand (0 MB kept now); c1 stages: at most 0 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 7, 78 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 11 binds (1.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 1 kept as spares, 0 destroyed; 1 spares now, 20 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4908 dirty pages from their submits' captures, 28034 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2372 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: mirror of pool 0x103d800000 (4 MB) given back: idle, 6015 submissions since its sync
[hos sead::MainThread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1445 ranges, 1320 MB), 142 pages written, 142 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (106 MB), 6 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 175 pages in 175 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: late pages' fix (600 recordings with one): of 856 captures with late pages 854 fixed (0.7 pages, 0.7 fixes a capture), waited as before: 0 no pool, 2 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 0 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 64 blocks (4096 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 0, of an earlier recording 0, not recorded yet 0, of this part 0; pages fixed from stamps 599, from snapshots 0; the shadows: 853 pages synced from captures, 127 made known by it, 726 left unknown because the replay or the collection wrote them since
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2322 dirty pages from their submits' captures, 29974 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2332 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 74 in captures, 0 only in the collection, 1 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 0.86, present 0.04, draw 0.04, descriptors 0.00, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.06, memory sync 0.42, queue submit 0.00, slot wait 0.16
[hos host] vk: a frame: mirrors 8 (76 MB): 1.73 MB synced in 4.7 copies, 0.00 MB written back
[hos host] vk: 840 presents, 164.3 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 7869 pipelines (1827 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 5.55 ms a frame (fences 2.63, timelines 2.92, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 28.1 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.8 ms a frame (the longest submission 3.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos sead::MainThread] vk: capture: a submit 0.09 ms (lock 0.00, scan 0.03 over 1445 ranges, 1320 MB), 140 pages written, 140 copied (0.5 MB)
[hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (102 MB), 9 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: mirror of pool 0x107cc00000 (12 MB) given back: idle, 6039 submissions since its sync
[hos host] vk: mirror of pool 0x107d800000 (8 MB) given back: idle, 6039 submissions since its sync
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos host] vk: sparse mirrors: 4, 64 MB with memory of 80 MB (15 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 13 binds (1.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 4 spares now, 44 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4489 dirty pages from their submits' captures, 29327 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2428 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: vertex format NVN 0x14 -> vk 81
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 1464 ranges, 1335 MB), 150 pages written, 150 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (116 MB), 10 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.07, present 0.05, draw 0.03, descriptors 0.00, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.07, memory sync 0.58, queue submit 0.00, slot wait 0.14
[hos host] vk: a frame: mirrors 8 (64 MB): 1.73 MB synced in 4.9 copies, 0.00 MB written back
[hos host] vk: 960 presents, 165.9 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 8106 pipelines (1903 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 5.01 ms a frame (fences 2.51, timelines 2.49, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 28.1 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 181 pages in 181 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6124 dirty pages from their submits' captures, 26843 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2397 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 91 in captures, 0 only in the collection, 1 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.11
[hos OptimizeChangeBanc] [debug] nvn: 2 copies of recordings whose control memory was reused let go (18707 live)
[hos host] vk: recorder: 154659 packets run in 10.0 s (busy 0.64 s, spun idle 0.10 s, slept 9659 times, 4872 of them under 200 us), 0 calls made directly (no room); 16150 batches over 1011 frames (16.0 a frame, 10 packets a batch), 8066 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 19 times, 1.41 ms (0.0 times, 0.00 ms a frame); every thread 791 times, 103.32 ms in all, the longest 3.87 ms; the queue's lock 380/57.35 ms (max 1.65), vkWaitForFences 360/43.22 ms (max 3.87), vkQueueBindSparse 13/1.15 ms (max 0.35), vmaDestroyBuffer 27/0.76 ms (max 0.08), vkDestroyPipeline 5/0.58 ms (max 0.51), vmaFreeMemory 4/0.13 ms (max 0.06)
[hos host] vk: recorder: of its busy 0.64 s, 0.49 s in parts' ends and submits (the longest 1.98 ms), 0.15 s of that blocked on the queue's lock (198 times, the longest 1.90 ms), so 0.35 s in the ends, submits and holds; 0.14 s running the other packets; 153 packets a frame
[hos sead::MainThread] nvn: GPU waits in 5 s: 807 sync waits (2232 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos Resource Memory] mem: the game's mapped pages 2304 MB (peak 2304); heap 6894 MB; the console's total 9809 MB
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.02, scan 0.06 over 1588 ranges, 1375 MB), 159 pages written, 159 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (114 MB), 9 spare buffers, 0 texture pages copied (0 MB)
[hos ModuleSystemWorker3] hid: game window in front (foreground window: process 32080, class "Longshot")
[hos ModuleSystemWorker3] hid: 1 controller through SDL
[hos host] vk: sparse mirrors: 4, 64 MB with memory of 80 MB (15 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 13 binds (1.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 4 spares now, 44 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4860 dirty pages from their submits' captures, 29882 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 228 dispatches, 2536 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 103 in captures, 0 only in the collection, 1 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: 6004 programs from the program cache
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.45, present 0.05, draw 0.03, descriptors 0.00, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.06, memory sync 1.00, queue submit 0.00, slot wait 0.14
[hos host] vk: a frame: mirrors 8 (64 MB): 1.76 MB synced in 5.3 copies, 0.00 MB written back
[hos host] vk: 1080 presents, 158.4 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 8991 pipelines (2077 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 5.40 ms a frame (fences 2.77, timelines 2.63, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 28.4 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 3.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos BaseProcExecutor1] vcores: 215849 takes, 39742 waits (mean 27 us, longest 12.6 ms, 0 over HOS_VCORE_LOG_MS), 20952 kicks, 20442 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.15
[hos ModuleSystemWorker2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 13 made): 7769385 lookups, 168816 missed (2.17 %)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.05 over 1661 ranges, 1392 MB), 153 pages written, 153 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (102 MB), 10 spare buffers, 0 texture pages copied (0 MB)
[hos host] audio: 1001 blocks played in the first 10 s, loudest sample 408 of 32767
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 0 KB; 0 made on demand (0 MB kept now); c1 stages: at most 0 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3991 dirty pages from their submits' captures, 29839 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2442 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 83 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 193 pages in 193 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: presenter thread: the copy's fence 2.69 ms, the present 0.74 ms a present
[hos host] vk: memory: working set 10388 MB (peak 10400), private 12789 MB, staging 16 MB in 1 arenas, 1 made so far (7 ms)
[hos host] vk: memory heaps: heap 0 (system): 12 blocks 592 MB, 37 allocations 540 MB; used 756 of 15579 MB; heap 1 (device): 83 blocks 1136 MB, 1244 allocations 293 MB; used 2744 of 11474 MB;
[hos host] vk: image blocks: 2 blocks 64 MB, 2 images 63 MB; 1 free ranges, the largest 0 MB
[hos host] vk: image blocks (render targets): 3 blocks 96 MB, 8 images 73 MB; 3 free ranges, the largest 17 MB
[hos host] vk: image blocks (textures, compacted): 24 blocks 768 MB, 1 images 7 MB; 24 free ranges, the largest 32 MB
[hos host] vk: render target images: 0 reused from the pool, 8 made new, 0 destroyed from it, 0 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 0 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 8 with the pool empty (0 of them with targets still in the dead queue), 0 with none in it a present yet, 0 with none of a kind they can use, 0 with all smaller (the largest 0% short on average), 0 with all bigger (over 1.5 times), 0 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 0 idle
[hos host] vk: VRAM: 2744 MB used of 11474 MB allowed (842 MB of it free in the blocks); textures 81 MB (budget 7158 MB), 9 textures
[hos host] vk: budget: VRAM 1136 of 11474 MB (2744 MB by the driver, 1608 MB unaccounted): blocks free 842, textures 81, mirrors 64, other images 63, other buffers 63, staging 16, shadows 4; RAM 704 MB (the process's private memory 12789 MB): capture spares 424, imports 124, captures 112, retired imports 20, capture table 20, other buffers 4, others 0
[hos host] vk: textures: decoded ASTC 0 MB, rendered 81 MB (1 drawn in one frame, kept: 128 KB), other 0 MB; used in the last 8 presents 81 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 0 MB of blocks hashed in 0 ms
[hos host] vk: 14 MB: 0x1004137000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 2 presents ago
[hos host] vk: 14 MB: 0x1005f37000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 3 presents ago
[hos host] vk: 14 MB: 0x1001f0ffe00 2560x1440x1 x1 levels 1 NVN 0x3f rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x10000348400 2560x1440x1 x1 levels 1 NVN 0x38 rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x1005037000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago
[hos host] vk: 7 MB: 0x10755b4000 1920x1080x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago
[hos host] vk: 1 MB: 0x10000000000 1920x1080x1 x1 levels 1 NVN 0x1 rendered, last used 1 presents ago
[hos host] vk: 0 MB: 0x1000021c200 960x540x1 x1 levels 1 NVN 0xd rendered, last used 1 presents ago
[hos Presentation Thread] vk: presents after the CPU's signal (600): 595 waited for it, 3.07 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos host] vk: status: 600 frames, the slowest 40.0 ms; VRAM 2.7 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 0 program records loaded (0.0 ms), 0 new graphics pipelines, pipelines from the driver's cache 0, from the start's cache 0, compiled 0; 0 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 3379 queued, 3379 loaded (median 1.15 ms, 99th percentile 17.66 ms, worst 34.27 ms, 8477 ms and 202.2 MB read in all); 4278 programs (583 of them compute variants without their code, 0 graphics programs no note names), 7257 graphics pipelines (0 not in a cache), 365 compute pipelines (19 not in a cache); a first use ran 0 itself and waited for 0 (0.0 ms); 3 pipelines made at create taken by draws; 1564 create calls with nothing kept for their stages, 0 before the index; units 4943 (3379 ready, 0 loading), 9293 programs alive; 0 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 7268 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 757 known, 0 learned in this run; 365 compute pipelines made by them at a create call, 19 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 0.85, present 0.03, draw 0.02, descriptors 0.00, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.01, dispatch 0.04, memory sync 0.45, queue submit 0.00, slot wait 0.15
[hos host] vk: a frame: mirrors 8 (64 MB): 1.73 MB synced in 4.7 copies, 0.00 MB written back
[hos host] vk: 1200 presents, 164.8 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 8991 pipelines (2077 ms), 0 programs (10 failed), 9 textures (81 MB)
[hos host] vk: host time in GPU resource waits: 6.62 ms a frame (fences 3.06, timelines 3.56, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 29.1 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.8 ms a frame (the longest submission 0.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 6.11 ms a frame): its jobs 6.10 ms a frame (submits 0.93, presents 4.95, others 0.22), idle 0.01; the game waited for room at 593 presents, 1.90 ms each (1.88 a frame; 11.82 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 597 captures at a CPU signal; the submits' waits read in 0.002 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.97 ms after its hand-over (44.30 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 5.20 + 0.12 (the window's copy is in the idle), after it 0.05 + 0.00; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 1.00 parts a frame recorded in the signal's job, 0.8% of them took nothing that came with the signal; of the others, 0.99 a frame took pages of the signal's capture into a mirror (62.9 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 5 of 597 frames, its first three in 5
[hos host] vk: frame ahead, what came with the signal: 597 signals with a capture, each of 78.5 pages in 2.8 pools; between a frame's last submit and that signal 0.37 texture and sampler registrations, 0.85 releases and 0.00 other CPU signals a signal (112 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.05+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos sead::MainThread] vk: capture: a submit 0.09 ms (lock 0.00, scan 0.03 over 1661 ranges, 1392 MB), 142 pages written, 142 copied (0.6 MB)
[hos host] vk: sparse mirrors: 4, 64 MB with memory of 80 MB (15 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 13 binds (1.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 4 spares now, 44 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4093 dirty pages from their submits' captures, 28849 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2372 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 0 texture pages copied (0 MB)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #5000: event 0x100d6c8000 mode 0 value 1249 (now 1247) (lr 0x806661f8)
[hos AudioFrameworkThread] audio: voice 6: 48000 Hz, 1 channels, format 6, parameter 0x1089637500 (32 bytes) at 106898.3 s
[hos AudioFrameworkThread] audio: voice 7: 48000 Hz, 1 channels, format 6, parameter 0x1089637740 (32 bytes) at 106898.3 s
[hos host] vk: ASTC not sampled by this GPU: blocks are decoded on upload
[hos host] vk: texture 0x7fcd55d200 1664x1280x1 x1 levels 1 NVN 0x87 -> vk 37 (ASTC decoded) (10 textures, 89 MB)
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 3, mip filter nearest, LOD bias -0.5; made with 0, an image of one level: not the row's)
[hos host] vk: texture 0x7fcd17d200 1464x640x1 x1 levels 1 NVN 0x87 -> vk 37 (ASTC decoded) (11 textures, 93 MB)
[hos host] vk: texture 0x7fcbe9b200 683x683x1 x1 levels 1 NVN 0x49 -> vk 139 (12 textures, 93 MB)
[hos host] vk: texture 0x7fdc8a7000 832x640x1 x1 levels 1 NVN 0x87 -> vk 37 (ASTC decoded) (13 textures, 95 MB)
[hos host] vk: texture 0x7fdccb7000 832x640x1 x1 levels 1 NVN 0x87 -> vk 37 (ASTC decoded) (14 textures, 97 MB)
[hos host] vk: texture 0x1001f0e1c00 240x94x1 x1 levels 1 NVN 0x25 -> vk 37 (15 textures, 97 MB)
[hos host] vk: texture 0x1001f0d4a00 198x40x1 x1 levels 1 NVN 0x25 -> vk 37 (16 textures, 97 MB)
[hos host] vk: texture 0x100c922a00 4x4x1 x1 levels 1 NVN 0x25 -> vk 37 (17 textures, 97 MB)
[hos host] vk: texture 0x107cd71600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (18 textures, 97 MB)
[hos host] vk: page table of guest memory at 0x39fc00000 (8 pools)
[hos host] vk: texture #19 at 0x10000000000 beside #6: 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture 0x10000000000 262x104x1 x1 levels 1 NVN 0x1 -> vk 9 (19 textures, 97 MB)
[hos host] vk: texture 0x1001f0ca800 262x104x1 x1 levels 1 NVN 0x1 -> vk 9 (20 textures, 97 MB)
[hos host] vk: texture 0x1001f0bd600 198x40x1 x1 levels 1 NVN 0x25 -> vk 37 (21 textures, 97 MB)
[hos host] vk: texture 0x107cd74600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (22 textures, 97 MB)
[hos host] vk: texture 0x1001f0b3400 262x104x1 x1 levels 1 NVN 0x1 -> vk 9 (23 textures, 97 MB)
[hos host] vk: texture 0x1001f0a6200 198x40x1 x1 levels 1 NVN 0x25 -> vk 37 (24 textures, 97 MB)
[hos host] vk: texture 0x107cd77600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (25 textures, 97 MB)
[hos host] vk: texture 0x1001f09c000 262x104x1 x1 levels 1 NVN 0x1 -> vk 9 (26 textures, 97 MB)
[hos host] vk: texture 0x1001f077e00 285x94x1 x1 levels 1 NVN 0x25 -> vk 37 (27 textures, 97 MB)
[hos host] vk: texture 0x1001f067c00 242x40x1 x1 levels 1 NVN 0x25 -> vk 37 (28 textures, 97 MB)
[hos host] vk: texture 0x107cd7d600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (29 textures, 97 MB)
[hos host] vk: texture #30 at 0x10000000000 beside #6: 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #30 at 0x10000000000 beside #19: 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture 0x10000000000 306x104x1 x1 levels 1 NVN 0x1 -> vk 9 (30 textures, 97 MB)
[hos host] vk: texture 0x1001f05da00 306x104x1 x1 levels 1 NVN 0x1 -> vk 9 (31 textures, 98 MB)
[hos host] vk: texture 0x1001f04d800 242x40x1 x1 levels 1 NVN 0x25 -> vk 37 (32 textures, 98 MB)
[hos host] vk: texture 0x107cd80600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (33 textures, 98 MB)
[hos host] vk: texture 0x1001f043600 306x104x1 x1 levels 1 NVN 0x1 -> vk 9 (34 textures, 98 MB)
[hos host] vk: texture 0x1001f033400 242x40x1 x1 levels 1 NVN 0x25 -> vk 37 (35 textures, 98 MB)
[hos host] vk: texture 0x107cd83600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (36 textures, 98 MB)
[hos host] vk: texture 0x1001f029200 306x104x1 x1 levels 1 NVN 0x1 -> vk 9 (37 textures, 98 MB)
[hos host] vk: texture 0x1001efd9000 631x94x1 x1 levels 1 NVN 0x25 -> vk 37 (38 textures, 98 MB)
[hos host] vk: texture 0x1001efb3e00 592x49x1 x1 levels 1 NVN 0x25 -> vk 37 (39 textures, 98 MB)
[hos host] vk: texture 0x107cd89600 640x46x1 x1 levels 1 NVN 0x1 -> vk 9 (40 textures, 98 MB)
[hos host] vk: texture #41 at 0x10000000000 beside #6: 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #41 at 0x10000000000 beside #19: 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #41 at 0x10000000000 beside #30: 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2681 dirty pages from their submits' captures, 29837 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 238 dispatches, 2358 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.09 ms (lock 0.00, scan 0.03 over 1662 ranges, 1624 MB), 141 pages written, 141 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (110 MB), 12 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 220 pages in 220 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos nn.npln.Worker] mutex 0x100ac110c8 (obj 000002DB48F189C0): nn.npln.Worker waiting 10 s; held by nn.npln.Worker (count 2, 1 waiting)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.00, present 0.05, draw 0.11, descriptors 0.04, texture lookups 0.03, uploads 0.02, pipelines 0.00, programs 0.02, commands 0.04, dispatch 0.05, memory sync 0.44, queue submit 0.00, slot wait 0.14
[hos host] vk: a frame: mirrors 10 (76 MB): 1.86 MB synced in 5.2 copies, 0.00 MB written back
[hos host] vk: 1320 presents, 163.7 fps; a frame: 11 draws (12 skipped), 2 dispatches, 3 passes, 15 barriers, 6 submits, 0.4 uploads (0.1 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 8991 pipelines (2077 ms), 0 programs (10 failed), 59 textures (99 MB)
[hos host] vk: host time in GPU resource waits: 6.26 ms a frame (fences 2.80, timelines 3.47, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 29.8 full barriers recorded, 30.7 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.3; 0x10: 0.3; 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.6, an op 15.3, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 1.0 ms a frame (the longest submission 2.7 ms); 0.4 textures made a frame: colour targets 0.3 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.01 ms
[hos Presentation Thread] vi: 1323 frames presented (163.8 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 818, 15.5-17.8 0, 17.8-22 0, 22-30 1, 30-40 0, over 40 1
[hos host] vk: late pages' fix (600 recordings with one): of 705 captures with late pages 705 fixed (0.9 pages, 0.9 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 0 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 64 blocks (4096 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 0, of an earlier recording 0, not recorded yet 0, of this part 0; pages fixed from stamps 600, from snapshots 0; the shadows: 605 pages synced from captures, 55 made known by it, 550 left unknown because the replay or the collection wrote them since
[hos AudioFrameworkThread] audio: voice 8: 48000 Hz, 1 channels, format 6, parameter 0x1089637980 (32 bytes) at 106898.7 s
[hos AudioFrameworkThread] audio: voice 9: 48000 Hz, 1 channels, format 6, parameter 0x1089637bc0 (32 bytes) at 106898.7 s
[hos host] vk: sparse mirrors: 6, 76 MB with memory of 100 MB (23 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 15 binds (2.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 2 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 2 spares now, 24 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 5610 dirty pages from their submits' captures, 29550 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 300 dispatches, 2578 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.03 over 1662 ranges, 1624 MB), 146 pages written, 146 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (112 MB), 10 spare buffers, 0 texture pages copied (0 MB)
[hos ModuleSystemWorker1] prompts: the font at 0x7fd37800e8: the icons in texts follow at once
[hos host] vk: texture storage read as submitted (1 times; 2 pages from the copies, 0 of them rewritten in guest memory since their submit; 6 pages kept, 6 spans, 317 MB)
[hos host] vk: texture #65 at 0x10000000000 beside #6: 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #65 at 0x10000000000 beside #19: 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #65 at 0x10000000000 beside #30: 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #65 at 0x10000000000 beside #41: 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #69 at 0x10000000000 beside #6: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #69 at 0x10000000000 beside #19: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #69 at 0x10000000000 beside #30: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #69 at 0x10000000000 beside #41: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #69 at 0x10000000000 beside #65: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #73 at 0x10000000000 beside #6: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #73 at 0x10000000000 beside #19: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #73 at 0x10000000000 beside #30: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #73 at 0x10000000000 beside #41: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #73 at 0x10000000000 beside #65: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #73 at 0x10000000000 beside #69: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #75 at 0x10000000000 beside #6: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #75 at 0x10000000000 beside #19: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #75 at 0x10000000000 beside #30: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #75 at 0x10000000000 beside #41: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #75 at 0x10000000000 beside #65: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #75 at 0x10000000000 beside #69: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #75 at 0x10000000000 beside #73: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] [debug] vk: a staging arena of 16 MB took 9.2 ms to make
[hos host] vk: staging arenas: 2 (16 MB in this one)
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 3, mip filter nearest, LOD bias 0; made with 0, an image of one level: not the row's)
[hos host] vk: import 3.9 ms: commit+properties 0.0, buffer 0.0, allocate 3.9, bind+address 0.0
[hos host] vk: pool 0x102e645400+0x20 imported in 3.9 ms as 0x102e400000..0x102f800000 (10 imports, 34 ms in all)
[hos host] vk: sampler 318 is nearest: min filter 2, mag filter 0, wrap 7/7/7, lod -1000..1000 bias -0.5, aniso 1, compare 0
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 2, mip filter nearest, LOD bias -0.5; made with 0, an image of one level: not the row's)
[hos host] vk: slow submit: 21.8 ms: draw 1.8 texture lookups 16.3 uploads 15.9
[hos host] vk: 3 textures retired (120 left, 139 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: images are destroyed on a thread of their own (RX_VK_REAPER=0: on the render thread)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 5946 dirty pages from their submits' captures, 30445 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 343 dispatches, 2758 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a frame: 5 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.31, present 0.05, draw 0.26, descriptors 0.12, texture lookups 0.18, uploads 0.16, pipelines 0.01, programs 0.03, commands 0.07, dispatch 0.06, memory sync 0.49, queue submit 0.00, slot wait 0.14
[hos host] vk: a frame: mirrors 11 (94 MB): 2.05 MB synced in 7.1 copies, 0.00 MB written back
[hos host] vk: 1440 presents, 160.8 fps; a frame: 49 draws (12 skipped), 2 dispatches, 9 passes, 22 barriers, 6 submits, 0.5 uploads (0.2 MB), 10 ub updates (6 overlays), GPU-bound wait 0.0 ms (recording slots only); 9026 pipelines (2081 ms), 0 programs (10 failed), 124 textures (140 MB)
[hos host] vk: host time in GPU resource waits: 5.69 ms a frame (fences 2.29, timelines 3.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 41.4 full barriers recorded, 38.1 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 14.0; 0x10: 1.5; 0x12: 6.5;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 21.1, an op 19.6, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 1.5 ms a frame (the longest submission 3.9 ms); 0.6 textures made a frame: colour targets 0.4 (0.2 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.15 MB unswizzled on the CPU in 0.08 ms
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 150 pages written, 150 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (146 MB), 10 spare buffers, 64 texture pages copied (0 MB)
[hos AudioFrameworkThread] audio: voice 10: 48000 Hz, 1 channels, format 6, parameter 0x1089637e00 (32 bytes) at 106899.6 s
[hos AudioFrameworkThread] audio: voice 11: 48000 Hz, 1 channels, format 6, parameter 0x1089638040 (32 bytes) at 106899.6 s
[hos ModuleSystemWorker4] [debug] nvn: 1 copies of recordings whose control memory was reused let go (20830 live)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 64 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 199 pages in 199 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.08
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 57 KB; 0 made on demand (0 MB kept now); c1 stages: at most 3 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 7, 94 MB with memory of 120 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 16 binds (2.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6473 dirty pages from their submits' captures, 29758 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2747 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 74 in captures, 0 only in the collection, 1 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: queues ordered by their events: 6000 submissions stopped at a wait, 5997 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 6024, not met 0; captures at a CPU event's wait 1482, 0 of them at a wait met on arrival
[hos host] vk: 3 textures retired (122 left, 142 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: texture #129 at 0x107cdcc600 beside #108: 256x32x1 x1 levels 1 layout 0x80000100 class 0x40001 NVN 0x1 target 1, the other 256x39x1 x1 levels 1 layout 0x80000100 class 0x40001 NVN 0x1 target 1
[hos host] vk: texture #131 at 0x107cdb9600 beside #104: 512x46x1 x1 levels 1 layout 0x80000200 class 0x40001 NVN 0x1 target 1, the other 256x39x1 x1 levels 1 layout 0x80000100 class 0x40001 NVN 0x1 target 1
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 148 pages written, 148 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (142 MB), 9 spare buffers, 64 texture pages copied (0 MB)
[hos host] vk: a frame: 9 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.22, present 0.05, draw 0.25, descriptors 0.11, texture lookups 0.02, uploads 0.00, pipelines 0.01, programs 0.03, commands 0.07, dispatch 0.07, memory sync 0.48, queue submit 0.00, slot wait 0.13
[hos host] vk: a frame: mirrors 11 (94 MB): 1.82 MB synced in 8.0 copies, 0.00 MB written back
[hos host] vk: 1560 presents, 164.9 fps; a frame: 68 draws (12 skipped), 2 dispatches, 16 passes, 31 barriers, 6 submits, 0.1 uploads (0.0 MB), 21 ub updates (11 overlays), GPU-bound wait 0.0 ms (recording slots only); 9034 pipelines (2082 ms), 0 programs (10 failed), 131 textures (142 MB)
[hos host] vk: host time in GPU resource waits: 5.13 ms a frame (fences 2.00, timelines 3.13, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 55.9 full barriers recorded, 47.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 15.0; 0x10: 3.0; 0x12: 13.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 30.0, an op 24.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 1.9 ms a frame (the longest submission 2.2 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6597 dirty pages from their submits' captures, 29211 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 378 dispatches, 2770 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 143 pages written, 143 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (100 MB), 10 spare buffers, 64 texture pages copied (0 MB)
[hos host] vk: sparse mirrors: 7, 94 MB with memory of 120 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 16 binds (2.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6606 dirty pages from their submits' captures, 27265 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2632 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 27 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 217 pages in 217 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.04, present 0.04, draw 0.17, descriptors 0.08, texture lookups 0.01, uploads 0.00, pipelines 0.01, programs 0.02, commands 0.04, dispatch 0.05, memory sync 0.45, queue submit 0.00, slot wait 0.14
[hos host] vk: a frame: mirrors 11 (94 MB): 1.81 MB synced in 7.7 copies, 0.00 MB written back
[hos host] vk: 1680 presents, 165.4 fps; a frame: 47 draws (12 skipped), 2 dispatches, 16 passes, 31 barriers, 6 submits, 0.0 uploads (0.0 MB), 21 ub updates (11 overlays), GPU-bound wait 0.0 ms (recording slots only); 9034 pipelines (2082 ms), 0 programs (10 failed), 134 textures (143 MB)
[hos host] vk: host time in GPU resource waits: 5.61 ms a frame (fences 2.19, timelines 3.42, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 56.1 full barriers recorded, 47.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 15.0; 0x10: 3.0; 0x12: 13.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 30.0, an op 25.7, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 1.7 ms a frame (the longest submission 1.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos sead::MainThread] barrier: 18848 looks: 2464 found the masks moved already, 16384 waited until they moved, 0 for the whole 1 ms (a wait 77 us on average); 19426 wakes
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 147 pages written, 147 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (88 MB), 9 spare buffers, 64 texture pages copied (0 MB)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6153 dirty pages from their submits' captures, 28364 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2672 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: presenter thread: the copy's fence 2.32 ms, the present 0.56 ms a present
[hos host] vk: memory: working set 10435 MB (peak 10435), private 12926 MB, staging 32 MB in 2 arenas, 2 made so far (16 ms)
[hos host] vk: memory heaps: heap 0 (system): 12 blocks 592 MB, 36 allocations 532 MB; used 776 of 15579 MB; heap 1 (device): 85 blocks 1224 MB, 1849 allocations 433 MB; used 2833 of 11474 MB;
[hos host] vk: image blocks: 2 blocks 64 MB, 2 images 63 MB; 1 free ranges, the largest 0 MB
[hos host] vk: image blocks (render targets): 3 blocks 96 MB, 12 images 84 MB; 3 free ranges, the largest 10 MB
[hos host] vk: image blocks (textures, compacted): 24 blocks 768 MB, 122 images 59 MB; 35 free ranges, the largest 32 MB
[hos host] vk: render target images: 0 reused from the pool, 12 made new, 0 destroyed from it, 0 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 0 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 12 with the pool empty (0 of them with targets still in the dead queue), 0 with none in it a present yet, 0 with none of a kind they can use, 0 with all smaller (the largest 0% short on average), 0 with all bigger (over 1.5 times), 0 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 0 idle
[hos host] vk: VRAM: 2833 MB used of 11474 MB allowed (790 MB of it free in the blocks); textures 143 MB (budget 7158 MB), 134 textures
[hos host] vk: budget: VRAM 1224 of 11474 MB (2833 MB by the driver, 1609 MB unaccounted): blocks free 790, textures 143, other buffers 96, mirrors 94, other images 63, staging 32, shadows 4; RAM 716 MB (the process's private memory 12926 MB): capture spares 396, imports 144, captures 102, other buffers 34, retired imports 20, capture table 20, others 0
[hos host] vk: textures: decoded ASTC 30 MB, rendered 106 MB (39 drawn in one frame, kept: 26702 KB), other 6 MB; used in the last 8 presents 101 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 10 MB of blocks hashed in 1 ms
[hos host] vk: 14 MB: 0x1004137000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 2 presents ago
[hos host] vk: 14 MB: 0x1001f0ffe00 2560x1440x1 x1 levels 1 NVN 0x3f rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x10000348400 2560x1440x1 x1 levels 1 NVN 0x38 rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x1005f37000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 3 presents ago
[hos host] vk: 14 MB: 0x1005037000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago
[hos host] vk: 8 MB: 0x7fcd55d200 1664x1280x1 x1 levels 1 NVN 0x87 ASTC, last used 417 presents ago
[hos host] vk: 7 MB: 0x10755b4000 1920x1080x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago
[hos host] vk: 7 MB: 0x7fc9cd8400 1440x1440x1 x1 levels 1 NVN 0x87 ASTC, last used 280 presents ago
[hos Presentation Thread] vk: presents after the CPU's signal (600): 599 waited for it, 3.49 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos host] vk: status: 600 frames, the slowest 27.6 ms; VRAM 2.8 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 0 program records loaded (0.0 ms), 0 new graphics pipelines, pipelines from the driver's cache 0, from the start's cache 0, compiled 0; 0 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 3398 queued, 3398 loaded (median 1.15 ms, 99th percentile 17.66 ms, worst 34.27 ms, 8522 ms and 203.4 MB read in all); 4300 programs (583 of them compute variants without their code, 0 graphics programs no note names), 7300 graphics pipelines (0 not in a cache), 365 compute pipelines (19 not in a cache); a first use ran 0 itself and waited for 0 (0.0 ms); 36 pipelines made at create taken by draws; 1564 create calls with nothing kept for their stages, 0 before the index; units 4962 (3398 ready, 0 loading), 9312 programs alive; 0 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 7312 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 757 known, 0 learned in this run; 365 compute pipelines made by them at a create call, 19 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.03, present 0.05, draw 0.16, descriptors 0.07, texture lookups 0.01, uploads 0.00, pipelines 0.01, programs 0.02, commands 0.04, dispatch 0.05, memory sync 0.46, queue submit 0.00, slot wait 0.14
[hos host] vk: a frame: mirrors 11 (94 MB): 1.81 MB synced in 7.7 copies, 0.00 MB written back
[hos host] vk: 1800 presents, 164.8 fps; a frame: 46 draws (12 skipped), 2 dispatches, 16 passes, 31 barriers, 6 submits, 0.0 uploads (0.0 MB), 21 ub updates (11 overlays), GPU-bound wait 0.0 ms (recording slots only); 9034 pipelines (2082 ms), 0 programs (10 failed), 134 textures (143 MB)
[hos host] vk: host time in GPU resource waits: 5.60 ms a frame (fences 2.15, timelines 3.46, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 56.0 full barriers recorded, 47.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 15.0; 0x10: 3.0; 0x12: 13.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 30.0, an op 25.8, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 1.7 ms a frame (the longest submission 1.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 6.10 ms a frame): its jobs 6.08 ms a frame (submits 0.78, presents 4.84, others 0.46), idle 0.02; the game waited for room at 595 presents, 1.63 ms each (1.62 a frame; 14.26 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 600 captures at a CPU signal; the submits' waits read in 0.002 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.69 ms after its hand-over (31.16 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.36 + 0.16 (the window's copy is in the idle), after it 0.03 + 0.00; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 1.00 parts a frame recorded in the signal's job, 0.2% of them took nothing that came with the signal; of the others, 1.00 a frame took pages of the signal's capture into a mirror (77.6 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 1 of 600 frames, its first three in 1
[hos host] vk: frame ahead, what came with the signal: 600 signals with a capture, each of 80.0 pages in 3.0 pools; between a frame's last submit and that signal 0.04 texture and sampler registrations, 0.01 releases and 0.00 other CPU signals a signal (16 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.03+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos sead::MainThread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 146 pages written, 146 copied (0.6 MB)
[hos host] vk: sparse mirrors: 7, 94 MB with memory of 120 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 16 binds (2.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6490 dirty pages from their submits' captures, 28901 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2728 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (110 MB), 9 spare buffers, 64 texture pages copied (0 MB)
[hos host] vk: late pages' fix (600 recordings with one): of 657 captures with late pages 657 fixed (0.9 pages, 0.9 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 0 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 64 blocks (4096 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 0, of an earlier recording 0, not recorded yet 0, of this part 0; pages fixed from stamps 600, from snapshots 0; the shadows: 493 pages synced from captures, 0 made known by it, 493 left unknown because the replay or the collection wrote them since
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 221 pages in 221 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] nvn: GPU waits in 5 s: 813 sync waits (2746 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 57 KB; 0 made on demand (0 MB kept now); c1 stages: at most 4 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6491 dirty pages from their submits' captures, 26506 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2576 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 142 pages written, 142 copied (0.6 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (124 MB), 9 spare buffers, 64 texture pages copied (0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos BaseProcExecutor1] vcores: 186020 takes, 20174 waits (mean 23 us, longest 5.2 ms, 0 over HOS_VCORE_LOG_MS), 8608 kicks, 8371 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.06, present 0.05, draw 0.19, descriptors 0.08, texture lookups 0.02, uploads 0.00, pipelines 0.01, programs 0.02, commands 0.06, dispatch 0.05, memory sync 0.47, queue submit 0.00, slot wait 0.13
[hos host] vk: a frame: mirrors 11 (94 MB): 1.73 MB synced in 7.7 copies, 0.00 MB written back
[hos host] vk: 1920 presents, 165.0 fps; a frame: 46 draws (12 skipped), 2 dispatches, 16 passes, 31 barriers, 6 submits, 0.0 uploads (0.0 MB), 21 ub updates (11 overlays), GPU-bound wait 0.0 ms (recording slots only); 9034 pipelines (2082 ms), 0 programs (10 failed), 134 textures (143 MB)
[hos host] vk: host time in GPU resource waits: 5.61 ms a frame (fences 2.15, timelines 3.46, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0
[hos host] vk: a frame: 56.2 full barriers recorded, 47.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 15.0; 0x10: 3.0; 0x12: 13.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 30.0, an op 24.5, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 1.7 ms a frame (the longest submission 2.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: sparse mirrors: 7, 94 MB with memory of 120 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 16 binds (2.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 9298 dirty pages from their submits' captures, 24574 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 376 dispatches, 2633 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[a64] unimplemented import ZN2nn5movie8FinalizeEv (lr 0x80df26c0)
[hos MovieOperateThread] video: a movie of 690 pictures on dav1d: 1.6 ms to decode each (worst 41.1), one shown every 16.7 ms; 15 skipped to keep up with its sound; 422 submissions put off while 16 waited
[hos MovieOperateThread] video: the longest wait between two of its pictures 289 ms
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 1665 ranges, 1627 MB), 139 pages written, 139 copied (0.5 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (96 MB), 11 spare buffers, 64 texture pages copied (0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.09
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6717 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 207 pages in 207 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vi: the movie is over: back to the game's frame rate cap 60
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1619 dirty pages from their submits' captures, 5244 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 251 dispatches, 738 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a frame: 1 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.08, present 0.06, draw 0.13, descriptors 0.05, texture lookups 0.01, uploads 0.00, pipelines 0.01, programs 0.01, commands 0.05, dispatch 0.03, memory sync 0.56, queue submit 0.00, slot wait 0.16
[hos host] vk: a frame: mirrors 11 (94 MB): 0.72 MB synced in 5.4 copies, 0.00 MB written back
[hos host] vk: 2040 presents, 115.0 fps; a frame: 20 draws (5 skipped), 1 dispatches, 7 passes, 12 barriers, 6 submits, 0.0 uploads (0.0 MB), 8 ub updates (4 overlays), GPU-bound wait 0.0 ms (recording slots only); 9173 pipelines (2095 ms), 0 programs (10 failed), 138 textures (143 MB)
[hos host] vk: host time in GPU resource waits: 5.26 ms a frame (fences 3.02, timelines 2.24, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: skips a frame by vk_draw.cpp line: 11682: 4.7
[hos host] vk: a frame: 32.9 full barriers recorded, 25.8 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.7; 0x10: 1.0; 0x12: 4.9;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 10.6, an op 9.9, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 1.1 ms a frame (the longest submission 1.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Presentation Thread] vi: 2060 frames presented (147.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 685, 15.5-17.8 50, 17.8-22 1, 22-30 0, 30-40 1, over 40 0
[hos sead::MainThread] vk: capture: a submit 0.08 ms (lock 0.00, scan 0.04 over 1672 ranges, 1626 MB), 92 pages written, 92 copied (0.4 MB)
[hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (108 MB), 8 spare buffers, 3531 texture pages copied (13 MB)
[hos host] vk: 1 textures retired (137 left, 139 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.00, writes collected 0.05
[hos host] vk: sparse mirrors: 7, 36 MB with memory of 120 MB (83 MB spared); blocks given memory later 37, given back 930; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 17 binds (2.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1 dirty pages from their submits' captures, 303 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 228 dispatches, 304 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 77 runs, 0.0 MB compared; 77 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 77 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9807 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Resource Memory] mem: the game's mapped pages 2580 MB (peak 2580); heap 6894 MB; the console's total 9809 MB
[hos bake::ModuleMgr] vcores: bake::ModuleMgr (priority 28) waited 54.1 ms for a core: given by aglTextureDataMgr::Compaction (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos GeneralFormatAsync
] vcores: GeneralFormatAsync
(priority 24) waited 64.1 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos BancModule] vcores: BancModule (priority 24) waited 110.6 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos GeneralFormatAsync
] vcores: GeneralFormatAsync
(priority 24) waited 133.7 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 1702 ranges, 1672 MB), 127 pages written, 127 copied (0.5 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (116 MB), 10 spare buffers, 6391 texture pages copied (24 MB)
[hos BancModule] vcores: BancModule (priority 24) waited 183.0 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos GeneralFormatAsync
] vcores: GeneralFormatAsync
(priority 24) waited 124.9 ms for a core: given by nn.npln.Callback (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos BancModule] vcores: BancModule (priority 24) waited 50.6 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 0.94, present 0.06, draw 0.07, descriptors 0.02, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.01, commands 0.03, dispatch 0.00, memory sync 0.47, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 11 (30 MB): 0.02 MB synced in 4.1 copies, 0.00 MB written back
[hos host] vk: 2160 presents, 60.0 fps; a frame: 5 draws (0 skipped), 0 dispatches, 2 passes, 1 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 9410 pipelines (2131 ms), 0 programs (10 failed), 137 textures (139 MB)
[hos host] vk: host time in GPU resource waits: 4.16 ms a frame (fences 4.16, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 18.9 full barriers recorded, 13.8 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 0.0, an op 1.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos GeneralFormatAsync
] vcores: GeneralFormatAsync
(priority 24) waited 281.0 ms for a core: given by BaseProcCreateAndDeleteThread (priority 23), which held it 4.5 ms and was never asked; the cores it may use now: core 1: it;
[hos BancModule] vcores: BancModule (priority 24) waited 256.0 ms for a core: given by BaseProcCreateAndDeleteThread (priority 23), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 634.5 ms for a core: given by BancModule (priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 578.7 ms for a core: given by tera::FileResourceMgr (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos bake::ModuleMgr] vcores: bake::ModuleMgr (priority 28) waited 603.1 ms for a core: given by PreActorModelCompose (priority 26), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos BaseProcCreateAndDeleteThread] nvn: 100000 recordings made, 21028 live copies (1.9 MB), 53 texture copies; 2 command buffers' segments (0.2 MB)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 5 dirty pages from their submits' captures, 325 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 231 dispatches, 330 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 76 runs, 0.0 MB compared; 76 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 93 in captures, 1 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.10
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 1495 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.02, scan 0.09 over 1710 ranges, 1618 MB), 97 pages written, 97 copied (0.4 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 6401 texture pages copied (25 MB)
[hos host] vk: recorder: 348216 packets run in 10.0 s (busy 0.62 s, spun idle 0.12 s, slept 11151 times, 5850 of them under 200 us), 0 calls made directly (no room); 19639 batches over 1232 frames (15.9 a frame, 18 packets a batch), 9816 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 7 times, 0.86 ms (0.0 times, 0.00 ms a frame); every thread 650 times, 26.52 ms in all, the longest 0.97 ms; the queue's lock 365/13.60 ms (max 0.97), vkWaitForFences 267/11.80 ms (max 0.35), vkQueueBindSparse 5/0.84 ms (max 0.57), vkDestroyImageView 3/0.12 ms (max 0.08), vmaDestroyBuffer 6/0.11 ms (max 0.03), vmaFreeMemory 2/0.03 ms (max 0.02)
[hos host] vk: recorder: of its busy 0.62 s, 0.43 s in parts' ends and submits (the longest 1.01 ms), 0.03 s of that blocked on the queue's lock (64 times, the longest 0.96 ms), so 0.40 s in the ends, submits and holds; 0.19 s running the other packets; 283 packets a frame
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 57 KB; 0 made on demand (0 MB kept now); c1 stages: at most 4 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 7, 30 MB with memory of 120 MB (89 MB spared); blocks given memory later 37, given back 1026; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 18 binds (2.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 7 dirty pages from their submits' captures, 346 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 229 dispatches, 353 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 78 runs, 0.0 MB compared; 78 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 119 in captures, 0 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB)
[hos Resource Control] vcores: 180481 takes, 67693 waits (mean 116 us, longest 634.5 ms, 12 over HOS_VCORE_LOG_MS), 30073 kicks, 28790 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 6746 programs from the program cache
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.53, present 0.07, draw 0.09, descriptors 0.03, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.01, commands 0.04, dispatch 0.00, memory sync 1.02, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 11 (30 MB): 0.02 MB synced in 4.6 copies, 0.00 MB written back
[hos host] vk: 2280 presents, 60.0 fps; a frame: 5 draws (0 skipped), 0 dispatches, 2 passes, 1 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 9972 pipelines (2219 ms), 0 programs (10 failed), 137 textures (139 MB)
[hos host] vk: host time in GPU resource waits: 4.17 ms a frame (fences 4.17, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 18.7 full barriers recorded, 13.7 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 0.0, an op 1.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 0.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Resource Control] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 31 made): 10195546 lookups, 134132 missed (1.32 %)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.14
[hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.02, scan 0.10 over 1866 ranges, 1622 MB), 108 pages written, 108 copied (0.4 MB)
[hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 6515 texture pages copied (25 MB)
[hos ModuleSystemWorker3] shadow: the sun's nearest cascade keeps the game's fit below Shadow resolution Highest (1600 texels)
[hos sead::MainThread] nvn: the game reports counter type 1 (lr 0x800876dc)
[hos host] vk: texture 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e -> vk 137 (ASTC decoded) (139 textures, 357 MB)
[hos host] [debug] vk: a staging arena of 57 MB took 8.5 ms to make
[hos host] vk: staging arenas: 3 (57 MB in this one)
[hos host] vk: vertex format NVN 0xa -> vk 100
[hos host] vk: slow upload 179.1 ms: #146 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC (BC3), 57825 of 57825 KB
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 5, mip filter linear, LOD bias 0; made with 0)
[hos host] vk: texture #149 at 0x10000000000 beside #6: 256x144x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: import 2.8 ms: commit+properties 0.1, buffer 0.0, allocate 2.7, bind+address 0.0
[hos host] vk: pool 0x113b681b00+0x90 imported in 2.8 ms as 0x113b400000..0x113cc00000 (11 imports, 37 ms in all)
[hos host] vk: sampler 409 is nearest: min filter 2, mag filter 0, wrap 7/7/7, lod 0..15 bias 0, aniso 1, compare 0
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 2, mip filter nearest, LOD bias 0; made with 0, an image of one level: not the row's)
[hos host] vk: mesh draw kind 0: 0 3 512, indirect 0 count 0 max 0 stride 0
[hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256]
[hos host] vk: import 3.5 ms: commit+properties 0.3, buffer 0.0, allocate 3.2, bind+address 0.0
[hos host] vk: pool 0x7fab0e5000+0x14300 imported in 3.5 ms as 0x7fab000000..0x7faf000000 (14 imports, 42 ms in all)
[hos host] vk: mesh draw kind 0: 0 3 512, indirect 0 count 0 max 0 stride 0
[hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256]
[hos host] vk: mesh draw kind 0: 0 3 512, indirect 0 count 0 max 0 stride 0
[hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256]
[hos host] vk: import 4.0 ms: commit+properties 0.2, buffer 0.0, allocate 3.8, bind+address 0.0
[hos host] vk: pool 0x104eafe800+0x60 imported in 4.0 ms as 0x104e000000..0x104f800000 (15 imports, 46 ms in all)
[hos host] vk: import 2.4 ms: commit+properties 0.1, buffer 0.0, allocate 2.3, bind+address 0.0
[hos host] vk: pool 0x7fd3c04000+0x69800 imported in 2.4 ms as 0x7fd3c00000..0x7fd4c00000 (16 imports, 49 ms in all)
[hos host] vk: mesh draw kind 0: 0 2 512, indirect 0 count 0 max 0 stride 0
[hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256]
[hos host] vk: import 4.8 ms: commit+properties 0.2, buffer 0.0, allocate 4.6, bind+address 0.0
[hos host] vk: pool 0x7fb7df5000+0x35900 imported in 4.8 ms as 0x7fb5c00000..0x7fb8800000 (17 imports, 53 ms in all)
[hos host] vk: mesh draw kind 0: 0 4 512, indirect 0 count 0 max 0 stride 0
[hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256]
[hos host] vk: mesh draw kind 0: 0 4 512, indirect 0 count 0 max 0 stride 0
[hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256]
[hos host] vk: mesh draw kind 0: 0 5 512, indirect 0 count 0 max 0 stride 0
[hos host] vk: mesh draw kind 0: 0 15 512, indirect 0 count 0 max 0 stride 0
[hos host] vk: pool 0x7fb8f04000+0x1b2800 imported in 5.9 ms as 0x7fb8800000..0x7fbbc00000 (18 imports, 59 ms in all)
[hos host] vk: pool 0x7fd1fbe200+0x76c imported in 4.2 ms as 0x7fd1800000..0x7fd2400000 (19 imports, 64 ms in all)
[hos host] vk: pool 0x7fb4943f00+0xa2ca8 imported in 2.4 ms as 0x7fb4400000..0x7fb5c00000 (20 imports, 66 ms in all)
[hos host] vk: pool 0x7fd6367b00+0x13390 imported in 3.0 ms as 0x7fd4c00000..0x7fd6400000 (21 imports, 69 ms in all)
[hos host] vk: pool 0x7fbe136600+0x7560 imported in 4.1 ms as 0x7fbd800000..0x7fbf800000 (22 imports, 73 ms in all)
[hos host] vk: pool 0x7fb23dce00+0x1ad9c imported in 3.8 ms as 0x7fb2000000..0x7fb4400000 (23 imports, 77 ms in all)
[hos host] vk: pool 0x7fb1e73900+0xb61c imported in 3.6 ms as 0x7fb0800000..0x7fb2000000 (24 imports, 80 ms in all)
[hos host] vk: pool 0x7fbcda0b00+0x8e68 imported in 3.3 ms as 0x7fbcc00000..0x7fbd800000 (25 imports, 84 ms in all)
[hos host] vk: pool 0x7fd694e300+0x150 imported in 3.8 ms as 0x7fd6400000..0x7fd9400000 (27 imports, 89 ms in all)
[hos host] vk: pool 0x7fc20d8900+0x11288 imported in 8.0 ms as 0x7fc1400000..0x7fc3000000 (28 imports, 97 ms in all)
[hos host] vk: pool 0x7fc120db00+0x1b8dc imported in 2.8 ms as 0x7fbf800000..0x7fc1400000 (29 imports, 100 ms in all)
[hos host] vk: pool 0x7faf0a7f00+0x93e4 imported in 4.1 ms as 0x7faf000000..0x7faf400000 (30 imports, 104 ms in all)
[hos host] vk: pool 0x7fbc4fd200+0x1f8a64 imported in 2.8 ms as 0x7fbc000000..0x7fbcc00000 (31 imports, 106 ms in all)
[hos host] vk: pool 0x7fd254ff00+0x76c64 imported in 4.0 ms as 0x7fd2400000..0x7fd2800000 (32 imports, 110 ms in all)
[hos host] vk: pool 0x7faefcb100+0x874c8 imported in 7.9 ms as 0x7fab000000..0x7faf400000 (33 imports, 118 ms in all)
[hos host] vk: pool 0x7faf000000 remade larger (2 retired)
[hos host] vk: pool 0x7fab000000 remade larger (3 retired)
[hos host] vk: pool 0x7fab000000..0x7faf400000 took in 2 pools
[hos host] vk: pool 0x7fab000000..0x7faf400000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: pool 0x7fdf775200+0x960 imported in 13.4 ms as 0x7fdf000000..0x7fdfc00000 (34 imports, 132 ms in all)
[hos host] vk: pool 0x7fd3babd00+0x7b5c imported in 4.0 ms as 0x7fd3800000..0x7fd3c00000 (35 imports, 136 ms in all)
[hos host] vk: sampler anisotropy: the game asks for 2 (min filter 5, mip filter linear, LOD bias -0.5; made with 0)
[hos host] vk: sampler anisotropy: the game asks for 0 (min filter 1, mip filter nearest, LOD bias 0; made with 0)
[hos host] vk: texture slot 0x10b4 (handle 0x18c010b4): no image for it; NVN target 1 format 0x8c 683x683x1 levels 10 gpu 0x10d58f2200, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc30 (handle 0x18c00c30): no image for it; NVN target 1 format 0x8b 512x512x1 levels 10 gpu 0x10c1776600, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xe5d (handle 0x18c00e5d): no image for it; NVN target 1 format 0x8b 512x512x1 levels 10 gpu 0x10cd3e7a00, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0x1001 (handle 0x18c01001): no image for it; NVN target 1 format 0x8a 341x341x1 levels 9 gpu 0x10d385da00, view yes (format 0x8a target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0x1011 (handle 0x18c01011): no image for it; NVN target 1 format 0x8b 512x512x1 levels 10 gpu 0x10d39c3600, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0x10e9 (handle 0x18c010e9): no image for it; NVN target 1 format 0x8d 171x171x1 levels 8 gpu 0x10d6b62a00, view yes (format 0x8d target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0x1060 (handle 0x18c01060): no image for it; NVN target 1 format 0x88 512x512x1 levels 10 gpu 0x10d4d6c800, view yes (format 0x88 target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc33 (handle 0x18c00c33): no image for it; NVN target 1 format 0x8d 1024x1024x1 levels 11 gpu 0x10c19fd600, view yes (format 0x8d target 1), wanted view type 1 use 0
[hos host] vk: texture #192 at 0x10000000000 beside #6: 128x128x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #192 at 0x10000000000 beside #149: 128x128x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 256x144x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: pool 0x108c6f2540+0x40 imported in 5.4 ms as 0x108c000000..0x108f000000 (36 imports, 141 ms in all)
[hos host] vk: texture slot 0xe37 (handle 0x18c00e37): no image for it; NVN target 1 format 0x8d 85x341x1 levels 9 gpu 0x10cc819400, view yes (format 0x8d target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xe38 (handle 0x18c00e38): no image for it; NVN target 1 format 0x7b 128x512x1 levels 10 gpu 0x10cc81e600, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: pool 0x10e5be9600+0x2028000 imported in 6.6 ms as 0x10e5800000..0x10e8000000 (37 imports, 148 ms in all)
[hos host] vk: texture #197 at 0x113755c000 beside #159: 4096x4096x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0xa target 1, the other 4096x4096x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: BC image 1120x1516 (vk 143) as a raw colour attachment: the driver says -11
[hos host] vk: BC image 2240x3032 (vk 141) as a raw colour attachment: the driver says -11
[hos host] vk: texture slot 0xcfd (handle 0x18c00cfd): no image for it; NVN target 1 format 0x8b 1024x1024x1 levels 11 gpu 0x10c6aaa000, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc3a (handle 0x18c00c3a): no image for it; NVN target 1 format 0x8b 1024x1024x1 levels 11 gpu 0x10c1b0b200, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: BC image 1528x920 (vk 143) as a raw colour attachment: the driver says -11
[hos host] vk: staging arenas: 4 (16 MB in this one)
[hos host] vk: BC image 3056x1840 (vk 141) as a raw colour attachment: the driver says -11
[hos host] vk: texture slot 0xc70 (handle 0x18c00c70): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c350d200, view yes (format 0x8d target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc71 (handle 0x18c00c71): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c352c000, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xafa (handle 0x18c00afa): no image for it; NVN target 1 format 0x8b 1024x1024x1 levels 11 gpu 0x10bb07d600, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xafc (handle 0x18c00afc): no image for it; NVN target 1 format 0x7f 768x768x1 levels 10 gpu 0x10bb319c00, view yes (format 0x7f target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xad6 (handle 0x18c00ad6): no image for it; NVN target 1 format 0x8b 341x341x1 levels 9 gpu 0x10ba37e000, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xad7 (handle 0x18c00ad7): no image for it; NVN target 1 format 0x7b 512x512x1 levels 10 gpu 0x10ba393000, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc18 (handle 0x18c00c18): no image for it; NVN target 1 format 0x8c 512x512x1 levels 10 gpu 0x10c10f8a00, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc19 (handle 0x18c00c19): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c111d000, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd57 (handle 0x18c00d57): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c8380e00, view yes (format 0x8d target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd58 (handle 0x18c00d58): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c839fc00, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc84 (handle 0x18c00c84): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c3e51600, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc49 (handle 0x18c00c49): no image for it; NVN target 1 format 0x8c 683x683x1 levels 10 gpu 0x10c249ee00, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc4a (handle 0x18c00c4a): no image for it; NVN target 1 format 0x7b 1024x1024x1 levels 11 gpu 0x10c24e2e00, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc4b (handle 0x18c00c4b): no image for it; NVN target 1 format 0x7e 512x512x1 levels 10 gpu 0x10c25c7000, view yes (format 0x7e target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd83 (handle 0x18c00d83): no image for it; NVN target 1 format 0x8b 512x512x1 levels 10 gpu 0x10c939de00, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd84 (handle 0x18c00d84): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c93c8600, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xcee (handle 0x18c00cee): no image for it; NVN target 1 format 0x8b 683x683x1 levels 10 gpu 0x10c5e2aa00, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xcef (handle 0x18c00cef): no image for it; NVN target 1 format 0x7b 1024x1024x1 levels 11 gpu 0x10c5e76000, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xcf0 (handle 0x18c00cf0): no image for it; NVN target 1 format 0x7e 512x512x1 levels 10 gpu 0x10c5f5a200, view yes (format 0x7e target 1), wanted view type 1 use 0
[hos host] [debug] vk: a staging arena of 16 MB took 12.4 ms to make
[hos host] vk: texture slot 0xcfa (handle 0x18c00cfa): no image for it; NVN target 1 format 0x8c 1024x1024x1 levels 11 gpu 0x10c67e1400, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xcfc (handle 0x18c00cfc): no image for it; NVN target 1 format 0x7f 768x768x1 levels 10 gpu 0x10c6a69400, view yes (format 0x7f target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0x919 (handle 0x18c00919): no image for it; NVN target 1 format 0x7b 512x512x1 levels 10 gpu 0x10b2e54200, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xa49 (handle 0x18900a49): no image for it; NVN target 1 format 0x7b 64x64x1 levels 7 gpu 0x10b7a23400, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xaf4 (handle 0x18c00af4): no image for it; NVN target 1 format 0x92 512x512x1 levels 10 gpu 0x10bae49a00, view yes (format 0x92 target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xaf5 (handle 0x18c00af5): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10bae5aa00, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xaf6 (handle 0x18c00af6): no image for it; NVN target 1 format 0x7d 768x768x1 levels 10 gpu 0x10baedee00, view yes (format 0x7d target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xcf4 (handle 0x18c00cf4): no image for it; NVN target 1 format 0x8c 1024x1024x1 levels 11 gpu 0x10c6267600, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: slow upload 32.3 ms: #244 0x10c62f3c00 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB
[hos host] vk: texture slot 0xcf6 (handle 0x18c00cf6): no image for it; NVN target 1 format 0x83 768x768x1 levels 10 gpu 0x10c64ef600, view yes (format 0x83 target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xcf7 (handle 0x18c00cf7): no image for it; NVN target 1 format 0x8c 1024x1024x1 levels 11 gpu 0x10c6518800, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xcf9 (handle 0x18c00cf9): no image for it; NVN target 1 format 0x7f 768x768x1 levels 10 gpu 0x10c67a0800, view yes (format 0x7f target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc3c (handle 0x18c00c3c): no image for it; NVN target 1 format 0x7e 768x768x1 levels 10 gpu 0x10c1da7800, view yes (format 0x7e target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc0c (handle 0x18c00c0c): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c0b56400, view yes (format 0x8d target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc0e (handle 0x18c00c0e): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c0b9fa00, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc10 (handle 0x18c00c10): no image for it; NVN target 1 format 0x7f 512x512x1 levels 10 gpu 0x10c0ca8200, view yes (format 0x7f target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xdc1 (handle 0x18c00dc1): no image for it; NVN target 1 format 0x8c 512x512x1 levels 10 gpu 0x10ca474400, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd7f (handle 0x18c00d7f): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c9193a00, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xdc2 (handle 0x18900dc2): no image for it; NVN target 1 format 0x93 341x341x1 levels 9 gpu 0x10ca498a00, view yes (format 0x93 target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xdc3 (handle 0x18900dc3): no image for it; NVN target 1 format 0x7b 512x512x1 levels 10 gpu 0x10ca4a0800, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd81 (handle 0x18900d81): no image for it; NVN target 1 format 0x7e 512x512x1 levels 10 gpu 0x10c9278c00, view yes (format 0x7e target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xdc4 (handle 0x18c00dc4): no image for it; NVN target 1 format 0x7b 512x512x1 levels 10 gpu 0x10ca4dba00, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd7b (handle 0x18c00d7b): no image for it; NVN target 1 format 0x8b 1024x1024x1 levels 11 gpu 0x10c8dc7c00, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd7c (handle 0x18900d7c): no image for it; NVN target 1 format 0x8b 683x683x1 levels 10 gpu 0x10c8e68800, view yes (format 0x8b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xd7e (handle 0x18900d7e): no image for it; NVN target 1 format 0x7b 1024x1024x1 levels 11 gpu 0x10c90af800, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc1b (handle 0x18c00c1b): no image for it; NVN target 1 format 0x8e 512x512x1 levels 10 gpu 0x10c11cca00, view yes (format 0x8e target 1), wanted view type 1 use 0
[hos host] vk: slow upload 15.2 ms: #268 0x10c11e2c00 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB
[hos host] vk: texture slot 0xc1f (handle 0x18c00c1f): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c1480a00, view yes (format 0x8d target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc20 (handle 0x18c00c20): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c149f800, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc60 (handle 0x18c00c60): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c2ae4e00, view yes (format 0x8d target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc61 (handle 0x18c00c61): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c2b03c00, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc37 (handle 0x18c00c37): no image for it; NVN target 1 format 0x8c 171x171x1 levels 8 gpu 0x10c1af0c00, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc38 (handle 0x18c00c38): no image for it; NVN target 1 format 0x7b 256x256x1 levels 9 gpu 0x10c1af6e00, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc29 (handle 0x18c00c29): no image for it; NVN target 1 format 0x8c 512x512x1 levels 10 gpu 0x10c1683c00, view yes (format 0x8c target 1), wanted view type 1 use 0
[hos host] vk: texture slot 0xc2a (handle 0x18c00c2a): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c16a8200, view yes (format 0x7b target 1), wanted view type 1 use 0
[hos host] vk: pool 0x104da60000+0xc0 imported in 4.1 ms as 0x104cc00000..0x104dc00000 (38 imports, 152 ms in all)
[hos host] vk: slow upload 25.5 ms: #388 0x10c2c3fc00 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB
[hos host] [debug] vk: a staging arena of 16 MB took 7.5 ms to make
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 3, mip filter nearest, LOD bias 0; made with 0)
[hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 64x64)
[hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 32x32)
[hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 16x16)
[hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 8x8)
[hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 4x4)
[hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 2x2)
[hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 1x1)
[hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 64x64)
[hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 32x32)
[hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 16x16)
[hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 8x8)
[hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 4x4)
[hos host] vk: texture #456 at 0x10000000000 beside #6: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #456 at 0x10000000000 beside #149: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 256x144x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: texture #456 at 0x10000000000 beside #192: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 128x128x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: slow submit: 606.3 ms: draw 581.0 descriptors 418.7 texture lookups 363.0 uploads 318.6 programs 13.3 commands 129.1 dispatch 6.4
[hos host] vk: pool 0x117883d100+0x700 imported in 18.2 ms as 0x1178400000..0x117c400000 (42 imports, 173 ms in all)
[hos host] vk: pool 0x117e724000+0x18000 imported in 4.3 ms as 0x117e400000..0x117e800000 (43 imports, 178 ms in all)
[hos host] vk: pool 0x117d058000+0x44b040 imported in 2.2 ms as 0x117d000000..0x117d800000 (45 imports, 181 ms in all)
[hos host] vk: pool 0x117cabc000+0x230040 imported in 2.8 ms as 0x117c800000..0x117d000000 (47 imports, 185 ms in all)
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 1, mip filter nearest, LOD bias 0; made with 0, an image of one level: not the row's)
[hos host] vk: pool 0x117c4a0000+0xa0040 imported in 2.1 ms as 0x117c400000..0x117c800000 (48 imports, 187 ms in all)
[hos host] vk: pool 0x117c800000 remade larger (4 retired)
[hos host] vk: pool 0x117c400000 remade larger (5 retired)
[hos host] vk: pool 0x117c400000..0x117d000000 took in 2 pools
[hos host] vk: pool 0x117c400000..0x117d000000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: pool 0x117d800000 remade larger (6 retired)
[hos host] vk: pool 0x117d800000..0x117e000000: mirror made with the GPU's contents of 1 retired pool(s)
[hos host] vk: pool 0x117e3d8000+0x8b040 imported in 3.0 ms as 0x117e000000..0x117e800000 (51 imports, 193 ms in all)
[hos host] vk: pool 0x117e400000 remade larger (7 retired)
[hos host] vk: pool 0x117e000000..0x117e800000: mirror made with the GPU's contents of 1 retired pool(s)
[hos host] vk: pool 0x1146b29e00+0x80 imported in 8.3 ms as 0x1145400000..0x1147c00000 (52 imports, 202 ms in all)
[hos host] vk: pool 0x115bd19580+0x1b580 imported in 9.7 ms as 0x1159c00000..0x115c000000 (53 imports, 211 ms in all)
[hos host] vk: sampler 410 is nearest: min filter 4, mag filter 0, wrap 7/7/7, lod 0..15 bias 0, aniso 1, compare 0
[hos host] vk: slow submit: 114.4 ms: texture lookups 1.7 uploads 1.7 dispatch 114.3
[hos host] vk: the indirect mesh draws' ring (4 MB) made and filled with 1s
[hos host] vk: 6 background workers (shader translation, pipelines)
[hos host] vk: quick pipelines only for draws that must wait (RX_VK_QUICK_MS=required)
[hos host] vk: quick pipeline (no optimizer) for cb71f79f6129aab8 in 13.2 ms, drawn with until its full build is done (1 so far, 13 ms in all, each kept for the session)
[hos host] vk: RX_VK_RECORDER=thread: vkCmdCopyBufferToImage did not fit in the room a packet is given (81920 bytes): made directly once the recorder had run what was before it (logged once)
[hos host] vk: pool 0x117e800000 remade larger (8 retired)
[hos host] vk: pool 0x117e000000..0x117ec00000 took in 2 pools
[hos host] vk: pool 0x117e000000..0x117ec00000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: a clear after 1 draws of its pass (2560x1440, 1 colour targets + depth, last program 3bde059fb051af1f): a pass of its own
[hos host] vk: sampler 404 is nearest: min filter 0, mag filter 0, wrap 7/7/7, lod 0..15 bias 0, aniso 1, compare 0
[hos host] vk: sampler anisotropy: the game asks for 1 (min filter 0, mip filter nearest, LOD bias 0; made with 0, an image of one level: not the row's)
[hos host] vk: a clear after 3 draws of its pass (640x360, 1 colour targets + depth, last program 619664ed509bbd99): a pass of its own
[hos host] vk: pool 0x102fde0a00+0x180 imported in 12.8 ms as 0x102fc00000..0x1033000000 (55 imports, 226 ms in all)
[hos host] vk: a clear after 3 draws of its pass (640x360, 2 colour targets + depth, last program 2479abad219cb19a): a pass of its own
[hos host] vk: sampler 414 is nearest: min filter 5, mag filter 0, wrap 7/7/7, lod 0..15 bias 0, aniso 1, compare 0
[hos host] vk: texture #479 at 0x10004dd1a00 beside #474: 1600x1600x1 x3 levels 1 layout 0x4 class 0x1007c NVN 0x32 target 4, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: table entry bound again at frame 2333: prog 507e946d64f2d55e ssbo 0 of 3 (c0+0x880) addr 0x108c702000 size 0x2550000 (before 0)
[hos host] vk: pipeline from the caches for a draw ahead of its build (1 so far, each kept for the session)
[hos host] vk: pool 0x7fe04e8a00+0x6888 imported in 5.3 ms as 0x7fdfc00000..0x7fe0800000 (56 imports, 231 ms in all)
[hos host] vk: pool 0x7fd37e4400+0x760 imported in 3.3 ms as 0x7fd3400000..0x7fd3800000 (57 imports, 234 ms in all)
[hos host] vk: write-back diff: a shadow for pool 0x113b400000 (24 MB; 28 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x1159c00000 (36 MB; 64 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117d000000 (8 MB; 72 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117d800000 (8 MB; 80 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117c400000 (4 MB; 84 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117c400000 (12 MB; 96 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117e400000 (4 MB; 100 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x1178400000 (64 MB; 164 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117d800000 (4 MB; 168 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117e800000 (4 MB; 172 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117c800000 (8 MB; 180 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x117e000000 (8 MB; 188 MB of shadows)
[hos host] vk: slow submit: 134.5 ms: draw 14.0 descriptors 11.4 texture lookups 33.1 uploads 9.7 programs 1.3
[hos host] vk: depth-compare sampler: the game's compare mode 1 function 4 -> compare on op 3 (filters min 1 mag 1)
[hos rmbl::VibrationThread] hid: 4096 vibration values, strongest 0.00 since the last count
[hos host] vk: texture 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 -> vk 124 (499 textures, 1748 MB)
[hos host] vk: upload: 12288x12288 NVN 0x32 (288 MB) is all zero: cleared on the GPU (checked in 59.0 ms, 288 MB read)
[hos host] vk: slow upload 59.1 ms: #506 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32, 294912 of 294912 KB
[hos host] vk: texture 0x10008ac2600 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (500 textures, 1751 MB)
[hos host] vk: texture #509 at 0x10008342200 beside #492: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: texture #511 at 0x10004921600 beside #494: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #512 at 0x10004831400 beside #493: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #513 at 0x10004741200 beside #495: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #514 at 0x10009242400 beside #510: 1280x720x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #520 at 0x10004831400 beside #493: 160x90x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x3 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #520 at 0x10004831400 beside #512: 160x90x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x3 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: NVN2 program 3d33bb73d3deee34 (stage 7) lowered with extended registers (577 REGMAP words)
[hos host] vk: NVN2 program 463f42630fce83a0 (stage 7) lowered with extended registers (1539 REGMAP words)
[hos host] [debug] vk: a staging arena of 16 MB took 11.8 ms to make
[hos host] vk: NVN2 program 66c494fe4cfa930b (stage 7) lowered with extended registers (1460 REGMAP words)
[hos host] vk: NVN2 program 354dde838352c262 (stage 7) lowered with extended registers (320 REGMAP words)
[hos host] vk: NVN2 program e3fc8fcf540f7a48 (stage 7) lowered with extended registers (1503 REGMAP words)
[hos host] vk: staging arenas: 16 (16 MB in this one)
[hos host] vk: NVN2 program d38b0d52a94ab5a8 (stage 7) lowered with extended registers (337 REGMAP words)
[hos host] vk: NVN2 program b509ae330829ce02 (stage 7) lowered with extended registers (264 REGMAP words)
[hos host] vk: texture #532 at 0x10001e00400 beside #466: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: NVN2 program bd0ee311e887f8e4 (stage 0) lowered with extended registers (1132 REGMAP words)
[hos host] vk: depth-compare sampler: the game's compare mode 0 function 0 -> compare on op 3 (filters min 1 mag 1)
[hos host] vk: upload: 4096x4096 NVN 0x32 (32 MB) is all zero: cleared on the GPU (checked in 1.7 ms, 32 MB read)
[hos host] vk: NVN2 program 7d8136fe8a9978f0 (stage 0) lowered with extended registers (1148 REGMAP words)
[hos host] vk: pool 0x7fdac97800+0xb40 imported in 6.1 ms as 0x7fd9800000..0x7fdc400000 (58 imports, 240 ms in all)
[hos host] vk: NVN2 program b764b6275a5f4b66 (stage 0) lowered with extended registers (1129 REGMAP words)
[hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own
[hos host] vk: texture #537 at 0x100021c0600 beside #536: 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 960x540x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #542 at 0x100023caa00 beside #540: 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1, the other 240x135x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #544 at 0x100023caa00 beside #540: 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1, the other 240x135x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #544 at 0x100023caa00 beside #542: 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1, the other 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #546 at 0x100023caa00 beside #540: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 240x135x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #546 at 0x100023caa00 beside #542: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #546 at 0x100023caa00 beside #544: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: sampler 423 is nearest: min filter 0, mag filter 0, wrap 5/5/5, lod 0..15 bias 0, aniso 1, compare 0
[hos host] vk: sampler 424 is nearest: min filter 2, mag filter 0, wrap 1/1/1, lod 0..15 bias 0, aniso 1, compare 0
[hos host] vk: texture #556 at 0x10004380c00 beside #469: 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40002 NVN 0x5 target 1 (rendered)
[hos host] vk: a clear after 11 draws of its pass (2560x1440, 1 colour targets + depth, last program 5986d2784af21135): a pass of its own
[hos host] vk: texture #557 at 0x10001e00400 beside #466: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: texture #557 at 0x10001e00400 beside #532: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: pool 0x1012e97e00+0xf00000 imported in 5.4 ms as 0x1012c00000..0x1014000000 (59 imports, 246 ms in all)
[hos host] vk: pool 0x100fb07600+0x1c2000 imported in 2.2 ms as 0x100f800000..0x1010000000 (60 imports, 248 ms in all)
[hos host] vk: pool 0x7fff4bda00+0x10 imported in 3.6 ms as 0x7ffe000000..0x7fff800000 (61 imports, 251 ms in all)
[hos host] vk: sampler 267 is nearest: min filter 0, mag filter 0, wrap 5/5/5, lod -1000..1000 bias 0, aniso 1, compare 0
[hos host] vk: network 9dc98871eb7fc101: dispatch 64x36x1, 1 textures 1 images: 2560x1440x1/122 2560x1440x1/122
[hos host] vk: upscaler: Off: network 9dc98871eb7fc101 left out, its input stretched linearly (2560x1440 -> 2560x1440), packed output
[hos host] vk: slow submit: 194.3 ms: draw 118.2 descriptors 109.3 texture lookups 141.8 uploads 88.2 programs 8.4 dispatch 20.8
[hos host] vk: write-back diff: a shadow for pool 0x100f800000 (8 MB; 196 MB of shadows)
[hos host] vk: a frame's uploads took 423 ms: ASTC on the GPU 27 (107.0 MB, 290 ms); unswizzled on the GPU 2 (47.2 MB, 12 ms); on the CPU 118 (503.5 MB, 121 ms); [1365x1365 x90 NVN 0x8e ASTC on the GPU 56.5 MB 179 ms] [12288x12288 x1 NVN 0x32 on the CPU 288.0 MB 59 ms] [1536x1536 x1 NVN 0x7b ASTC on the GPU 2.0 MB 32 ms]
[hos host] vk: 204 new ASTC textures left for a later frame (this frame's uploads took their budget; 204 so far)
[hos sead::MainThread] nvn: GPU waits in 5 s: 173 sync waits (1629 ms), 0 polls timed out (0 ms), 1 queue finishes (8 ms)
[hos Presentation Thread] vi: 2334 frames presented (48.6 fps over the last 5.6 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 11, 15.5-17.8 247, 17.8-22 15, 22-30 0, 30-40 0, over 40 1
[hos host] vk: readback pages refreshed from guest memory: 7 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: table entry bound again at frame 2334: prog a03bdab95992c48e ssbo 0 of 3 (c0+0x880) addr 0x108c702000 size 0x2550000 (before 0)
[hos host] vk: pool 0x117d000000..0x117e000000 took in 2 pools
[hos host] vk: pool 0x117d000000..0x117e000000: 12 KB the CPU wrote since retired pool 0x117d800000's last sync kept from guest memory, not its mirror
[hos host] vk: pool 0x117d000000..0x117e000000: 4 KB the CPU wrote since retired pool 0x117d000000's last sync kept from guest memory, not its mirror
[hos host] vk: pool 0x117d000000..0x117e000000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: pool 0x117cfbb040+0x9b040 imported in 2.5 ms as 0x117c400000..0x117e000000 (63 imports, 255 ms in all)
[hos host] vk: pool 0x117c400000..0x117e000000 took in 2 pools
[hos host] vk: pool 0x117c400000..0x117e000000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: pool 0x117c3fe000+0xa0040 imported in 5.4 ms as 0x1178400000..0x117e000000 (64 imports, 261 ms in all)
[hos host] vk: pool 0x1178400000..0x117e000000 took in 2 pools
[hos host] vk: pool 0x1178400000..0x117e000000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: pool 0x117df8b040+0x44b040 imported in 7.0 ms as 0x1178400000..0x117ec00000 (65 imports, 268 ms in all)
[hos host] vk: pool 0x1178400000..0x117ec00000 took in 2 pools
[hos host] vk: pool 0x1178400000..0x117ec00000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: slow submit: 35.9 ms: dispatch 35.8
[hos host] vk: upscaler: NVIDIA DLAA after the game's bloom and tone map (DLAA before the bloom: off)
[hos host] vk: a clear after 3 draws of its pass (640x360, 1 colour targets + depth, last program 619664ed509bbd99): a pass of its own
[hos host] vk: a clear after 3 draws of its pass (640x360, 2 colour targets + depth, last program 2479abad219cb19a): a pass of its own
[hos host] vk: write-back diff: a shadow for pool 0x117c400000 (12 MB; 116 MB of shadows)
[hos host] vk: write-back diff: a shadow for pool 0x1178400000 (64 MB; 180 MB of shadows)
[hos host] vk: slow submit: 20.4 ms:
[hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own
[hos host] vk: 92 new ASTC textures left for a later frame (this frame's uploads took their budget; 296 so far)
[hos ModuleSystemWorker4] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 2.0 ms; 1 in all, 0 MB, 2 ms)
[hos host] vk: slow submit: 16.9 ms: draw 16.2 descriptors 16.0 texture lookups 15.8 uploads 15.7
[hos] guest backtrace: lr 0x804fc1ec 0x83b6f194 0x83b6f080
[hos] guest backtrace: lr 0x8348c5e0 0x83b6f08c
[hos] guest backtrace: lr 0x804fc1ec 0x83b6f194 0x83b6f080
[hos] guest backtrace: lr 0x8348c5e0 0x83b6f08c
[hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own
[hos host] vk: 62 new ASTC textures left for a later frame (this frame's uploads took their budget; 358 so far)
[hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.8 ms, 1 frames after their copies were recorded, the first #191 0x113739f000 128x128 levels 8 NVN 0x3f (1 so far, 528 KB, 1 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos sead::MainThread] nvn: the game sets present interval 1 -> 4
[hos sead::MainThread] nvn: the game sets present interval 4 -> 1
[hos host] vk: slow submit: 17.0 ms: draw 16.8 descriptors 16.1 texture lookups 15.6 uploads 15.5
[hos host] vk: texture #466 0x10001e00400 2560x1440 NVN 0x34: the GPU wrote its storage (submission 16346): uploaded again
[hos host] vk: upload over a drawn image: #466 0x10001e00400 2560x1440x1 x1 NVN 0x34 at present 2335 (texture frame 2336), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: texture 0x10001e00400 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own
[hos host] vk: 136 new ASTC textures left for a later frame (this frame's uploads took their budget; 494 so far)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 20 bindings followed, 32 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1843 dirty pages from their submits' captures, 2893 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 282 dispatches, 2687 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 363 runs, 247.8 MB compared; 52310 pages read into the shadows; 432 MB of shadows (108 MB of pages known, in 123 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 213 in captures, 0 only in the collection, 4 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.9 ms; 2 in all, 0 MB, 3 ms)
[hos sead::MainThread] nvn: the game sets present interval 1 -> 4
[hos sead::MainThread] nvn: the game sets present interval 4 -> 1
[hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own
[hos host] vk: 22 new ASTC textures left for a later frame (this frame's uploads took their budget; 516 so far)
[hos host] vk: slow upload 15.6 ms: #282 0x10c6e4fe00 512x512x1 x1 levels 10 NVN 0x7f ASTC (BC3), 123 of 123 KB
[hos host] vk: slow submit: 17.5 ms: draw 17.3 descriptors 17.0 texture lookups 16.5 uploads 16.4
[hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own
[hos host] vk: 30 new ASTC textures left for a later frame (this frame's uploads took their budget; 546 so far)
[hos ModuleSystemWorker1] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.8 ms; 3 in all, 0 MB, 4 ms)
[hos host] vk: mirror of pool 0x7faf000000 (4 MB) given back: retired, 60 submissions since its sync
[hos host] vk: mirror of pool 0x7fab000000 (64 MB) given back: retired, 60 submissions since its sync
[hos host] vk: mirror of pool 0x117c800000 (8 MB) given back: retired, 56 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x117c400000 (4 MB) given back: retired, 56 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x117d800000 (4 MB) given back: retired, 56 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x117e400000 (4 MB) given back: retired, 56 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x117e800000 (4 MB) given back: retired, 56 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x117e000000 (8 MB) given back: retired, 56 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x117d800000 (8 MB) given back: retired, 54 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x117d000000 (8 MB) given back: retired, 54 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x117d000000 (16 MB) given back: retired, 45 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x117c400000 (12 MB) given back: retired, 45 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x117c400000 (28 MB) given back: retired, 45 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x1178400000 (64 MB) given back: retired, 45 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x117e000000 (12 MB) given back: retired, 45 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x1178400000 (92 MB) given back: retired, 45 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: 20 mirrors given back (353 MB in all); 874 MB of mirrors now, 0 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: slow submit: 15.2 ms: draw 15.2 descriptors 14.8 texture lookups 14.3 uploads 14.2
[hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own
[hos host] vk: 17 new ASTC textures left for a later frame (this frame's uploads took their budget; 563 so far)
[hos host] vk: 1 textures retired (549 left, 1901 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: NVN2 program 5599b891dfe20fdf (stage 0) lowered with extended registers (1132 REGMAP words)
[hos host] vk: NVN2 program e5f2865bdce56993 (stage 0) lowered with extended registers (1129 REGMAP words)
[hos host] vk: NVN2 program 57b9a80d47da5078 (stage 0) lowered with extended registers (1148 REGMAP words)
[hos ModuleSystemWorker4] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 4.0 ms; 4 in all, 0 MB, 8 ms)
[hos ModuleSystemWorker3] vk: readbacks read without their wait: the culling camera block (0x3980) by default (newest 0, 1 sizes)
[hos host] vk: reaper: 8 images (7 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 3 (10 MB)
[hos ModuleSystemWorker5] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.2 ms; 5 in all, 0 MB, 8 ms)
[hos BaseProcExecutor3] map: Market_Child (child)
[hos BaseProcExecutor3] cheats: inventory probe: Weapon_Sword_Kokiri=0 Weapon_Sword_Master=0 Weapon_Sword_Giant=0 Weapon_Sword_Biggoron=0 Weapon_Shield_Deku=0 Weapon_Shield_Hylian=0 Weapon_Shield_Mirror=0 Armor_Adult_Goron=0 Armor_Adult_Zora=0 Armor_Adult_HeavyBoot=0 Armor_Adult_HoverBoot=0 Item_DekuStick=0 Item_DekuNut=0 Item_Bomb=0 Item_Bombchu=0
[hos BaseProcExecutor3] cheats: inventory probe: Weapon_Range_SlingShot=0 Item_DekuSeed=0 Weapon_Range_Bow=0 Item_Arrow_Normal=0 Item_Arrow_Fire=0 Item_Arrow_Ice=0 Item_Arrow_Light=0 Item_DinFire=0 Item_FaroreWind=0 Item_NayruLove=0 Item_Ocarina_Fairy=0 Item_Ocarina_Time=0 Item_Hookshot=0 Item_Longshot=0 Item_Boomerang=0 Item_GlassTruth=0 Item_MagicBean=0
[hos BaseProcExecutor3] cheats: inventory probe: Item_Hammer=0 Item_Bottle=0 Armor_Child_Mask_Keaton=0 Armor_Child_Mask_Skull=0 Armor_Child_Mask_Spooky=0 Armor_Child_Mask_Bunny=0 Armor_Child_Mask_Goron=0 Armor_Child_Mask_Zora=0 Armor_Child_Mask_Gerudo=0 Armor_Child_Mask_Truth=0 Item_Gauntlet_Goron=0 Armor_Adult_Gauntlet_Silver=0 Armor_Adult_Gauntlet_Gold=0
[hos BaseProcExecutor3] cheats: inventory probe: Item_Scale_Silver=0 Item_Scale_Gold=0 Item_BombBag_Small=0 Item_BombBag_Medium=0 Item_BombBag_Large=0 Item_Quiver_Small=0 Item_Quiver_Medium=0 Item_Quiver_Large=0 Item_BulletBag_Small=0 Item_BulletBag_Medium=0 Item_BulletBag_Large=0 Item_Wallet_Medium=0 Item_Wallet_Large=0 Item_FairyBless_MP=0 Item_FairyBless_MPDouble=0
[hos BaseProcExecutor3] cheats: inventory probe: Item_FairyBless_DefenseDouble=0 Item_Stone_Forest=0 Item_Stone_Fire=0 Item_Stone_Water=0 Item_Medal_Forest=0 Item_Medal_Fire=0 Item_Medal_Water=0 Item_Medal_Shadow=0 Item_Medal_Spirit=0 Item_Medal_Light=0 Item_Melody_ZeldaLullaby=0 Item_Melody_EponaSong=0 Item_Melody_SariaSong=0 Item_Melody_SunSong=0
[hos BaseProcExecutor3] cheats: inventory probe: Item_Melody_TimeSong=0 Item_Melody_StormSong=0 Item_Melody_ForestMinuet=0 Item_Melody_FireBolero=0 Item_Melody_WaterSerenade=0 Item_Melody_ShadowNocturne=0 Item_Melody_SpiritRequiem=0 Item_Melody_LightPrelude=0 Item_AgonyStone=0 Item_Token_Gerudo=0 Item_Token_Gold=0 Item_Sword_Kokiri=0 Item_Sword_Giant=0
[hos BaseProcExecutor3] cheats: inventory probe: Item_Sword_Biggoron=0 Item_Shield_Mirror=0 Item_Range_Bow=0 Item_Range_SlingShot=0 Weapon_Range_Hookshot=0 Weapon_Range_Longshot=0 Weapon_Boomerang=0 Weapon_Hammer=0 Item_HeavyBoot=0 Item_HoverBoot=0 Item_Mask_Keaton=0 Item_Gauntlet_Silver=0 Armor_GlassTruth=0 Armor_Child_GlassTruth=0 Armor_Adult_GlassTruth=0
[hos BaseProcExecutor3] cheats: inventory probe: Armor_Child_Gauntlet_Goron=0 Item_Wallet_Small=1
[hos host] vk: slow upload 27.8 ms: #294 0x10bad3a200 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB
[hos host] vk: slow submit: 32.1 ms: draw 31.7 descriptors 30.2 texture lookups 28.8 uploads 28.6
[hos host] vk: 94 new ASTC textures left for a later frame (this frame's uploads took their budget; 657 so far)
[hos ModuleSystemWorker1] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.9 ms; 6 in all, 0 MB, 9 ms)
[hos host] vk: texture #809 at 0x10001e00400 beside #466: 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: texture #809 at 0x10001e00400 beside #532: 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #809 at 0x10001e00400 beside #557: 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1 (rendered)
[hos host] vk: slow submit: 72.7 ms: draw 68.6 descriptors 48.9 texture lookups 46.3 uploads 18.6 programs 15.5 commands 2.8
[hos host] vk: texture #831 at 0x10001fe0800 beside #811: 1280x720x1 x1 levels 1 layout 0x4 class 0x40002 NVN 0x5 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: texture #833 at 0x10004a11800 beside #478: 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x40010 NVN 0x2e target 1 (rendered)
[hos host] vk: texture #840 at 0x10005192200 beside #835: 1600x1600x1 x3 levels 1 layout 0x4 class 0x1007c NVN 0x32 target 4, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: slow submit: 18.1 ms: draw 2.9 texture lookups 11.7 programs 1.4
[hos host] vk: texture #874 at 0x100092d4400 beside #868: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: texture #876 at 0x10004ce1e00 beside #870: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #877 at 0x10004bf1c00 beside #869: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #878 at 0x10004b01a00 beside #871: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #879 at 0x1000a1d4600 beside #875: 1280x720x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #884 at 0x10004bf1c00 beside #869: 160x90x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x3 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: texture #884 at 0x10004bf1c00 beside #877: 160x90x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x3 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: NVN2 program b81f83fe05a2a32a (stage 0) lowered with extended registers (939 REGMAP words)
[hos host] vk: NVN2 program 2d6b101e5868b4ea (stage 0) lowered with extended registers (1132 REGMAP words)
[hos host] vk: NVN2 program 40d0c0768e874600 (stage 7) lowered with extended registers (331 REGMAP words)
[hos host] vk: NVN2 program c567dec57b3bbad1 (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: NVN2 program 56cc6c3a14ce827a (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: NVN2 program 73b819fbe71d2814 (stage 0) lowered with extended registers (1128 REGMAP words)
[hos host] discord: the status: Hyrule Market - Child / ❤️ 0.5/3 · 🕷️ 0 · 💗 0
[hos host] vk: NVN2 program 167110f0f411da2e (stage 0) lowered with extended registers (1122 REGMAP words)
[hos host] vk: NVN2 program c5f227a6efbe5b41 (stage 0) lowered with extended registers (1141 REGMAP words)
[hos host] vk: NVN2 program a19c1ffa6e31ecd4 (stage 0) lowered with extended registers (1145 REGMAP words)
[hos host] vk: NVN2 program 2de89066ce43a6c1 (stage 0) lowered with extended registers (1134 REGMAP words)
[hos host] vk: NVN2 program 9039474cf8f30ef4 (stage 0) lowered with extended registers (940 REGMAP words)
[hos host] vk: NVN2 program 5c55ab4d22261ede (stage 0) lowered with extended registers (1132 REGMAP words)
[hos host] vk: NVN2 program a19c1ffa6e31ecd4 (stage 0) lowered with extended registers (1145 REGMAP words)
[hos host] vk: NVN2 program 690e54aa085d7bd5 (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: NVN2 program f33fc74589ea6ff0 (stage 0) lowered with extended registers (1131 REGMAP words)
[hos host] vk: NVN2 program 196c6bfc68efb1dd (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: texture 0x10d14b8e00 171x341x1 x1 levels 9 NVN 0x8d -> vk 37 (ASTC decoded) (1000 textures, 2606 MB)
[hos host] vk: NVN2 program 73b819fbe71d2814 (stage 0) lowered with extended registers (1128 REGMAP words)
[hos host] vk: NVN2 program 2ef7974fd3b630c7 (stage 0) lowered with extended registers (943 REGMAP words)
[hos host] vk: NVN2 program 809483568f9bbfc9 (stage 0) lowered with extended registers (1137 REGMAP words)
[hos host] vk: NVN2 program a19c1ffa6e31ecd4 (stage 0) lowered with extended registers (1145 REGMAP words)
[hos host] vk: NVN2 program 2855427a5b00e4d4 (stage 0) lowered with extended registers (1142 REGMAP words)
[hos host] vk: NVN2 program 2d6b101e5868b4ea (stage 0) lowered with extended registers (1132 REGMAP words)
[hos host] vk: NVN2 program f14336ce5b132a86 (stage 0) lowered with extended registers (1122 REGMAP words)
[hos host] vk: NVN2 program b07021ee80078f45 (stage 0) lowered with extended registers (1117 REGMAP words)
[hos host] vk: NVN2 program 73b819fbe71d2814 (stage 0) lowered with extended registers (1128 REGMAP words)
[hos host] vk: NVN2 program 167110f0f411da2e (stage 0) lowered with extended registers (1122 REGMAP words)
[hos host] vk: NVN2 program 78d7ec0f190748f5 (stage 0) lowered with extended registers (1117 REGMAP words)
[hos host] vk: NVN2 program 56cc6c3a14ce827a (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: NVN2 program a6d021d3dce2b9f9 (stage 0) lowered with extended registers (1128 REGMAP words)
[hos host] vk: NVN2 program 389c379053aab5fe (stage 0) lowered with extended registers (1129 REGMAP words)
[hos host] vk: NVN2 program 13a1f5b177ab2712 (stage 0) lowered with extended registers (1130 REGMAP words)
[hos host] vk: NVN2 program 2d6b101e5868b4ea (stage 0) lowered with extended registers (1132 REGMAP words)
[hos host] vk: NVN2 program 2ef7974fd3b630c7 (stage 0) lowered with extended registers (943 REGMAP words)
[hos host] vk: NVN2 program b1fc459b591f7806 (stage 0) lowered with extended registers (939 REGMAP words)
[hos host] vk: NVN2 program 9039474cf8f30ef4 (stage 0) lowered with extended registers (940 REGMAP words)
[hos host] vk: texture #1169 at 0x100023a0e00 beside #828: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: NVN2 program 57b9a80d47da5078 (stage 0) lowered with extended registers (1148 REGMAP words)
[hos host] vk: NVN2 program a9cad530cd625cc4 (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: NVN2 program 196c6bfc68efb1dd (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: a draw (program b426f821e80124dd) samples a target the pass wrote, after 1 draws of the pass (2560x1440): a pass of its own
[hos host] vk: texture #1201 at 0x10001fe0800 beside #811: 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: texture #1201 at 0x10001fe0800 beside #831: 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40002 NVN 0x5 target 1 (rendered)
[hos host] vk: texture #1208 at 0x100021aea00 beside #1206: 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1, the other 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #1210 at 0x100021aea00 beside #1206: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #1210 at 0x100021aea00 beside #1208: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #1211 at 0x100020e5a00 beside #1202: 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #1212 at 0x10001fe0800 beside #811: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered)
[hos host] vk: texture #1212 at 0x10001fe0800 beside #831: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40002 NVN 0x5 target 1 (rendered)
[hos host] vk: texture #1212 at 0x10001fe0800 beside #1201: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered)
[hos host] vk: texture #1213 at 0x10004921600 beside #494: 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x25 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered)
[hos host] vk: pool 0x1012538600+0xf00000 imported in 2.7 ms as 0x1012400000..0x1014000000 (79 imports, 286 ms in all)
[hos host] vk: pool 0x1010bfcc00+0x1c2000 imported in 2.2 ms as 0x1010800000..0x1011000000 (80 imports, 289 ms in all)
[hos host] vk: slow submit: 122.7 ms: draw 89.6 descriptors 45.1 texture lookups 60.6 uploads 8.3 programs 42.5 commands 1.2 dispatch 2.7
[hos host] vk: 391 new ASTC textures left for a later frame (this frame's uploads took their budget; 1048 so far)
[hos host] vk: slow upload 25.2 ms: #587 0x10c4fe3e00 1024x1024x1 x1 levels 11 NVN 0x8b ASTC (BC3), 642 of 642 KB
[hos host] vk: slow submit: 40.2 ms: draw 37.1 descriptors 31.0 texture lookups 27.4 uploads 26.9 programs 2.3 commands 2.8
[hos host] vk: NVN2 program 875fa04074350384 (stage 0) lowered with extended registers (1142 REGMAP words)
[hos host] vk: NVN2 program cb9e651e1a2900cc (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: NVN2 program 196c6bfc68efb1dd (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: NVN2 program 2e41b85d34167b3b (stage 0) lowered with extended registers (1113 REGMAP words)
[hos host] vk: NVN2 program d88eb83da249de23 (stage 0) lowered with extended registers (940 REGMAP words)
[hos host] vk: NVN2 program 42be3be9f8a7413b (stage 0) lowered with extended registers (944 REGMAP words)
[hos host] vk: NVN2 program 500b27d6180ac176 (stage 0) lowered with extended registers (1134 REGMAP words)
[hos host] vk: NVN2 program 9039474cf8f30ef4 (stage 0) lowered with extended registers (940 REGMAP words)
[hos host] vk: slow submit: 21.0 ms: draw 18.7 descriptors 7.3 texture lookups 6.1 uploads 1.4 programs 10.4
[hos host] vk: 395 new ASTC textures left for a later frame (this frame's uploads took their budget; 1443 so far)
[hos ModuleSystemWorker1] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 3.4 ms; 7 in all, 0 MB, 12 ms)
[hos host] vk: slow upload 20.9 ms: #593 0x10b7b09c00 512x512x1 x1 levels 10 NVN 0x8d ASTC (BC3), 123 of 123 KB
[hos host] vk: slow submit: 31.7 ms: draw 30.1 descriptors 25.0 texture lookups 22.0 uploads 21.7 programs 1.8 commands 2.4
[hos host] vk: 332 new ASTC textures left for a later frame (this frame's uploads took their budget; 1775 so far)
[hos host] vk: slow submit: 30.6 ms: draw 24.2 descriptors 17.7 texture lookups 14.0 uploads 13.4 programs 2.5 commands 2.7
[hos host] vk: NVN2 program 2251fa487a08a403 (stage 0) lowered with extended registers (1136 REGMAP words)
[hos host] vk: NVN2 program c30cc196df83447f (stage 0) lowered with extended registers (1128 REGMAP words)
[hos host] vk: NVN2 program c2a34f499fd72642 (stage 0) lowered with extended registers (1127 REGMAP words)
[hos host] vk: NVN2 program 00528550fc06c30c (stage 0) lowered with extended registers (1093 REGMAP words)
[hos host] vk: NVN2 program 847f64de27c1dc0c (stage 0) lowered with extended registers (1129 REGMAP words)
[hos host] vk: slow submit: 38.9 ms: draw 37.7 descriptors 19.7 texture lookups 17.5 uploads 4.2 programs 16.1 commands 1.3
[hos host] vk: 433 new ASTC textures left for a later frame (this frame's uploads took their budget; 2208 so far)
[hos ModuleSystemWorker3] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 26.6 ms; 8 in all, 0 MB, 39 ms)
[hos host] vk: slow upload 18.2 ms: #375 0x10d5789a00 341x341x1 x1 levels 9 NVN 0x8c ASTC, 73 of 73 KB
[hos host] vk: slow submit: 37.2 ms: draw 31.9 descriptors 24.9 texture lookups 20.7 uploads 19.9 programs 2.4 commands 3.4
[hos host] vk: texture 0x10afe8cc00 576x576x1 x1 levels 10 NVN 0x7b -> vk 37 (ASTC decoded) (1500 textures, 2876 MB)
[hos host] vk: replay reads of guest memory as submitted: 20000 (12380 pages from the captures)
[hos host] vk: slow submit: 23.2 ms: draw 22.1 descriptors 14.7 texture lookups 12.7 uploads 2.7 programs 5.7 commands 1.3
[hos host] vk: 522 new ASTC textures left for a later frame (this frame's uploads took their budget; 2730 so far)
[hos host] vk: texture #8 0x10755b4000 1920x1080x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 383 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 85 textures retired (1545 left, 2925 MB): budget 85, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 7 MB
[hos host] vk: late pages' fix: stamps for 0x10387f0000 (pool 0x1038400000; 128 blocks, 8192 KB)
[hos host] vk: late pages' fix: stamps for 0x103d7f0000 (pool 0x103d400000; 192 blocks, 12288 KB)
[hos host] vk: late pages' fix: stamps for 0x117d5f0000 (pool 0x1178400000; 256 blocks, 16384 KB)
[hos host] vk: late pages' fix: stamps for 0x117dbf0000 (pool 0x1178400000; 320 blocks, 20480 KB)
[hos host] vk: late pages' fix: stamps for 0x117e5f0000 (pool 0x1178400000; 384 blocks, 24576 KB)
[hos host] vk: slow submit: 25.3 ms: draw 22.7 descriptors 17.7 texture lookups 16.1 uploads 15.7 programs 1.2 commands 2.7
[hos host] vk: slow submit: 16.7 ms: draw 12.4 descriptors 5.9 texture lookups 5.6 programs 4.6 commands 1.3
[hos host] vk: 431 new ASTC textures left for a later frame (this frame's uploads took their budget; 3161 so far)
[hos ModuleSystemWorker2] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.5 ms; 9 in all, 0 MB, 39 ms)
[hos host] vk: slow submit: 19.8 ms: draw 17.8 descriptors 14.7 texture lookups 13.2 uploads 12.9 commands 1.6
[hos host] vk: slow submit: 17.7 ms: draw 11.9 descriptors 5.2 texture lookups 4.6 programs 4.0 commands 2.0
[hos host] vk: 478 new ASTC textures left for a later frame (this frame's uploads took their budget; 3639 so far)
[hos host] vk: late pages' fix: stamps for 0x10383f0000 (pool 0x1038000000; 448 blocks, 28672 KB)
[hos host] vk: slow upload 23.6 ms: #1535 0x10bd32f400 576x576x1 x1 levels 10 NVN 0x7e ASTC (BC3), 187 of 187 KB
[hos host] vk: slow submit: 34.9 ms: draw 32.6 descriptors 27.3 texture lookups 25.0 uploads 24.6 programs 4.0
[hos host] vk: 359 new ASTC textures left for a later frame (this frame's uploads took their budget; 3998 so far)
[hos ModuleSystemWorker1] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 5.0 ms; 10 in all, 0 MB, 44 ms)
[hos host] vk: slow upload 20.8 ms: #1808 0x10b3bbcc00 576x576x1 x1 levels 10 NVN 0x7e ASTC (BC3), 187 of 187 KB
[hos host] vk: slow submit: 30.2 ms: draw 28.8 descriptors 25.9 texture lookups 23.2 uploads 22.7 programs 1.1 commands 1.3
[hos host] vk: 344 new ASTC textures left for a later frame (this frame's uploads took their budget; 4342 so far)
[hos host] vk: slow submit: 24.4 ms: draw 22.4 descriptors 16.3 texture lookups 12.5 uploads 12.2 programs 1.9 commands 3.0
[hos host] vk: 430 new ASTC textures left for a later frame (this frame's uploads took their budget; 4772 so far)
[hos ModuleSystemWorker2] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.3 ms; 11 in all, 0 MB, 45 ms)
[hos host] vk: slow upload 21.5 ms: #1560 0x10b2853c00 576x576x1 x1 levels 10 NVN 0x7e ASTC (BC3), 187 of 187 KB
[hos host] vk: slow submit: 33.1 ms: draw 29.9 descriptors 26.3 texture lookups 22.9 uploads 22.3 programs 1.1 commands 1.7
[hos host] vk: 323 new ASTC textures left for a later frame (this frame's uploads took their budget; 5095 so far)
[hos host] vk: slow submit: 19.4 ms: draw 18.7 descriptors 15.6 texture lookups 13.9 uploads 13.5 commands 1.6
[hos host] vk: 356 new ASTC textures left for a later frame (this frame's uploads took their budget; 5451 so far)
[hos ModuleSystemWorker3] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 2.6 ms; 12 in all, 0 MB, 47 ms)
[hos host] vk: slow upload 17.1 ms: #1361 0x10cab48200 576x576x1 x1 levels 10 NVN 0x7b ASTC, 301 of 301 KB
[hos host] vk: slow submit: 27.3 ms: draw 23.5 descriptors 20.4 texture lookups 18.5 uploads 18.1 programs 1.5 commands 1.1
[hos host] vk: 302 new ASTC textures left for a later frame (this frame's uploads took their budget; 5753 so far)
[hos host] vk: slow upload 19.8 ms: #1378 0x10af4d8a00 768x768x1 x1 levels 10 NVN 0x7e ASTC (BC3), 324 of 324 KB
[hos host] vk: slow submit: 26.9 ms: draw 26.0 descriptors 23.2 texture lookups 21.0 uploads 20.5 programs 1.3 commands 1.1
[hos host] vk: 281 new ASTC textures left for a later frame (this frame's uploads took their budget; 6034 so far)
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #10000: event 0x100d6c8028 mode 0 value 2855 (now 2817) (lr 0x806674b0)
[hos host] vk: slow submit: 22.2 ms: draw 20.5 descriptors 16.4 texture lookups 14.3 uploads 13.9 programs 1.1 commands 2.1
[hos host] vk: 323 new ASTC textures left for a later frame (this frame's uploads took their budget; 6357 so far)
[hos host] vk: replay reads of guest memory as submitted: 40000 (26086 pages from the captures)
[hos host] vk: slow upload 21.2 ms: #1780 0x10afcdb200 768x768x1 x1 levels 10 NVN 0x7e ASTC (BC3), 324 of 324 KB
[hos host] vk: slow submit: 27.7 ms: draw 26.6 descriptors 24.4 texture lookups 22.4 uploads 22.0 commands 1.0
[hos host] vk: 269 new ASTC textures left for a later frame (this frame's uploads took their budget; 6626 so far)
[hos host] vk: slow submit: 20.2 ms: draw 19.0 descriptors 16.5 texture lookups 14.0 uploads 13.5 commands 1.2
[hos host] vk: 244 new ASTC textures left for a later frame (this frame's uploads took their budget; 6870 so far)
[hos host] vk: slow upload 15.2 ms: #1436 0x10b5369400 576x576x1 x1 levels 10 NVN 0x7b ASTC, 301 of 301 KB
[hos host] vk: slow submit: 23.6 ms: draw 23.4 descriptors 20.1 texture lookups 17.0 uploads 16.3 programs 1.0 commands 1.8
[hos host] vk: 223 new ASTC textures left for a later frame (this frame's uploads took their budget; 7093 so far)
[hos host] vk: presenter thread: the copy's fence 3.34 ms, the present 0.60 ms a present
[hos host] vk: slow submit: 27.1 ms: draw 26.7 descriptors 21.2 texture lookups 18.4 uploads 15.8 programs 3.5 commands 1.4
[hos host] vk: 215 new ASTC textures left for a later frame (this frame's uploads took their budget; 7308 so far)
[hos host] vk: slow submit: 18.1 ms: draw 17.0 descriptors 14.6 texture lookups 12.5 uploads 12.0 commands 1.2
[hos host] vk: 187 new ASTC textures left for a later frame (this frame's uploads took their budget; 7495 so far)
[hos host] vk: slow submit: 22.0 ms: draw 20.2 descriptors 16.9 texture lookups 14.9 uploads 14.4 programs 1.7 commands 1.1
[hos host] vk: 172 new ASTC textures left for a later frame (this frame's uploads took their budget; 7667 so far)
[hos host] vk: slow submit: 19.4 ms: draw 17.2 descriptors 14.9 texture lookups 13.1 uploads 12.6
[hos host] vk: 150 new ASTC textures left for a later frame (this frame's uploads took their budget; 7817 so far)
[hos host] vk: slow submit: 21.7 ms: draw 19.2 descriptors 16.4 texture lookups 14.3 uploads 13.8 programs 1.1 commands 1.2
[hos host] vk: 124 new ASTC textures left for a later frame (this frame's uploads took their budget; 7941 so far)
[hos host] vk: slow submit: 26.7 ms: draw 25.5 descriptors 22.9 texture lookups 20.6 uploads 20.2 commands 1.3
[hos host] vk: 90 new ASTC textures left for a later frame (this frame's uploads took their budget; 8031 so far)
[hos host] vk: slow submit: 22.7 ms: draw 21.4 descriptors 18.6 texture lookups 16.3 uploads 15.9 commands 1.2
[hos host] vk: 71 new ASTC textures left for a later frame (this frame's uploads took their budget; 8102 so far)
[hos host] vk: slow submit: 18.8 ms: draw 19.0 descriptors 15.7 texture lookups 13.9 uploads 13.5 commands 1.6
[hos host] vk: 98 new ASTC textures left for a later frame (this frame's uploads took their budget; 8200 so far)
[hos host] vk: sparse mirrors: 67, 933 MB with memory of 1184 MB (250 MB spared); blocks given memory later 71, given back 1026; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 103 binds (16.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 13 taken from the spares, 74 made new; 20 kept as spares, 0 destroyed; 7 spares now, 288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 537, 2529.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3543 dirty pages from their submits' captures, 50023 from earlier ones', 14 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 834 dispatches, 35231 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5135 runs, 1068.8 MB compared; 1183 pages read into the shadows; 196 MB of shadows (112 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8811 in captures, 174 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool: 0x7fdf000000 (12 MB, captured, state 2): 7; 0x7fd4c00000 (24 MB, captured, state 2): 7;
[hos host] vk: imports: 70 live (1272 MB), 18 retired still imported (380 MB), 0 retired given back (0 MB)
[hos host] vk: slow submit: 20.4 ms: draw 18.8 descriptors 15.8 texture lookups 13.2 uploads 12.8 programs 1.1 commands 1.4
[hos host] vk: 51 new ASTC textures left for a later frame (this frame's uploads took their budget; 8251 so far)
[hos host] vk: replay reads of guest memory as submitted: 60000 (40507 pages from the captures)
[hos host] vk: slow submit: 23.4 ms: draw 22.9 descriptors 19.0 texture lookups 15.0 uploads 13.9 programs 1.5 commands 1.8
[hos host] vk: 52 new ASTC textures left for a later frame (this frame's uploads took their budget; 8303 so far)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5796 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos OptimizeChangeBanc] nvn: 150000 recordings made, 53206 live copies (26.3 MB), 259 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: slow submit: 23.1 ms: draw 21.0 descriptors 17.2 texture lookups 15.2 uploads 14.8 commands 2.1
[hos host] vk: 75 new ASTC textures left for a later frame (this frame's uploads took their budget; 8378 so far)
[hos host] vk: late pages found at the replay (600 takes): 695 pages in 10 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: slow submit: 21.4 ms: draw 19.5 descriptors 17.0 texture lookups 14.6 uploads 14.3 commands 1.2
[hos host] vk: 32 new ASTC textures left for a later frame (this frame's uploads took their budget; 8410 so far)
[hos host] vk: slow submit: 20.0 ms: draw 19.1 descriptors 16.5 texture lookups 15.1 uploads 15.0 commands 1.4
[hos host] vk: 12 new ASTC textures left for a later frame (this frame's uploads took their budget; 8422 so far)
[hos host] vk: slow upload 29.1 ms: #1979 0x10d026cc00 341x341x1 x1 levels 9 NVN 0x8d ASTC, 62 of 62 KB
[hos host] vk: slow submit: 46.9 ms: draw 46.2 descriptors 38.1 texture lookups 34.7 uploads 31.7 programs 5.8 commands 1.7
[hos host] vk: 4 new ASTC textures left for a later frame (this frame's uploads took their budget; 8426 so far)
[hos host] vk: program picks: 115897 kept (42840 by the last picks' textures, a uniform buffer bound since), 24784 new, 59318 with other textures (of 200000)
[hos host] vk: replay reads of guest memory as submitted: 80000 (54583 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 100000 (68581 pages from the captures)
[hos host] vk: memory: working set 12370 MB (peak 12387), private 19295 MB, staging 256 MB in 16 arenas, 17 made so far (56 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 624 MB, 36 allocations 558 MB; used 2296 of 15579 MB; heap 1 (device): 168 blocks 5104 MB, 17371 allocations 4829 MB; used 6915 of 7823 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 16 blocks 512 MB, 108 images 405 MB; 71 free ranges, the largest 16 MB
[hos host] vk: image blocks (textures, compacted): 62 blocks 1984 MB, 1685 images 1979 MB; 82 free ranges, the largest 4 MB
[hos host] vk: render target images: 57 reused from the pool, 174 made new, 64 destroyed from it, 10 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 1 missed after their image was taken by size; it holds 12 (64 MB)
[hos host] vk: render targets not reused by size: 154 with the pool empty (0 of them with targets still in the dead queue), 2 with none in it a present yet, 0 with none of a kind they can use, 0 with all smaller (the largest 0% short on average), 10 with all bigger (over 1.5 times), 8 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 64 idle
[hos host] vk: VRAM: 6915 MB used of 7823 MB allowed (275 MB of it free in the blocks); textures 3062 MB (budget 2900 MB), 1787 textures
[hos host] vk: budget: VRAM 5104 of 7823 MB (6915 MB by the driver, 1810 MB unaccounted): textures 3062, mirrors 933, blocks free 275, staging 256, other buffers 205, shadows 196, other images 111, target pool 64; RAM 2236 MB (the process's private memory 19295 MB): imports 1272, capture spares 416, retired imports 380, captures 118, capture table 26, other buffers 24, others 0
[hos host] vk: textures: decoded ASTC 1796 MB, rendered 861 MB (73 drawn in one frame, kept: 51753 KB), other 365 MB; used in the last 8 presents 2682 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 358 MB of blocks hashed in 38 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0xa, last used 67 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 13 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 1 (2 KB): 8x8x1 NVN 0x3f: 1 (2 KB);
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.16
[hos Presentation Thread] vk: presents after the CPU's signal (600): 429 waited for it, 4.94 ms each, 1 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 1079.0 ms; VRAM 6.8 of 7.6 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 23 program records loaded (8.3 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 4810 queued, 4810 loaded (median 1.25 ms, 99th percentile 17.14 ms, worst 34.27 ms, 11581 ms and 267.7 MB read in all); 6641 programs (1430 of them compute variants without their code, 0 graphics programs no note names), 9515 graphics pipelines (1 not in a cache), 576 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 985 pipelines made at create taken by draws; 2185 create calls with nothing kept for their stages, 0 before the index; units 6995 (4810 ready, 0 loading), 12768 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 10218 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 574 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before)
[hos host] vk: 7535 programs from the program cache
[hos host] vk: a frame: 221 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 27.09, present 0.13, draw 18.69, descriptors 13.27, texture lookups 11.88, uploads 9.46, pipelines 0.26, programs 1.95, commands 2.63, dispatch 2.37, memory sync 1.04, queue submit 0.02, slot wait 0.30
[hos host] vk: a frame: mirrors 87 (933 MB): 13.41 MB synced in 559.7 copies, 15.36 MB written back
[hos host] vk: 2400 presents, 25.8 fps; a frame: 1579 draws (0 skipped), 79 dispatches, 94 passes, 98 barriers, 6 submits, 13.2 uploads (6.2 MB), 59 ub updates (39 overlays), GPU-bound wait 0.0 ms (recording slots only); 11303 pipelines (2506 ms), 0 programs (10 failed), 1787 textures (3062 MB)
[hos host] vk: host time in GPU resource waits: 23.88 ms a frame (fences 11.48, timelines 12.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 426.4 full barriers recorded, 173.2 left out (nothing recorded since the one before), 10.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.5; 0x8: 1.8; 0x9: 0.6; 0x10: 1.4; 0x11: 27.8; 0x12: 48.4; 0x13: 1.1; 0x21: 2.2; 0x28: 9.1; 0x92: 1.7; 0x97: 1.6;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 19.7 from 13.6 batches at the game's barriers (21.5 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 92.6, an op 1255.2, the end 4.4, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.1; 5: 11.3; 48: 3.4;
[hos host] vk: GPU time 8.0 ms a frame (the longest submission 43.8 ms); 15.8 textures made a frame: colour targets 1.7 (6.5 MB), depth 0.8 (6.3 MB), ASTC or BC 13.3 (17.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 2.77 MB unswizzled on the CPU in 0.47 ms
[hos host] vk: the render thread (600 presents, 17.37 ms a frame): its jobs 9.45 ms a frame (submits 4.71, presents 2.85, others 1.89), idle 7.92; the game waited for room at 213 presents, 1.91 ms each (0.68 a frame; 10.90 jobs queued at a present); 35 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 27.83 ms each; 1 presents held for the CPU's signal, 72.11 ms each (0 let go by RX_VK_HOLD_MS); 488 captures at a CPU signal; the submits' waits read in 0.012 ms a frame; readbacks: 313 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 9.47 ms after its hand-over (1082.38 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 10.71 + 0.27 (the window's copy is in the idle), after it 0.92 + 1.40; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 876.78 (0.44 ms)
[hos host] vk: frame ahead (600 presents): 1.73 parts a frame recorded in the signal's job, 56.8% of them took nothing that came with the signal; of the others, 0.75 a frame took pages of the signal's capture into a mirror (51.9 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 234 of 489 frames, its first three in 234
[hos host] vk: frame ahead, what came with the signal: 489 signals with a capture, each of 71.7 pages in 2.4 pools; between a frame's last submit and that signal 0.59 texture and sampler registrations, 1.05 releases and 0.00 other CPU signals a signal (37 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.33+0.47, 0.06+0.03, 0.08+0.02, 0.06+0.13, 0.05+0.01, 0.02+0.03, 0.02+0.40, 0.11+0.09, 0.09+0.15, 0.01+0.06, 0.03+0.00, 0.07+0.01
[hos host] vk: pipeline cache: play file 7 saved (363 KB): 8 pipelines, 6 from the driver's cache, 2 compiled; 10231 more from the start's cache
[hos host] vk: write-backs left out (600 submits): 682, 3247.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2063 dirty pages from their submits' captures, 41181 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 742 dispatches, 31867 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5485 runs, 540.0 MB compared; 174 pages read into the shadows; 196 MB of shadows (113 MB of pages known, in 133 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 9478 in captures, 1 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 120000 (82210 pages from the captures)
[hos BaseProcExecutor5] vcores: 124237 takes, 54276 waits (mean 113 us, longest 37.0 ms, 0 over HOS_VCORE_LOG_MS), 25760 kicks, 24949 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: render targets drawn over from another start, let go: 209 so far (968 MB), 199 of them (907 MB) between the passes
[hos host] vk: 13 textures retired (1842 left, 3100 MB): budget 0, drawn over 3, dead targets 10
[hos host] vk: pool 0x103d000000..0x103d800000: mirror made with the GPU's contents of 1 retired pool(s)
[hos host] vk: late pages' fix: stamps for 0x103d3e0000 (pool 0x103d000000; 512 blocks, 32768 KB)
[hos host] vk: mirror of pool 0x103d400000 (4 MB) given back: retired, 58 submissions since its sync, GPU-written
[hos host] vk: 1 textures retired (1862 left, 3111 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: replay reads of guest memory as submitted: 140000 (94738 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.29 ms (lock 0.05, scan 0.10 over 1061 ranges, 3248 MB), 308 pages written, 308 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.010 (4 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (166 MB), 10 spare buffers, 6553 texture pages copied (25 MB)
[hos host] vk: late pages' fix: stamps for 0x103d7f0000 (pool 0x103d000000; 512 blocks, 32768 KB)
[hos host] vk: 2 textures retired (1912 left, 3191 MB): budget 0, drawn over 1, dead targets 1
[hos host] vk: sparse mirrors: 68, 937 MB with memory of 1192 MB (254 MB spared); blocks given memory later 71, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 107 binds (17.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 13 taken from the spares, 76 made new; 21 kept as spares, 0 destroyed; 8 spares now, 292 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 682, 3247.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2224 dirty pages from their submits' captures, 41922 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 744 dispatches, 32151 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5930 runs, 579.0 MB compared; 961 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 10162 in captures, 3 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 71 live (1280 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB)
[hos host] vk: slow submit: 15.5 ms: draw 15.0 descriptors 3.3 programs 9.4 commands 1.7
[hos host] vk: replay reads of guest memory as submitted: 160000 (106714 pages from the captures)
[hos host] vk: budget: gave back 256 MB from staging at 90% of the allowance
[hos host] vk: 2 textures retired (1974 left, 3300 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: program picks: 267399 kept (99085 by the last picks' textures, a uniform buffer bound since), 24880 new, 107720 with other textures (of 400000)
[hos host] vk: texture 0x1000b9a3800 1280x720x1 x1 levels 1 NVN 0x29 -> vk 97 (2000 textures, 3422 MB)
[hos host] vk: texture 0x10be4dca00 1024x1024x1 x1 levels 11 NVN 0x8e -> vk 137 (ASTC decoded) (2000 textures, 3451 MB)
[hos host] vk: replay reads of guest memory as submitted: 180000 (118167 pages from the captures)
[hos host] vk: 12 textures retired (2036 left, 3500 MB): budget 1, drawn over 0, dead targets 11
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 72 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1334, sb 559, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 682, 3247.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1970 dirty pages from their submits' captures, 41215 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 737 dispatches, 31833 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5855 runs, 561.1 MB compared; 124 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 10261 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 200000 (130074 pages from the captures)
[hos host] vk: texture #558 0x10005280e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 126 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #495 0x10004741200 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 126 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #493 0x10004831400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 126 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #479 0x10004dd1a00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 126 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 118 textures retired (1937 left, 3483 MB): budget 117, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 30 MB
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #5000: event 0x100d6c802c mode 0 value 5003 (lr 0x806616b0)
[hos OptimizeChangeBanc] nvn: 200000 recordings made, 65901 live copies (24.6 MB), 533 texture copies; 4 command buffers' segments (4.2 MB)
[hos sead::MainThread] vk: readbacks of GPU results: 816 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 18919 kept their virtual core (nothing there to wait for), 2841 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 42 textures retired (1920 left, 3124 MB): budget 0, drawn over 8, dead targets 34
[hos host] vk: replay reads of guest memory as submitted: 220000 (142585 pages from the captures)
[hos host] vk: texture #2308 0x100091e2e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2494 (drawn at 2433)
[hos Presentation Thread] vi: 2495 frames presented (32.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 2, 15.5-17.8 3, 17.8-22 39, 22-30 67, 30-40 26, over 40 24
[hos host] vk: readback pages refreshed from guest memory: 73115 pages so far (0 held: may be written ahead in their submission)
[hos sead::MainThread] nvn: GPU waits in 5 s: 78 sync waits (640 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: sparse mirrors: 70, 938 MB with memory of 1200 MB (261 MB spared); blocks given memory later 71, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 109 binds (17.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 14 taken from the spares, 77 made new; 21 kept as spares, 0 destroyed; 7 spares now, 288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 727, 3458.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1340 dirty pages from their submits' captures, 44461 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 809 dispatches, 33950 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6139 runs, 596.3 MB compared; 116 pages read into the shadows; 200 MB of shadows (112 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11382 in captures, 2 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool: 0x7fe2800000 (4 MB, captured, state 2): 2;
[hos host] vk: imports: 73 live (1288 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 612 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 134 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] nvn: EventSignal (CPU) #5000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.03, writes collected 0.21
[hos host] vk: replay reads of guest memory as submitted: 240000 (154302 pages from the captures)
[hos host] vk: mirror of pool 0x1072c00000 (68 MB) given back: idle, 6048 submissions since its sync
[hos host] vk: slow submit: 23.3 ms: draw 21.7 descriptors 18.6 texture lookups 17.2 uploads 14.9 programs 1.3 commands 1.3
[hos sead::MainThread] vk: capture: a submit 0.24 ms (lock 0.04, scan 0.08 over 1328 ranges, 3419 MB), 267 pages written, 267 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.003 (1 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (106 MB), 11 spare buffers, 6552 texture pages copied (25 MB)
[hos host] vk: program picks: 416049 kept (153866 by the last picks' textures, a uniform buffer bound since), 24943 new, 159007 with other textures (of 600000)
[hos host] vk: slow submit: 15.9 ms: memory sync 15.8 [memory sync: maps 0.0 (lock waits 0.0), capture 15.6, collect 0.2]
[hos host] vk: replay reads of guest memory as submitted: 260000 (166568 pages from the captures)
[hos host] vk: a frame: 503 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 17.40, present 0.22, draw 10.55, descriptors 4.40, texture lookups 1.07, uploads 0.28, pipelines 0.46, programs 1.93, commands 2.94, dispatch 1.10, memory sync 1.40, queue submit 0.00, slot wait 0.28
[hos host] vk: a frame: mirrors 91 (938 MB): 5.54 MB synced in 1029.5 copies, 18.09 MB written back
[hos host] vk: 2520 presents, 42.1 fps; a frame: 2864 draws (0 skipped), 162 dispatches, 164 passes, 171 barriers, 6 submits, 2.8 uploads (0.8 MB), 88 ub updates (60 overlays), GPU-bound wait 0.1 ms (recording slots only); 11914 pipelines (2757 ms), 0 programs (10 failed), 1991 textures (3159 MB)
[hos host] vk: host time in GPU resource waits: 6.88 ms a frame (fences 4.03, timelines 2.85, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 592.8 full barriers recorded, 294.6 left out (nothing recorded since the one before), 17.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.8; 0x8: 3.5; 0x9: 1.0; 0x10: 4.8; 0x11: 46.1; 0x12: 82.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 34.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 161.8, an op 2320.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 13.5 ms a frame (the longest submission 3.6 ms); 5.2 textures made a frame: colour targets 1.2 (7.2 MB), depth 1.4 (6.5 MB), ASTC or BC 2.6 (2.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.33 MB unswizzled on the CPU in 0.07 ms
[hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2213 dirty pages from their submits' captures, 45673 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 873 dispatches, 35034 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6291 runs, 612.6 MB compared; 76 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11556 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 280000 (178165 pages from the captures)
[hos host] vk: 19 textures retired (1973 left, 3145 MB): budget 19, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 300000 (189793 pages from the captures)
[hos host] vk: 16 textures retired (1957 left, 3139 MB): budget 16, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: CommandBufferWaitEvent #15000: event 0x100d6c8028 mode 0 value 6655 (now 6636) (lr 0x806674b0)
[hos host] vk: sparse mirrors: 70, 938 MB with memory of 1136 MB (197 MB spared); blocks given memory later 71, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 110 binds (17.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 14 taken from the spares, 78 made new; 22 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 726, 3457.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2897 dirty pages from their submits' captures, 44391 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 867 dispatches, 34537 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5788 runs, 588.2 MB compared; 73 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12184 in captures, 1 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 74 live (1292 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 320000 (201516 pages from the captures)
[hos host] vk: 17 textures retired (1952 left, 3138 MB): budget 7, drawn over 0, dead targets 10
[hos host] vk: 16 textures retired (1936 left, 3133 MB): budget 16, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 340000 (213968 pages from the captures)
[hos host] vk: texture #2123 0x10008b23400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2587 (drawn at 2413)
[hos host] vk: 22 textures retired (1915 left, 3122 MB): budget 21, drawn over 0, dead targets 1
[hos host] vk: program picks: 559264 kept (206757 by the last picks' textures, a uniform buffer bound since), 24959 new, 215776 with other textures (of 800000)
[hos host] vk: texture #2124 0x10008bb3600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2594 (drawn at 2413)
[hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4475 dirty pages from their submits' captures, 44014 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 969 dispatches, 35291 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5705 runs, 601.8 MB compared; 74 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11964 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 54 textures retired (1878 left, 3093 MB): budget 44, drawn over 0, dead targets 10
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05
[hos host] vk: texture #2148 0x10008c43800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2600 (drawn at 2413)
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.04 over 1312 ranges, 3448 MB), 227 pages written, 227 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.043; GPU wait for late pages 0.026 (1 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (106 MB), 8 spare buffers, 6552 texture pages copied (25 MB)
[hos host] vk: replay reads of guest memory as submitted: 360000 (226547 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 849 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 23 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 7339 kept their virtual core (nothing there to wait for), 2901 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 18 textures retired (1860 left, 3083 MB): budget 18, drawn over 0, dead targets 0
[hos host] vk: 55 textures retired (1809 left, 3035 MB): budget 55, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 380000 (238983 pages from the captures)
[hos host] vk: handle table: element 25 of program 2fd8c7ca964e5a1d names texture #822 (format 83) for a depth comparison: bound as nothing
[hos host] vk: handle table: element 26 of program 2fd8c7ca964e5a1d names texture #823 (format 83) for a depth comparison: bound as nothing
[hos host] vk: handle table: element 28 of program 2fd8c7ca964e5a1d names texture #823 (format 83) for a depth comparison: bound as nothing
[hos host] vk: handle table: element 25 of program b1b048510819ac68 names texture #822 (format 83) for a depth comparison: bound as nothing
[hos host] vk: handle table: element 26 of program b1b048510819ac68 names texture #823 (format 83) for a depth comparison: bound as nothing
[hos host] vk: handle table: element 28 of program b1b048510819ac68 names texture #823 (format 83) for a depth comparison: bound as nothing
[hos host] vk: handle table: element 25 of program b1b048510819ac68 names texture #822 (format 83) for a depth comparison: bound as nothing
[hos host] vk: handle table: element 26 of program b1b048510819ac68 names texture #823 (format 83) for a depth comparison: bound as nothing
[hos host] vk: recorder: 4441759 packets run in 10.0 s (busy 1.26 s, spun idle 0.10 s, slept 9008 times, 2340 of them under 200 us), 4 calls made directly (no room); 11921 batches over 384 frames (31.0 a frame, 373 packets a batch), 6025 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 297 times, 34.33 ms (0.8 times, 0.09 ms a frame); every thread 734 times, 79.15 ms in all, the longest 4.52 ms; vkWaitForFences 329/31.96 ms (max 4.52), vkEndCommandBuffer 152/14.88 ms (max 0.74), vkDestroyImageView 84/14.70 ms (max 0.64), vkQueueBindSparse 102/8.12 ms (max 0.94), the queue's lock 19/2.91 ms (max 0.56), vkDestroyPipeline 15/2.48 ms (max 0.40)
[hos host] vk: recorder: of its busy 1.26 s, 0.30 s in parts' ends and submits (the longest 4.57 ms), 0.06 s of that blocked on the queue's lock (108 times, the longest 2.68 ms), so 0.24 s in the ends, submits and holds; 0.96 s running the other packets; 11567 packets a frame
[hos host] vk: 6 textures retired (1832 left, 3048 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 83 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1334, sb 559, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 78, 940 MB with memory of 1168 MB (227 MB spared); blocks given memory later 84, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 121 binds (18.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 14 taken from the spares, 86 made new; 22 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3883 dirty pages from their submits' captures, 46412 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1078 dispatches, 35828 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5814 runs, 603.3 MB compared; 90 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12353 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 82 live (1324 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 400000 (251457 pages from the captures)
[hos host] vk: 16 textures retired (1828 left, 3048 MB): budget 6, drawn over 0, dead targets 10
[hos host] vk: a frame: 340 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 13.19, present 0.17, draw 8.79, descriptors 3.82, texture lookups 0.25, uploads 0.02, pipelines 0.38, programs 1.42, commands 2.48, dispatch 1.05, memory sync 0.44, queue submit 0.00, slot wait 0.27
[hos host] vk: a frame: mirrors 103 (941 MB): 5.68 MB synced in 1044.0 copies, 17.77 MB written back
[hos host] vk: 2640 presents, 56.0 fps; a frame: 2425 draws (0 skipped), 149 dispatches, 165 passes, 167 barriers, 6 submits, 0.4 uploads (0.1 MB), 86 ub updates (59 overlays), GPU-bound wait 0.1 ms (recording slots only); 11914 pipelines (2757 ms), 0 programs (10 failed), 1829 textures (3049 MB)
[hos host] vk: host time in GPU resource waits: 4.86 ms a frame (fences 2.24, timelines 2.62, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 545.7 full barriers recorded, 288.0 left out (nothing recorded since the one before), 21.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.8; 0x8: 3.0; 0x9: 1.0; 0x10: 4.8; 0x11: 41.8; 0x12: 84.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.6 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 158.4, an op 1850.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 11.5 ms a frame (the longest submission 3.5 ms); 1.8 textures made a frame: colour targets 1.0 (6.1 MB), depth 0.5 (2.7 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: 16 textures retired (1827 left, 3045 MB): budget 16, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 420000 (263799 pages from the captures)
[hos AudioUserTaskThread] vcores: 135485 takes, 33541 waits (mean 103 us, longest 23.5 ms, 0 over HOS_VCORE_LOG_MS), 16686 kicks, 16782 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 666 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 15 pages in 15 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 60 textures retired (1796 left, 3031 MB): budget 49, drawn over 0, dead targets 11
[hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 35 made): 25122194 lookups, 191012 missed (0.76 %)
[hos host] vk: readback pages refreshed from guest memory: 164054 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: mirror of pool 0x7fd3400000 (4 MB) given back: idle, 6007 submissions since its sync
[hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3708 dirty pages from their submits' captures, 48557 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1197 dispatches, 36190 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5813 runs, 604.4 MB compared; 78 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12683 in captures, 0 only in the collection, 27 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 440000 (276171 pages from the captures)
[hos host] vk: 18 textures retired (1791 left, 3024 MB): budget 8, drawn over 0, dead targets 10
[hos host] vk: program picks: 699778 kept (256124 by the last picks' textures, a uniform buffer bound since), 24979 new, 275242 with other textures (of 1000000)
[hos host] vk: 8 textures retired (1794 left, 3029 MB): budget 8, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 460000 (289091 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 1310 ranges, 3479 MB), 246 pages written, 246 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (142 MB), 12 spare buffers, 6555 texture pages copied (25 MB)
[hos host] vk: 3 textures retired (1822 left, 3054 MB): budget 3, drawn over 0, dead targets 0
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 600 of 3000 waited for their write-back, 0.08 ms each; 0 looked at the stopped queues first, 24 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 4327 kept their virtual core (nothing there to wait for), 3097 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: queues ordered by their events: 12000 submissions stopped at a wait, 11999 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 18341, not met 0; captures at a CPU event's wait 2479, 0 of them at a wait met on arrival
[hos host] vk: 3 textures retired (1846 left, 2867 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: replay reads of guest memory as submitted: 480000 (301643 pages from the captures)
[hos host] vk: sparse mirrors: 83, 945 MB with memory of 1188 MB (242 MB spared); blocks given memory later 85, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 129 binds (18.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 15 taken from the spares, 91 made new; 23 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 747, 3560.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4254 dirty pages from their submits' captures, 48890 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1221 dispatches, 36545 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5552 runs, 599.2 MB compared; 67 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12578 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 88 live (1348 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB)
[hos host] vk: 32 textures retired (1851 left, 2861 MB): budget 0, drawn over 0, dead targets 32
[hos host] vk: replay reads of guest memory as submitted: 500000 (314201 pages from the captures)
[hos host] vk: 20 textures retired (1859 left, 2862 MB): budget 0, drawn over 0, dead targets 20
[hos host] vk: 25 textures retired (1881 left, 2899 MB): budget 25, drawn over 0, dead targets 0
[hos host] vk: mirror of pool 0x1042c00000 (4 MB) given back: idle, 6003 submissions since its sync
[hos host] vk: write-backs left out (600 submits): 727, 3458.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4420 dirty pages from their submits' captures, 48370 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1280 dispatches, 35945 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5313 runs, 580.3 MB compared; 80 pages read into the shadows; 200 MB of shadows (112 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12469 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 520000 (326814 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #10000: event 0x100d6c8028 mode 0 value 9999 (lr 0x80667410)
[hos host] vk: replay reads of guest memory as submitted: 540000 (339417 pages from the captures)
[hos host] vk: program picks: 837584 kept (306999 by the last picks' textures, a uniform buffer bound since), 25017 new, 337398 with other textures (of 1200000)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #20000: event 0x100d6c8028 mode 0 value 10455 (now 10436) (lr 0x806674b0)
[hos host] vk: a frame: 273 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 13.90, present 0.20, draw 9.00, descriptors 3.84, texture lookups 0.60, uploads 0.08, pipelines 0.39, programs 1.52, commands 2.58, dispatch 0.97, memory sync 0.45, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 110 (945 MB): 6.23 MB synced in 1080.3 copies, 17.64 MB written back
[hos host] vk: 2760 presents, 53.4 fps; a frame: 2438 draws (0 skipped), 149 dispatches, 184 passes, 175 barriers, 6 submits, 1.8 uploads (0.4 MB), 85 ub updates (58 overlays), GPU-bound wait 0.1 ms (recording slots only); 11914 pipelines (2757 ms), 0 programs (10 failed), 1928 textures (2908 MB)
[hos host] vk: host time in GPU resource waits: 5.43 ms a frame (fences 2.36, timelines 3.07, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 568.9 full barriers recorded, 293.5 left out (nothing recorded since the one before), 30.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.3; 0x8: 3.0; 0x9: 1.0; 0x10: 4.3; 0x11: 42.4; 0x12: 92.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 166.0, an op 1822.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 11.7 ms a frame (the longest submission 3.2 ms); 6.3 textures made a frame: colour targets 2.6 (14.7 MB), depth 2.2 (7.8 MB), ASTC or BC 1.6 (1.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.17 MB unswizzled on the CPU in 0.03 ms
[hos host] vk: texture #866 0x10009154000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2763 (drawn at 2351)
[hos Presentation Thread] vi: 2766 frames presented (54.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 29, 15.5-17.8 111, 17.8-22 107, 22-30 20, 30-40 3, over 40 1
[hos host] vk: sparse mirrors: 86, 945 MB with memory of 1200 MB (254 MB spared); blocks given memory later 86, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 135 binds (19.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 15 taken from the spares, 95 made new; 24 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: replay reads of guest memory as submitted: 560000 (351936 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 747, 3560.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4440 dirty pages from their submits' captures, 51312 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1385 dispatches, 37467 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5359 runs, 596.5 MB compared; 70 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12287 in captures, 0 only in the collection, 20 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 92 live (1364 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 229 sync waits (614 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos sead::MainThread] vk: readbacks of GPU results: 510 of 3000 waited for their write-back, 0.09 ms each; 0 looked at the stopped queues first, 20 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 2934 kept their virtual core (nothing there to wait for), 2698 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 18 textures retired (1942 left, 2952 MB): budget 0, drawn over 0, dead targets 18
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 1276 ranges, 3527 MB), 259 pages written, 259 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (128 MB), 10 spare buffers, 6564 texture pages copied (25 MB)
[hos host] vk: texture 0x10aec87200 576x576x1 x1 levels 10 NVN 0x7b -> vk 37 (ASTC decoded) (2000 textures, 3049 MB)
[hos host] vk: reaper: 817 images (1760 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 116 (258 MB)
[hos host] vk: texture #38 0x1001efd9000 631x94x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1514 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #80 0x1001eccba00 1466x282x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 821 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #7 0x1000021c200 960x540x1 x1 NVN 0xd the GPU drew goes with no copy of its texels (finalized, unused for 813 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #9 0x10000348400 2560x1440x1 x1 NVN 0x38 the GPU drew goes with no copy of its texels (finalized, unused for 813 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 34 textures retired (1995 left, 2964 MB): budget 33, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 16 MB
[hos host] vk: texture 0x1000a535400 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 2987 MB)
[hos host] vk: texture 0x1000a5c5600 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 3000 MB)
[hos host] vk: texture 0x1000a655800 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 3012 MB)
[hos host] vk: replay reads of guest memory as submitted: 580000 (364547 pages from the captures)
[hos host] vk: texture 0x1000a985800 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 3037 MB)
[hos host] vk: texture 0x1000a295600 640x360x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3062 MB)
[hos host] vk: 13 textures retired (1999 left, 3088 MB): budget 11, drawn over 2, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 600000 (377139 pages from the captures)
[hos host] vk: texture 0x10d18f6800 768x768x1 x1 levels 10 NVN 0x49 -> vk 139 (2000 textures, 3089 MB)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 86 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1334, sb 559, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 704, 3352.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4518 dirty pages from their submits' captures, 51415 from earlier ones', 12 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1400 dispatches, 36553 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 4814 runs, 541.8 MB compared; 73 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 133 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11861 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool: 0x1054000000 (4 MB, captured, state 2): 6; 0x1053800000 (4 MB, captured, state 2): 6;
[hos host] vk: 3 textures retired (1998 left, 3088 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: texture 0x10008c43800 1920x1920x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3102 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 632 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 74 pages in 20 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture 0x1000a4d5e00 640x360x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3019 MB)
[hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 20000 so far (x1 0x117daa3070 x2 0 x3 0x117daa3040 x4 0x9b040)
[hos host] vk: texture 0x10008bb3600 1693x1693x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3045 MB)
[hos host] vk: texture 0x1000a775c00 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 2977 MB)
[hos host] vk: 16 textures retired (1990 left, 2961 MB): budget 5, drawn over 0, dead targets 11
[hos host] vk: texture 0x100099c4a00 384x384x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 2979 MB)
[hos OptimizeChangeBanc] nvn: 250000 recordings made, 68472 live copies (23.7 MB), 1083 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: replay reads of guest memory as submitted: 620000 (389759 pages from the captures)
[hos host] vk: program picks: 979481 kept (364878 by the last picks' textures, a uniform buffer bound since), 25038 new, 395480 with other textures (of 1400000)
[hos host] vk: texture 0x10009273e00 384x384x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3025 MB)
[hos host] vk: 26 textures retired (1973 left, 3014 MB): budget 26, drawn over 0, dead targets 0
[hos host] vk: mirror of pool 0x7f54000000 (4 MB) given back: idle, 6006 submissions since its sync
[hos host] vk: texture 0x10d2b5de00 1536x1536x1 x1 levels 11 NVN 0x8d -> vk 137 (ASTC decoded) (2000 textures, 3044 MB)
[hos host] vk: slow submit: 24.4 ms: draw 22.2 descriptors 16.1 texture lookups 13.3 uploads 12.8 programs 4.1 commands 1.5
[hos host] vk: replay reads of guest memory as submitted: 640000 (402940 pages from the captures)
[hos host] vk: 2 textures retired (2030 left, 3066 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 298110 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 95, 950 MB with memory of 1236 MB (285 MB spared); blocks given memory later 89, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 148 binds (21.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 17 taken from the spares, 103 made new; 25 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 705, 3354.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 5403 dirty pages from their submits' captures, 53888 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1544 dispatches, 37776 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5279 runs, 579.2 MB compared; 142 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12154 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 102 live (1404 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 437 of 3000 waited for their write-back, 0.10 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 2372 kept their virtual core (nothing there to wait for), 3004 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: (its written part: 4 of 4 KB readable, 1014 bytes not zero; levels L0)
[hos host] vk: texture 0x107cd82600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it
[hos host] vk: 10 textures retired (2031 left, 3062 MB): budget 9, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 660000 (415527 pages from the captures)
[hos host] vk: 32 textures retired (1999 left, 3047 MB): budget 31, drawn over 1, dead targets 0
[hos host] vk: texture 0x10cde3fe00 171x171x1 x1 levels 8 NVN 0x8e -> vk 37 (ASTC decoded) (2000 textures, 3047 MB)
[hos host] vk: 26 textures retired (1976 left, 3030 MB): budget 26, drawn over 0, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.22 ms (lock 0.02, scan 0.05 over 688 ranges, 3337 MB), 332 pages written, 332 copied (1.3 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (126 MB), 9 spare buffers, 6576 texture pages copied (25 MB)
[hos host] vk: replay reads of guest memory as submitted: 680000 (428168 pages from the captures)
[hos host] vk: texture 0x10009875c00 2560x1440x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3065 MB)
[hos host] vk: write-backs left out (600 submits): 704, 3352.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3537 dirty pages from their submits' captures, 52433 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1471 dispatches, 36828 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5269 runs, 561.9 MB compared; 96 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12180 in captures, 1 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 5 textures retired (1974 left, 2977 MB): budget 0, drawn over 3, dead targets 2
[hos host] vk: texture 0x1000a9b5c00 1280x720x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3034 MB)
[hos host] vk: replay reads of guest memory as submitted: 700000 (440582 pages from the captures)
[hos host] vk: texture 0x10009e15800 2560x1440x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3019 MB)
[hos host] vk: 21 textures retired (1984 left, 2989 MB): budget 0, drawn over 0, dead targets 21
[hos host] vk: 8179 programs from the program cache
[hos host] vk: a frame: 366 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 17.33, present 0.24, draw 10.58, descriptors 4.54, texture lookups 0.80, uploads 0.17, pipelines 0.46, programs 1.74, commands 3.03, dispatch 1.20, memory sync 1.00, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 120 (950 MB): 6.94 MB synced in 1149.0 copies, 17.54 MB written back
[hos host] vk: 2880 presents, 45.4 fps; a frame: 2851 draws (0 skipped), 157 dispatches, 202 passes, 182 barriers, 6 submits, 1.6 uploads (0.3 MB), 82 ub updates (57 overlays), GPU-bound wait 0.1 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1984 textures (2995 MB)
[hos host] vk: host time in GPU resource waits: 6.49 ms a frame (fences 2.55, timelines 3.94, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 595.6 full barriers recorded, 305.3 left out (nothing recorded since the one before), 37.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 44.8; 0x12: 99.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 34.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 173.1, an op 2099.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 13.3 ms a frame (the longest submission 3.9 ms); 6.4 textures made a frame: colour targets 2.1 (11.8 MB), depth 2.9 (9.9 MB), ASTC or BC 1.4 (1.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.12 MB unswizzled on the CPU in 0.02 ms
[hos host] vk: texture 0x10009b74a00 384x384x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3044 MB)
[hos host] vk: program picks: 1118828 kept (416106 by the last picks' textures, a uniform buffer bound since), 25070 new, 456101 with other textures (of 1600000)
[hos host] vk: texture #4177 0x10009815c00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2888 (drawn at 2827)
[hos host] vk: 23 textures retired (1972 left, 3052 MB): budget 23, drawn over 0, dead targets 0
[hos host] vk: texture #4209 0x10009665c00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2891 (drawn at 2830)
[hos host] vk: texture 0x1000b2b5600 1280x720x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3139 MB)
[hos sead::MainThread] vk: readbacks of GPU results: 406 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1845 kept their virtual core (nothing there to wait for), 3019 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: sparse mirrors: 95, 950 MB with memory of 1236 MB (285 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 158 binds (21.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 17 taken from the spares, 103 made new; 25 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 747, 3560.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3706 dirty pages from their submits' captures, 53527 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1448 dispatches, 38187 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5531 runs, 580.7 MB compared; 86 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13259 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 102 live (1404 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB)
[hos host] vk: replay reads of guest memory as submitted: 720000 (452502 pages from the captures)
[hos tera::FileResourceMgr] vcores: 145002 takes, 37873 waits (mean 121 us, longest 48.0 ms, 0 over HOS_VCORE_LOG_MS), 18392 kicks, 18543 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.15
[hos host] vk: late pages' fix (600 recordings with one): of 803 captures with late pages 797 fixed (47.0 pages, 47.0 fixes a capture), waited as before: 0 no pool, 6 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 242 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 2 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 130, of an earlier recording 110, not recorded yet 0, of this part 0; pages fixed from stamps 36833, from snapshots 587; the shadows: 210447 pages synced from captures, 203319 made known by it, 7069 left unknown because the replay or the collection wrote them since
[hos host] vk: replay reads of guest memory as submitted: 740000 (464479 pages from the captures)
[hos host] vk: texture #1870 0x1001f023c00 870x133x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 196 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 13 textures retired (1943 left, 3041 MB): budget 11, drawn over 2, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: texture #4208 0x100095d5a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2921 (drawn at 2860)
[hos host] vk: texture #4305 0x10009785a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2925 (drawn at 2864)
[hos host] vk: 8 textures retired (1963 left, 3234 MB): budget 8, drawn over 0, dead targets 0
[hos host] vk: write-backs left out (600 submits): 749, 3563.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3370 dirty pages from their submits' captures, 49692 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1309 dispatches, 36374 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5621 runs, 612.4 MB compared; 71 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13029 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: texture #4351 0x10009935600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2935 (drawn at 2874)
[hos host] vk: replay reads of guest memory as submitted: 760000 (476403 pages from the captures)
[hos host] vk: 4 textures retired (1971 left, 3354 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: texture #4397 0x10009c95400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2945 (drawn at 2884)
[hos host] vk: mirror of pool 0x7fe1800000 (4 MB) given back: idle, 6015 submissions since its sync
[hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 764 ranges, 3369 MB), 269 pages written, 269 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (112 MB), 11 spare buffers, 6576 texture pages copied (25 MB)
[hos host] vk: 84 textures retired (1899 left, 3382 MB): budget 84, drawn over 0, dead targets 0
[hos host] vk: texture #4441 0x10009db5200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2952 (drawn at 2891)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 409 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1928 kept their virtual core (nothing there to wait for), 3192 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 637 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 5 pages in 5 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #4512 0x1000a3e5000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2957 (drawn at 2896)
[hos host] vk: replay reads of guest memory as submitted: 780000 (488380 pages from the captures)
[hos host] vk: commit: 2476 MB left under Windows' limit as it is now, and the page files can grow by 80896 MB more: counted as room
[hos sead::MainThread] nvn: CommandBufferWaitEvent #25000: event 0x100d6c8028 mode 0 value 14255 (now 14236) (lr 0x806674b0)
[hos host] vk: presenter thread: the copy's fence 2.34 ms, the present 0.84 ms a present
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 86 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1334, sb 559, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 95, 950 MB with memory of 1236 MB (285 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 160 binds (21.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 17 taken from the spares, 104 made new; 26 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3174 dirty pages from their submits' captures, 51226 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1327 dispatches, 37697 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5683 runs, 626.1 MB compared; 85 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13427 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 103 live (1408 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB)
[hos host] vk: 168 textures retired (1754 left, 3461 MB): budget 168, drawn over 0, dead targets 0
[hos host] vk: texture #4588 0x1000a354a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2973 (drawn at 2912)
[hos host] vk: mirror of pool 0x7fe2800000 (4 MB) given back: idle, 6063 submissions since its sync
[hos host] vk: program picks: 1259712 kept (464759 by the last picks' textures, a uniform buffer bound since), 25070 new, 515217 with other textures (of 1800000)
[hos host] vk: replay reads of guest memory as submitted: 800000 (500296 pages from the captures)
[hos host] vk: 17 textures retired (1748 left, 3384 MB): budget 0, drawn over 17, dead targets 0
[hos host] vk: 29 textures retired (1725 left, 3328 MB): budget 18, drawn over 11, dead targets 0
[hos host] vk: mirror of pool 0x7fe2400000 (4 MB) given back: idle, 6018 submissions since its sync
[hos sead::MainThread] nvn: CommandBufferSignalEvent #15000: event 0x100d6c8034 mode 0 value 14999 (lr 0x806616b0)
[hos host] vk: replay reads of guest memory as submitted: 820000 (512282 pages from the captures)
[hos host] vk: memory: working set 13072 MB (peak 13092), private 20589 MB, staging 32 MB in 2 arenas, 19 made so far (56 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 624 MB, 36 allocations 548 MB; used 2416 of 15579 MB; heap 1 (device): 196 blocks 5687 MB, 17640 allocations 4987 MB; used 7504 of 7868 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 42 blocks 1344 MB, 191 images 1068 MB; 171 free ranges, the largest 25 MB
[hos host] vk: image blocks (textures, compacted): 69 blocks 2208 MB, 1545 images 1955 MB; 320 free ranges, the largest 11 MB
[hos host] vk: render target images: 1416 reused from the pool, 688 made new, 495 destroyed from it, 138 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 25 missed after their image was taken by size; it holds 0 (0 MB), VRAM short: halved
[hos host] vk: render targets not reused by size: 324 with the pool empty (0 of them with targets still in the dead queue), 50 with none in it a present yet, 13 with none of a kind they can use, 79 with all smaller (the largest 57% short on average), 86 with all bigger (over 1.5 times), 136 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 10 over its bound and 485 idle
[hos host] vk: VRAM: 7504 MB used of 7868 MB allowed (699 MB of it free in the blocks); textures 3483 MB (budget 3145 MB), 1740 textures
[hos host] vk: budget: VRAM 5687 of 7868 MB (7504 MB by the driver, 1817 MB unaccounted): textures 3483, mirrors 949, blocks free 699, other buffers 211, shadows 200, other images 111, staging 32; RAM 2346 MB (the process's private memory 20589 MB): imports 1408, capture spares 422, retired imports 364, captures 108, capture table 26, other buffers 18, others 0
[hos host] vk: textures: decoded ASTC 1600 MB, rendered 1457 MB (65 drawn in one frame, kept: 43444 KB), other 278 MB; used in the last 8 presents 2571 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 457 MB of blocks hashed in 49 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 1 presents ago
[hos host] vk: 21 MB: 0x10006fce600 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x10005192200 1600x1600x1 x3 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 24 (8683 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 64x62x1 NVN 0x25: 1 (15 KB); 73x33x1 NVN 0x25: 1 (9 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 128x126x1 NVN 0x1: 1 (15 KB); 137x97x1 NVN 0x1: 1 (12 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 232x33x1 NVN 0x25: 1 (29 KB); 262x104x1 NVN 0x1: 1 (26 KB); 296x97x1 NVN 0x1: 1 (28 KB); 592x49x1 NVN 0x25: 2 (226 KB);
[hos host] vk: 34 textures retired (1706 left, 3376 MB): budget 22, drawn over 12, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos Presentation Thread] vk: presents after the CPU's signal (600): 488 waited for it, 2.99 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 48.7 ms; VRAM 7.3 of 7.6 GB
[hos host] vk: readback pages refreshed from guest memory: 461078 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 24 program records loaded (9.5 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5253 queued, 5253 loaded (median 1.38 ms, 99th percentile 16.77 ms, worst 34.27 ms, 13012 ms and 296.2 MB read in all); 7285 programs (1431 of them compute variants without their code, 0 graphics programs no note names), 10490 graphics pipelines (1 not in a cache), 577 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1275 pipelines made at create taken by draws; 2569 create calls with nothing kept for their stages, 0 before the index; units 7822 (5253 ready, 0 loading), 14420 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 11458 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 575 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 307 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.51, present 0.23, draw 7.52, descriptors 3.19, texture lookups 0.88, uploads 0.01, pipelines 0.34, programs 1.23, commands 2.18, dispatch 0.93, memory sync 0.36, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 121 (949 MB): 6.05 MB synced in 1061.9 copies, 17.44 MB written back
[hos host] vk: 3000 presents, 59.0 fps; a frame: 2152 draws (0 skipped), 149 dispatches, 186 passes, 173 barriers, 6 submits, 0.1 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1703 textures (3389 MB)
[hos host] vk: host time in GPU resource waits: 3.43 ms a frame (fences 1.82, timelines 1.61, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 561.9 full barriers recorded, 295.0 left out (nothing recorded since the one before), 29.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 41.3; 0x12: 93.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 34.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 163.8, an op 1434.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 11.1 ms a frame (the longest submission 3.2 ms); 3.6 textures made a frame: colour targets 0.8 (4.7 MB), depth 2.8 (13.1 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: the render thread (600 presents, 19.86 ms a frame): its jobs 18.75 ms a frame (submits 4.02, presents 3.53, others 11.21), idle 1.10; the game waited for room at 14 presents, 0.21 ms each (0.00 a frame; 11.04 jobs queued at a present); 58 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 8.12 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 492 captures at a CPU signal; the submits' waits read in 0.062 ms a frame; readbacks: 2768 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 17.58 ms after its hand-over (40.94 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.20 + 0.21 (the window's copy is in the idle), after it 4.01 + 1.24; after a frame's last recording that reads or writes the game's memory come 1.04 recordings (0.00 ms of GPU time), after its last that reads it 1.04 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 11.08 parts a frame recorded in the signal's job, 62.3% of them took nothing that came with the signal; of the others, 4.18 a frame took pages of the signal's capture into a mirror (127.2 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 20 of 492 frames, its first three in 14
[hos host] vk: frame ahead, what came with the signal: 492 signals with a capture, each of 209.2 pages in 23.4 pools; between a frame's last submit and that signal 0.53 texture and sampler registrations, 0.74 releases and 0.00 other CPU signals a signal (71 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.72+0.25, 0.29+0.06, 0.44+0.07, 0.64+0.12, 0.33+0.10, 0.18+0.18, 0.14+0.16, 0.08+0.09, 0.09+0.13, 0.05+0.07, 0.03+0.02, 0.01+0.01
[hos host] vk: VRAM compaction 1 begins: the textures' blocks hold 2208 MB, 262 MB of it free space
[hos host] vk: write-backs left out (600 submits): 769, 3664.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3445 dirty pages from their submits' captures, 49704 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1277 dispatches, 36678 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5279 runs, 589.6 MB compared; 76 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13250 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vi: real-time step on: ratio 1.051, game's frame rate 60
[hos host] vk: render targets drawn over from another start, let go: 2078 so far (9928 MB), 1865 of them (8920 MB) between the passes
[hos host] vk: 111 textures retired (1609 left, 3397 MB): budget 104, drawn over 7, dead targets 0
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 413 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1637 kept their virtual core (nothing there to wait for), 2715 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 840000 (524297 pages from the captures)
[hos host] vk: 49 textures retired (1606 left, 3433 MB): budget 41, drawn over 8, dead targets 0
[hos Presentation Thread] vi: 3025 frames presented (51.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 22, 15.5-17.8 95, 17.8-22 97, 22-30 37, 30-40 6, over 40 2
[hos sead::MainThread] nvn: GPU waits in 5 s: 219 sync waits (652 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 5 textures retired (1629 left, 3056 MB): budget 0, drawn over 5, dead targets 0
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 763 ranges, 3369 MB), 257 pages written, 257 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (104 MB), 10 spare buffers, 6579 texture pages copied (25 MB)
[hos host] vk: sparse mirrors: 93, 949 MB with memory of 1228 MB (278 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 162 binds (22.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 17 taken from the spares, 104 made new; 28 kept as spares, 0 destroyed; 11 spares now, 368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3947 dirty pages from their submits' captures, 48885 from earlier ones', 8 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1310 dispatches, 36181 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5532 runs, 609.1 MB compared; 75 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13016 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool: 0x1053800000 (4 MB, captured, state 2): 7; 0x1051000000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 103 live (1408 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB)
[hos host] vk: replay reads of guest memory as submitted: 860000 (536382 pages from the captures)
[hos host] vk: 1 textures retired (1614 left, 2857 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: VRAM compaction 1 over: 274 textures moved (146 MB) in 23 passes, 0 left where they were; 8 blocks (256 MB) given back to the driver, 256 MB in all so far
[hos host] vk: 40 textures retired (1627 left, 2920 MB): budget 0, drawn over 0, dead targets 40
[hos host] vk: replay reads of guest memory as submitted: 880000 (548381 pages from the captures)
[hos host] vk: program picks: 1398101 kept (506854 by the last picks' textures, a uniform buffer bound since), 25081 new, 576817 with other textures (of 2000000)
[hos sead::MainThread] vk: readbacks of GPU results: 401 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1802 kept their virtual core (nothing there to wait for), 3062 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2899 dirty pages from their submits' captures, 52014 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1367 dispatches, 38066 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5668 runs, 632.2 MB compared; 84 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13619 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 900000 (560340 pages from the captures)
[hos ModuleSystemWorker4] nvn: 300000 recordings made, 65514 live copies (21.2 MB), 1886 texture copies; 3 command buffers' segments (4.1 MB)
[hos host] vk: slow submit: 27.9 ms: draw 27.1 descriptors 19.3 texture lookups 17.0 uploads 14.2 programs 6.4
[hos host] vk: 27 textures retired (1742 left, 2804 MB): budget 0, drawn over 0, dead targets 27
[hos host] vk: texture #4560 0x1000a1a4e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3090 (drawn at 2902)
[hos host] vk: replay reads of guest memory as submitted: 920000 (572989 pages from the captures)
[hos host] vk: 50 textures retired (1731 left, 2766 MB): budget 0, drawn over 0, dead targets 50
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: mirror of pool 0x7f55400000 (4 MB) given back: idle, 6041 submissions since its sync
[hos host] vk: sparse mirrors: 95, 948 MB with memory of 1236 MB (287 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 166 binds (23.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 20 taken from the spares, 104 made new; 29 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 793, 3773.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3165 dirty pages from their submits' captures, 51703 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1359 dispatches, 37629 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5549 runs, 627.4 MB compared; 91 pages read into the shadows; 200 MB of shadows (112 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13988 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool: 0x1053800000 (4 MB, captured, state 2): 2;
[hos host] vk: imports: 106 live (1420 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB)
[hos host] vk: replay reads of guest memory as submitted: 940000 (584974 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 674 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 4 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #5286 0x1001f0d3800 382x97 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 3118 (drawn at 3053)
[hos host] vk: 19 textures retired (1749 left, 2800 MB): budget 0, drawn over 0, dead targets 19
[hos host] vk: a frame: 330 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.63, present 0.43, draw 7.83, descriptors 3.60, texture lookups 1.14, uploads 0.22, pipelines 0.33, programs 1.30, commands 2.05, dispatch 0.88, memory sync 0.39, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 124 (948 MB): 5.97 MB synced in 1050.1 copies, 17.44 MB written back
[hos host] vk: 3120 presents, 56.6 fps; a frame: 2049 draws (0 skipped), 143 dispatches, 183 passes, 172 barriers, 6 submits, 2.2 uploads (0.3 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1750 textures (2809 MB)
[hos host] vk: host time in GPU resource waits: 4.54 ms a frame (fences 2.04, timelines 2.50, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 560.6 full barriers recorded, 291.0 left out (nothing recorded since the one before), 26.3 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 41.6; 0x12: 92.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 162.9, an op 1369.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 11.2 ms a frame (the longest submission 3.7 ms); 9.4 textures made a frame: colour targets 2.2 (11.8 MB), depth 5.4 (17.2 MB), ASTC or BC 1.9 (0.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.18 MB unswizzled on the CPU in 0.05 ms
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 756 ranges, 3381 MB), 252 pages written, 252 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (114 MB), 10 spare buffers, 6593 texture pages copied (25 MB)
[hos sead::MainThread] vk: readbacks of GPU results: 396 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1623 kept their virtual core (nothing there to wait for), 2985 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 960000 (596871 pages from the captures)
[hos host] vk: 33 textures retired (1779 left, 3008 MB): budget 0, drawn over 4, dead targets 29
[hos host] vk: 1 textures retired (1795 left, 3133 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 86 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 924, 4400.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2756 dirty pages from their submits' captures, 57920 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1375 dispatches, 42506 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6166 runs, 721.5 MB compared; 95 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15754 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 980000 (608764 pages from the captures)
[hos host] vk: 1 textures retired (1778 left, 3012 MB): budget 1, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: CommandBufferWaitEvent #30000: event 0x100d6c8028 mode 0 value 18055 (now 18036) (lr 0x806674b0)
[hos host] vk: recorder: 7455220 packets run in 10.0 s (busy 2.14 s, spun idle 0.14 s, slept 12323 times, 4210 of them under 200 us), 0 calls made directly (no room); 18677 batches over 537 frames (34.8 a frame, 399 packets a batch), 9288 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 387 times, 83.16 ms (0.7 times, 0.15 ms a frame); every thread 1085 times, 150.76 ms in all, the longest 0.93 ms; vkDestroyImageView 488/97.65 ms (max 0.93), vkWaitForFences 481/41.69 ms (max 0.70), the queue's lock 28/3.50 ms (max 0.47), vmaFreeMemory 8/2.75 ms (max 0.67), vkQueueBindSparse 44/2.42 ms (max 0.28), a part's submit 8/0.95 ms (max 0.29)
[hos host] vk: recorder: of its busy 2.14 s, 0.51 s in parts' ends and submits (the longest 1.29 ms), 0.18 s of that blocked on the queue's lock (295 times, the longest 1.26 ms), so 0.33 s in the ends, submits and holds; 1.63 s running the other packets; 13883 packets a frame
[hos host] vk: mirror of pool 0x7fbbc00000 (4 MB) given back: idle, 6064 submissions since its sync
[hos host] vk: 6 textures retired (1773 left, 3008 MB): budget 0, drawn over 6, dead targets 0
[hos host] vk: texture #5802 0x10009c04800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3171 (drawn at 3110)
[hos host] vk: program picks: 1530028 kept (535367 by the last picks' textures, a uniform buffer bound since), 25108 new, 644863 with other textures (of 2200000)
[hos host] vk: mirror of pool 0x105dc00000 (4 MB) given back: idle, 6063 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 1000000 (620678 pages from the captures)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 406 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1549 kept their virtual core (nothing there to wait for), 3059 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #5962 0x10009e44400 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 3186 (drawn at 3125)
[hos host] vk: readback pages refreshed from guest memory: 628224 pages so far (0 held: may be written ahead in their submission)
[hos OptimizeChangeBanc] vcores: 152668 takes, 27567 waits (mean 81 us, longest 18.1 ms, 0 over HOS_VCORE_LOG_MS), 13588 kicks, 13803 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: sparse mirrors: 93, 944 MB with memory of 1228 MB (283 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 168 binds (23.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 20 taken from the spares, 104 made new; 31 kept as spares, 0 destroyed; 11 spares now, 368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 968, 4609.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 780 dirty pages from their submits' captures, 60672 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1403 dispatches, 43961 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7314 runs, 769.2 MB compared; 94 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17608 in captures, 0 only in the collection, 44 only by the replay itself
[hos host] vk: read live, by pool: 0x1053800000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 106 live (1420 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB)
[hos host] vk: replay reads of guest memory as submitted: 1020000 (632705 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: (its written part: 4 of 4 KB readable, 1512 bytes not zero; levels L0)
[hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it
[hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 35 made): 24285364 lookups, 146793 missed (0.60 %)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 828 ranges, 3367 MB), 254 pages written, 254 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 143 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 1040000 (644927 pages from the captures)
[hos host] vk: texture #5736 0x10009814a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3223 (drawn at 3104)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 269 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 5 textures retired (1740 left, 2767 MB): budget 0, drawn over 5, dead targets 0
[hos host] vk: write-backs left out (600 submits): 990, 4714.3 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 343 dirty pages from their submits' captures, 60303 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1375 dispatches, 44042 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7610 runs, 788.4 MB compared; 106 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17813 in captures, 0 only in the collection, 45 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 1060000 (657510 pages from the captures)
[hos host] vk: a frame: 251 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.36, present 0.20, draw 5.69, descriptors 2.45, texture lookups 0.42, uploads 0.02, pipelines 0.25, programs 1.00, commands 1.57, dispatch 0.82, memory sync 0.38, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 124 (944 MB): 5.41 MB synced in 992.0 copies, 17.42 MB written back
[hos host] vk: 3240 presents, 61.1 fps; a frame: 1546 draws (0 skipped), 144 dispatches, 166 passes, 164 barriers, 6 submits, 0.2 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1740 textures (2767 MB)
[hos host] vk: host time in GPU resource waits: 2.15 ms a frame (fences 1.74, timelines 0.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 521.9 full barriers recorded, 280.3 left out (nothing recorded since the one before), 23.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.0; 0x12: 85.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 35.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 155.3, an op 1045.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 10.2 ms a frame (the longest submission 5.2 ms); 2.9 textures made a frame: colour targets 0.9 (5.0 MB), depth 1.9 (6.5 MB), ASTC or BC 0.1 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 397 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1736 kept their virtual core (nothing there to wait for), 3128 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #6271 0x10008cd3a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3246 (drawn at 3185)
[hos host] vk: 1 textures retired (1737 left, 2757 MB): budget 0, drawn over 1, dead targets 0
[hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #20000: event 0x100d6c8040 mode 0 value 19999 (lr 0x806621e8)
[hos host] vk: replay reads of guest memory as submitted: 1080000 (670024 pages from the captures)
[hos host] vk: texture #6432 0x10008cd3000 384x384 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (1 so far)
[hos host] vk: replay reads of guest memory as submitted: 1100000 (682464 pages from the captures)
[hos host] vk: mirror of pool 0x1042400000 (4 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: 32 mirrors given back (366 MB in all); 942 MB of mirrors now, 84 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: sparse mirrors: 92, 942 MB with memory of 1224 MB (281 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 169 binds (23.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 20 taken from the spares, 104 made new; 32 kept as spares, 0 destroyed; 12 spares now, 372 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1078, 5133.3 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 616 dirty pages from their submits' captures, 61812 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1342 dispatches, 46427 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7948 runs, 853.0 MB compared; 138 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18655 in captures, 1 only in the collection, 49 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 106 live (1420 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB)
[hos host] vk: texture #6270 0x10008c43800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3295 (drawn at 3234)
[hos host] vk: 13 textures retired (1762 left, 2798 MB): budget 0, drawn over 0, dead targets 13
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 393 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 9 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1592 kept their virtual core (nothing there to wait for), 3016 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: program picks: 1660735 kept (563714 by the last picks' textures, a uniform buffer bound since), 25123 new, 714141 with other textures (of 2400000)
[hos host] vk: mirror of pool 0x105cc00000 (4 MB) given back: idle, 6029 submissions since its sync
[hos host] vk: texture #6269 0x10008bb3600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3305 (drawn at 3244)
[hos host] vk: replay reads of guest memory as submitted: 1120000 (695021 pages from the captures)
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 40000 so far (x1 0x117cfbb040 x2 0 x3 0x117cfbb040 x4 0x9b040)
[hos host] vk: 9 textures retired (1791 left, 2860 MB): budget 0, drawn over 0, dead targets 9
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 823 ranges, 3373 MB), 244 pages written, 244 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (108 MB), 11 spare buffers, 147 texture pages copied (0 MB)
[hos Presentation Thread] vi: 3322 frames presented (59.2 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 41, 15.5-17.8 207, 17.8-22 45, 22-30 3, 30-40 1, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 1140000 (707637 pages from the captures)
[hos host] vk: 1 textures retired (1795 left, 2911 MB): budget 0, drawn over 1, dead targets 0
[hos sead::MainThread] nvn: GPU waits in 5 s: 138 sync waits (277 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos Presentation Thread] nvn: QueueSubmitCommands #20000: queue 0000101004133020 (its #3334), 1 handles (lr 0x80590358)
[hos host] vk: write-backs left out (600 submits): 989, 4712.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1569 dirty pages from their submits' captures, 57195 from earlier ones', 16 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1326 dispatches, 43414 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6974 runs, 765.2 MB compared; 102 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17154 in captures, 1 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool: 0x1054c00000 (4 MB, captured, state 2): 10; 0x7fa2800000 (8 MB, captured, state 2): 4; 0x1055c00000 (4 MB, captured, state 2): 2;
[hos rmbl::VibrationThread] hid: 8192 vibration values, strongest 0.00 since the last count
[hos host] vk: reaper: 1415 images (4027 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 167 (248 MB)
[hos host] vk: 25 textures retired (1862 left, 2960 MB): budget 0, drawn over 3, dead targets 22
[hos host] vk: texture #6105 0x10009b74000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3350 (drawn at 3143)
[hos host] vk: (its written part: 4 of 4 KB readable, 1502 bytes not zero; levels L0)
[hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it
[hos host] vk: replay reads of guest memory as submitted: 1160000 (719829 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 196 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 383 of 3000 waited for their write-back, 0.10 ms each; 0 looked at the stopped queues first, 26 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1472 kept their virtual core (nothing there to wait for), 2880 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 1 textures retired (1879 left, 2969 MB): budget 0, drawn over 0, dead targets 1
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #35000: event 0x100d6c8028 mode 0 value 21855 (now 21836) (lr 0x806674b0)
[hos host] vk: a frame: 236 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.57, present 0.20, draw 5.82, descriptors 2.65, texture lookups 0.80, uploads 0.17, pipelines 0.24, programs 1.00, commands 1.51, dispatch 0.81, memory sync 0.41, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 127 (943 MB): 5.10 MB synced in 963.1 copies, 17.44 MB written back
[hos host] vk: 3360 presents, 59.0 fps; a frame: 1506 draws (0 skipped), 137 dispatches, 157 passes, 159 barriers, 6 submits, 1.1 uploads (0.6 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1882 textures (2974 MB)
[hos host] vk: host time in GPU resource waits: 2.84 ms a frame (fences 1.19, timelines 1.65, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 514.6 full barriers recorded, 273.3 left out (nothing recorded since the one before), 18.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.0; 0x12: 80.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 36.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 150.3, an op 982.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 10.5 ms a frame (the longest submission 5.9 ms); 5.4 textures made a frame: colour targets 2.0 (10.6 MB), depth 2.6 (8.1 MB), ASTC or BC 0.8 (0.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.52 MB unswizzled on the CPU in 0.12 ms
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 111 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 95, 943 MB with memory of 1236 MB (292 MB spared); blocks given memory later 113, given back 1056; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 175 binds (23.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 24 taken from the spares, 104 made new; 33 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 859, 4087.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2320 dirty pages from their submits' captures, 50554 from earlier ones', 67 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1195 dispatches, 38998 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6487 runs, 701.5 MB compared; 76 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15292 in captures, 1 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool: 0x105e800000 (4 MB, captured, state 2): 67;
[hos host] vk: imports: 108 live (1428 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB)
[hos host] vk: replay reads of guest memory as submitted: 1180000 (731927 pages from the captures)
[hos host] vk: 33 textures retired (1864 left, 2927 MB): budget 0, drawn over 20, dead targets 13
[hos OptimizeChangeBanc] nvn: 350000 recordings made, 67513 live copies (21.4 MB), 3065 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: replay reads of guest memory as submitted: 1200000 (743993 pages from the captures)
[hos host] vk: 24 textures retired (1960 left, 2987 MB): budget 0, drawn over 8, dead targets 16
[hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.02, scan 0.04 over 902 ranges, 3384 MB), 280 pages written, 280 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (148 MB), 9 spare buffers, 147 texture pages copied (0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.12
[hos host] vk: (its written part: 14 of 14 KB readable, 4280 bytes not zero; levels L0)
[hos host] vk: texture 0x107cd9f600 384x39x1 x1 levels 1 NVN 0x1 (14 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 14 KB of it
[hos host] vk: readback pages refreshed from guest memory: 800005 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: program picks: 1806500 kept (624498 by the last picks' textures, a uniform buffer bound since), 25135 new, 768364 with other textures (of 2600000)
[hos host] vk: texture 0x1000b4f4a00 1280x720x1 x1 levels 1 NVN 0x29 -> vk 97 (2000 textures, 3070 MB)
[hos host] vk: write-backs left out (600 submits): 813, 3874.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2486 dirty pages from their submits' captures, 50880 from earlier ones', 23 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1300 dispatches, 38172 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6374 runs, 639.7 MB compared; 109 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14402 in captures, 1 only in the collection, 26 only by the replay itself
[hos host] vk: read live, by pool: 0x1057c00000 (4 MB, captured, state 2): 10; 0x1059c00000 (4 MB, captured, state 2): 8; 0x1054400000 (4 MB, captured, state 2): 4; 0x105e400000 (4 MB, captured, state 2): 1;
[hos host] vk: texture 0x1000ae04a00 1280x720x1 x1 levels 1 NVN 0x29 -> vk 97 (2000 textures, 3069 MB)
[hos host] vk: 13 textures retired (1988 left, 3002 MB): budget 0, drawn over 0, dead targets 13
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 333 of 3000 waited for their write-back, 0.18 ms each; 0 looked at the stopped queues first, 22 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1458 kept their virtual core (nothing there to wait for), 3150 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture 0x10d3801200 341x341x1 x1 levels 9 NVN 0x46 -> vk 133 (2000 textures, 3013 MB)
[hos host] vk: 97 textures retired (1965 left, 3028 MB): budget 81, drawn over 2, dead targets 14
[hos host] vk: replay reads of guest memory as submitted: 1220000 (756005 pages from the captures)
[hos host] vk: 75 textures retired (1900 left, 3028 MB): budget 73, drawn over 0, dead targets 2
[hos sead::MainThread] nvn: 400000 recordings made, 71335 live copies (22.0 MB), 3305 texture copies; 5 command buffers' segments (4.3 MB)
[hos host] vk: 68 textures retired (1834 left, 3028 MB): budget 68, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 1240000 (768100 pages from the captures)
[hos host] vk: sparse mirrors: 99, 948 MB with memory of 1252 MB (303 MB spared); blocks given memory later 117, given back 1059; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 184 binds (25.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 25 taken from the spares, 107 made new; 33 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 837, 3982.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2764 dirty pages from their submits' captures, 54070 from earlier ones', 38 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1401 dispatches, 40270 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7017 runs, 693.2 MB compared; 129 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15285 in captures, 2 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool: 0x1057000000 (4 MB, captured, state 2): 18; 0x1053c00000 (4 MB, captured, state 2): 11; 0x1054800000 (4 MB, captured, state 2): 6; 0x1054000000 (4 MB, captured, state 2): 2; 0x1059800000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 108 live (1428 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB)
[hos host] vk: texture #5525 0x1001f077c00 64x62x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 205 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #5528 0x1001f034600 128x126x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 205 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 39 textures retired (1811 left, 3115 MB): budget 38, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 2 finalized, 0 MB
[hos host] vk: 39 textures retired (1791 left, 3043 MB): budget 24, drawn over 0, dead targets 15
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 338 of 3000 waited for their write-back, 0.14 ms each; 0 looked at the stopped queues first, 22 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1164 kept their virtual core (nothing there to wait for), 2676 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: queues ordered by their events: 18000 submissions stopped at a wait, 17996 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 37717, not met 0; captures at a CPU event's wait 3051, 0 of them at a wait met on arrival
[hos host] vk: 18 textures retired (1797 left, 3088 MB): budget 0, drawn over 1, dead targets 17
[hos host] vk: replay reads of guest memory as submitted: 1260000 (780185 pages from the captures)
[hos nn.npln.Shared] vcores: 161771 takes, 37763 waits (mean 83 us, longest 49.3 ms, 0 over HOS_VCORE_LOG_MS), 19294 kicks, 19315 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 5 textures retired (1814 left, 3348 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: a frame: 194 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.87, present 0.26, draw 7.00, descriptors 3.38, texture lookups 1.57, uploads 0.49, pipelines 0.29, programs 1.08, commands 1.78, dispatch 0.96, memory sync 1.30, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 133 (949 MB): 5.76 MB synced in 1040.3 copies, 17.71 MB written back
[hos host] vk: 3480 presents, 52.7 fps; a frame: 1674 draws (0 skipped), 158 dispatches, 162 passes, 166 barriers, 6 submits, 2.0 uploads (1.6 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1795 textures (3080 MB)
[hos host] vk: host time in GPU resource waits: 5.38 ms a frame (fences 1.92, timelines 3.46, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 546.9 full barriers recorded, 277.7 left out (nothing recorded since the one before), 22.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 42.8; 0x12: 85.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 30.8 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 157.2, an op 1169.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.8 ms); 8.7 textures made a frame: colour targets 3.3 (18.2 MB), depth 3.6 (22.5 MB), ASTC or BC 1.8 (1.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.44 MB unswizzled on the CPU in 0.41 ms
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 178 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 22 pages in 12 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 924, 4400.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1193 dirty pages from their submits' captures, 61091 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1498 dispatches, 44013 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7122 runs, 739.6 MB compared; 97 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15994 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 91 textures retired (1719 left, 3125 MB): budget 91, drawn over 0, dead targets 0
[hos sead::MainThread] vk: capture: a submit 0.21 ms (lock 0.03, scan 0.06 over 1319 ranges, 3589 MB), 303 pages written, 303 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (120 MB), 11 spare buffers, 146 texture pages copied (0 MB)
[hos host] vk: pool 0x1018e0bc00+0xf00000 imported in 5.9 ms as 0x1018c00000..0x101a000000 (128 imports, 335 ms in all)
[hos host] vk: replay reads of guest memory as submitted: 1280000 (792306 pages from the captures)
[hos host] vk: pool 0x10184ab800+0xf00000 imported in 3.4 ms as 0x1018400000..0x101a000000 (129 imports, 338 ms in all)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.15
[hos host] vk: mirror of pool 0x1051000000 (4 MB) given back: idle, 6033 submissions since its sync
[hos host] vk: mirror of pool 0x1053000000 (4 MB) given back: idle, 6048 submissions since its sync
[hos host] vk: (its written part: 4 of 4 KB readable, 1495 bytes not zero; levels L0)
[hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 364 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1308 kept their virtual core (nothing there to wait for), 3044 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 10 textures retired (1774 left, 3020 MB): budget 0, drawn over 0, dead targets 10
[hos host] vk: replay reads of guest memory as submitted: 1300000 (805048 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #25000: event 0x100d6c8000 mode 0 value 24999 (lr 0x80291440)
[hos host] vk: VRAM compaction 2 begins: the textures' blocks hold 2464 MB, 203 MB of it free space
[hos host] vk: 15 textures retired (1786 left, 3028 MB): budget 3, drawn over 0, dead targets 12
[hos host] vk: slow upload 20.4 ms: #8318 0x10eead2a00 576x576x1 x1 levels 10 NVN 0x8d ASTC (BC3), 149 of 149 KB
[hos host] vk: slow submit: 25.5 ms: draw 25.5 descriptors 24.0 texture lookups 22.6 uploads 22.0
[hos host] vk: sparse mirrors: 98, 947 MB with memory of 1248 MB (300 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 188 binds (25.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 25 taken from the spares, 108 made new; 35 kept as spares, 0 destroyed; 10 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 902, 4295.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 680 dirty pages from their submits' captures, 59745 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1437 dispatches, 42707 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6685 runs, 718.0 MB compared; 98 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16324 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 109 live (1456 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB)
[hos host] vk: mirror of pool 0x1050800000 (4 MB) given back: idle, 6009 submissions since its sync
[hos host] vk: program picks: 1938293 kept (658368 by the last picks' textures, a uniform buffer bound since), 25165 new, 836541 with other textures (of 2800000)
[hos host] vk: 65 textures retired (1776 left, 3065 MB): budget 51, drawn over 0, dead targets 14
[hos host] vk: replay reads of guest memory as submitted: 1320000 (816929 pages from the captures)
[hos host] vk: (its written part: 4 of 4 KB readable, 1495 bytes not zero; levels L0)
[hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it
[hos host] vk: 4 textures retired (1795 left, 3157 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: texture #8089 0x1000a894800 1280x720 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3557 (drawn at 3496)
[hos host] vk: 4 textures retired (1808 left, 3288 MB): budget 4, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: CommandBufferWaitEvent #40000: event 0x100d6c8028 mode 0 value 25655 (now 25636) (lr 0x806674b0)
[hos host] vk: presenter thread: the copy's fence 1.73 ms, the present 0.75 ms a present
[hos host] vk: budget: gave back 34 MB from target pool at 90% of the allowance
[hos host] vk: 34 textures retired (1745 left, 3031 MB): budget 33, drawn over 1, dead targets 0
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 374 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1293 kept their virtual core (nothing there to wait for), 3059 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: mirror of pool 0x7fd9400000 (4 MB) given back: idle, 6023 submissions since its sync
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 111 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 924, 4400.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 784 dirty pages from their submits' captures, 59160 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1378 dispatches, 42934 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6847 runs, 735.1 MB compared; 92 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15617 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 1340000 (828851 pages from the captures)
[hos host] vk: texture #8032 0x10008e84000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3579 (drawn at 3518)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 1319 ranges, 3589 MB), 270 pages written, 270 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (122 MB), 10 spare buffers, 147 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 198 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 35 textures retired (1753 left, 3026 MB): budget 30, drawn over 5, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: memory: working set 13243 MB (peak 13245), private 20169 MB, staging 48 MB in 3 arenas, 20 made so far (56 ms)
[hos host] vk: memory heaps: heap 0 (system): 15 blocks 640 MB, 36 allocations 572 MB; used 2500 of 15579 MB; heap 1 (device): 175 blocks 5015 MB, 17652 allocations 4722 MB; used 6833 of 7823 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 21 blocks 672 MB, 152 images 557 MB; 122 free ranges, the largest 28 MB
[hos host] vk: image blocks (textures, compacted): 69 blocks 2208 MB, 1657 images 2189 MB; 310 free ranges, the largest 4 MB
[hos host] vk: render target images: 3177 reused from the pool, 1530 made new, 1376 destroyed from it, 660 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 85 missed after their image was taken by size; it holds 33 (135 MB)
[hos host] vk: render targets not reused by size: 418 with the pool empty (0 of them with targets still in the dead queue), 123 with none in it a present yet, 16 with none of a kind they can use, 263 with all smaller (the largest 48% short on average), 233 with all bigger (over 1.5 times), 477 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 65 over its bound and 1311 idle
[hos host] vk: VRAM: 6833 MB used of 7823 MB allowed (292 MB of it free in the blocks); textures 3066 MB (budget 3028 MB), 1768 textures
[hos host] vk: budget: VRAM 5011 of 7823 MB (6833 MB by the driver, 1821 MB unaccounted): textures 3066, mirrors 945, blocks free 292, other buffers 211, shadows 200, target pool 135, other images 111, staging 48; RAM 2438 MB (the process's private memory 20169 MB): imports 1456, capture spares 440, retired imports 384, captures 114, capture table 26, other buffers 18, others 0
[hos host] vk: textures: decoded ASTC 1608 MB, rendered 914 MB (104 drawn in one frame, kept: 94765 KB), other 480 MB; used in the last 8 presents 2536 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 528 MB of blocks hashed in 65 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 4 presents ago
[hos host] vk: 21 MB: 0x10006fce600 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x10005192200 1600x1600x1 x3 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 17 (8407 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 208x105x1 NVN 0x1: 1 (21 KB); 262x104x1 NVN 0x1: 1 (26 KB); 592x49x1 NVN 0x25: 1 (113 KB); 656x113x1 NVN 0x1: 1 (72 KB); 1466x282x1 NVN 0x25: 5 (8074 KB);
[hos host] vk: 28 textures retired (1740 left, 2971 MB): budget 0, drawn over 3, dead targets 25
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos Presentation Thread] vk: presents after the CPU's signal (600): 324 waited for it, 2.64 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: replay reads of guest memory as submitted: 1360000 (840734 pages from the captures)
[hos host] vk: status: 600 frames, the slowest 43.5 ms; VRAM 6.7 of 7.6 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5253 queued, 5253 loaded (median 1.38 ms, 99th percentile 16.77 ms, worst 34.27 ms, 13012 ms and 296.2 MB read in all); 7285 programs (1431 of them compute variants without their code, 0 graphics programs no note names), 10490 graphics pipelines (1 not in a cache), 577 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1369 pipelines made at create taken by draws; 2569 create calls with nothing kept for their stages, 0 before the index; units 7822 (5253 ready, 0 loading), 14420 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 11458 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 575 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 176 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.97, present 0.54, draw 6.00, descriptors 2.85, texture lookups 1.16, uploads 0.38, pipelines 0.25, programs 0.98, commands 1.52, dispatch 0.86, memory sync 0.38, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 134 (945 MB): 5.68 MB synced in 1030.8 copies, 17.50 MB written back
[hos host] vk: 3600 presents, 60.0 fps; a frame: 1402 draws (0 skipped), 146 dispatches, 162 passes, 166 barriers, 6 submits, 1.5 uploads (0.7 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1742 textures (2979 MB)
[hos host] vk: host time in GPU resource waits: 2.82 ms a frame (fences 2.04, timelines 0.79, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 518.4 full barriers recorded, 275.6 left out (nothing recorded since the one before), 24.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 41.0; 0x12: 87.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 30.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 157.2, an op 920.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 9.5 ms a frame (the longest submission 3.0 ms); 6.5 textures made a frame: colour targets 2.1 (11.7 MB), depth 3.2 (13.2 MB), ASTC or BC 1.2 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.37 MB unswizzled on the CPU in 0.08 ms
[hos host] vk: queues ordered by their events: 18952 submissions stopped at a wait, 18952 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 41016, not met 0; captures at a CPU event's wait 3083, 0 of them at a wait met on arrival
[hos host] vk: the render thread (600 presents, 17.33 ms a frame): its jobs 16.03 ms a frame (submits 5.34, presents 4.83, others 5.86), idle 1.30; the game waited for room at 50 presents, 0.19 ms each (0.02 a frame; 21.55 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 326 captures at a CPU signal; the submits' waits read in 0.044 ms a frame; readbacks: 1938 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.21 ms after its hand-over (36.00 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.68 + 0.35 (the window's copy is in the idle), after it 1.84 + 1.06; after a frame's last recording that reads or writes the game's memory come 1.02 recordings (0.00 ms of GPU time), after its last that reads it 1.02 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.60 parts a frame recorded in the signal's job, 66.9% of them took nothing that came with the signal; of the others, 2.84 a frame took pages of the signal's capture into a mirror (79.2 pages a frame) and 0.05 pages read as guest memory had them (0.4 pages a frame); a frame's first part took nothing in 37 of 326 frames, its first three in 21
[hos host] vk: frame ahead, what came with the signal: 326 signals with a capture, each of 201.5 pages in 32.3 pools; between a frame's last submit and that signal 0.03 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (8 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.60+0.09, 0.33+0.12, 0.23+0.10, 0.33+0.22, 0.11+0.21, 0.10+0.16, 0.09+0.11, 0.03+0.03, 0.02+0.02, 0.00+0.01, 0.00+0.00, 0.00+0.00
[hos ModuleSystemWorker3] tex: the game's texture heap: 45 compaction rounds asked for in 3600 frames without one running
[hos Presentation Thread] vi: 3604 frames presented (56.3 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 64, 15.5-17.8 145, 17.8-22 45, 22-30 19, 30-40 7, over 40 2
[hos host] vk: render targets drawn over from another start, let go: 5303 so far (25283 MB), 4728 of them (22703 MB) between the passes
[hos host] vk: readback pages refreshed from guest memory: 982054 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: VRAM compaction 2 over: 454 textures moved (337 MB) in 42 passes, 1 left where they were; 8 blocks (256 MB) given back to the driver, 512 MB in all so far
[hos sead::MainThread] nvn: GPU waits in 5 s: 132 sync waits (486 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: sparse mirrors: 97, 945 MB with memory of 1244 MB (298 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 191 binds (25.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 26 taken from the spares, 108 made new; 37 kept as spares, 0 destroyed; 11 spares now, 368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 858, 4085.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1041 dirty pages from their submits' captures, 55813 from earlier ones', 97 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1314 dispatches, 40467 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6374 runs, 681.2 MB compared; 102 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14606 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 85; 0x1057800000 (4 MB, captured, state 2): 10; 0x1055800000 (4 MB, captured, state 2): 2;
[hos host] vk: imports: 109 live (1456 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB)
[hos host] vk: 87 textures retired (1724 left, 3027 MB): budget 75, drawn over 1, dead targets 11
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 354 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1250 kept their virtual core (nothing there to wait for), 3102 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 1380000 (852769 pages from the captures)
[hos host] vk: 2 textures retired (1733 left, 2996 MB): budget 0, drawn over 0, dead targets 2
[hos host] vk: 68 textures retired (1754 left, 3025 MB): budget 68, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 1400000 (865183 pages from the captures)
[hos host] vk: texture #8536 0x1000a2c4200 1280x720 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 3647 (drawn at 3586)
[hos host] vk: 16 textures retired (1775 left, 3024 MB): budget 13, drawn over 1, dead targets 2
[hos host] vk: write-backs left out (600 submits): 836, 3980.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2182 dirty pages from their submits' captures, 54215 from earlier ones', 16 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1339 dispatches, 39406 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6173 runs, 663.6 MB compared; 81 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14256 in captures, 0 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool: 0x1055000000 (4 MB, captured, state 2): 11; 0x1055400000 (4 MB, captured, state 2): 5;
[hos host] vk: program picks: 2066726 kept (685887 by the last picks' textures, a uniform buffer bound since), 25172 new, 908101 with other textures (of 3000000)
[hos host] vk: slow submit: 18.6 ms: draw 16.9 descriptors 14.3 texture lookups 12.8 uploads 12.4 commands 1.3
[hos host] vk: 80 textures retired (1776 left, 3035 MB): budget 51, drawn over 3, dead targets 26
[hos host] vk: replay reads of guest memory as submitted: 1420000 (879032 pages from the captures)
[hos host] vk: mirror of pool 0x105c000000 (4 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: 9 textures retired (1779 left, 3020 MB): budget 9, drawn over 0, dead targets 0
[hos ModuleSystemWorker2] barrier: 47779 looks: 15011 found the masks moved already, 31435 waited until they moved, 1333 for the whole 1 ms (a wait 302 us on average); 39649 wakes
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 351 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1180 kept their virtual core (nothing there to wait for), 2916 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 1319 ranges, 3589 MB), 258 pages written, 258 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (132 MB), 11 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: 22 textures retired (1768 left, 3026 MB): budget 21, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 1440000 (892706 pages from the captures)
[hos host] vk: sparse mirrors: 98, 946 MB with memory of 1248 MB (301 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 194 binds (26.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 28 taken from the spares, 108 made new; 38 kept as spares, 0 destroyed; 10 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2733 dirty pages from their submits' captures, 50487 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1274 dispatches, 37156 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5864 runs, 631.1 MB compared; 91 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13046 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 109 live (1456 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB)
[hos host] vk: 7 textures retired (1768 left, 3023 MB): budget 1, drawn over 6, dead targets 0
[hos host] vk: mirror of pool 0x7fb0000000 (8 MB) given back: idle, 6003 submissions since its sync
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 1460000 (906372 pages from the captures)
[hos host] vk: 21 textures retired (1759 left, 2994 MB): budget 0, drawn over 21, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 1480000 (920106 pages from the captures)
[hos host] vk: 1 textures retired (1760 left, 2996 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: a frame: 219 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.26, present 0.28, draw 6.90, descriptors 3.08, texture lookups 0.73, uploads 0.23, pipelines 0.31, programs 1.10, commands 1.91, dispatch 0.89, memory sync 0.39, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 136 (939 MB): 5.85 MB synced in 1040.8 copies, 17.53 MB written back
[hos host] vk: 3720 presents, 55.6 fps; a frame: 1779 draws (0 skipped), 150 dispatches, 175 passes, 170 barriers, 6 submits, 1.8 uploads (0.5 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1761 textures (3000 MB)
[hos host] vk: host time in GPU resource waits: 4.10 ms a frame (fences 1.68, timelines 2.42, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 547.3 full barriers recorded, 286.0 left out (nothing recorded since the one before), 29.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.8; 0x12: 91.1; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.2 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 161.4, an op 1223.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 11.2 ms a frame (the longest submission 5.6 ms); 5.5 textures made a frame: colour targets 1.3 (7.9 MB), depth 2.4 (7.0 MB), ASTC or BC 1.7 (1.5 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.22 MB unswizzled on the CPU in 0.04 ms
[hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2898 dirty pages from their submits' captures, 51512 from earlier ones', 2 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1297 dispatches, 37759 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5976 runs, 633.1 MB compared; 72 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13465 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd3800000 (4 MB, captured, state 2): 2;
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 360 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 10 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1253 kept their virtual core (nothing there to wait for), 3099 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: recorder: 5917301 packets run in 10.0 s (busy 1.86 s, spun idle 0.13 s, slept 11712 times, 4139 of them under 200 us), 0 calls made directly (no room); 17669 batches over 572 frames (30.9 a frame, 335 packets a batch), 8455 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 476 times, 86.65 ms (0.8 times, 0.15 ms a frame); every thread 1175 times, 139.25 ms in all, the longest 0.83 ms; vkDestroyImageView 596/98.90 ms (max 0.83), vkWaitForFences 505/35.06 ms (max 0.44), vkQueueBindSparse 28/1.82 ms (max 0.35), vmaFreeMemory 13/1.16 ms (max 0.37), a part's submit 5/0.77 ms (max 0.21), the queue's lock 9/0.62 ms (max 0.22)
[hos host] vk: recorder: of its busy 1.86 s, 0.49 s in parts' ends and submits (the longest 1.66 ms), 0.19 s of that blocked on the queue's lock (325 times, the longest 1.62 ms), so 0.30 s in the ends, submits and holds; 1.37 s running the other packets; 10345 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 1500000 (933825 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 189 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 102 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 57 textures retired (1766 left, 2976 MB): budget 0, drawn over 25, dead targets 32
[hos host] vk: texture #9520 0x10009304600 384x384 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (2 so far)
[hos host] vk: replay reads of guest memory as submitted: 1520000 (947496 pages from the captures)
[hos host] vk: program picks: 2195084 kept (718626 by the last picks' textures, a uniform buffer bound since), 25172 new, 979743 with other textures (of 3200000)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 117 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 98, 946 MB with memory of 1248 MB (301 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 196 binds (26.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 29 taken from the spares, 108 made new; 39 kept as spares, 0 destroyed; 10 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2583 dirty pages from their submits' captures, 53540 from earlier ones', 5 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1320 dispatches, 38765 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6344 runs, 638.3 MB compared; 81 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14263 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool: 0x1057400000 (4 MB, captured, state 2): 5;
[hos host] vk: imports: 109 live (1456 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #45000: event 0x100d6c8028 mode 0 value 29455 (now 29436) (lr 0x806674b0)
[hos aal::LowPrioWorkerThread] vcores: 151722 takes, 27390 waits (mean 78 us, longest 18.3 ms, 0 over HOS_VCORE_LOG_MS), 13317 kicks, 13500 yields, 0 starved grants, 0 escapes in 5 s
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 1320 ranges, 3589 MB), 256 pages written, 256 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (128 MB), 11 spare buffers, 149 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 1540000 (961071 pages from the captures)
[hos BaseProcExecutor3] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 35 made): 21031640 lookups, 148040 missed (0.70 %)
[hos host] vk: mirror of pool 0x1041c00000 (4 MB) given back: idle, 6063 submissions since its sync
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 358 of 3000 waited for their write-back, 0.08 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1152 kept their virtual core (nothing there to wait for), 2944 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #30000: event 0x100d6c8010 mode 0 value 30000 (lr 0x80660adc)
[hos host] vk: (its written part: 4 of 4 KB readable, 1495 bytes not zero; levels L0)
[hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it
[hos host] vk: 11 textures retired (1817 left, 3021 MB): budget 0, drawn over 0, dead targets 11
[hos host] vk: replay reads of guest memory as submitted: 1560000 (974468 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 1166032 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 836, 3980.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2646 dirty pages from their submits' captures, 55823 from earlier ones', 84 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1308 dispatches, 40669 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6946 runs, 676.5 MB compared; 79 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14690 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 84;
[hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 60000 so far (x1 0x117b427040 x2 0x6001d40 x3 0x117b427000 x4 0xf33080)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: mirror of pool 0x105e800000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 1580000 (987723 pages from the captures)
[hos host] vk: pool 0x7f9d35d600+0x2cb8 imported in 2.1 ms as 0x7f9c400000..0x7f9e400000 (130 imports, 340 ms in all)
[hos ModuleSystemWorker2] nvn: 450000 recordings made, 72154 live copies (23.3 MB), 4043 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: slow submit: 30.4 ms: draw 29.7 descriptors 20.2 texture lookups 18.2 uploads 15.6 programs 8.1
[hos host] vk: mirror of pool 0x1050c00000 (4 MB) given back: idle, 6030 submissions since its sync
[hos host] vk: 37 textures retired (1957 left, 3029 MB): budget 0, drawn over 37, dead targets 0
[hos host] vk: texture 0x10bf605200 768x768x1 x1 levels 10 NVN 0x49 -> vk 139 (2000 textures, 3055 MB)
[hos host] vk: replay reads of guest memory as submitted: 1600000 (1000547 pages from the captures)
[hos host] vk: 1 textures retired (2014 left, 3059 MB): budget 0, drawn over 0, dead targets 1
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 344 of 3000 waited for their write-back, 0.13 ms each; 0 looked at the stopped queues first, 8 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1133 kept their virtual core (nothing there to wait for), 2707 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: sparse mirrors: 101, 978 MB with memory of 1292 MB (313 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 205 binds (28.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 32 taken from the spares, 111 made new; 42 kept as spares, 0 destroyed; 10 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 814, 3876.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2358 dirty pages from their submits' captures, 52550 from earlier ones', 35 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1191 dispatches, 39168 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6344 runs, 651.8 MB compared; 74 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14862 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x1053800000 (4 MB, captured, state 2): 13; 0x1056800000 (4 MB, captured, state 2): 12; 0x1053400000 (4 MB, captured, state 2): 10;
[hos host] vk: imports: 111 live (1496 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB)
[hos host] vk: 16 textures retired (2103 left, 3107 MB): budget 0, drawn over 0, dead targets 16
[hos host] vk: a frame: 218 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.73, present 0.20, draw 7.42, descriptors 3.25, texture lookups 0.68, uploads 0.32, pipelines 0.33, programs 1.26, commands 2.06, dispatch 0.94, memory sync 0.44, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 144 (978 MB): 6.28 MB synced in 1063.5 copies, 17.74 MB written back
[hos host] vk: 3840 presents, 52.8 fps; a frame: 1839 draws (0 skipped), 164 dispatches, 176 passes, 171 barriers, 6 submits, 3.3 uploads (0.7 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2118 textures (3112 MB)
[hos host] vk: host time in GPU resource waits: 5.10 ms a frame (fences 1.99, timelines 3.11, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 569.8 full barriers recorded, 286.1 left out (nothing recorded since the one before), 31.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 42.6; 0x12: 90.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.5 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 161.9, an op 1292.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 11.3 ms a frame (the longest submission 5.3 ms); 6.0 textures made a frame: colour targets 1.6 (8.8 MB), depth 1.1 (4.2 MB), ASTC or BC 3.3 (2.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.33 MB unswizzled on the CPU in 0.06 ms
[hos host] vk: replay reads of guest memory as submitted: 1620000 (1012068 pages from the captures)
[hos host] vk: 33 textures retired (2099 left, 3107 MB): budget 32, drawn over 1, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 1300 ranges, 3628 MB), 265 pages written, 265 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (142 MB), 11 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: program picks: 2327693 kept (751825 by the last picks' textures, a uniform buffer bound since), 25226 new, 1047080 with other textures (of 3400000)
[hos host] vk: 11 textures retired (2096 left, 3106 MB): budget 11, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 1640000 (1023369 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 814, 3876.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1924 dirty pages from their submits' captures, 52272 from earlier ones', 16 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1272 dispatches, 38694 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6568 runs, 661.0 MB compared; 83 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13902 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x1056c00000 (4 MB, captured, state 2): 9; 0x105ac00000 (4 MB, captured, state 2): 6; 0x1042c00000 (4 MB, captured, state 2): 1;
[hos Presentation Thread] vi: 3878 frames presented (54.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 36, 15.5-17.8 131, 17.8-22 79, 22-30 21, 30-40 5, over 40 2
[hos host] vk: 11 textures retired (2082 left, 3089 MB): budget 0, drawn over 0, dead targets 11
[hos host] vk: mirror of pool 0x7fa2800000 (8 MB) given back: idle, 6026 submissions since its sync
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 328 of 3000 waited for their write-back, 0.11 ms each; 0 looked at the stopped queues first, 20 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1229 kept their virtual core (nothing there to wait for), 3379 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 1660000 (1034773 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 172 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 113 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] nvn: GPU waits in 5 s: 207 sync waits (585 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 1680000 (1046288 pages from the captures)
[hos host] vk: sparse mirrors: 103, 974 MB with memory of 1296 MB (321 MB spared); blocks given memory later 118, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 210 binds (28.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 34 taken from the spares, 112 made new; 43 kept as spares, 0 destroyed; 9 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1886 dirty pages from their submits' captures, 52109 from earlier ones', 1042 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1322 dispatches, 38315 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6316 runs, 637.3 MB compared; 93 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13763 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool: 0x7f7dc00000 (8 MB, captured, state 2): 1014; 0x1053800000 (4 MB, captured, state 2): 15; 0x105dc00000 (4 MB, captured, state 2): 7; 0x105a800000 (4 MB, captured, state 2): 6;
[hos host] vk: imports: 112 live (1500 MB), 3 retired still imported (44 MB), 17 retired given back (360 MB)
[hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 67.2 ms for a core: given by NavMeshCAStepDepWorker1 (priority 22), which held it 0.6 ms and was never asked; the cores it may use now: core 1: it;
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 79.9 ms for a core: given by tera::FileResourceMgr (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos host] vk: reaper: 1531 images (3611 MB) destroyed in 10 s, 15 of them where they were handed over (its cap or VRAM pressure); the queue at most 90 (229 MB)
[hos host] vk: replay reads of guest memory as submitted: 1700000 (1057771 pages from the captures)
[hos host] vk: 17 textures retired (2098 left, 3101 MB): budget 7, drawn over 0, dead targets 10
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 325 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1107 kept their virtual core (nothing there to wait for), 2989 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 34 textures retired (2079 left, 3107 MB): budget 34, drawn over 0, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.22 ms (lock 0.03, scan 0.05 over 801 ranges, 3436 MB), 342 pages written, 342 copied (1.3 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 24 kept (108 MB), 11 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 118 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 836, 3980.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2865 dirty pages from their submits' captures, 55429 from earlier ones', 8132 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1392 dispatches, 40439 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6657 runs, 673.3 MB compared; 90 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14464 in captures, 1 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9c400000 (32 MB, captured, state 2): 6892; 0x7f7dc00000 (8 MB, captured, state 2): 1236; 0x1051000000 (4 MB, captured, state 2): 2; 0x1053000000 (4 MB, captured, state 2): 1; 0x1042c00000 (4 MB, captured, state 2): 1;
[hos host] vk: replay reads of guest memory as submitted: 1720000 (1070124 pages from the captures)
[hos host] vk: 29 textures retired (2058 left, 3107 MB): budget 29, drawn over 0, dead targets 0
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #50000: event 0x100d6c8028 mode 0 value 33255 (now 33236) (lr 0x806674b0)
[hos host] vk: a frame: 244 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.47, present 0.23, draw 7.14, descriptors 2.98, texture lookups 0.34, uploads 0.03, pipelines 0.33, programs 1.26, commands 2.03, dispatch 0.99, memory sync 1.02, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 146 (973 MB): 6.17 MB synced in 1059.2 copies, 17.73 MB written back
[hos host] vk: 3960 presents, 55.1 fps; a frame: 1814 draws (0 skipped), 162 dispatches, 181 passes, 171 barriers, 6 submits, 0.4 uploads (0.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2049 textures (3075 MB)
[hos host] vk: host time in GPU resource waits: 3.02 ms a frame (fences 1.50, timelines 1.52, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 566.0 full barriers recorded, 290.6 left out (nothing recorded since the one before), 29.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 42.2; 0x12: 91.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.2 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 162.1, an op 1323.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 10.7 ms a frame (the longest submission 4.7 ms); 3.3 textures made a frame: colour targets 1.6 (9.4 MB), depth 1.4 (5.0 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.06 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: program picks: 2463396 kept (785830 by the last picks' textures, a uniform buffer bound since), 25247 new, 1111356 with other textures (of 3600000)
[hos host] vk: replay reads of guest memory as submitted: 1740000 (1081484 pages from the captures)
[hos host] vk: texture #10330 0x10009154a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3980 (drawn at 3919)
[hos host] vk: readback pages refreshed from guest memory: 1353050 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 103, 973 MB with memory of 1296 MB (322 MB spared); blocks given memory later 145, given back 1095; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 222 binds (29.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 34 taken from the spares, 112 made new; 43 kept as spares, 0 destroyed; 9 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 901, 4293.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1050 dirty pages from their submits' captures, 57601 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1367 dispatches, 42249 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6827 runs, 720.7 MB compared; 103 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15344 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 2 retired still imported (24 MB), 18 retired given back (380 MB)
[hos sead::MainThread] vk: readbacks of GPU results: 363 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1050 kept their virtual core (nothing there to wait for), 2790 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 1760000 (1092733 pages from the captures)
[hos host] vk: mirror of pool 0x7fe2800000 (4 MB) given back: idle, 6028 submissions since its sync
[hos host] vk: 10 textures retired (2052 left, 3076 MB): budget 0, drawn over 0, dead targets 10
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.16
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 164 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 1780000 (1103957 pages from the captures)
[hos host] vk: mirror of pool 0x7fe2400000 (4 MB) given back: idle, 6055 submissions since its sync
[hos host] vk: write-backs left out (600 submits): 903, 4297.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 518 dirty pages from their submits' captures, 57403 from earlier ones', 19 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1401 dispatches, 42064 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6652 runs, 716.1 MB compared; 103 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15551 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x1057400000 (4 MB, captured, state 2): 12; 0x1053400000 (4 MB, captured, state 2): 5; 0x7fe1800000 (4 MB, captured, state 2): 2;
[hos host] vk: replay reads of guest memory as submitted: 1800000 (1115448 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 802 ranges, 3436 MB), 257 pages written, 257 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (110 MB), 12 spare buffers, 148 texture pages copied (0 MB)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 347 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1206 kept their virtual core (nothing there to wait for), 3146 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos NavMeshDepWorker2] vcores: 151037 takes, 34875 waits (mean 88 us, longest 79.9 ms, 2 over HOS_VCORE_LOG_MS), 17510 kicks, 17630 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 11 textures retired (2058 left, 3075 MB): budget 0, drawn over 1, dead targets 10
[hos host] vk: replay reads of guest memory as submitted: 1820000 (1127805 pages from the captures)
[hos host] vk: 3 textures retired (2062 left, 3094 MB): budget 0, drawn over 3, dead targets 0
[hos sead::MainThread] nvn: CommandBufferSignalEvent #35000: event 0x100d6c8008 mode 0 value 34995 (lr 0x80660f34)
[hos host] vk: 10 textures retired (2055 left, 3082 MB): budget 0, drawn over 10, dead targets 0
[hos host] vk: sparse mirrors: 101, 972 MB with memory of 1288 MB (315 MB spared); blocks given memory later 145, given back 1095; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 224 binds (29.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 34 taken from the spares, 112 made new; 45 kept as spares, 0 destroyed; 11 spares now, 372 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 879, 4188.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1461 dirty pages from their submits' captures, 56887 from earlier ones', 120 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1409 dispatches, 41871 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6367 runs, 687.6 MB compared; 93 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14643 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x1055800000 (4 MB, captured, state 2): 23; 0x105e400000 (4 MB, captured, state 2): 21; 0x1054c00000 (4 MB, captured, state 2): 15; 0x1057000000 (4 MB, captured, state 2): 11; 0x1053c00000 (4 MB, captured, state 2): 11; 0x1057c00000 (4 MB, captured, state 2): 10; 0x1057800000 (4 MB, captured, state 2): 10; 0x1054800000 (4 MB, captured, state 2): 6;
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: replay reads of guest memory as submitted: 1840000 (1140358 pages from the captures)
[hos host] vk: 13 textures retired (2060 left, 3116 MB): budget 0, drawn over 3, dead targets 10
[hos host] vk: program picks: 2598639 kept (817264 by the last picks' textures, a uniform buffer bound since), 25257 new, 1176103 with other textures (of 3800000)
[hos host] vk: texture #9520 0x10009304600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4075 (drawn at 3949)
[hos sead::MainThread] nvn: 500000 recordings made, 70835 live copies (21.1 MB), 4274 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: a frame: 265 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.42, present 0.22, draw 6.42, descriptors 2.68, texture lookups 0.43, uploads 0.03, pipelines 0.29, programs 1.11, commands 1.82, dispatch 0.87, memory sync 0.36, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 148 (978 MB): 5.77 MB synced in 1046.6 copies, 17.57 MB written back
[hos host] vk: 4080 presents, 56.8 fps; a frame: 1775 draws (0 skipped), 150 dispatches, 177 passes, 170 barriers, 6 submits, 0.6 uploads (0.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2101 textures (3058 MB)
[hos host] vk: host time in GPU resource waits: 3.48 ms a frame (fences 1.97, timelines 1.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 548.0 full barriers recorded, 287.8 left out (nothing recorded since the one before), 29.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 41.5; 0x12: 90.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.8 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 160.9, an op 1247.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 10.4 ms a frame (the longest submission 4.1 ms); 3.9 textures made a frame: colour targets 1.9 (11.5 MB), depth 1.5 (7.7 MB), ASTC or BC 0.5 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.08 MB unswizzled on the CPU in 0.01 ms
[hos sead::MainThread] vk: readbacks of GPU results: 309 of 3000 waited for their write-back, 0.31 ms each; 0 looked at the stopped queues first, 38 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1064 kept their virtual core (nothing there to wait for), 2776 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 1860000 (1152873 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 749, 3563.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4295 dirty pages from their submits' captures, 53565 from earlier ones', 34 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1521 dispatches, 38774 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5877 runs, 619.7 MB compared; 78 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12746 in captures, 1 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x105b400000 (4 MB, captured, state 2): 13; 0x1059c00000 (4 MB, captured, state 2): 11; 0x1054000000 (4 MB, captured, state 2): 10;
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 1880000 (1165097 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 802 ranges, 3436 MB), 276 pages written, 276 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (116 MB), 12 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 1900000 (1177437 pages from the captures)
[hos host] vk: 10 textures retired (2130 left, 3120 MB): budget 0, drawn over 0, dead targets 10
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 118 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 103, 978 MB with memory of 1300 MB (321 MB spared); blocks given memory later 145, given back 1095; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 226 binds (29.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 36 taken from the spares, 112 made new; 45 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4640 dirty pages from their submits' captures, 53533 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1517 dispatches, 38831 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6100 runs, 606.6 MB compared; 82 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13138 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 176 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 4 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] vk: readbacks of GPU results: 353 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1227 kept their virtual core (nothing there to wait for), 3125 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 1 textures retired (2149 left, 3136 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: replay reads of guest memory as submitted: 1920000 (1189679 pages from the captures)
[hos host] vk: 13 textures retired (2176 left, 3132 MB): budget 0, drawn over 0, dead targets 13
[hos host] vk: program picks: 2729518 kept (853812 by the last picks' textures, a uniform buffer bound since), 25258 new, 1245223 with other textures (of 4000000)
[hos Presentation Thread] vi: 4154 frames presented (55.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 26, 15.5-17.8 153, 17.8-22 61, 22-30 34, 30-40 2, over 40 0
[hos sead::MainThread] nvn: CommandBufferWaitEvent #55000: event 0x100d6c8028 mode 0 value 37055 (now 37036) (lr 0x806674b0)
[hos host] vk: 144 textures retired (2074 left, 3171 MB): budget 142, drawn over 0, dead targets 2
[hos host] vk: replay reads of guest memory as submitted: 1940000 (1202143 pages from the captures)
[hos host] vk: presenter thread: the copy's fence 1.75 ms, the present 0.78 ms a present
[hos host] vk: readback pages refreshed from guest memory: 1540072 pages so far (0 held: may be written ahead in their submission)
[hos sead::MainThread] nvn: GPU waits in 5 s: 154 sync waits (400 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: write-backs left out (600 submits): 726, 3457.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4758 dirty pages from their submits' captures, 52798 from earlier ones', 24 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1500 dispatches, 38131 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6351 runs, 594.1 MB compared; 64 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13440 in captures, 0 only in the collection, 24 only by the replay itself
[hos host] vk: read live, by pool: 0x105a400000 (4 MB, captured, state 2): 24;
[hos host] vk: 2 textures retired (2056 left, 3052 MB): budget 0, drawn over 0, dead targets 2
[hos host] vk: replay reads of guest memory as submitted: 1960000 (1214975 pages from the captures)
[hos host] vk: 15 textures retired (2075 left, 3040 MB): budget 0, drawn over 0, dead targets 15
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 348 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1008 kept their virtual core (nothing there to wait for), 2832 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 1980000 (1227742 pages from the captures)
[hos host] vk: memory: working set 13219 MB (peak 13255), private 20304 MB, staging 48 MB in 3 arenas, 20 made so far (56 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 640 MB, 37 allocations 562 MB; used 2180 of 15579 MB; heap 1 (device): 181 blocks 5207 MB, 18473 allocations 4750 MB; used 7026 of 7823 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 22 blocks 704 MB, 139 images 523 MB; 113 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 74 blocks 2368 MB, 1955 images 2217 MB; 438 free ranges, the largest 14 MB
[hos host] vk: render target images: 4634 reused from the pool, 2122 made new, 1981 destroyed from it, 866 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 118 missed after their image was taken by size; it holds 25 (128 MB)
[hos host] vk: render targets not reused by size: 619 with the pool empty (0 of them with targets still in the dead queue), 161 with none in it a present yet, 22 with none of a kind they can use, 313 with all smaller (the largest 47% short on average), 355 with all bigger (over 1.5 times), 652 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 65 over its bound and 1916 idle
[hos host] vk: VRAM: 7026 MB used of 7823 MB allowed (456 MB of it free in the blocks); textures 3072 MB (budget 3171 MB), 2073 textures
[hos host] vk: budget: VRAM 5207 of 7823 MB (7026 MB by the driver, 1819 MB unaccounted): textures 3072, mirrors 978, blocks free 456, other buffers 211, shadows 200, target pool 128, other images 111, staging 48; RAM 2108 MB (the process's private memory 20304 MB): imports 1500, capture spares 434, captures 112, capture table 26, retired imports 20, other buffers 16, others 0
[hos host] vk: textures: decoded ASTC 1808 MB, rendered 884 MB (71 drawn in one frame, kept: 53518 KB), other 331 MB; used in the last 8 presents 2654 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 629 MB of blocks hashed in 86 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 4 presents ago
[hos host] vk: 21 MB: 0x10006fce600 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x10005192200 1600x1600x1 x3 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 25 (8617 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 64x62x1 NVN 0x25: 1 (15 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 128x126x1 NVN 0x1: 1 (15 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 208x105x1 NVN 0x1: 1 (21 KB); 262x104x1 NVN 0x1: 1 (26 KB); 310x33x1 NVN 0x25: 1 (39 KB); 312x42x1 NVN 0x25: 1 (51 KB); 328x33x1 NVN 0x25: 1 (42 KB); 374x97x1 NVN 0x1: 1 (35 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 441 waited for it, 2.52 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: status: 600 frames, the slowest 53.4 ms; VRAM 6.9 of 7.6 GB
[hos host] vk: sparse mirrors: 104, 978 MB with memory of 1304 MB (325 MB spared); blocks given memory later 145, given back 1095; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 227 binds (30.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 45 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3861 dirty pages from their submits' captures, 55695 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1559 dispatches, 39602 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6549 runs, 611.4 MB compared; 65 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13964 in captures, 0 only in the collection, 27 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5253 queued, 5253 loaded (median 1.38 ms, 99th percentile 16.77 ms, worst 34.27 ms, 13012 ms and 296.2 MB read in all); 7285 programs (1431 of them compute variants without their code, 0 graphics programs no note names), 10490 graphics pipelines (1 not in a cache), 577 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1463 pipelines made at create taken by draws; 2569 create calls with nothing kept for their stages, 0 before the index; units 7822 (5253 ready, 0 loading), 14420 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 11458 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 575 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 276 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 14.42, present 0.26, draw 8.84, descriptors 3.82, texture lookups 0.97, uploads 0.05, pipelines 0.40, programs 1.41, commands 2.55, dispatch 1.05, memory sync 0.44, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 149 (978 MB): 6.75 MB synced in 1150.1 copies, 17.91 MB written back
[hos host] vk: 4200 presents, 51.7 fps; a frame: 2366 draws (0 skipped), 179 dispatches, 196 passes, 180 barriers, 6 submits, 0.9 uploads (0.2 MB), 82 ub updates (57 overlays), GPU-bound wait 0.1 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2075 textures (3090 MB)
[hos host] vk: host time in GPU resource waits: 4.87 ms a frame (fences 2.01, timelines 2.86, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 598.0 full barriers recorded, 299.1 left out (nothing recorded since the one before), 34.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 42.9; 0x12: 99.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.9 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 170.9, an op 1708.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 12.2 ms a frame (the longest submission 3.6 ms); 7.1 textures made a frame: colour targets 1.6 (9.4 MB), depth 4.7 (13.8 MB), ASTC or BC 0.8 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.15 MB unswizzled on the CPU in 0.02 ms
[hos host] vk: the render thread (600 presents, 18.40 ms a frame): its jobs 17.28 ms a frame (submits 3.75, presents 4.87, others 8.67), idle 1.12; the game waited for room at 67 presents, 0.26 ms each (0.03 a frame; 33.09 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 449 captures at a CPU signal; the submits' waits read in 0.050 ms a frame; readbacks: 2389 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 16.63 ms after its hand-over (44.61 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.48 + 0.31 (the window's copy is in the idle), after it 3.11 + 1.32; after a frame's last recording that reads or writes the game's memory come 1.01 recordings (0.00 ms of GPU time), after its last that reads it 1.01 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 9.73 parts a frame recorded in the signal's job, 68.5% of them took nothing that came with the signal; of the others, 3.05 a frame took pages of the signal's capture into a mirror (94.7 pages a frame) and 0.06 pages read as guest memory had them (18.2 pages a frame); a frame's first part took nothing in 39 of 449 frames, its first three in 28
[hos host] vk: frame ahead, what came with the signal: 449 signals with a capture, each of 208.0 pages in 33.1 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.01 releases and 0.00 other CPU signals a signal (2 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.64+0.21, 0.25+0.07, 0.40+0.12, 0.36+0.10, 0.16+0.19, 0.07+0.30, 0.11+0.20, 0.05+0.06, 0.05+0.04, 0.01+0.02, 0.00+0.01, 0.00+0.00
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: render targets drawn over from another start, let go: 7457 so far (35065 MB), 6658 of them (31380 MB) between the passes
[hos host] vk: 30 textures retired (2068 left, 3007 MB): budget 0, drawn over 2, dead targets 28
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.03 over 815 ranges, 2842 MB), 281 pages written, 281 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (92 MB), 11 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 2000000 (1240256 pages from the captures)
[hos host] vk: 1 textures retired (2087 left, 3073 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: replay reads of guest memory as submitted: 2020000 (1252712 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 769, 3664.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4461 dirty pages from their submits' captures, 53629 from earlier ones', 10 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1441 dispatches, 39443 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6734 runs, 629.2 MB compared; 81 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14446 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x105a000000 (4 MB, captured, state 2): 10;
[hos host] vk: program picks: 2862222 kept (892622 by the last picks' textures, a uniform buffer bound since), 25271 new, 1312506 with other textures (of 4200000)
[hos host] vk: texture #11558 0x1000b1f5800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4235 (drawn at 4167)
[hos host] vk: texture #11517 0x1000b4c5800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4237 (drawn at 4170)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 359 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1173 kept their virtual core (nothing there to wait for), 3179 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 2040000 (1265683 pages from the captures)
[hos host] vk: queues ordered by their events: 24000 submissions stopped at a wait, 23998 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 57281, not met 0; captures at a CPU event's wait 3575, 0 of them at a wait met on arrival
[hos host] vk: texture #11860 0x1000a355400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4266 (drawn at 4205)
[hos host] vk: replay reads of guest memory as submitted: 2060000 (1278917 pages from the captures)
[hos host] vk: 11 textures retired (2076 left, 3012 MB): budget 0, drawn over 0, dead targets 11
[hos host] vk: sparse mirrors: 104, 978 MB with memory of 1304 MB (325 MB spared); blocks given memory later 145, given back 1097; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 229 binds (30.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 45 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 749, 3563.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4136 dirty pages from their submits' captures, 49907 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1277 dispatches, 37351 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6537 runs, 611.0 MB compared; 77 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14061 in captures, 0 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: recorder: 6932152 packets run in 10.0 s (busy 1.98 s, spun idle 0.13 s, slept 11938 times, 4113 of them under 200 us), 0 calls made directly (no room); 18295 batches over 542 frames (33.8 a frame, 379 packets a batch), 8915 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 217 times, 46.58 ms (0.4 times, 0.09 ms a frame); every thread 871 times, 95.42 ms in all, the longest 0.94 ms; vkDestroyImageView 287/51.84 ms (max 0.94), vkWaitForFences 473/33.09 ms (max 0.54), vkQueueBindSparse 34/2.47 ms (max 0.36), vmaDestroyBuffer 24/1.93 ms (max 0.41), vmaFreeMemory 10/1.72 ms (max 0.48), vkDestroyBuffer 10/1.39 ms (max 0.58)
[hos host] vk: recorder: of its busy 1.98 s, 0.48 s in parts' ends and submits (the longest 1.49 ms), 0.17 s of that blocked on the queue's lock (286 times, the longest 1.43 ms), so 0.31 s in the ends, submits and holds; 1.49 s running the other packets; 12790 packets a frame
[hos host] vk: 4 textures retired (2076 left, 3000 MB): budget 0, drawn over 4, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 2080000 (1292531 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 7 pages in 7 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] vk: readbacks of GPU results: 350 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1151 kept their virtual core (nothing there to wait for), 2945 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 80000 so far (x1 0x117b427000 x2 0 x3 0x117b427000 x4 0xf33080)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 815 ranges, 2842 MB), 272 pages written, 272 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (94 MB), 12 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 2100000 (1306304 pages from the captures)
[hos aal::LowPrioWorkerThread] vcores: 135512 takes, 26311 waits (mean 97 us, longest 34.1 ms, 0 over HOS_VCORE_LOG_MS), 12745 kicks, 13030 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 118 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 770, 3666.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3645 dirty pages from their submits' captures, 50859 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1318 dispatches, 37876 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6275 runs, 621.1 MB compared; 67 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14053 in captures, 0 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool:
[hos rmbl::VibrationThread] hid: vibration #11732: handle 00000003, low 0.04 @ 160 Hz, high 0.02 @ 252 Hz
[hos rmbl::VibrationThread] hid: vibration #11733: handle 00010003, low 0.04 @ 160 Hz, high 0.02 @ 252 Hz
[hos rmbl::VibrationThread] hid: vibration #11734: handle 00000003, low 0.05 @ 182 Hz, high 0.03 @ 182 Hz
[hos rmbl::VibrationThread] hid: vibration #11735: handle 00010003, low 0.05 @ 182 Hz, high 0.03 @ 182 Hz
[hos rmbl::VibrationThread] hid: vibration #11736: handle 00000003, low 0.06 @ 55 Hz, high 0.02 @ 231 Hz
[hos rmbl::VibrationThread] hid: vibration #11737: handle 00010003, low 0.06 @ 55 Hz, high 0.02 @ 231 Hz
[hos host] vk: mirror of pool 0x7f7dc00000 (8 MB) given back: idle, 6031 submissions since its sync
[hos host] vk: 10 textures retired (2081 left, 3002 MB): budget 0, drawn over 0, dead targets 10
[hos host] vk: mirror of pool 0x7f9c400000 (32 MB) given back: idle, 6063 submissions since its sync
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #40000: event 0x100d6c801c mode 0 value 39997 (lr 0x802911b8)
[hos BaseProcExecutor0] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 35 made): 24913091 lookups, 138131 missed (0.55 %)
[hos host] vk: replay reads of guest memory as submitted: 2120000 (1320208 pages from the captures)
[hos host] vk: a frame: 282 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.37, present 0.23, draw 7.65, descriptors 3.18, texture lookups 0.31, uploads 0.01, pipelines 0.36, programs 1.28, commands 2.23, dispatch 1.00, memory sync 0.41, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 149 (952 MB): 6.23 MB synced in 1098.0 copies, 17.87 MB written back
[hos host] vk: 4320 presents, 54.9 fps; a frame: 2088 draws (0 skipped), 173 dispatches, 180 passes, 172 barriers, 6 submits, 0.2 uploads (0.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2081 textures (3002 MB)
[hos host] vk: host time in GPU resource waits: 4.17 ms a frame (fences 2.17, timelines 1.99, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 567.7 full barriers recorded, 288.6 left out (nothing recorded since the one before), 29.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 42.1; 0x12: 92.1; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 162.8, an op 1491.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 12.3 ms a frame (the longest submission 5.8 ms); 3.0 textures made a frame: colour targets 1.4 (8.0 MB), depth 1.5 (5.2 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: program picks: 2993533 kept (928942 by the last picks' textures, a uniform buffer bound since), 25278 new, 1381188 with other textures (of 4400000)
[hos host] vk: replay reads of guest memory as submitted: 2140000 (1333903 pages from the captures)
[hos host] vk: texture #11831 0x1000a6b5600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4337 (drawn at 4195)
[hos host] vk: readback pages refreshed from guest memory: 1727578 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 102, 952 MB with memory of 1264 MB (311 MB spared); blocks given memory later 145, given back 1097; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 231 binds (30.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 47 kept as spares, 0 destroyed; 10 spares now, 396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 770, 3666.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3292 dirty pages from their submits' captures, 51029 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1313 dispatches, 38110 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6067 runs, 618.2 MB compared; 89 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13978 in captures, 0 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 344 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1203 kept their virtual core (nothing there to wait for), 3149 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 2160000 (1347732 pages from the captures)
[hos host] vk: 10 textures retired (2118 left, 3074 MB): budget 0, drawn over 0, dead targets 10
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #60000: event 0x100d6c8028 mode 0 value 40855 (now 40836) (lr 0x806674b0)
[hos host] vk: late pages' fix (600 recordings with one): of 1196 captures with late pages 1196 fixed (54.8 pages, 55.0 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 386 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 3 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 164, of an earlier recording 219, not recorded yet 0, of this part 0; pages fixed from stamps 64424, from snapshots 1094; the shadows: 573707 pages synced from captures, 559260 made known by it, 14447 left unknown because the replay or the collection wrote them since
[hos host] vk: replay reads of guest memory as submitted: 2180000 (1361410 pages from the captures)
[hos host] vk: texture #12104 0x10009034600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4369 (drawn at 4308)
[hos host] vk: texture #11080 0x100090c4800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4369 (drawn at 4308)
[hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3115 dirty pages from their submits' captures, 50979 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1307 dispatches, 38112 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6017 runs, 632.1 MB compared; 89 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13744 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 12 textures retired (2115 left, 3071 MB): budget 0, drawn over 0, dead targets 12
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 815 ranges, 2842 MB), 252 pages written, 252 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (110 MB), 12 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 2200000 (1375073 pages from the captures)
[hos host] vk: mirror of pool 0x105c000000 (4 MB) given back: idle, 6050 submissions since its sync
[hos host] vk: 48 mirrors given back (413 MB in all); 952 MB of mirrors now, 68 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: 24 textures retired (2105 left, 3018 MB): budget 0, drawn over 0, dead targets 24
[hos host] vk: texture #11935 0x10009b75000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4394 (drawn at 4212)
[hos sead::MainThread] vk: readbacks of GPU results: 357 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1155 kept their virtual core (nothing there to wait for), 3197 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #11903 0x10009e45400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4397 (drawn at 4207)
[hos host] vk: texture #11900 0x10009ae4e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4398 (drawn at 4214)
[hos host] vk: texture #11902 0x10009db5200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4398 (drawn at 4208)
[hos host] vk: 20 textures retired (2106 left, 3036 MB): budget 0, drawn over 0, dead targets 20
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 2220000 (1388777 pages from the captures)
[hos host] vk: mirror of pool 0x105d400000 (4 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: 27 textures retired (2094 left, 3171 MB): budget 26, drawn over 1, dead targets 0
[hos host] vk: sparse mirrors: 100, 952 MB with memory of 1256 MB (303 MB spared); blocks given memory later 145, given back 1097; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 233 binds (30.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 49 kept as spares, 0 destroyed; 12 spares now, 404 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3439 dirty pages from their submits' captures, 50858 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1313 dispatches, 37899 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5836 runs, 634.3 MB compared; 62 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14016 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: budget: gave back 48 MB from staging at 93% of the allowance
[hos host] vk: 395 textures retired (1719 left, 3186 MB): budget 394, drawn over 1, dead targets 0
[hos host] vk: program picks: 3123531 kept (963626 by the last picks' textures, a uniform buffer bound since), 25278 new, 1451190 with other textures (of 4600000)
[hos Presentation Thread] vi: 4421 frames presented (53.2 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 5, 15.5-17.8 89, 17.8-22 155, 22-30 18, 30-40 0, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 2240000 (1402313 pages from the captures)
[hos host] vk: 20 textures retired (1696 left, 2837 MB): budget 0, drawn over 0, dead targets 20
[hos host] vk: texture #12029 0x10009304a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4430 (drawn at 4233)
[hos sead::MainThread] nvn: GPU waits in 5 s: 233 sync waits (652 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: texture #12028 0x10009274800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4439 (drawn at 4234)
[hos host] vk: 2 textures retired (1695 left, 2827 MB): budget 0, drawn over 1, dead targets 1
[hos host] vk: a frame: 283 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.45, present 0.24, draw 7.69, descriptors 3.37, texture lookups 0.97, uploads 0.14, pipelines 0.35, programs 1.23, commands 2.17, dispatch 0.92, memory sync 0.39, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 149 (952 MB): 5.91 MB synced in 1059.6 copies, 17.58 MB written back
[hos host] vk: 4440 presents, 53.6 fps; a frame: 1993 draws (0 skipped), 153 dispatches, 174 passes, 170 barriers, 6 submits, 0.4 uploads (0.5 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1697 textures (2841 MB)
[hos host] vk: host time in GPU resource waits: 4.43 ms a frame (fences 1.31, timelines 3.12, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 545.2 full barriers recorded, 287.3 left out (nothing recorded since the one before), 26.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 42.5; 0x12: 89.9; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.9 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 161.0, an op 1401.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 12.4 ms a frame (the longest submission 4.7 ms); 3.7 textures made a frame: colour targets 1.4 (8.3 MB), depth 2.0 (11.6 MB), ASTC or BC 0.2 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.44 MB unswizzled on the CPU in 0.13 ms
[hos host] vk: replay reads of guest memory as submitted: 2260000 (1416192 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 348 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1034 kept their virtual core (nothing there to wait for), 2550 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 814, 3876.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3123 dirty pages from their submits' captures, 52914 from earlier ones', 8 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1361 dispatches, 39066 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6043 runs, 649.2 MB compared; 78 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14343 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool: 0x1055000000 (4 MB, captured, state 2): 6; 0x1057400000 (4 MB, captured, state 2): 1; 0x1055400000 (4 MB, captured, state 2): 1;
[hos rmbl::VibrationThread] hid: 12288 vibration values, strongest 0.06 since the last count
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 2280000 (1429977 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: reaper: 1243 images (3107 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 363 (189 MB)
[hos OptimizeChangeBanc] nvn: 550000 recordings made, 70620 live copies (22.4 MB), 5131 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: 1 textures retired (1708 left, 2848 MB): budget 0, drawn over 1, dead targets 0
[hos BaseProcCreateAndDeleteThread] vcores: BaseProcCreateAndDeleteThread (priority 23) waited 69.0 ms for a core: given by NavMeshCAStepDepWorker1 (priority 22), which held it 1.4 ms and was never asked; the cores it may use now: core 1: it;
[hos GeneralFormatAsync
] vcores: GeneralFormatAsync
(priority 24) waited 67.7 ms for a core: given by BaseProcCreateAndDeleteThread (priority 23), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 75.5 ms for a core: given by GeneralFormatAsync
(priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 54.2 ms for a core: given by tera::FileResourceMgr (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.03 over 712 ranges, 2719 MB), 298 pages written, 298 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (224 MB), 8 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 2300000 (1444306 pages from the captures)
[hos host] vk: texture #1212 0x10001fe0800 640x360 NVN 0x14: the GPU wrote its storage (submission 31172): uploaded again
[hos host] vk: upload over a drawn image: #1212 0x10001fe0800 640x360x1 x1 NVN 0x14 at present 4478 (texture frame 4479), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: texture 0x10001fe0800 640x360x1 x1 levels 1 NVN 0x14 (960 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 960 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: 56 textures retired (1732 left, 2877 MB): budget 0, drawn over 0, dead targets 56
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 118 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 100, 952 MB with memory of 1256 MB (303 MB spared); blocks given memory later 149, given back 1097; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 235 binds (30.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 49 kept as spares, 0 destroyed; 12 spares now, 404 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 814, 3876.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2950 dirty pages from their submits' captures, 56089 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1363 dispatches, 40024 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6298 runs, 649.6 MB compared; 84 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14568 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: (its written part: 4 of 4 KB readable, 1495 bytes not zero; levels L0)
[hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it
[hos host] vk: replay reads of guest memory as submitted: 2320000 (1458472 pages from the captures)
[hos ModuleSystemWorker5] vk: readbacks of GPU results: 356 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1040 kept their virtual core (nothing there to wait for), 3056 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.18
[hos host] vk: 11 textures retired (1763 left, 2904 MB): budget 0, drawn over 0, dead targets 11
[hos host] vk: texture #12707 0x10009034600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4511 (drawn at 4450)
[hos host] vk: replay reads of guest memory as submitted: 2340000 (1472083 pages from the captures)
[hos OptimizeChangeBanc] nvn: 600000 recordings made, 72275 live copies (23.6 MB), 5174 texture copies; 4 command buffers' segments (4.2 MB)
[hos host] vk: 11 textures retired (1769 left, 2911 MB): budget 0, drawn over 0, dead targets 11
[hos host] vk: readback pages refreshed from guest memory: 1914680 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 835, 3979.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2522 dirty pages from their submits' captures, 52803 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1229 dispatches, 39279 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6480 runs, 666.9 MB compared; 125 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14659 in captures, 2 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: program picks: 3253887 kept (996799 by the last picks' textures, a uniform buffer bound since), 25279 new, 1520833 with other textures (of 4800000)
[hos host] vk: replay reads of guest memory as submitted: 2360000 (1485841 pages from the captures)
[hos host] vk: texture #12803 0x1000a7a4600 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 4533 (drawn at 4472)
[hos host] vk: texture slots registered again at another size: 200000 (of 276207 registrations; slot 10968 480x270 -> 128x128)
[hos sead::MainThread] vk: readbacks of GPU results: 351 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1100 kept their virtual core (nothing there to wait for), 2996 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #12828 0x1001f0f5c00 285x98 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 4548 (drawn at 4462)
[hos host] vk: replay reads of guest memory as submitted: 2380000 (1499301 pages from the captures)
[hos host] vk: 1 textures retired (1836 left, 3119 MB): budget 0, drawn over 1, dead targets 0
[hos sead::MainThread] vk: capture: a submit 0.24 ms (lock 0.05, scan 0.06 over 1315 ranges, 3034 MB), 316 pages written, 316 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (114 MB), 9 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: texture #12900 0x10008f14200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4558 (drawn at 4497)
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #65000: event 0x100d6c8028 mode 0 value 44655 (now 44636) (lr 0x806674b0)
[hos host] vk: a frame: 228 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.16, present 0.21, draw 6.72, descriptors 2.89, texture lookups 0.45, uploads 0.07, pipelines 0.31, programs 1.11, commands 1.91, dispatch 0.98, memory sync 1.53, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 149 (952 MB): 5.78 MB synced in 1037.3 copies, 17.58 MB written back
[hos host] vk: 4560 presents, 52.6 fps; a frame: 1683 draws (0 skipped), 156 dispatches, 167 passes, 167 barriers, 6 submits, 0.8 uploads (0.4 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1785 textures (2995 MB)
[hos host] vk: host time in GPU resource waits: 3.72 ms a frame (fences 1.38, timelines 2.33, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 539.4 full barriers recorded, 282.9 left out (nothing recorded since the one before), 24.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 42.7; 0x12: 86.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 35.2 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 158.2, an op 1112.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 11.6 ms a frame (the longest submission 6.7 ms); 4.5 textures made a frame: colour targets 2.1 (9.6 MB), depth 1.9 (6.7 MB), ASTC or BC 0.5 (0.5 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.18 MB unswizzled on the CPU in 0.04 ms
[hos host] vk: mirror of pool 0x1053400000 (4 MB) given back: idle, 6009 submissions since its sync
[hos host] vk: sparse mirrors: 99, 951 MB with memory of 1252 MB (300 MB spared); blocks given memory later 149, given back 1100; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 237 binds (30.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 50 kept as spares, 0 destroyed; 13 spares now, 408 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 947, 4506.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2330 dirty pages from their submits' captures, 56041 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1188 dispatches, 42471 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7590 runs, 757.4 MB compared; 87 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17085 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: replay reads of guest memory as submitted: 2400000 (1512676 pages from the captures)
[hos host] vk: texture #12898 0x10008df3e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4569 (drawn at 4508)
[hos tera::FileResourceMgr] vcores: 155254 takes, 41666 waits (mean 97 us, longest 75.5 ms, 4 over HOS_VCORE_LOG_MS), 21743 kicks, 21974 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: texture #12897 0x10008d63c00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4572 (drawn at 4511)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #45000: event 0x100d6c8030 mode 0 value 44999 (lr 0x806612a8)
[hos host] vk: mirror of pool 0x1053000000 (4 MB) given back: idle, 6008 submissions since its sync
[hos host] vk: texture #13015 0x10008e83a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4579 (drawn at 4518)
[hos host] vk: 1 textures retired (1777 left, 2939 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: mirror of pool 0x1051000000 (4 MB) given back: idle, 6018 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 2420000 (1526032 pages from the captures)
[hos host] vk: mirror of pool 0x1042400000 (4 MB) given back: idle, 6063 submissions since its sync
[hos host] vk: 2 textures retired (1777 left, 2941 MB): budget 0, drawn over 2, dead targets 0
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 364 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1321 kept their virtual core (nothing there to wait for), 3287 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.13
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 2440000 (1539596 pages from the captures)
[hos host] vk: texture #13114 0x10009783600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4607 (drawn at 4546)
[hos host] vk: 2 textures retired (1778 left, 2945 MB): budget 0, drawn over 1, dead targets 1
[hos host] vk: texture #13112 0x100094b3200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4612 (drawn at 4551)
[hos host] vk: write-backs left out (600 submits): 1034, 4923.8 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1363 dirty pages from their submits' captures, 57564 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1169 dispatches, 44248 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8090 runs, 824.2 MB compared; 105 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18212 in captures, 0 only in the collection, 47 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 1 textures retired (1777 left, 2945 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 2460000 (1553041 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 2480000 (1566538 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 1315 ranges, 3030 MB), 237 pages written, 237 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (98 MB), 11 spare buffers, 148 texture pages copied (0 MB)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 357 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1069 kept their virtual core (nothing there to wait for), 2771 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: program picks: 3382464 kept (1021580 by the last picks' textures, a uniform buffer bound since), 25280 new, 1592255 with other textures (of 5000000)
[hos host] vk: sparse mirrors: 97, 948 MB with memory of 1244 MB (295 MB spared); blocks given memory later 149, given back 1100; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 241 binds (31.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 38 taken from the spares, 112 made new; 53 kept as spares, 0 destroyed; 15 spares now, 416 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1056, 5028.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 199 dirty pages from their submits' captures, 59576 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1224 dispatches, 45057 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7892 runs, 836.2 MB compared; 111 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17894 in captures, 0 only in the collection, 48 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: replay reads of guest memory as submitted: 2500000 (1580051 pages from the captures)
[hos host] vk: a frame: 191 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.23, present 0.16, draw 5.03, descriptors 2.06, texture lookups 0.17, uploads 0.03, pipelines 0.23, programs 0.97, commands 1.39, dispatch 0.76, memory sync 0.42, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 150 (948 MB): 4.89 MB synced in 940.7 copies, 17.47 MB written back
[hos host] vk: 4680 presents, 60.8 fps; a frame: 1328 draws (0 skipped), 139 dispatches, 157 passes, 157 barriers, 6 submits, 0.3 uploads (0.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1791 textures (2957 MB)
[hos host] vk: host time in GPU resource waits: 1.73 ms a frame (fences 1.07, timelines 0.66, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 513.1 full barriers recorded, 275.1 left out (nothing recorded since the one before), 20.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.0; 0x12: 78.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 38.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 148.1, an op 888.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 10.5 ms a frame (the longest submission 6.9 ms); 0.6 textures made a frame: colour targets 0.2 (1.1 MB), depth 0.1 (0.5 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: a submission's parts ended: 16201 where the game had ended its render pass, 3799 at 512 draws
[hos host] vk: replay reads of guest memory as submitted: 2520000 (1593577 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 357 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1085 kept their virtual core (nothing there to wait for), 3011 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos Presentation Thread] vi: 4703 frames presented (56.4 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 29, 15.5-17.8 163, 17.8-22 69, 22-30 21, 30-40 0, over 40 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 8 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1011, 4817.3 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 309 dirty pages from their submits' captures, 57340 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1169 dispatches, 43453 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7346 runs, 783.0 MB compared; 123 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16594 in captures, 0 only in the collection, 45 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 2540000 (1607044 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 156 sync waits (339 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 2560000 (1620597 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.05 over 1312 ranges, 3026 MB), 239 pages written, 239 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (108 MB), 10 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 2580000 (1634219 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 2102340 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 97, 948 MB with memory of 1244 MB (295 MB spared); blocks given memory later 149, given back 1101; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 242 binds (31.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 38 taken from the spares, 112 made new; 53 kept as spares, 0 destroyed; 15 spares now, 416 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 991, 4716.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 418 dirty pages from their submits' captures, 56377 from earlier ones', 281 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1172 dispatches, 42732 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7358 runs, 798.6 MB compared; 112 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16651 in captures, 0 only in the collection, 46 only by the replay itself
[hos host] vk: read live, by pool: 0x1052c00000 (4 MB, captured, state 2): 281;
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 360 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1103 kept their virtual core (nothing there to wait for), 2993 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #70000: event 0x100d6c8028 mode 0 value 48455 (now 48436) (lr 0x806674b0)
[hos host] vk: presenter thread: the copy's fence 1.40 ms, the present 0.77 ms a present
[hos host] vk: texture #13267 0x100092d3600 1280x720 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 4766 (drawn at 4705)
[hos host] vk: replay reads of guest memory as submitted: 2600000 (1648440 pages from the captures)
[hos host] vk: texture #12787 0x10008b53a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4771 (drawn at 4478)
[hos host] vk: texture #13446 0x100091e2e00 384x384 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (3 so far)
[hos host] vk: 5 textures retired (1833 left, 3170 MB): budget 5, drawn over 0, dead targets 0
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 100000 so far (x1 0x117d97b040 x2 0x2 x3 0x117d97b040 x4 0x8b040)
[hos host] vk: replay reads of guest memory as submitted: 2620000 (1662674 pages from the captures)
[hos host] vk: 148 textures retired (1695 left, 3166 MB): budget 148, drawn over 0, dead targets 0
[hos host] vk: memory: working set 13261 MB (peak 13264), private 20408 MB, staging 32 MB in 2 arenas, 22 made so far (56 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 640 MB, 36 allocations 564 MB; used 2180 of 15579 MB; heap 1 (device): 189 blocks 5463 MB, 17594 allocations 4826 MB; used 7282 of 7823 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 30 blocks 960 MB, 155 images 819 MB; 123 free ranges, the largest 10 MB
[hos host] vk: image blocks (textures, compacted): 74 blocks 2368 MB, 1553 images 2044 MB; 434 free ranges, the largest 32 MB
[hos host] vk: render target images: 5323 reused from the pool, 2734 made new, 2577 destroyed from it, 970 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 140 missed after their image was taken by size; it holds 11 (63 MB)
[hos host] vk: render targets not reused by size: 832 with the pool empty (0 of them with targets still in the dead queue), 194 with none in it a present yet, 24 with none of a kind they can use, 410 with all smaller (the largest 49% short on average), 476 with all bigger (over 1.5 times), 798 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 76 over its bound and 2501 idle
[hos host] vk: VRAM: 7282 MB used of 7823 MB allowed (636 MB of it free in the blocks); textures 3259 MB (budget 3171 MB), 1701 textures
[hos host] vk: budget: VRAM 5463 of 7823 MB (7282 MB by the driver, 1819 MB unaccounted): textures 3259, mirrors 948, blocks free 636, other buffers 211, shadows 200, other images 111, target pool 63, staging 32; RAM 2110 MB (the process's private memory 20408 MB): imports 1500, capture spares 462, captures 102, capture table 26, retired imports 20, others 0
[hos host] vk: textures: decoded ASTC 1638 MB, rendered 1205 MB (81 drawn in one frame, kept: 68153 KB), other 309 MB; used in the last 8 presents 2306 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 650 MB of blocks hashed in 89 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 4 presents ago
[hos host] vk: 21 MB: 0x10006fce600 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 14 MB: 0x10004dd1a00 1600x1600x1 x3 levels 1 NVN 0x32 rendered, last used 322 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 23 (8517 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 73x33x1 NVN 0x25: 1 (9 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 137x97x1 NVN 0x1: 1 (12 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 208x105x1 NVN 0x1: 1 (21 KB); 262x104x1 NVN 0x1: 1 (26 KB); 285x98x1 NVN 0x1: 1 (27 KB); 328x33x1 NVN 0x25: 1 (42 KB); 374x97x1 NVN 0x1: 1 (35 KB); 452x106x1 NVN 0x1: 1 (46 KB);
[hos host] vk: texture #12771 0x10004dd1a00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 322 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #12820 0x10005280e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 322 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #12793 0x10004741200 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 322 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #12791 0x10004831400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 322 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 87 textures retired (1614 left, 3171 MB): budget 86, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 30 MB
[hos Presentation Thread] vk: presents after the CPU's signal (600): 405 waited for it, 2.42 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: status: 600 frames, the slowest 31.4 ms; VRAM 7.1 of 7.7 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5253 queued, 5253 loaded (median 1.38 ms, 99th percentile 16.77 ms, worst 34.27 ms, 13012 ms and 296.2 MB read in all); 7285 programs (1431 of them compute variants without their code, 0 graphics programs no note names), 10490 graphics pipelines (1 not in a cache), 577 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1472 pipelines made at create taken by draws; 2569 create calls with nothing kept for their stages, 0 before the index; units 7822 (5253 ready, 0 loading), 14420 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 11458 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 575 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 150 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 7.99, present 0.19, draw 4.57, descriptors 1.93, texture lookups 0.41, uploads 0.00, pipelines 0.21, programs 0.83, commands 1.26, dispatch 0.76, memory sync 0.42, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 150 (948 MB): 4.84 MB synced in 933.9 copies, 17.30 MB written back
[hos host] vk: 4800 presents, 59.9 fps; a frame: 1169 draws (0 skipped), 129 dispatches, 163 passes, 158 barriers, 6 submits, 0.1 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1615 textures (3186 MB)
[hos host] vk: host time in GPU resource waits: 2.01 ms a frame (fences 1.34, timelines 0.67, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 520.4 full barriers recorded, 280.1 left out (nothing recorded since the one before), 20.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 41.0; 0x12: 79.3; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 38.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 149.3, an op 726.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 10.4 ms a frame (the longest submission 6.0 ms); 1.6 textures made a frame: colour targets 0.7 (4.0 MB), depth 0.8 (4.9 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 17.80 ms a frame): its jobs 16.97 ms a frame (submits 3.91, presents 5.98, others 7.08), idle 0.84; the game waited for room at 106 presents, 0.16 ms each (0.03 a frame; 38.73 jobs queued at a present); 3 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 7.87 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 407 captures at a CPU signal; the submits' waits read in 0.043 ms a frame; readbacks: 2260 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 16.28 ms after its hand-over (28.67 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.65 + 0.33 (the window's copy is in the idle), after it 2.67 + 0.71; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.86 parts a frame recorded in the signal's job, 66.9% of them took nothing that came with the signal; of the others, 2.93 a frame took pages of the signal's capture into a mirror (88.3 pages a frame) and 0.01 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 15 of 406 frames, its first three in 13
[hos host] vk: frame ahead, what came with the signal: 406 signals with a capture, each of 185.9 pages in 35.7 pools; between a frame's last submit and that signal 0.03 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (9 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.39+0.15, 0.26+0.09, 0.31+0.10, 0.37+0.09, 0.11+0.07, 0.08+0.07, 0.08+0.06, 0.04+0.03, 0.03+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00
[hos host] vk: write-backs left out (600 submits): 1077, 5131.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1025 dirty pages from their submits' captures, 57908 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1238 dispatches, 44872 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7286 runs, 842.2 MB compared; 110 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17015 in captures, 0 only in the collection, 48 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: readbacks of GPU results: 343 of 3000 waited for their write-back, 0.09 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1207 kept their virtual core (nothing there to wait for), 3145 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 2640000 (1677050 pages from the captures)
[hos host] vk: render targets drawn over from another start, let go: 8813 so far (41733 MB), 7826 of them (37136 MB) between the passes
[hos host] vk: 67 textures retired (1553 left, 3171 MB): budget 57, drawn over 10, dead targets 0
[hos host] vk: mirror of pool 0x7fe2800000 (4 MB) given back: idle, 6043 submissions since its sync
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 3509204 kept (1049807 by the last picks' textures, a uniform buffer bound since), 25280 new, 1665515 with other textures (of 5200000)
[hos host] vk: mirror of pool 0x7fe2400000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 2660000 (1691473 pages from the captures)
[hos ModuleSystemWorker4] nvn: 650000 recordings made, 73843 live copies (21.6 MB), 5456 texture copies; 3 command buffers' segments (4.1 MB)
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #50000: event 0x100d6c803c mode 0 value 49999 (lr 0x80661bfc)
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 1312 ranges, 3008 MB), 226 pages written, 226 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (108 MB), 12 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: recorder: 5746095 packets run in 10.0 s (busy 1.85 s, spun idle 0.13 s, slept 11371 times, 3885 of them under 200 us), 0 calls made directly (no room); 17338 batches over 564 frames (30.7 a frame, 331 packets a batch), 8137 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 120 times, 24.37 ms (0.2 times, 0.04 ms a frame); every thread 806 times, 77.07 ms in all, the longest 0.93 ms; vkWaitForFences 500/35.69 ms (max 0.58), vkDestroyImageView 246/33.67 ms (max 0.71), vmaDestroyBuffer 25/3.80 ms (max 0.93), vmaFreeMemory 13/3.10 ms (max 0.70), the queue's lock 4/0.50 ms (max 0.39), vkQueueBindSparse 16/0.27 ms (max 0.07)
[hos host] vk: recorder: of its busy 1.85 s, 0.51 s in parts' ends and submits (the longest 1.35 ms), 0.21 s of that blocked on the queue's lock (360 times, the longest 1.31 ms), so 0.29 s in the ends, submits and holds; 1.35 s running the other packets; 10188 packets a frame
[hos host] vk: mirror of pool 0x1042800000 (4 MB) given back: idle, 6047 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 2680000 (1706231 pages from the captures)
[hos host] vk: sparse mirrors: 94, 945 MB with memory of 1232 MB (286 MB spared); blocks given memory later 149, given back 1102; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 246 binds (31.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 38 taken from the spares, 112 made new; 56 kept as spares, 0 destroyed; 18 spares now, 428 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1057, 5030.3 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 880 dirty pages from their submits' captures, 57007 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1207 dispatches, 44268 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6939 runs, 821.8 MB compared; 118 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17316 in captures, 1 only in the collection, 49 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 356 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1104 kept their virtual core (nothing there to wait for), 2992 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos rmbl::VibrationThread] vcores: 160569 takes, 25082 waits (mean 70 us, longest 15.0 ms, 0 over HOS_VCORE_LOG_MS), 12588 kicks, 12577 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 2700000 (1721011 pages from the captures)
[hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 36 made): 21384318 lookups, 150505 missed (0.70 %)
[hos host] vk: mirror of pool 0x1050800000 (4 MB) given back: idle, 6028 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 2720000 (1735760 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: mirror of pool 0x7fa2800000 (8 MB) given back: idle, 6064 submissions since its sync
[hos host] vk: mirror of pool 0x7fd9400000 (4 MB) given back: idle, 6048 submissions since its sync
[hos host] vk: write-backs left out (600 submits): 1122, 5342.8 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1094 dirty pages from their submits' captures, 58913 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1220 dispatches, 45679 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7138 runs, 869.8 MB compared; 123 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17333 in captures, 0 only in the collection, 51 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: mirror of pool 0x7fbbc00000 (4 MB) given back: idle, 6032 submissions since its sync
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 358 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1117 kept their virtual core (nothing there to wait for), 2979 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 2 textures retired (1514 left, 2723 MB): budget 0, drawn over 1, dead targets 1
[hos host] vk: a frame: 112 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 7.45, present 0.22, draw 4.33, descriptors 1.92, texture lookups 0.27, uploads 0.03, pipelines 0.20, programs 0.74, commands 1.18, dispatch 0.79, memory sync 0.43, queue submit 0.00, slot wait 0.19
[hos host] vk: a frame: mirrors 151 (934 MB): 4.64 MB synced in 903.0 copies, 17.07 MB written back
[hos host] vk: 4920 presents, 60.1 fps; a frame: 921 draws (0 skipped), 114 dispatches, 161 passes, 161 barriers, 6 submits, 0.1 uploads (0.2 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1513 textures (2722 MB)
[hos host] vk: host time in GPU resource waits: 2.79 ms a frame (fences 1.93, timelines 0.87, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 479.3 full barriers recorded, 273.8 left out (nothing recorded since the one before), 23.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.0; 0x12: 82.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 38.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 152.2, an op 555.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 10.0 ms a frame (the longest submission 8.7 ms); 2.5 textures made a frame: colour targets 1.6 (9.7 MB), depth 0.8 (5.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.14 MB unswizzled on the CPU in 0.03 ms
[hos host] vk: replay reads of guest memory as submitted: 2740000 (1750080 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 1307 ranges, 3008 MB), 222 pages written, 222 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.010; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (124 MB), 11 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 4 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: mirror of pool 0x7fd3800000 (4 MB) given back: idle, 6041 submissions since its sync
[hos host] vk: mirror of pool 0x105e400000 (4 MB) given back: idle, 6013 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 2760000 (1763383 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 89, 932 MB with memory of 1208 MB (275 MB spared); blocks given memory later 149, given back 1103; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 252 binds (31.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 39 taken from the spares, 112 made new; 62 kept as spares, 0 destroyed; 23 spares now, 452 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1165, 5550.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 458 dirty pages from their submits' captures, 59778 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1285 dispatches, 46743 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7005 runs, 900.7 MB compared; 154 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17772 in captures, 0 only in the collection, 52 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: texture #13642 0x10008cd3a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4956 (drawn at 4895)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #75000: event 0x100d6c8028 mode 0 value 52255 (now 52236) (lr 0x806674b0)
[hos ModuleSystemWorker5] vk: readbacks of GPU results: 364 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1056 kept their virtual core (nothing there to wait for), 2784 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #13631 0x10008c43800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4962 (drawn at 4901)
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34770): uploaded again
[hos host] vk: upload over a drawn image: #12876 0x10001fe0800 1280x720x1 x1 NVN 0x5 at present 4967 (texture frame 4968), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: (its written part: 842 of 842 KB readable, 0 bytes not zero; levels L0)
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): what the CPU wrote, 842 KB of it
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34777): uploaded again
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34784): uploaded again
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34791): uploaded again
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it
[hos host] vk: texture #13616 0x10008bb3600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4973 (drawn at 4912)
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34805): uploaded again
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it
[hos host] vk: (its written part: 842 of 842 KB readable, 0 bytes not zero; levels L0)
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): what the CPU wrote, 842 KB of it
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34812): uploaded again
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it
[hos host] vk: texture #13573 0x10008a93200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4976 (drawn at 4915)
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34826): uploaded again
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it
[hos host] vk: (its written part: 842 of 842 KB readable, 0 bytes not zero; levels L0)
[hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): what the CPU wrote, 842 KB of it
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34833): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34847): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34854): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34868): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34875): uploaded again
[hos host] vk: texture #13562 0x10008a03000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4985 (drawn at 4924)
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34889): uploaded again
[hos host] vk: texture #13356 0x10008972e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4986 (drawn at 4925)
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34896): uploaded again
[hos host] vk: replay reads of guest memory as submitted: 2780000 (1776037 pages from the captures)
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34910): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34917): uploaded again
[hos host] vk: mirror of pool 0x7fd3c00000 (16 MB) given back: idle, 6056 submissions since its sync
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34924): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34931): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34945): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34952): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34966): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34973): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34987): uploaded again
[hos sead::MainThread] nvn: EventSignal (CPU) #10000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34994): uploaded again
[hos host] vk: mirror of pool 0x105dc00000 (4 MB) given back: idle, 6011 submissions since its sync
[hos host] vk: 64 mirrors given back (450 MB in all); 917 MB of mirrors now, 104 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35008): uploaded again
[hos Presentation Thread] vi: 5003 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 29, 15.5-17.8 249, 17.8-22 20, 22-30 1, 30-40 1, over 40 0
[hos host] vk: readback pages refreshed from guest memory: 2289945 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35022): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35029): uploaded again
[hos host] vk: write-backs left out (600 submits): 1167, 5554.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 132 dirty pages from their submits' captures, 58485 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1236 dispatches, 46047 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6698 runs, 892.7 MB compared; 136 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17502 in captures, 0 only in the collection, 54 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35043): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35050): uploaded again
[hos host] vk: 10 textures retired (1542 left, 2779 MB): budget 0, drawn over 0, dead targets 10
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35064): uploaded again
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 354 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1229 kept their virtual core (nothing there to wait for), 3123 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35071): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35085): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35092): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35106): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35113): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35127): uploaded again
[hos host] vk: 1 textures retired (1551 left, 2795 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35141): uploaded again
[hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35148): uploaded again
[hos host] vk: texture #869 0x10004bf1c00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5027 (drawn at 4966)
[hos host] vk: texture #12878 0x10004a11800 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 5027 (drawn at 4966)
[hos host] vk: texture #870 0x10004ce1e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5027 (drawn at 4966)
[hos host] vk: replay reads of guest memory as submitted: 2800000 (1788751 pages from the captures)
[hos host] vk: mirror of pool 0x7fe1800000 (4 MB) given back: idle, 6057 submissions since its sync
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.05 over 1306 ranges, 3008 MB), 216 pages written, 216 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a frame: 49 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 5.97, present 0.16, draw 3.20, descriptors 1.42, texture lookups 0.17, uploads 0.13, pipelines 0.16, programs 0.57, commands 0.84, dispatch 0.72, memory sync 0.41, queue submit 0.00, slot wait 0.19
[hos host] vk: a frame: mirrors 151 (917 MB): 4.33 MB synced in 868.1 copies, 16.84 MB written back
[hos host] vk: 5040 presents, 60.0 fps; a frame: 636 draws (0 skipped), 101 dispatches, 144 passes, 156 barriers, 6 submits, 0.9 uploads (1.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1572 textures (2904 MB)
[hos host] vk: host time in GPU resource waits: 2.01 ms a frame (fences 1.86, timelines 0.15, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 449.1 full barriers recorded, 266.4 left out (nothing recorded since the one before), 14.3 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 41.0; 0x12: 77.1; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 38.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 146.5, an op 389.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0;
[hos host] vk: GPU time 9.3 ms a frame (the longest submission 5.9 ms); 1.4 textures made a frame: colour targets 0.7 (3.2 MB), depth 0.5 (2.5 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.96 MB unswizzled on the CPU in 0.11 ms
[hos sead::MainThread] nvn: GPU waits in 5 s: 67 sync waits (97 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: queues ordered by their events: 30000 submissions stopped at a wait, 29996 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 77142, not met 0; captures at a CPU event's wait 4012, 0 of them at a wait met on arrival
[hos host] vk: sparse mirrors: 86, 917 MB with memory of 1184 MB (266 MB spared); blocks given memory later 149, given back 1103; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 255 binds (32.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 39 taken from the spares, 112 made new; 65 kept as spares, 0 destroyed; 26 spares now, 476 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1166, 5552.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 141 dirty pages from their submits' captures, 57370 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1231 dispatches, 45440 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5967 runs, 881.5 MB compared; 139 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16882 in captures, 0 only in the collection, 53 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB)
[hos host] vk: reaper: 800 images (2518 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 147 (156 MB)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 351 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1080 kept their virtual core (nothing there to wait for), 3016 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 2820000 (1801600 pages from the captures)
[hos host] vk: 15 textures retired (1585 left, 3170 MB): budget 11, drawn over 4, dead targets 0
[hos host] vk: program picks: 3625076 kept (1071696 by the last picks' textures, a uniform buffer bound since), 25283 new, 1749640 with other textures (of 5400000)
[hos host] vk: 164 textures retired (1441 left, 3170 MB): budget 146, drawn over 8, dead targets 10
[hos host] vk: 55 textures retired (1397 left, 3171 MB): budget 49, drawn over 6, dead targets 0
[hos host] vk: texture #13862 0x10008492200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 5098 (drawn at 5008)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos sead::MainThread] nvn: CommandBufferSignalEvent #55000: event 0x100d6c8048 mode 0 value 54999 (lr 0x80666598)
[hos host] vk: replay reads of guest memory as submitted: 2840000 (1814546 pages from the captures)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 356 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1088 kept their virtual core (nothing there to wait for), 3008 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 1254, 5971.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 183 dirty pages from their submits' captures, 60591 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1296 dispatches, 48486 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6264 runs, 945.8 MB compared; 144 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17964 in captures, 1 only in the collection, 57 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.05 over 1299 ranges, 2998 MB), 215 pages written, 215 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.007; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (122 MB), 11 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: pool 0x102bc88200+0xf00000 imported in 2.7 ms as 0x102bc00000..0x102cc00000 (133 imports, 345 ms in all)
[hos host] vk: pool 0x1012087e00+0xf00000 imported in 2.5 ms as 0x1012000000..0x1014000000 (134 imports, 348 ms in all)
[hos host] vk: a 128 MB capture buffer made ahead in 82.5 ms, off the game's thread
[hos host] vk: a frame: 23 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 5.46, present 0.15, draw 2.41, descriptors 1.15, texture lookups 0.60, uploads 0.24, pipelines 0.10, programs 0.41, commands 0.59, dispatch 0.61, memory sync 0.62, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 152 (949 MB): 3.97 MB synced in 756.6 copies, 14.81 MB written back
[hos host] vk: 5160 presents, 58.3 fps; a frame: 430 draws (0 skipped), 81 dispatches, 120 passes, 135 barriers, 6 submits, 0.8 uploads (1.6 MB), 74 ub updates (52 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1338 textures (2674 MB)
[hos host] vk: host time in GPU resource waits: 1.80 ms a frame (fences 1.62, timelines 0.18, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 380.8 full barriers recorded, 231.9 left out (nothing recorded since the one before), 9.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.8; 0x8: 2.7; 0x9: 0.9; 0x10: 1.0; 0x11: 36.6; 0x12: 67.7; 0x13: 1.8; 0x21: 3.6; 0x28: 14.3; 0x92: 2.7; 0x97: 0.9;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 32.8 from 21.4 batches at the game's barriers (33.9 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 126.7, an op 273.0, the end 6.3, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.8; 5: 17.8; 48: 5.3;
[hos host] vk: GPU time 8.4 ms a frame (the longest submission 6.7 ms); 0.9 textures made a frame: colour targets 0.2 (1.4 MB), depth 0.6 (4.6 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.33 MB unswizzled on the CPU in 0.19 ms
[hos sead::MainThread] vcores: 161566 takes, 26774 waits (mean 61 us, longest 14.0 ms, 0 over HOS_VCORE_LOG_MS), 13412 kicks, 13306 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: sparse mirrors: 87, 949 MB with memory of 1216 MB (266 MB spared); blocks given memory later 149, given back 1103; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 256 binds (32.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 40 taken from the spares, 112 made new; 65 kept as spares, 0 destroyed; 25 spares now, 444 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 681, 3245.9 MB; 22 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 72 dirty pages from their submits' captures, 32581 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 26270 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 3410 runs, 515.4 MB compared; 572 pages read into the shadows; 232 MB of shadows (116 MB of pages known, in 137 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 9855 in captures, 2 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 113 live (1520 MB), 2 retired still imported (48 MB), 19 retired given back (384 MB)
[hos Resource Control] vcores: Resource Control (priority 20) waited 57.2 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 7.2 ms and was never asked; the cores it may use now: core 1: it;
[hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 65.1 ms for a core: given by Resource Memory (priority 21), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos bake::ModuleMgr] vcores: bake::ModuleMgr (priority 28) waited 65.3 ms for a core: given by PreActorModelCompose (priority 26), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it;
[hos host] vk: 1 textures retired (1338 left, 2673 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.13
[hos host] vk: a 128 MB capture buffer made ahead in 28.3 ms, off the game's thread
[hos Presentation Thread] vk: capture: a submit 0.61 ms (lock 0.02, scan 0.09 over 1157 ranges, 2993 MB), 755 pages written, 755 copied (2.9 MB)
[hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (330 MB), 8 spare buffers, 372 texture pages copied (1 MB)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #80000: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos host] vk: 1 textures retired (1337 left, 2673 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 459 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 5 pages in 5 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 22 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 13 dirty pages from their submits' captures, 400 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 311 dispatches, 413 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 78 runs, 0.0 MB compared; 78 pages read into the shadows; 232 MB of shadows (112 MB of pages known, in 131 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 1053 in captures, 19 only in the collection, 0 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: mirror of pool 0x7fde000000 (4 MB) given back: idle, 6028 submissions since its sync
[hos host] vk: 1 textures retired (1336 left, 2673 MB): budget 0, drawn over 1, dead targets 0
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 238 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1365 kept their virtual core (nothing there to wait for), 3243 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 1.82, present 0.15, draw 0.09, descriptors 0.03, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.01, commands 0.04, dispatch 0.00, memory sync 0.97, queue submit 0.00, slot wait 0.27
[hos host] vk: a frame: mirrors 152 (698 MB): 0.02 MB synced in 5.4 copies, 0.00 MB written back
[hos host] vk: 5280 presents, 57.4 fps; a frame: 5 draws (0 skipped), 0 dispatches, 2 passes, 1 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 12784 pipelines (3060 ms), 0 programs (10 failed), 1336 textures (2673 MB)
[hos host] vk: host time in GPU resource waits: 5.80 ms a frame (fences 4.96, timelines 0.84, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 33.7 full barriers recorded, 13.9 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x12: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 0.0, an op 1.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.1 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: pool 0x112a982c00+0x2685000 imported in 3.7 ms as 0x112a800000..0x112d400000 (135 imports, 351 ms in all)
[hos host] vk: pool 0x7f9933c100+0x5550 imported in 4.5 ms as 0x7f99000000..0x7f9a000000 (136 imports, 356 ms in all)
[hos host] vk: pool 0x7f8b1de000+0x22ea3c imported in 2.3 ms as 0x7f8ac00000..0x7f8b800000 (137 imports, 358 ms in all)
[hos host] vk: slow upload 30.4 ms: #14101 0x10fff15400 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB
[hos host] vk: pool 0x108c711000+0xd5b6000 imported in 13.1 ms as 0x108c000000..0x109a000000 (138 imports, 371 ms in all)
[hos host] vk: pool 0x108c000000..0x109a000000: mirror made with the GPU's contents of 1 retired pool(s)
[hos host] vk: pool 0x7f88551b00+0x22f74 imported in 18.7 ms as 0x7f88000000..0x7f88c00000 (139 imports, 390 ms in all)
[hos host] vk: pool 0x7fa16b0600+0x8420 imported in 3.4 ms as 0x7fa1400000..0x7fa1800000 (140 imports, 393 ms in all)
[hos host] vk: pool 0x7f9bc33800+0x30510 imported in 3.2 ms as 0x7f9ac00000..0x7f9c000000 (141 imports, 396 ms in all)
[hos host] vk: pool 0x7f87f0d300+0x19ea0 imported in 10.0 ms as 0x7f86c00000..0x7f88000000 (142 imports, 406 ms in all)
[hos host] vk: pool 0x7f83f34000+0xca00 imported in 7.4 ms as 0x7f83c00000..0x7f84000000 (143 imports, 414 ms in all)
[hos host] vk: slow submit: 136.2 ms: draw 220.7 descriptors 50.4 texture lookups 49.2 uploads 41.9 programs 3.9 commands 165.8
[hos host] vk: pool 0x117efc9ac0+0x540 imported in 2.3 ms as 0x117ec00000..0x117f000000 (146 imports, 419 ms in all)
[hos host] vk: pool 0x1181000000..0x1181800000 took in 2 pools
[hos host] vk: pool 0x1181000000..0x1181800000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: pool 0x11807c7000+0x595000 imported in 4.0 ms as 0x1180400000..0x1181000000 (150 imports, 426 ms in all)
[hos host] vk: pool 0x117e8ea000+0x6d3000 imported in 9.3 ms as 0x1178400000..0x117f000000 (151 imports, 435 ms in all)
[hos host] vk: pool 0x1178400000..0x117f000000 took in 2 pools
[hos host] vk: pool 0x1178400000..0x117f000000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: pool 0x118163c000+0x215a00 imported in 5.2 ms as 0x1181000000..0x1181c00000 (152 imports, 440 ms in all)
[hos host] vk: pool 0x1181000000..0x1181c00000 took in 2 pools
[hos host] vk: pool 0x1181000000..0x1181c00000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: slow submit: 121.6 ms: dispatch 50.1
[hos host] vk: pool 0x1180d5d000+0x430000 imported in 8.8 ms as 0x1180400000..0x1181c00000 (153 imports, 449 ms in all)
[hos host] vk: pool 0x1180400000..0x1181c00000 took in 2 pools
[hos host] vk: pool 0x1180400000..0x1181c00000: mirror made with the GPU's contents of 2 retired pool(s)
[hos host] vk: pool 0x117efcb000+0x17fb000 imported in 12.7 ms as 0x1178400000..0x1181c00000 (154 imports, 462 ms in all)
[hos host] vk: pool 0x1178400000..0x1181c00000 took in 2 pools
[hos host] vk: pool 0x7f96677700+0x84 imported in 23.3 ms as 0x7f95800000..0x7f96800000 (155 imports, 485 ms in all)
[hos host] vk: pool 0x7f97ad9100+0xb970 imported in 17.8 ms as 0x7f97800000..0x7f98400000 (156 imports, 503 ms in all)
[hos host] vk: pool 0x7f8b934900+0x350dc imported in 7.1 ms as 0x7f8b800000..0x7f8d000000 (157 imports, 510 ms in all)
[hos host] vk: pool 0x7fffdd5000+0x8fe00 imported in 11.7 ms as 0x7fff800000..0x8000000000 (158 imports, 522 ms in all)
[hos host] vk: pool 0x7f9a67e200+0xe0400 imported in 3.5 ms as 0x7f9a400000..0x7f9a800000 (159 imports, 525 ms in all)
[hos host] vk: pool 0x7f9ee14000+0xedd00 imported in 6.2 ms as 0x7f9ec00000..0x7f9f000000 (161 imports, 532 ms in all)
[hos host] vk: pool 0x7f98a95e00+0x14a0 imported in 2.0 ms as 0x7f98800000..0x7f99000000 (162 imports, 534 ms in all)
[hos host] vk: pool 0x7f96982800+0xc844 imported in 5.1 ms as 0x7f96800000..0x7f96c00000 (163 imports, 539 ms in all)
[hos host] vk: pool 0x7f96f8d500+0xc0 imported in 2.0 ms as 0x7f96c00000..0x7f97400000 (165 imports, 543 ms in all)
[hos host] vk: pool 0x7f95038000+0xc79c imported in 3.6 ms as 0x7f95000000..0x7f95400000 (166 imports, 547 ms in all)
[hos host] vk: pool 0x7f94c3ea00+0x1b6a8 imported in 14.7 ms as 0x7f94000000..0x7f95000000 (167 imports, 561 ms in all)
[hos host] vk: pool 0x7f86079c00+0x1b404 imported in 4.7 ms as 0x7f86000000..0x7f86400000 (168 imports, 566 ms in all)
[hos host] vk: pool 0x7f91aad400+0x23040 imported in 4.2 ms as 0x7f91000000..0x7f93000000 (169 imports, 570 ms in all)
[hos host] vk: pool 0x7f8577ef00+0x42a54 imported in 7.8 ms as 0x7f85400000..0x7f85800000 (170 imports, 578 ms in all)
[hos host] vk: pool 0x7f896f9a00+0xb340 imported in 2.2 ms as 0x7f89000000..0x7f8a000000 (172 imports, 582 ms in all)
[hos host] vk: pool 0x7fa92b5600+0x1490 imported in 7.2 ms as 0x7fa9000000..0x7fa9400000 (173 imports, 589 ms in all)
[hos host] vk: pool 0x7f86a2a100+0x86ef4 imported in 2.5 ms as 0x7f86400000..0x7f86c00000 (174 imports, 592 ms in all)
[hos host] vk: pool 0x7f9ff8a500+0x9d68 imported in 5.0 ms as 0x7f9fc00000..0x7fa0000000 (175 imports, 597 ms in all)
[hos host] vk: slow submit: 417.5 ms: draw 62.0 texture lookups 1.0 commands 60.6
[hos host] vk: pool 0x7f8f940700+0x9354 imported in 3.1 ms as 0x7f8f800000..0x7f8fc00000 (180 imports, 604 ms in all)
[hos host] vk: pool 0x7f9f800000 (4 MB): mirror made at its first binding (201 so)
[hos host] vk: texture #13584 0x10008b23400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 4915)
[hos host] vk: texture #13858 0x10004dd1a00 1600x1600 x3 NVN 0x32 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 5147)
[hos host] vk: upload over a drawn image: #13858 0x10004dd1a00 1600x1600x1 x3 NVN 0x32 at present 5290 (texture frame 5291), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: texture 0x10004dd1a00 1600x1600x1 x3 levels 1 NVN 0x32 (15600 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15600 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: texture #13865 0x10004831400 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 5147)
[hos host] vk: texture #13866 0x10004921600 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 5147)
[hos host] vk: texture #13893 0x10005280e00 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 5146)
[hos host] vk: upload over a drawn image: #13893 0x10005280e00 2560x1440x1 x1 NVN 0x25 at present 5290 (texture frame 5291), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: texture 0x10005280e00 2560x1440x1 x1 levels 1 NVN 0x25 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: texture 0x10f69cfc00 512x512x1 x1 levels 10 NVN 0x49 -> vk 139 (1500 textures, 3140 MB)
[hos host] vk: the driver's allowance is the count itself (7842 MB used of 7842): past it, taken as 7058 MB
[hos host] vk: slow submit: 104.5 ms: draw 70.1 descriptors 54.1 texture lookups 53.9 uploads 39.1 pipelines 3.5 programs 12.1
[hos host] vk: a frame's uploads took 82 ms: ASTC on the GPU 20 (15.2 MB, 38 ms); unswizzled on the GPU 2 (53.5 MB, 4 ms); on the CPU 73 (163.0 MB, 40 ms); [1024x1024 x1 NVN 0x7b ASTC on the GPU 0.9 MB 30 ms] [4096x4096 x1 NVN 0x4b on the CPU 16.0 MB 6 ms] [2048x2048 x3 NVN 0x50 on the CPU 12.0 MB 5 ms]
[hos sead::MainThread] nvn: GPU waits in 5 s: 11 sync waits (904 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos Presentation Thread] vi: 5292 frames presented (50.1 fps over the last 5.8 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 37, 15.5-17.8 216, 17.8-22 22, 22-30 4, 30-40 6, over 40 4
[hos host] vk: slow submit: 25.3 ms: draw 7.9 descriptors 5.7 texture lookups 4.5 uploads 4.0 programs 1.1
[hos host] vk: slow upload 24.5 ms: #14230 0x10f88e4c00 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB
[hos host] vk: slow submit: 28.8 ms: draw 28.4 descriptors 27.3 texture lookups 26.1 uploads 26.0
[hos host] vk: the driver's allowance is the count itself (7797 MB used of 7797): past it, taken as 7017 MB
[hos host] vk: replay reads of guest memory as submitted: 2860000 (1827086 pages from the captures)
[hos host] vk: pool 0x7faf5d5200+0x38fc0 imported in 2.4 ms as 0x7faf400000..0x7faf800000 (193 imports, 622 ms in all)
[hos host] vk: slow upload 16.4 ms: #14471 0x10fa125800 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB
[hos host] vk: slow upload 38.1 ms: #14677 0x11177bc400 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB
[hos host] vk: texture 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e -> vk 137 (ASTC decoded) (1941 textures, 4061 MB)
[hos host] [debug] vk: a staging arena of 57 MB took 55.8 ms to make
[hos host] vk: slow upload 195.5 ms: #14681 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC (BC3), 57825 of 57825 KB
[hos host] vk: texture 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 -> vk 137 (ASTC decoded) (1942 textures, 4279 MB)
[hos host] [debug] vk: a staging arena of 57 MB took 60.7 ms to make
[hos host] vk: slow upload 217.5 ms: #14682 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC (BC3), 57825 of 57825 KB
[hos host] vk: slow submit: 940.6 ms: draw 972.5 descriptors 869.6 texture lookups 852.6 uploads 512.7 programs 44.7 commands 56.7 queue submit 1.0
[hos host] vk: a frame's uploads took 514 ms: ASTC on the GPU 41 (166.9 MB, 509 ms); on the CPU 152 (25.2 MB, 5 ms); [1365x1365 x90 NVN 0x80 ASTC on the GPU 56.5 MB 218 ms] [1365x1365 x90 NVN 0x8e ASTC on the GPU 56.5 MB 195 ms] [1536x1536 x1 NVN 0x7b ASTC on the GPU 2.0 MB 38 ms]
[hos host] vk: 8 of 8 small textures the GPU drew written into their memory in 2.1 ms, 1 frames after their copies were recorded, the first #14061 0x1168866000 260x260 levels 9 NVN 0xd (9 so far, 2204 KB, 3 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: slow submit: 16.2 ms: draw 14.7 descriptors 14.5 texture lookups 14.4 uploads 14.4
[hos host] vk: the driver's allowance is the count itself (8424 MB used of 8424): past it, taken as 7581 MB
[hos host] vk: ASTC texture #14738 0x1132250e00 1536x1536 x1 levels 11: the image of #14200 0x10e174b600 (the same blocks, finalized by the game) copied, not decoded again (1 so far, 11 MB)
[hos host] vk: slow submit: 25.2 ms: draw 20.5 descriptors 13.9 programs 2.3 commands 3.1
[hos host] vk: 8 of 8 small textures the GPU drew written into their memory in 1.9 ms, 1 frames after their copies were recorded, the first #14252 0x1169e3e000 260x260 levels 9 NVN 0xd (17 so far, 3880 KB, 5 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: the driver's allowance is the count itself (8379 MB used of 8379): past it, taken as 7541 MB
[hos host] vk: mirror of pool 0x108c000000 (48 MB) given back: retired, 69 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x1181400000 (4 MB) given back: retired, 68 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x1181000000 (4 MB) given back: retired, 68 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x117ec00000 (4 MB) given back: retired, 68 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x1178400000 (104 MB) given back: retired, 68 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x1181800000 (4 MB) given back: retired, 68 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x1181000000 (8 MB) given back: retired, 68 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x1181000000 (12 MB) given back: retired, 67 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x1180400000 (12 MB) given back: retired, 68 submissions since its sync, GPU-only data
[hos host] vk: mirror of pool 0x1180400000 (24 MB) given back: retired, 67 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: mirror of pool 0x1178400000 (108 MB) given back: retired, 67 submissions since its sync, GPU-only data, GPU-written
[hos host] vk: replay reads of guest memory as submitted: 2880000 (1845318 pages from the captures)
[hos host] vk: slow submit: 37.4 ms: draw 29.7 descriptors 24.2 texture lookups 12.6 uploads 12.4 programs 2.0 commands 2.6
[hos host] vk: 7 of 7 small textures the GPU drew written into their memory in 1.5 ms, 1 frames after their copies were recorded, the first #14268 0x116b416000 260x260 levels 9 NVN 0xd (24 so far, 5346 KB, 6 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: slow upload 31.7 ms: #14361 0x10f9e0b800 1024x1024x1 x1 levels 11 NVN 0x8d ASTC (BC3), 475 of 475 KB
[hos host] vk: slow submit: 55.8 ms: draw 50.9 descriptors 44.7 texture lookups 33.4 uploads 33.2 programs 2.5 commands 2.7
[hos host] vk: the driver's allowance is the count itself (8379 MB used of 8379): past it, taken as 7541 MB
[hos host] vk: (its written part: 2 of 2 KB readable, 320 bytes not zero; levels L0)
[hos host] vk: texture 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 2 KB of it
[hos host] vk: (its written part: 1 of 1 KB readable, 864 bytes not zero; levels L0)
[hos host] vk: texture 0x10d93e1600 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 1 KB of it
[hos host] vk: replay reads of guest memory as submitted: 2900000 (1863408 pages from the captures)
[hos host] vk: slow submit: 39.0 ms: draw 35.0 descriptors 28.6 texture lookups 15.8 uploads 15.6 programs 2.6 commands 2.8
[hos host] vk: texture 0x10001e00400 640x360x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 4330 MB)
[hos host] vk: the driver's allowance is the count itself (8424 MB used of 8424): past it, taken as 7581 MB
[hos host] vk: (its written part: 2 of 2 KB readable, 320 bytes not zero; levels L0)
[hos host] vk: texture 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 2 KB of it
[hos host] vk: (its written part: 1 of 1 KB readable, 864 bytes not zero; levels L0)
[hos host] vk: texture 0x10d93e1600 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 1 KB of it
[hos host] vk: ASTC texture #14785 0x10da19b000 1536x1536 x1 levels 11: the image of #14738 0x1132250e00 (the same blocks, finalized by the game) copied, not decoded again (2 so far, 23 MB)
[hos host] vk: texture #1212 0x10001fe0800 640x360 x1 NVN 0x14 the GPU drew, finalized by the game, sampled again at frame 5298 (drawn at 4966)
[hos host] vk: upload over a drawn image: #1212 0x10001fe0800 640x360x1 x1 NVN 0x14 at present 5297 (texture frame 5298), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: slow submit: 46.9 ms: draw 49.0 descriptors 17.2 texture lookups 17.6 uploads 11.9 programs 30.1 commands 1.0
[hos host] vk: 191 new ASTC textures left for a later frame (this frame's uploads took their budget; 10072 so far)
[hos host] vk: the driver's allowance is the count itself (8424 MB used of 8424): past it, taken as 7581 MB
[hos host] vk: the driver's allowance is the count itself (8424 MB used of 8424): past it, taken as 7581 MB
[hos host] vk: slow submit: 21.5 ms:
[hos host] vk: slow upload 71.8 ms: #14697 0x10f5406c00 768x768x1 x1 levels 10 NVN 0x7b ASTC, 528 of 528 KB
[hos host] vk: slow submit: 81.3 ms: draw 78.2 descriptors 75.5 texture lookups 73.5 uploads 73.3 programs 1.4
[hos host] vk: a frame's uploads took 73 ms: ASTC on the GPU 13 (5.3 MB, 73 ms); [768x768 x1 NVN 0x7b ASTC on the GPU 0.5 MB 72 ms]
[hos host] vk: budget: gave back 192 MB from staging at 111% of the allowance
[hos host] vk: budget: gave back 213 MB from target pool at 111% of the allowance
[hos host] vk: write-back of #550 0x1137207e00: 0 KB the CPU wrote since the GPU kept
[hos host] vk: write-back of #549 0x1137207d80: 0 KB the CPU wrote since the GPU kept
[hos host] vk: textures the GPU drew, written back into their memory before eviction: 1 so far
[hos host] vk: texture #14057 0x10004380c00 640x360x1 x1 NVN 0x14 the GPU drew goes with no copy of its texels (finalized, unused for 153 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 1017 textures retired (995 left, 2865 MB): budget 1013, drawn over 4, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: slow present: 15.2 ms:
[hos BaseProcExecutor3] map: HyruleField_Child (child)
[hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.02, writes collected 0.16
[hos host] vk: texture 0x1100866400 455x455x1 x1 levels 9 NVN 0x49 -> vk 139 (1000 textures, 2872 MB)
[hos host] vk: slow submit: 47.2 ms: draw 45.3 descriptors 37.5 texture lookups 34.5 uploads 20.4 programs 6.0 commands 1.3
[hos host] vk: pool 0x1010917600+0xf00000 imported in 3.0 ms as 0x1010800000..0x1011c00000 (197 imports, 627 ms in all)
[hos host] vk: pool 0x7f93912f00+0x138ea0 imported in 2.4 ms as 0x7f93800000..0x7f94000000 (201 imports, 633 ms in all)
[hos ModuleSystemWorker3] nvn: 700000 recordings made, 66490 live copies (21.5 MB), 5678 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: slow submit: 15.9 ms: draw 4.9 descriptors 3.8 texture lookups 3.7 uploads 3.3
[hos host] vk: pool 0x1167269000+0x225d58 imported in 2.1 ms as 0x1167000000..0x1167800000 (204 imports, 638 ms in all)
[hos host] vk: recorder: 1883653 packets run in 10.0 s (busy 0.81 s, spun idle 0.10 s, slept 9189 times, 3478 of them under 200 us), 0 calls made directly (no room); 10302 batches over 459 frames (22.4 a frame, 183 packets a batch), 4709 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 199 times, 26.98 ms (0.4 times, 0.06 ms a frame); every thread 711 times, 64.68 ms in all, the longest 14.19 ms; vkWaitForFences 403/28.29 ms (max 0.44), the queue's lock 4/14.27 ms (max 14.19), vkDestroyImageView 70/7.36 ms (max 0.40), vkQueueBindSparse 112/5.24 ms (max 0.49), vmaDestroyBuffer 48/3.92 ms (max 0.97), vkEndCommandBuffer 46/3.81 ms (max 0.88)
[hos host] vk: recorder: of its busy 0.81 s, 0.23 s in parts' ends and submits (the longest 15.81 ms), 0.03 s of that blocked on the queue's lock (55 times, the longest 1.22 ms), so 0.20 s in the ends, submits and holds; 0.57 s running the other packets; 4104 packets a frame
[hos host] vk: slow upload 26.1 ms: #15090 0x10d39ede00 512x512x1 x1 levels 10 NVN 0x7b ASTC, 236 of 236 KB
[hos host] vk: slow submit: 57.9 ms: draw 46.7 descriptors 38.9 texture lookups 36.2 uploads 34.6 programs 5.3 commands 1.6
[hos host] vk: upload over a drawn image: #1212 0x10001fe0800 640x360x1 x1 NVN 0x14 at present 5301 (texture frame 5302), finalized, 842 KB dirty in 1 ranges
[hos host] vk: slow submit: 16.8 ms: draw 14.3 descriptors 7.4 texture lookups 4.8 uploads 3.6 programs 5.0 commands 1.4 dispatch 1.5
[hos host] vk: slow submit: 25.8 ms: draw 26.6 descriptors 18.5 texture lookups 15.4 uploads 14.9 programs 5.9 commands 1.6
[hos host] vk: mirror of pool 0x1010800000 (8 MB) given back: retired, 99 submissions since its sync, GPU-written
[hos host] vk: write-back diff: VRAM 8493 of 7643 MB: the shadows given back (120 MB)
[hos host] vk: replay reads of guest memory as submitted: 2920000 (1877798 pages from the captures)
[hos host] vk: slow upload 17.0 ms: #14628 0x11053b6000 341x341x1 x1 levels 9 NVN 0x7d ASTC, 83 of 83 KB
[hos host] vk: (its written part: 0 of 0 KB readable, 192 bytes not zero; levels L0)
[hos host] vk: texture 0x10debdc600 1240x1240x1 x3 levels 1 NVN 0x4d (4563 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 0 KB of it
[hos sead::MainThread] vk: capture: a submit 0.36 ms (lock 0.03, scan 0.09 over 712 ranges, 4221 MB), 871 pages written, 587 copied (2.3 MB)
[hos host] vk: (its written part: 0 of 0 KB readable, 288 bytes not zero; levels L0)
[hos host] vk: texture 0x10dff20200 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 0 KB of it
[hos host] vk: 35 textures retired (1390 left, 3132 MB): budget 35, drawn over 0, dead targets 0
[hos host] vk: a capture taken (ms): all 0.043; GPU wait for late pages 0.009 (3 of 600 takes), pools lock 0.000, retires and stamps 0.025, bytes copied 0.000; 25 kept (254 MB), 7 spare buffers, 7528 texture pages copied (29 MB)
[hos host] vk: readback pages refreshed from guest memory: 2438387 pages so far (0 held: may be written ahead in their submission)
[hos OptimizeChangeBanc] nvn: 750000 recordings made, 70160 live copies (26.0 MB), 5697 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: slow upload 17.6 ms: #14556 0x1105b1b800 1024x1024x1 x1 levels 11 NVN 0x8b ASTC (BC3), 642 of 642 KB
[hos host] vk: sparse mirrors: 147, 1382 MB with memory of 1932 MB (549 MB spared); blocks given memory later 6023, given back 7934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 377 binds (95.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 66 taken from the spares, 159 made new; 78 kept as spares, 0 destroyed; 12 spares now, 632 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 501, 2281.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 10373 dirty pages from their submits' captures, 35718 from earlier ones', 71598 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 783 dispatches, 24732 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 2807 runs, 78.7 MB compared; 6682 pages read into the shadows; 0 MB of shadows (0 MB of pages known, in 0 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 4596 in captures, 388 only in the collection, 12 only by the replay itself
[hos host] vk: read live, by pool: 0x7fab000000 (68 MB, captured, state 2): 32999; 0x7fb5c00000 (44 MB, captured, state 2): 10652; 0x108c000000 (224 MB, captured, state 2): 10605; 0x7fb8800000 (52 MB, captured, state 2): 8899; 0x1178400000 (152 MB, captured, state 2): 1257; 0x104e000000 (24 MB, captured, state 2): 1110; 0x1042800000 (4 MB, captured, state 2): 1002; 0x1042c00000 (4 MB, captured, state 2): 986;
[hos host] vk: imports: 172 live (2232 MB), 13 retired still imported (368 MB), 20 retired given back (404 MB)
[hos BaseProcExecutor1] vcores: 99428 takes, 42600 waits (mean 95 us, longest 65.3 ms, 3 over HOS_VCORE_LOG_MS), 19595 kicks, 18989 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 2940000 (1888903 pages from the captures)
[hos Resource Control] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 91310689 lookups, 167903 missed (0.18 %)
[hos host] vk: pool 0x11637ca000+0xa2438 imported in 2.7 ms as 0x1163400000..0x1163c00000 (206 imports, 642 ms in all)
[hos host] vk: mirror of pool 0x1163400000 (4 MB) given back: retired, 16 submissions since its sync
[hos host] vk: mirror of pool 0x7f55400000 (4 MB) given back: idle, 6037 submissions since its sync
[hos host] vk: 80 mirrors given back (778 MB in all); 1384 MB of mirrors now, 144 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: 3 textures retired (1415 left, 3149 MB): budget 1, drawn over 0, dead targets 2
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 68.4 ms for a core: given by GeneralFormatAsync
(priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 3: it;
[hos host] vk: replay reads of guest memory as submitted: 2960000 (1900796 pages from the captures)
[hos host] vk: over the driver's video-memory allowance (8.2 of 7.4 GB)
[hos host] discord: the status: Hyrule Field - Child / ❤️ 0.5/3 · 🕷️ 0 · 💗 0
[hos host] vk: pool 0x7fd9800000..0x7fdc400000 (44 MB) trimmed to its 2 live part(s), 8 MB: the rest outside every game pool for 1856 submissions (1 trims, 36 MB in all)
[hos host] vk: pool 0x7fdac00000..0x7fdb000000: 3900 KB the CPU wrote since retired pool 0x7fd9800000's last sync kept from guest memory, not its mirror
[hos host] vk: pool 0x7fdc000000..0x7fdc400000: 3948 KB the CPU wrote since retired pool 0x7fd9800000's last sync kept from guest memory, not its mirror
[hos host] vk: 228 pools mirrored in VRAM (1385 MB)
[hos host] vk: write-backs left out (600 submits): 799, 3550.5 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3815 dirty pages from their submits' captures, 47889 from earlier ones', 403 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 782 dispatches, 36653 chunks; 0 syncs as copies (no table room)
[hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 400; 0x7f90000000 (4 MB, captured, state 2): 2; 0x7fd3c00000 (16 MB, captured, state 2): 1;
[hos host] vk: program picks: 3771454 kept (1132652 by the last picks' textures, a uniform buffer bound since), 25971 new, 1802574 with other textures (of 5600000)
[hos host] vk: replay reads of guest memory as submitted: 2980000 (1911627 pages from the captures)
[hos host] vk: mirror of pool 0x7fd9800000 (44 MB) given back: retired, 2139 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 3000000 (1922573 pages from the captures)
[hos host] vk: texture #14163 0x10005191c00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 63 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 85 textures retired (1342 left, 3028 MB): budget 85, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16527 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 1987 pages in 10 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: presenter thread: the copy's fence 2.62 ms, the present 0.72 ms a present
[hos host] vk: texture #14173 0x10004ce1800 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5363 (drawn at 5301)
[hos host] vk: texture #14172 0x10004bf1600 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 69 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14174 0x10004b01400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 69 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14171 0x10004741200 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 69 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 169 textures retired (1174 left, 2661 MB): budget 167, drawn over 2, dead targets 0; GPU-drawn with no copy kept: 0 live, 3 finalized, 5 MB
[hos host] vk: replay reads of guest memory as submitted: 3020000 (1933689 pages from the captures)
[hos host] vk: sparse mirrors: 147, 1385 MB with memory of 1896 MB (510 MB spared); blocks given memory later 6124, given back 7970; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 393 binds (96.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 69 taken from the spares, 159 made new; 81 kept as spares, 0 destroyed; 12 spares now, 668 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1581 dirty pages from their submits' captures, 44268 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 695 dispatches, 35219 chunks; 0 syncs as copies (no table room)
[hos host] vk: read live, by pool:
[hos host] vk: imports: 173 live (2200 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos host] vk: 33 textures retired (1141 left, 2626 MB): budget 33, drawn over 0, dead targets 0
[hos host] vk: ASTC texture #15398 0x1132c9c600 768x768 x1 levels 10: the image of #14678 0x11179b7e00 (the same blocks, finalized by the game) copied, not decoded again (3 so far, 24 MB)
[hos host] vk: ASTC texture #15399 0x1132391200 683x683 x1 levels 10: the image of #14919 0x11182db400 (the same blocks, finalized by the game) copied, not decoded again (4 so far, 25 MB)
[hos host] vk: ASTC texture #15400 0x11322ad000 1024x1024 x1 levels 11: the image of #14920 0x1118314400 (the same blocks, finalized by the game) copied, not decoded again (5 so far, 30 MB)
[hos host] vk: texture #14276 0x116c342000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 5384 (drawn at 5293)
[hos host] vk: ASTC texture #15413 0x113248fc00 512x512 x1 levels 10: the image of #14139 0x11181f6c00 (the same blocks, finalized by the game) copied, not decoded again (6 so far, 31 MB)
[hos host] vk: ASTC texture #15414 0x113240b800 768x768 x1 levels 10: the image of #14140 0x1118215a00 (the same blocks, finalized by the game) copied, not decoded again (7 so far, 34 MB)
[hos host] vk: replay reads of guest memory as submitted: 3040000 (1946610 pages from the captures)
[hos host] vk: ASTC texture #15416 0x10dd585200 768x768 x1 levels 10: the image of #15398 0x1132c9c600 (the same blocks, finalized by the game) copied, not decoded again (8 so far, 34 MB)
[hos host] vk: 75 textures retired (1079 left, 2533 MB): budget 73, drawn over 2, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.23 ms (lock 0.03, scan 0.06 over 779 ranges, 4298 MB), 308 pages written, 308 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.297; GPU wait for late pages 0.249 (65 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (80 MB), 12 spare buffers, 3429 texture pages copied (13 MB)
[hos host] vk: replay reads of guest memory as submitted: 3060000 (1958408 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.32, writes collected 0.12
[hos host] vk: memory: working set 15103 MB (peak 15142), private 22834 MB, staging 80 MB in 5 arenas, 37 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 752 MB, 37 allocations 710 MB; used 3388 of 15579 MB; heap 1 (device): 210 blocks 6628 MB, 23415 allocations 4288 MB; used 8493 of 8493 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 19 blocks 608 MB, 101 images 310 MB; 68 free ranges, the largest 29 MB
[hos host] vk: image blocks (textures, compacted): 99 blocks 3168 MB, 972 images 1290 MB; 469 free ranges, the largest 31 MB
[hos host] vk: render target images: 5657 reused from the pool, 3102 made new, 3000 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB), VRAM critical: given back
[hos host] vk: render targets not reused by size: 1017 with the pool empty (0 of them with targets still in the dead queue), 252 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 189 over its bound and 2811 idle
[hos host] vk: VRAM: 8493 MB used of 7643 MB allowed (2339 MB of it free in the blocks); textures 2533 MB (budget 1800 MB), 1079 textures
[hos host] vk: budget: VRAM 6628 of 7643 MB (8493 MB by the driver, 1864 MB unaccounted): textures 2533, blocks free 2339, mirrors 1385, other buffers 177, other images 111, staging 80; RAM 3355 MB (the process's private memory 22834 MB): imports 2200, capture spares 612, retired imports 416, captures 82, capture table 29, other buffers 16, others 0
[hos host] vk: textures: decoded ASTC 1301 MB, rendered 859 MB (80 drawn in one frame, kept: 46554 KB), other 330 MB; used in the last 8 presents 2091 MB
[hos host] vk: decoded ASTC used again: 12 textures (44 MB) copied from ones the game finalized, not decoded; 12 of those retired, 0 listed; 980 MB of blocks hashed in 160 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 253 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 253 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 20 (9459 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 262x104x1 NVN 0x1: 1 (26 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 656x113x1 NVN 0x1: 1 (72 KB); 1280x720x1 NVN 0x1: 1 (900 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 266 waited for it, 12.34 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 950.9 ms; VRAM 8.2 of 7.4 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5932 queued, 5932 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14130 ms and 314.7 MB read in all); 8028 programs (1613 of them compute variants without their code, 0 graphics programs no note names), 11132 graphics pipelines (1 not in a cache), 704 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1959 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8554 (5932 ready, 0 loading), 15572 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 12564 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 702 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: 8733 programs from the program cache
[hos host] vk: a frame: 255 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.1 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 35.47, present 0.37, draw 24.04, descriptors 15.78, texture lookups 11.95, uploads 8.48, pipelines 0.37, programs 2.48, commands 4.75, dispatch 1.53, memory sync 2.52, queue submit 0.01, slot wait 0.32
[hos host] vk: a frame: mirrors 228 (1385 MB): 15.04 MB synced in 1034.6 copies, 17.79 MB written back
[hos host] vk: 5400 presents, 24.2 fps; a frame: 2515 draws (0 skipped), 220 dispatches, 191 passes, 149 barriers, 6 submits, 8.9 uploads (5.7 MB), 84 ub updates (59 overlays), GPU-bound wait 0.1 ms (recording slots only); 12966 pipelines (3200 ms), 0 programs (10 failed), 1084 textures (2536 MB)
[hos host] vk: host time in GPU resource waits: 36.44 ms a frame (fences 9.06, timelines 27.38, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 863.6 full barriers recorded, 324.7 left out (nothing recorded since the one before), 15.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 2.7; 0x9: 0.9; 0x10: 3.2; 0x11: 44.1; 0x12: 69.6; 0x13: 1.8; 0x21: 3.6; 0x28: 14.7; 0x92: 2.8; 0x97: 0.9; 0x9b: 0.2;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 42.6 from 27.5 batches at the game's barriers (44.9 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 141.0, an op 1818.6, the end 6.5, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.8; 5: 23.8; 48: 7.3;
[hos host] vk: GPU time 14.2 ms a frame (the longest submission 8.4 ms); 11.4 textures made a frame: colour targets 2.1 (4.5 MB), depth 0.5 (2.3 MB), ASTC or BC 8.7 (16.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 2.40 MB unswizzled on the CPU in 0.32 ms
[hos host] vk: the render thread (600 presents, 21.82 ms a frame): its jobs 16.78 ms a frame (submits 5.61, presents 5.21, others 5.95), idle 5.04; the game waited for room at 44 presents, 0.74 ms each (0.05 a frame; 16.73 jobs queued at a present); 7 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 12.55 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 289 captures at a CPU signal; the submits' waits read in 0.027 ms a frame; readbacks: 1705 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.59 ms after its hand-over (953.07 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 7.56 + 0.43 (the window's copy is in the idle), after it 1.75 + 1.21; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 31.62 (0.17 ms)
[hos host] vk: frame ahead (600 presents): 6.10 parts a frame recorded in the signal's job, 65.7% of them took nothing that came with the signal; of the others, 2.06 a frame took pages of the signal's capture into a mirror (66.8 pages a frame) and 0.06 pages read as guest memory had them (132.5 pages a frame); a frame's first part took nothing in 115 of 289 frames, its first three in 115
[hos host] vk: frame ahead, what came with the signal: 290 signals with a capture, each of 561.2 pages in 14.3 pools; between a frame's last submit and that signal 0.34 texture and sampler registrations, 2.56 releases and 0.00 other CPU signals a signal (15 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.54+0.16, 0.47+0.12, 0.17+0.06, 0.17+0.06, 0.19+0.27, 0.07+0.19, 0.02+0.01, 0.06+0.13, 0.03+0.12, 0.02+0.07, 0.00+0.01, 0.01+0.00
[hos host] vk: pipeline cache: play file 8 saved (189 KB): 5 pipelines, 5 from the driver's cache, 0 compiled; 1786 more from the start's cache
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1603 dirty pages from their submits' captures, 44917 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 719 dispatches, 35564 chunks; 0 syncs as copies (no table room)
[hos host] vk: read live, by pool:
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 120000 so far (x1 0x1178b9a000 x2 0x2 x3 0x1178b9a000 x4 0x654040)
[hos host] vk: render targets drawn over from another start, let go: 9527 so far (45576 MB), 8517 of them (40855 MB) between the passes
[hos host] vk: replay reads of guest memory as submitted: 3080000 (1970731 pages from the captures)
[hos host] vk: program picks: 3925414 kept (1200685 by the last picks' textures, a uniform buffer bound since), 25976 new, 1848609 with other textures (of 5800000)
[hos host] vk: 6 textures retired (1082 left, 2534 MB): budget 5, drawn over 1, dead targets 0
[hos host] vk: VRAM compaction 3 begins: the textures' blocks hold 3168 MB, 1874 MB of it free space
[hos host] vk: replay reads of guest memory as submitted: 3100000 (1981642 pages from the captures)
[hos host] vk: late pages' fix: stamps for 0x100d7f0000 (pool 0x100d400000; 64 blocks, 4096 KB)
[hos host] vk: late pages' fix: stamps for 0x10387f0000 (pool 0x1038400000; 128 blocks, 8192 KB)
[hos host] vk: late pages' fix: stamps for 0x103d3f0000 (pool 0x103d000000; 192 blocks, 12288 KB)
[hos host] vk: late pages' fix: stamps for 0x103d7f0000 (pool 0x103d000000; 256 blocks, 16384 KB)
[hos host] vk: mirror of pool 0x7fdfc00000 (12 MB) given back: idle, 6008 submissions since its sync
[hos host] vk: late pages' fix: stamps for 0x117cff0000 (pool 0x1178400000; 320 blocks, 20480 KB)
[hos host] vk: late pages' fix: stamps for 0x117d9f0000 (pool 0x1178400000; 384 blocks, 24576 KB)
[hos host] vk: late pages' fix: stamps for 0x117dff0000 (pool 0x1178400000; 448 blocks, 28672 KB)
[hos host] vk: write-back diff: 5 shadows made again (76 MB)
[hos host] vk: texture #15466 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (4 so far)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 146, 1382 MB with memory of 1884 MB (501 MB spared); blocks given memory later 6149, given back 8008; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 401 binds (97.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 69 taken from the spares, 159 made new; 82 kept as spares, 0 destroyed; 13 spares now, 680 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 1856 dirty pages from their submits' captures, 44232 from earlier ones', 12349 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 709 dispatches, 35345 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 3392 runs, 191.2 MB compared; 8489 pages read into the shadows; 280 MB of shadows (34 MB of pages known, in 71 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 4674 in captures, 0 only in the collection, 12 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd6400000 (48 MB, captured, state 2): 12008; 0x7fa2000000 (8 MB, captured, state 2): 341;
[hos host] vk: imports: 173 live (2200 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos host] vk: replay reads of guest memory as submitted: 3120000 (1992862 pages from the captures)
[hos host] vk: mirror of pool 0x105d400000 (4 MB) given back: idle, 6003 submissions since its sync
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 908 of 3000 waited for their write-back, 0.11 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1723 kept their virtual core (nothing there to wait for), 2629 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos sead::MainThread] nvn: GPU waits in 5 s: 152 sync waits (2612 ms), 0 polls timed out (0 ms), 6 queue finishes (89 ms)
[hos Presentation Thread] vi: 5445 frames presented (30.5 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 0, 17.8-22 105, 22-30 23, 30-40 4, over 40 21
[hos host] vk: replay reads of guest memory as submitted: 3140000 (2006290 pages from the captures)
[hos host] vk: mirror of pool 0x7fd4c00000 (24 MB) given back: idle, 6042 submissions since its sync
[hos rmbl::VibrationThread] hid: 16384 vibration values, strongest 0.00 since the last count
[hos host] vk: replay reads of guest memory as submitted: 3160000 (2019122 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 2504931 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 6509 dirty pages from their submits' captures, 52867 from earlier ones', 157 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 770 dispatches, 36328 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7703 runs, 591.3 MB compared; 58 pages read into the shadows; 280 MB of shadows (34 MB of pages known, in 71 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14953 in captures, 0 only in the collection, 24 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa1800000 (4 MB, captured, state 2): 157;
[hos host] vk: VRAM compaction 3 over: 332 textures moved (202 MB) in 29 passes, 0 left where they were; 53 blocks (1696 MB) given back to the driver, 2208 MB in all so far
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #85000: event 0x100d6c800c mode 0 value 59554 (now 59535) (lr 0x8066179c)
[hos host] vk: replay reads of guest memory as submitted: 3180000 (2031996 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.01, scan 0.04 over 787 ranges, 4279 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.144; GPU wait for late pages 0.105 (32 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 24 kept (94 MB), 12 spare buffers, 5063 texture pages copied (19 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11881 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 4078197 kept (1270980 by the last picks' textures, a uniform buffer bound since), 25984 new, 1895818 with other textures (of 6000000)
[hos host] vk: replay reads of guest memory as submitted: 3200000 (2045465 pages from the captures)
[hos host] vk: texture #15522 0x10002199c00 120x67 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (5 so far)
[hos host] vk: texture #15523 0x100021a9e00 60x33 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (6 so far)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.14, writes collected 0.05
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #60000: event 0x100d6c802c mode 0 value 59999 (lr 0x80667688)
[hos host] vk: sparse mirrors: 145, 1370 MB with memory of 1860 MB (489 MB spared); blocks given memory later 6180, given back 8008; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 407 binds (97.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 159 made new; 84 kept as spares, 0 destroyed; 14 spares now, 704 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 4487 dirty pages from their submits' captures, 45248 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 765 dispatches, 35626 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7357 runs, 586.5 MB compared; 47 pages read into the shadows; 280 MB of shadows (34 MB of pages known, in 72 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15085 in captures, 0 only in the collection, 26 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 174 live (2204 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos host] vk: replay reads of guest memory as submitted: 3220000 (2057354 pages from the captures)
[hos host] vk: a frame: 261 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 14.62, present 0.67, draw 8.45, descriptors 3.75, texture lookups 0.29, uploads 0.05, pipelines 0.33, programs 1.49, commands 2.24, dispatch 1.02, memory sync 0.98, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 229 (1370 MB): 6.60 MB synced in 1104.2 copies, 18.60 MB written back
[hos host] vk: 5520 presents, 47.1 fps; a frame: 2540 draws (0 skipped), 214 dispatches, 207 passes, 157 barriers, 6 submits, 0.7 uploads (0.2 MB), 84 ub updates (58 overlays), GPU-bound wait 0.0 ms (recording slots only); 12966 pipelines (3200 ms), 0 programs (10 failed), 1156 textures (2606 MB)
[hos host] vk: host time in GPU resource waits: 11.45 ms a frame (fences 4.06, timelines 7.39, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 850.3 full barriers recorded, 349.0 left out (nothing recorded since the one before), 17.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.8; 0x8: 3.0; 0x9: 1.0; 0x10: 4.8; 0x11: 43.4; 0x12: 72.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 148.4, an op 1920.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 15.7 ms a frame (the longest submission 10.5 ms); 0.8 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.1 (0.0 MB), ASTC or BC 0.7 (0.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.07 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: mirror of pool 0x7f54000000 (4 MB) given back: idle, 6005 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 3240000 (2069045 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 3146 dirty pages from their submits' captures, 44385 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 731 dispatches, 35325 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6928 runs, 578.1 MB compared; 42 pages read into the shadows; 280 MB of shadows (34 MB of pages known, in 71 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15008 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 2;
[hos host] vk: replay reads of guest memory as submitted: 3260000 (2082490 pages from the captures)
[hos nn.npln.Callback] vcores: 130930 takes, 27364 waits (mean 102 us, longest 68.4 ms, 1 over HOS_VCORE_LOG_MS), 12625 kicks, 12716 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: mirror of pool 0x7fb0000000 (8 MB) given back: idle, 6048 submissions since its sync
[hos host] vk: mirror of pool 0x7fbc000000 (12 MB) given back: idle, 6048 submissions since its sync
[hos host] vk: mirror of pool 0x7fdf000000 (12 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: mirror of pool 0x7fb2000000 (36 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: mirror of pool 0x7fde800000 (8 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: mirror of pool 0x7fd2400000 (4 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: mirror of pool 0x7fd1800000 (12 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: mirror of pool 0x7fbcc00000 (12 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: mirror of pool 0x7fb4400000 (24 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: mirror of pool 0x7fb0800000 (24 MB) given back: idle, 6046 submissions since its sync
[hos host] vk: mirror of pool 0x7fd3400000 (4 MB) given back: idle, 6044 submissions since its sync
[hos host] vk: mirror of pool 0x7fc1400000 (28 MB) given back: idle, 6044 submissions since its sync
[hos host] vk: mirror of pool 0x7fbf800000 (28 MB) given back: idle, 6044 submissions since its sync
[hos host] vk: mirror of pool 0x7fbd800000 (32 MB) given back: idle, 6044 submissions since its sync
[hos host] vk: 99 mirrors given back (998 MB in all); 1173 MB of mirrors now, 152 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: replay reads of guest memory as submitted: 3280000 (2095551 pages from the captures)
[hos ModuleSystemWorker4] longjmp to 0x81681690 (value 1, 0 slots above dropped)
[hos host] vk: reaper: 1904 images (3681 MB) destroyed in 10 s, 55 of them where they were handed over (its cap or VRAM pressure); the queue at most 792 (1023 MB)
[hos host] vk: pool 0x115c37d000+0x700000 imported in 3.4 ms as 0x115c000000..0x115d000000 (209 imports, 647 ms in all)
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.04 over 694 ranges, 4311 MB), 250 pages written, 250 copied (1.0 MB)
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 5102 texture pages copied (19 MB)
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: sparse mirrors: 132, 1187 MB with memory of 1636 MB (448 MB spared); blocks given memory later 6210, given back 8008; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 413 binds (100.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 99 kept as spares, 0 destroyed; 29 spares now, 952 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 727, 3233.2 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 2830 dirty pages from their submits' captures, 45989 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 698 dispatches, 37322 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5686 runs, 570.7 MB compared; 2065 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 88 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14571 in captures, 115 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10751 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 152 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 1 textures retired (1300 left, 2771 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: ui: the letters of zh-Hant: DFPT_C8, C:\WINDOWS\Fonts\msjh.ttc
[hos host] vk: a frame: 100 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 6.56, present 0.13, draw 3.68, descriptors 1.72, texture lookups 0.33, uploads 0.11, pipelines 0.14, programs 0.55, commands 0.97, dispatch 0.47, memory sync 0.53, queue submit 0.00, slot wait 0.21
[hos host] vk: a frame: mirrors 231 (1187 MB): 3.93 MB synced in 785.3 copies, 7.06 MB written back
[hos host] vk: 5640 presents, 54.5 fps; a frame: 976 draws (0 skipped), 62 dispatches, 159 passes, 107 barriers, 6 submits, 1.3 uploads (0.4 MB), 324 ub updates (147 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1300 textures (2771 MB)
[hos host] vk: host time in GPU resource waits: 7.94 ms a frame (fences 5.13, timelines 2.81, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 425.2 full barriers recorded, 188.4 left out (nothing recorded since the one before), 51.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 3.2; 0x8: 1.1; 0x9: 0.4; 0x10: 2.5; 0x11: 15.8; 0x12: 70.2; 0x13: 1.4; 0x21: 2.8; 0x28: 5.9; 0x92: 3.7; 0x97: 0.4;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 16.9 from 11.0 batches at the game's barriers (18.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 102.1, an op 649.3, the end 3.8, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 0.7; 5: 9.5; 48: 2.9;
[hos host] vk: GPU time 9.0 ms a frame (the longest submission 4.8 ms); 1.7 textures made a frame: colour targets 0.8 (0.7 MB), depth 0.1 (0.4 MB), ASTC or BC 0.8 (1.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.16 MB unswizzled on the CPU in 0.03 ms
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 135 dirty pages from their submits' captures, 47256 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42999 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 495 runs, 7.0 MB compared; 141 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7948 in captures, 31 only in the collection, 70 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.05 over 691 ranges, 4189 MB), 189 pages written, 189 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.015; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (88 MB), 12 spare buffers, 5614 texture pages copied (21 MB)
[hos host] vk: program picks: 4224141 kept (1335362 by the last picks' textures, a uniform buffer bound since), 26010 new, 1949848 with other textures (of 6200000)
[hos host] vk: 19 textures retired (1281 left, 2769 MB): budget 0, drawn over 19, dead targets 0
[hos host] vk: mirror of pool 0x7f8a800000 (4 MB) given back: idle, 6041 submissions since its sync
[hos host] vk: mirror of pool 0x7f90c00000 (4 MB) given back: idle, 6041 submissions since its sync
[hos host] vk: mirror of pool 0x7f9c000000 (4 MB) given back: idle, 6041 submissions since its sync
[hos host] vk: mirror of pool 0x7f93800000 (8 MB) given back: idle, 6041 submissions since its sync
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: 1 textures retired (1280 left, 2769 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 2559140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 128, 1175 MB with memory of 1616 MB (440 MB spared); blocks given memory later 6210, given back 8024; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 415 binds (101.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 103 kept as spares, 0 destroyed; 33 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 174 dirty pages from their submits' captures, 47068 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42979 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 501 runs, 5.9 MB compared; 143 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7792 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos Presentation Thread] vi: 5719 frames presented (54.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 16, 15.5-17.8 134, 17.8-22 100, 22-30 20, 30-40 2, over 40 2
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11716 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: mirror of pool 0x7fa3400000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f90000000 (4 MB) given back: idle, 6002 submissions since its sync
[hos host] vk: mirror of pool 0x7f8ec00000 (4 MB) given back: idle, 6064 submissions since its sync
[hos host] vk: mirror of pool 0x7f93000000 (8 MB) given back: idle, 6063 submissions since its sync
[hos host] vk: mirror of pool 0x7f8e000000 (4 MB) given back: idle, 6063 submissions since its sync
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 692 ranges, 4179 MB), 178 pages written, 178 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (90 MB), 10 spare buffers, 5680 texture pages copied (22 MB)
[hos ModuleSystemWorker3] barrier: 80369 looks: 31217 found the masks moved already, 47019 waited until they moved, 2133 for the whole 1 ms (a wait 396 us on average); 64261 wakes
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.02, present 0.10, draw 0.74, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.10, memory sync 0.40, queue submit 0.00, slot wait 0.15
[hos host] vk: a frame: mirrors 231 (1164 MB): 2.60 MB synced in 603.3 copies, 0.08 MB written back
[hos host] vk: 5760 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.67 ms a frame (fences 5.36, timelines 0.31, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.4 full barriers recorded, 95.5 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.8 ms a frame (the longest submission 21.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos sead::MainThread] nvn: 800000 recordings made, 71671 live copies (18.3 MB), 5794 texture copies; 2 command buffers' segments (4.3 MB)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 78 dirty pages from their submits' captures, 47163 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42982 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7790 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: mirror of pool 0x7fdc000000 (4 MB) given back: idle, 6016 submissions since its sync
[hos host] vk: mirror of pool 0x7fdac00000 (4 MB) given back: idle, 6016 submissions since its sync
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: recorder: 4153686 packets run in 10.0 s (busy 1.21 s, spun idle 0.11 s, slept 10004 times, 3274 of them under 200 us), 0 calls made directly (no room); 13778 batches over 511 frames (27.0 a frame, 301 packets a batch), 7006 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 120 times, 9.00 ms (0.2 times, 0.02 ms a frame); every thread 821 times, 95.16 ms in all, the longest 1.29 ms; vmaDestroyBuffer 150/39.34 ms (max 1.29), vkWaitForFences 442/36.52 ms (max 0.48), the queue's lock 60/7.00 ms (max 0.40), vkDestroyImageView 69/4.45 ms (max 0.30), vmaFreeMemory 18/3.74 ms (max 0.89), vkQueueBindSparse 59/1.56 ms (max 0.35)
[hos host] vk: recorder: of its busy 1.21 s, 0.34 s in parts' ends and submits (the longest 1.33 ms), 0.09 s of that blocked on the queue's lock (137 times, the longest 1.29 ms), so 0.25 s in the ends, submits and holds; 0.88 s running the other packets; 8129 packets a frame
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 7920 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos BaseProcExecutor1] vcores: 96582 takes, 17647 waits (mean 31 us, longest 11.4 ms, 0 over HOS_VCORE_LOG_MS), 8974 kicks, 8810 yields, 0 starved grants, 0 escapes in 5 s
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 693 ranges, 4161 MB), 191 pages written, 191 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (98 MB), 11 spare buffers, 5680 texture pages copied (22 MB)
[hos host] vk: sparse mirrors: 121, 1163 MB with memory of 1584 MB (420 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 419 binds (101.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 110 kept as spares, 0 destroyed; 40 spares now, 1004 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 108 dirty pages from their submits' captures, 47800 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43590 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7906 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.08, present 0.10, draw 0.78, descriptors 0.40, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.17
[hos host] vk: a frame: mirrors 231 (1163 MB): 2.60 MB synced in 603.4 copies, 0.08 MB written back
[hos host] vk: 5880 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.15 ms a frame (fences 5.15, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 8394062 lookups, 74097 missed (0.88 %)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 835 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1217 kept their virtual core (nothing there to wait for), 3135 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 7920 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 70, 490.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 62 dirty pages from their submits' captures, 46515 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 630 dispatches, 42358 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 486 runs, 5.7 MB compared; 139 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7679 in captures, 0 only in the collection, 69 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 696 ranges, 4127 MB), 182 pages written, 182 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 5680 texture pages copied (22 MB)
[hos host] vk: presenter thread: the copy's fence 4.71 ms, the present 0.63 ms a present
[hos host] vk: mirror of pool 0x7f90400000 (4 MB) given back: idle, 6011 submissions since its sync
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: memory: working set 15229 MB (peak 15236), private 21161 MB, staging 80 MB in 5 arenas, 37 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 37 allocations 672 MB; used 3416 of 15579 MB; heap 1 (device): 160 blocks 5155 MB, 20538 allocations 4626 MB; used 7149 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 93 images 283 MB; 53 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 47 blocks 1504 MB, 1178 images 1485 MB; 221 free ranges, the largest 13 MB
[hos host] vk: render target images: 5676 reused from the pool, 3142 made new, 3046 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1025 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2854 idle
[hos host] vk: VRAM: 7149 MB used of 11474 MB allowed (528 MB of it free in the blocks); textures 2769 MB (budget 7158 MB), 1280 textures
[hos host] vk: budget: VRAM 5155 of 11474 MB (7149 MB by the driver, 1993 MB unaccounted): textures 2769, mirrors 1162, blocks free 528, shadows 296, other buffers 207, other images 111, staging 80; RAM 3346 MB (the process's private memory 21161 MB): imports 2228, capture spares 538, retired imports 416, captures 90, other buffers 44, capture table 30, others 0
[hos host] vk: textures: decoded ASTC 1503 MB, rendered 878 MB (90 drawn in one frame, kept: 22286 KB), other 345 MB; used in the last 8 presents 257 MB
[hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 0 listed; 1028 MB of blocks hashed in 170 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 853 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 436 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 436 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 436 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 436 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 436 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 1 presents ago
[hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 14 (10200 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 5 (8074 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 287 waited for it, 4.02 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 120, 1162 MB with memory of 1580 MB (417 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 420 binds (101.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 111 kept as spares, 0 destroyed; 41 spares now, 1008 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 84 dirty pages from their submits' captures, 47156 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42982 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 501 runs, 5.9 MB compared; 143 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7798 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos host] vk: status: 600 frames, the slowest 115.8 ms; VRAM 7.0 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5944 queued, 5944 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14147 ms and 315.2 MB read in all); 8040 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11149 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1994 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8566 (5944 ready, 0 loading), 15674 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 12586 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.01, present 0.10, draw 0.72, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.19, dispatch 0.11, memory sync 0.39, queue submit 0.00, slot wait 0.17
[hos host] vk: a frame: mirrors 231 (1162 MB): 2.60 MB synced in 603.3 copies, 0.08 MB written back
[hos host] vk: 6000 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 4.42 ms a frame (fences 4.42, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.3 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 6.3 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 17.92 ms a frame): its jobs 11.39 ms a frame (submits 1.55, presents 5.73, others 4.11), idle 6.53; the game waited for room at 14 presents, 4.26 ms each (0.10 a frame; 6.47 jobs queued at a present); 6 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 3.08 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 362 captures at a CPU signal; the submits' waits read in 0.019 ms a frame; readbacks: 820 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 10.71 ms after its hand-over (128.64 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 7.82 + 0.18 (the window's copy is in the idle), after it 0.79 + 0.40; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 6.32 parts a frame recorded in the signal's job, 75.2% of them took nothing that came with the signal; of the others, 1.56 a frame took pages of the signal's capture into a mirror (53.3 pages a frame) and 0.01 pages read as guest memory had them (34.6 pages a frame); a frame's first part took nothing in 180 of 362 frames, its first three in 180
[hos host] vk: frame ahead, what came with the signal: 362 signals with a capture, each of 65.2 pages in 6.7 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.19 releases and 0.00 other CPU signals a signal (1 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.32+0.18, 0.09+0.03, 0.04+0.03, 0.16+0.03, 0.01+0.00, 0.01+0.00, 0.04+0.02, 0.07+0.05, 0.02+0.04, 0.02+0.02, 0.00+0.01, 0.01+0.00
[hos Presentation Thread] vi: real-time step on: ratio 0.980, game's frame rate 60
[hos host] vk: mirror of pool 0x7fa1800000 (4 MB) given back: idle, 6049 submissions since its sync
[hos host] vk: 112 mirrors given back (1024 MB in all); 1159 MB of mirrors now, 640 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14656 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: render targets drawn over from another start, let go: 9600 so far (45692 MB), 8590 of them (40971 MB) between the passes
[hos Presentation Thread] vi: 6019 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 6, 15.5-17.8 288, 17.8-22 6, 22-30 0, 30-40 0, over 40 0
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.05 over 690 ranges, 4065 MB), 192 pages written, 192 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (92 MB), 11 spare buffers, 7728 texture pages copied (30 MB)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #90000: event 0x100d6c8008 mode 0 value 62610 (now 62607) (lr 0x806661f8)
[hos host] vk: readback pages refreshed from guest memory: 2571140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 88 dirty pages from their submits' captures, 47205 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 646 dispatches, 43003 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7795 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] nvn: GPU waits in 5 s: 131 sync waits (791 ms), 0 polls timed out (0 ms), 6 queue finishes (21 ms)
[hos host] vk: mirror of pool 0x7f90800000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 8492 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.11, present 0.11, draw 0.75, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.46, queue submit 0.00, slot wait 0.19
[hos host] vk: a frame: mirrors 231 (1157 MB): 2.60 MB synced in 603.4 copies, 0.08 MB written back
[hos host] vk: 6120 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 4.91 ms a frame (fences 4.90, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 6.1 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 689 ranges, 4051 MB), 196 pages written, 196 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 4246 texture pages copied (16 MB)
[hos host] vk: queues ordered by their events: 36000 submissions stopped at a wait, 35992 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 90568, not met 0; captures at a CPU event's wait 4646, 0 of them at a wait met on arrival
[hos BaseProcExecutor1] vcores: 94374 takes, 16547 waits (mean 29 us, longest 5.6 ms, 0 over HOS_VCORE_LOG_MS), 8502 kicks, 8302 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: sparse mirrors: 118, 1157 MB with memory of 1572 MB (415 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 422 binds (101.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 113 kept as spares, 0 destroyed; 43 spares now, 1016 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 97 dirty pages from their submits' captures, 47759 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 641 dispatches, 43569 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7907 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: mirror of pool 0x7f8f000000 (8 MB) given back: idle, 6035 submissions since its sync
[hos host] vk: mirror of pool 0x7f8e800000 (4 MB) given back: idle, 6035 submissions since its sync
[hos host] vk: mirror of pool 0x1065800000 (8 MB) given back: idle, 6034 submissions since its sync
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12240 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: mirror of pool 0x7fa6800000 (4 MB) given back: idle, 6012 submissions since its sync
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 98 dirty pages from their submits' captures, 47860 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 655 dispatches, 43616 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7917 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 689 ranges, 4051 MB), 186 pages written, 186 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (106 MB), 10 spare buffers, 5086 texture pages copied (19 MB)
[hos host] vk: mirror of pool 0x7fa2000000 (8 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.05, present 0.10, draw 0.74, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 231 (1142 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back
[hos host] vk: 6240 presents, 59.9 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 4.85 ms a frame (fences 4.85, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 202.9 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 6.0 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: mirror of pool 0x7fb8800000 (52 MB) given back: idle, 6005 submissions since its sync
[hos host] vk: mirror of pool 0x7f81400000 (4 MB) given back: idle, 6064 submissions since its sync
[hos host] vk: mirror of pool 0x7fd3c00000 (16 MB) given back: idle, 6064 submissions since its sync
[hos host] vk: mirror of pool 0x7f8a000000 (4 MB) given back: idle, 6064 submissions since its sync
[hos host] vk: mirror of pool 0x7f9ac00000 (20 MB) given back: idle, 6061 submissions since its sync
[hos host] vk: mirror of pool 0x7f99000000 (16 MB) given back: idle, 6061 submissions since its sync
[hos host] vk: mirror of pool 0x7f91000000 (32 MB) given back: idle, 6061 submissions since its sync
[hos host] vk: mirror of pool 0x1142400000 (4 MB) given back: idle, 6061 submissions since its sync
[hos host] vk: mirror of pool 0x7f86000000 (4 MB) given back: idle, 6061 submissions since its sync
[hos host] vk: mirror of pool 0x7f84400000 (4 MB) given back: idle, 6061 submissions since its sync
[hos host] vk: mirror of pool 0x7f97400000 (4 MB) given back: idle, 6060 submissions since its sync
[hos host] vk: mirror of pool 0x7f61000000 (4 MB) given back: idle, 6060 submissions since its sync
[hos host] vk: mirror of pool 0x7fff800000 (8 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7fa9000000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f83c00000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f84000000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7fd6400000 (48 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f84800000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f9a400000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f85800000 (8 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f9ec00000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f98800000 (8 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f86c00000 (20 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f88000000 (12 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f88c00000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f9f000000 (8 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f9fc00000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f8ac00000 (12 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f8b800000 (24 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f8f800000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7fa1400000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f94000000 (16 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f95000000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f9c400000 (32 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: mirror of pool 0x7f95400000 (4 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7f96800000 (4 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7f96c00000 (8 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7fab000000 (68 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7faf400000 (4 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7f89000000 (16 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7f86400000 (8 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7fb5c00000 (44 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7f85400000 (4 MB) given back: idle, 6058 submissions since its sync
[hos host] vk: mirror of pool 0x7f62800000 (8 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x1042800000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x1042c00000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x7f98400000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x7f9f800000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x7f97800000 (12 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x7fa0400000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x7fa0800000 (8 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x104cc00000 (16 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x7fa1c00000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x7f95800000 (16 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x104e000000 (24 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x1066000000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x1165c00000 (8 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x1042400000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x1167000000 (8 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: mirror of pool 0x1042000000 (4 MB) given back: idle, 6057 submissions since its sync
[hos host] vk: 178 mirrors given back (1525 MB in all); 658 MB of mirrors now, 640 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 128 dirty pages from their submits' captures, 47728 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 641 dispatches, 43574 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7911 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 25649 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vi: 6311 frames presented (58.3 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 12, 15.5-17.8 271, 17.8-22 6, 22-30 2, 30-40 0, over 40 1
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.05 over 686 ranges, 4007 MB), 215 pages written, 215 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.032; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.029, bytes copied 0.000; 25 kept (116 MB), 11 spare buffers, 2640 texture pages copied (10 MB)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 161 dirty pages from their submits' captures, 47081 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42962 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7698 in captures, 64 only in the collection, 70 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.32, present 0.10, draw 0.77, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.11, memory sync 0.51, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.2 copies, 0.08 MB written back
[hos host] vk: 6360 presents, 56.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 7.18 ms a frame (fences 5.29, timelines 1.89, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.4 full barriers recorded, 95.5 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.5 ms a frame (the longest submission 5.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5280 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: recorder: 936684 packets run in 10.0 s (busy 0.47 s, spun idle 0.09 s, slept 8532 times, 3209 of them under 200 us), 0 calls made directly (no room); 9958 batches over 591 frames (16.8 a frame, 94 packets a batch), 4969 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 16 times, 0.50 ms (0.0 times, 0.00 ms a frame); every thread 578 times, 61.55 ms in all, the longest 0.44 ms; vkWaitForFences 553/60.85 ms (max 0.44), vmaFreeMemory 8/0.39 ms (max 0.17), vmaDestroyBuffer 9/0.20 ms (max 0.04), vkQueueBindSparse 8/0.11 ms (max 0.02)
[hos host] vk: recorder: of its busy 0.47 s, 0.19 s in parts' ends and submits (the longest 0.85 ms), 0.00 s of that blocked on the queue's lock (5 times, the longest 0.81 ms), so 0.19 s in the ends, submits and holds; 0.28 s running the other packets; 1585 packets a frame
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 685 ranges, 3983 MB), 180 pages written, 180 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (118 MB), 9 spare buffers, 2640 texture pages copied (10 MB)
[hos host] vk: readback pages refreshed from guest memory: 2583140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 105 dirty pages from their submits' captures, 47801 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43591 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 500 runs, 5.8 MB compared; 143 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7911 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos BaseProcExecutor2] vcores: 91943 takes, 16685 waits (mean 25 us, longest 4.6 ms, 0 over HOS_VCORE_LOG_MS), 8629 kicks, 8444 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.05, present 0.10, draw 0.74, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back
[hos host] vk: 6480 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.12 ms a frame (fences 5.11, timelines 0.01, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.1 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.8 ms a frame (the longest submission 12.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos ModuleSystemWorker4] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 2844646 lookups, 6547 missed (0.23 %)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10560 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 683 ranges, 3969 MB), 190 pages written, 190 copied (0.7 MB)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 88 dirty pages from their submits' captures, 47155 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42987 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 501 runs, 5.9 MB compared; 143 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7806 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (110 MB), 10 spare buffers, 2640 texture pages copied (10 MB)
[hos host] vk: presenter thread: the copy's fence 4.97 ms, the present 0.56 ms a present
[hos host] vk: program picks: 4317354 kept (1344410 by the last picks' textures, a uniform buffer bound since), 26010 new, 2056635 with other textures (of 6400000)
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 73, 511.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 48471 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 657 dispatches, 44205 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 511 runs, 6.0 MB compared; 146 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8025 in captures, 0 only in the collection, 73 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 682 ranges, 3967 MB), 192 pages written, 192 copied (0.8 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9742 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (132 MB), 10 spare buffers, 3500 texture pages copied (13 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: memory: working set 15225 MB (peak 15259), private 21030 MB, staging 80 MB in 5 arenas, 37 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 36 allocations 686 MB; used 3416 of 15579 MB; heap 1 (device): 158 blocks 5027 MB, 12476 allocations 4123 MB; used 7022 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 93 images 283 MB; 53 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 47 blocks 1504 MB, 1178 images 1485 MB; 221 free ranges, the largest 13 MB
[hos host] vk: render target images: 5676 reused from the pool, 3142 made new, 3046 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1025 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2854 idle
[hos host] vk: VRAM: 7022 MB used of 11474 MB allowed (904 MB of it free in the blocks); textures 2769 MB (budget 7158 MB), 1280 textures
[hos host] vk: budget: VRAM 5027 of 11474 MB (7022 MB by the driver, 1994 MB unaccounted): textures 2769, blocks free 904, mirrors 658, shadows 296, other buffers 207, other images 111, staging 80; RAM 3381 MB (the process's private memory 21030 MB): imports 2228, capture spares 574, retired imports 416, captures 132, capture table 31, others 0
[hos host] vk: textures: decoded ASTC 1503 MB, rendered 878 MB (90 drawn in one frame, kept: 22286 KB), other 345 MB; used in the last 8 presents 257 MB
[hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 2 listed; 1028 MB of blocks hashed in 170 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1453 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1036 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1036 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1036 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1036 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1036 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 1 presents ago
[hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 14 (10200 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 5 (8074 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 138 waited for it, 1.98 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 185.9 ms; VRAM 6.9 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5944 queued, 5944 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14147 ms and 315.2 MB read in all); 8040 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11149 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1994 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8566 (5944 ready, 0 loading), 15674 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 12586 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.06, present 0.09, draw 0.82, descriptors 0.44, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.22, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.21
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back
[hos host] vk: 6600 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.14 ms a frame (fences 5.14, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 202.8 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 16.91 ms a frame): its jobs 8.52 ms a frame (submits 1.75, presents 6.23, others 0.54), idle 8.38; the game waited for room at 7 presents, 1.03 ms each (0.01 a frame; 5.79 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 215 captures at a CPU signal; the submits' waits read in 0.010 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.96 ms after its hand-over (176.08 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 10.33 + 0.17 (the window's copy is in the idle), after it 0.11 + 0.19; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 92.1% of them took nothing that came with the signal; of the others, 0.16 a frame took pages of the signal's capture into a mirror (0.9 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 192 of 215 frames, its first three in 192
[hos host] vk: frame ahead, what came with the signal: 215 signals with a capture, each of 11.3 pages in 2.1 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.04+0.17, 0.07+0.02, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: render targets drawn over from another start, let go: 9600 so far (45692 MB), 8590 of them (40971 MB) between the passes
[hos Presentation Thread] vi: 6611 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 4, 15.5-17.8 291, 17.8-22 5, 22-30 0, 30-40 0, over 40 0
[hos host] vk: write-backs left out (600 submits): 73, 511.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 116 dirty pages from their submits' captures, 48458 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 657 dispatches, 44218 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 511 runs, 6.0 MB compared; 146 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8028 in captures, 0 only in the collection, 73 only by the replay itself
[hos host] vk: read live, by pool:
[hos rmbl::VibrationThread] hid: 20480 vibration values, strongest 0.00 since the last count
[hos sead::MainThread] nvn: QueueSubmitCommands #40000: queue 000010100D6BF020 (its #6665), 1 handles (lr 0x80666be0)
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 681 ranges, 3961 MB), 184 pages written, 184 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 3500 texture pages copied (13 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5280 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 70, 490.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 72 dirty pages from their submits' captures, 46553 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 637 dispatches, 42385 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 490 runs, 5.8 MB compared; 140 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7684 in captures, 0 only in the collection, 70 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 14 retired still imported (388 MB), 21 retired given back (432 MB)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.09, present 0.10, draw 0.75, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.11, memory sync 0.41, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.4 copies, 0.08 MB written back
[hos host] vk: 6720 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.16 ms a frame (fences 5.16, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.7 ms a frame (the longest submission 11.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos BaseProcExecutor0] vcores: 94018 takes, 17090 waits (mean 26 us, longest 3.3 ms, 0 over HOS_VCORE_LOG_MS), 8877 kicks, 8648 yields, 0 starved grants, 0 escapes in 5 s
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.05 over 680 ranges, 3943 MB), 188 pages written, 188 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 24 kept (96 MB), 12 spare buffers, 3500 texture pages copied (13 MB)
[hos host] vk: readback pages refreshed from guest memory: 2595140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10560 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 92 dirty pages from their submits' captures, 47766 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 641 dispatches, 43569 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7914 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.02, present 0.09, draw 0.75, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.21, dispatch 0.11, memory sync 0.38, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back
[hos host] vk: 6840 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.13 ms a frame (fences 5.13, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.1 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.6 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #65000: event 0x100d6c8000 mode 0 value 64999 (lr 0x80664f14)
[hos sead::MainThread] barrier: 102787 looks: 37251 found the masks moved already, 63403 waited until they moved, 2133 for the whole 1 ms (a wait 313 us on average); 99077 wakes
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 91 dirty pages from their submits' captures, 47815 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43594 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7914 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 14 retired still imported (388 MB), 21 retired given back (432 MB)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #95000: event 0x100d6c8000 mode 0 value 65110 (now 65104) (lr 0x80664f90)
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.05 over 680 ranges, 3943 MB), 178 pages written, 178 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.008; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (88 MB), 11 spare buffers, 3981 texture pages copied (15 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3172 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos Presentation Thread] vi: 6911 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 4, 15.5-17.8 292, 17.8-22 4, 22-30 0, 30-40 0, over 40 0
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 47808 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43592 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7772 in captures, 133 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.17, present 0.10, draw 0.76, descriptors 0.39, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.21, dispatch 0.11, memory sync 0.41, queue submit 0.00, slot wait 0.21
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back
[hos host] vk: 6960 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.19 ms a frame (fences 5.19, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.0 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.3 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 680 ranges, 3941 MB), 192 pages written, 192 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (114 MB), 9 spare buffers, 3981 texture pages copied (15 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10560 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: recorder: 950517 packets run in 10.0 s (busy 0.46 s, spun idle 0.10 s, slept 8741 times, 3379 of them under 200 us), 0 calls made directly (no room); 10016 batches over 600 frames (16.7 a frame, 95 packets a batch), 5004 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 8 times, 0.41 ms (0.0 times, 0.00 ms a frame); every thread 565 times, 56.03 ms in all, the longest 0.29 ms; vkWaitForFences 550/55.38 ms (max 0.29), vkDestroyBuffer 8/0.41 ms (max 0.29), vmaDestroyBuffer 7/0.23 ms (max 0.04)
[hos host] vk: recorder: of its busy 0.46 s, 0.19 s in parts' ends and submits (the longest 0.64 ms), 0.00 s of that blocked on the queue's lock (4 times, the longest 0.58 ms), so 0.19 s in the ends, submits and holds; 0.27 s running the other packets; 1584 packets a frame
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 108 dirty pages from their submits' captures, 47798 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43589 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7910 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 2 retired still imported (48 MB), 33 retired given back (772 MB)
[hos BaseProcExecutor1] vcores: 93853 takes, 17366 waits (mean 26 us, longest 4.1 ms, 0 over HOS_VCORE_LOG_MS), 8999 kicks, 8759 yields, 0 starved grants, 0 escapes in 5 s
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.05 over 680 ranges, 3941 MB), 184 pages written, 184 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (124 MB), 11 spare buffers, 3981 texture pages copied (15 MB)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 78 dirty pages from their submits' captures, 47164 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42978 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7799 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5280 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.09, present 0.09, draw 0.72, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.19, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.3 copies, 0.08 MB written back
[hos host] vk: 7080 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.23 ms a frame (fences 5.23, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.4 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos SceneLowPriority] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 3179825 lookups, 7044 missed (0.22 %)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: readback pages refreshed from guest memory: 2607140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 72 dirty pages from their submits' captures, 47219 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 646 dispatches, 43000 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7790 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 0 retired still imported (0 MB), 35 retired given back (820 MB)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.01, scan 0.05 over 634 ranges, 3175 MB), 185 pages written, 185 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (118 MB), 10 spare buffers, 2128 texture pages copied (8 MB)
[hos host] vk: presenter thread: the copy's fence 5.16 ms, the present 0.57 ms a present
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 8512 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: memory: working set 15220 MB (peak 15259), private 21025 MB, staging 32 MB in 2 arenas, 37 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 37 allocations 692 MB; used 3000 of 15579 MB; heap 1 (device): 158 blocks 5027 MB, 12473 allocations 4075 MB; used 7022 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 93 images 283 MB; 53 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 47 blocks 1504 MB, 1178 images 1485 MB; 221 free ranges, the largest 13 MB
[hos host] vk: render target images: 5676 reused from the pool, 3142 made new, 3046 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1025 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2854 idle
[hos host] vk: VRAM: 7022 MB used of 11474 MB allowed (952 MB of it free in the blocks); textures 2769 MB (budget 7158 MB), 1280 textures
[hos host] vk: budget: VRAM 5027 of 11474 MB (7022 MB by the driver, 1994 MB unaccounted): textures 2769, blocks free 952, mirrors 658, shadows 296, other buffers 207, other images 111, staging 32; RAM 2961 MB (the process's private memory 21025 MB): imports 2228, capture spares 586, captures 116, capture table 31, others 0
[hos host] vk: textures: decoded ASTC 1503 MB, rendered 878 MB (90 drawn in one frame, kept: 22286 KB), other 345 MB; used in the last 8 presents 257 MB
[hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 6 listed; 1028 MB of blocks hashed in 170 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 2053 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1636 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1636 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1636 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1636 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1636 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 1 presents ago
[hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 14 (10200 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 5 (8074 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 153 waited for it, 0.19 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 19.7 ms; VRAM 6.9 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5944 queued, 5944 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14147 ms and 315.2 MB read in all); 8040 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11149 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1994 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8566 (5944 ready, 0 loading), 15674 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 12586 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.01, present 0.09, draw 0.74, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.19, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.4 copies, 0.08 MB written back
[hos host] vk: 7200 presents, 59.9 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.19 ms a frame (fences 5.19, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.1 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.4 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: queues ordered by their events: 40617 submissions stopped at a wait, 40613 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 96927, not met 0; captures at a CPU event's wait 5027, 0 of them at a wait met on arrival
[hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 8.14 ms a frame (submits 1.70, presents 5.89, others 0.55), idle 8.53; the game waited for room at 0 presents, 0.00 ms each (0.00 a frame; 5.15 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 222 captures at a CPU signal; the submits' waits read in 0.009 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.63 ms after its hand-over (16.17 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 10.56 + 0.17 (the window's copy is in the idle), after it 0.11 + 0.19; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 96.8% of them took nothing that came with the signal; of the others, 0.06 a frame took pages of the signal's capture into a mirror (0.2 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 210 of 222 frames, its first three in 210
[hos host] vk: frame ahead, what came with the signal: 222 signals with a capture, each of 3.3 pages in 1.8 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.04+0.16, 0.07+0.02, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos ModuleSystemWorker1] nvn: 850000 recordings made, 71504 live copies (18.1 MB), 5794 texture copies; 5 command buffers' segments (4.3 MB)
[hos host] vk: render targets drawn over from another start, let go: 9600 so far (45692 MB), 8590 of them (40971 MB) between the passes
[hos Presentation Thread] vi: 7212 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 3, 15.5-17.8 294, 17.8-22 4, 22-30 0, 30-40 0, over 40 0
[hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 83 dirty pages from their submits' captures, 47825 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43588 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7899 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 618 ranges, 2911 MB), 180 pages written, 180 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.008; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (130 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3232 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 73, 511.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 80 dirty pages from their submits' captures, 48441 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 650 dispatches, 44161 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 507 runs, 5.9 MB compared; 145 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8007 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 0 retired still imported (0 MB), 35 retired given back (820 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.02, present 0.10, draw 0.73, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.02, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.39, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.3 copies, 0.08 MB written back
[hos host] vk: 7320 presents, 60.1 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.23 ms a frame (fences 5.23, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 202.9 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos nn.npln.Worker] vcores: 93350 takes, 17048 waits (mean 26 us, longest 2.7 ms, 0 over HOS_VCORE_LOG_MS), 8716 kicks, 8550 yields, 0 starved grants, 0 escapes in 5 s
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 594 ranges, 2807 MB), 177 pages written, 177 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (114 MB), 10 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: write-backs left out (600 submits): 73, 511.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 93 dirty pages from their submits' captures, 48481 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 657 dispatches, 44204 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 511 runs, 6.0 MB compared; 146 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8025 in captures, 0 only in the collection, 73 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 0 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 77 kept their virtual core (nothing there to wait for), 2995 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 70, 490.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 59 dirty pages from their submits' captures, 46517 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 630 dispatches, 42353 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 494 runs, 5.8 MB compared; 141 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7681 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 176 live (2228 MB), 0 retired still imported (0 MB), 35 retired given back (820 MB)
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.01, present 0.09, draw 0.75, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.38, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.2 copies, 0.08 MB written back
[hos host] vk: 7440 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 5.06 ms a frame (fences 5.06, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 5.7 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 592 ranges, 2572 MB), 170 pages written, 170 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (102 MB), 10 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: reaper: 73 images (117 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 15 (58 MB)
[hos host] vk: pool 0x103cc00000..0x103d800000 took in 2 pools
[hos host] vk: pool 0x7fa308ce00+0xe38 imported in 3.7 ms as 0x7fa3000000..0x7fa3400000 (212 imports, 653 ms in all)
[hos host] vk: slow upload 36.7 ms: #15784 0x10f0851400 768x1536x1 x1 levels 11 NVN 0x7f ASTC (BC3), 515 of 515 KB
[hos host] vk: slow upload 15.8 ms: #15793 0x10f0b09a00 1536x1536x1 x1 levels 11 NVN 0x8d ASTC (BC3), 1026 of 1026 KB
[hos host] vk: a frame's uploads took 59 ms: ASTC on the GPU 10 (17.0 MB, 57 ms); on the CPU 17 (7.7 MB, 1 ms); [768x1536 x1 NVN 0x7f ASTC on the GPU 0.5 MB 37 ms] [1536x1536 x1 NVN 0x8d ASTC on the GPU 1.0 MB 16 ms]
[hos sead::MainThread] nvn: GPU waits in 5 s: 11 sync waits (322 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: mirror of pool 0x103d000000 (8 MB) given back: retired, 63 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x103cc00000 (4 MB) given back: retired, 63 submissions since its sync, GPU-written
[hos host] vk: the CPU wrote 0x103cddc000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cddd000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cff8000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cff9000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cffa000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cdde000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cffb000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cddf000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cffc000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cde0000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cffd000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cde1000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cffe000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cfff000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cde2000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103d000000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cde3000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103d001000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103cde4000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: the CPU wrote 0x103d002000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 4411136 kept (1353274 by the last picks' textures, a uniform buffer bound since), 26032 new, 2162831 with other textures (of 6600000)
[hos host] vk: late pages found at the replay (600 takes): 19 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos sead::MainThread] nvn: EventSignal (CPU) #15000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: readback pages refreshed from guest memory: 2619140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 157, 656.5 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 100 so far
[hos host] vk: mirror syncs (600 submits): 146 dirty pages from their submits' captures, 60189 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 52304 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 885 runs, 84.2 MB compared; 1969 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18081 in captures, 47 only in the collection, 67 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vi: 7509 frames presented (59.3 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 4, 15.5-17.8 291, 17.8-22 1, 22-30 0, 30-40 0, over 40 1
[hos host] vk: queues ordered by their events: 42000 submissions stopped at a wait, 41994 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 98813, not met 0; captures at a CPU event's wait 5150, 0 of them at a wait met on arrival
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 171 pages written, 171 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (114 MB), 11 spare buffers, 15 texture pages copied (0 MB)
[hos ModuleSystemWorker5] tex: the game's texture heap: 115 compaction rounds asked for in 7200 frames without one running
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.93, present 0.12, draw 1.45, descriptors 1.06, texture lookups 0.70, uploads 0.49, pipelines 0.03, programs 0.09, commands 0.19, dispatch 0.16, memory sync 0.40, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 237 (696 MB): 3.89 MB synced in 770.0 copies, 1.35 MB written back
[hos host] vk: 7560 presents, 58.2 fps; a frame: 175 draws (0 skipped), 7 dispatches, 130 passes, 82 barriers, 6 submits, 0.6 uploads (0.2 MB), 455 ub updates (196 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1347 textures (2917 MB)
[hos host] vk: host time in GPU resource waits: 4.15 ms a frame (fences 3.28, timelines 0.87, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 220.2 full barriers recorded, 101.0 left out (nothing recorded since the one before), 68.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.4; 0x10: 4.2; 0x11: 1.0; 0x12: 68.8; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 79.3, an op 50.3, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.8 ms a frame (the longest submission 7.2 ms); 0.7 textures made a frame: colour targets 0.4 (0.4 MB), depth 0.0 (0.1 MB), ASTC or BC 0.2 (1.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.07 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: late pages' fix: stamps for 0x103cdf0000 (pool 0x103cc00000; 384 blocks, 24576 KB)
[hos host] vk: late pages' fix: stamps for 0x103d1f0000 (pool 0x103cc00000; 448 blocks, 28672 KB)
[hos host] vk: late pages' fix: stamps for 0x103d5f0000 (pool 0x103cc00000; 512 blocks, 32768 KB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 4 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 210, 730.6 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 259 so far
[hos host] vk: mirror syncs (600 submits): 114 dirty pages from their submits' captures, 64604 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 982 dispatches, 55728 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 1050 runs, 104.5 MB compared; 140 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23415 in captures, 0 only in the collection, 70 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos host] vk: 11 textures retired (1338 left, 2877 MB): budget 0, drawn over 11, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: recorder: 945269 packets run in 10.0 s (busy 0.47 s, spun idle 0.10 s, slept 8771 times, 3455 of them under 200 us), 0 calls made directly (no room); 10199 batches over 597 frames (17.1 a frame, 93 packets a batch), 5018 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 16 times, 1.41 ms (0.0 times, 0.00 ms a frame); every thread 550 times, 50.82 ms in all, the longest 0.52 ms; vkWaitForFences 515/48.09 ms (max 0.37), vkDestroyImageView 9/1.01 ms (max 0.38), vkQueueBindSparse 7/0.80 ms (max 0.52), vmaDestroyBuffer 13/0.51 ms (max 0.13), vkEndCommandBuffer 2/0.35 ms (max 0.24), the queue's lock 1/0.03 ms (max 0.03)
[hos host] vk: recorder: of its busy 0.47 s, 0.19 s in parts' ends and submits (the longest 0.63 ms), 0.00 s of that blocked on the queue's lock (4 times, the longest 0.59 ms), so 0.19 s in the ends, submits and holds; 0.28 s running the other packets; 1583 packets a frame
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 173 pages written, 173 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.001 (1 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (98 MB), 11 spare buffers, 19 texture pages copied (0 MB)
[hos ModuleSystemWorker2] vcores: 93575 takes, 17031 waits (mean 26 us, longest 5.9 ms, 0 over HOS_VCORE_LOG_MS), 8727 kicks, 8566 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: write-backs left out (600 submits): 216, 706.0 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 423 so far
[hos host] vk: mirror syncs (600 submits): 118 dirty pages from their submits' captures, 67468 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57543 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 1080 runs, 152.9 MB compared; 144 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 24840 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 2 textures retired (1361 left, 2878 MB): budget 0, drawn over 1, dead targets 1
[hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.06, present 0.11, draw 0.75, descriptors 0.40, texture lookups 0.05, uploads 0.01, pipelines 0.02, programs 0.06, commands 0.19, dispatch 0.13, memory sync 0.37, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 237 (696 MB): 3.67 MB synced in 797.8 copies, 2.20 MB written back
[hos host] vk: 7680 presents, 60.0 fps; a frame: 172 draws (0 skipped), 8 dispatches, 131 passes, 83 barriers, 6 submits, 0.2 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.91 ms a frame (fences 3.86, timelines 0.05, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 221.8 full barriers recorded, 101.6 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.1; 0x10: 4.1; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.2, an op 51.1, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.1 ms a frame (the longest submission 15.1 ms); 0.2 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 6713644 lookups, 7689 missed (0.11 %)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #100000: event 0x100d6c8000 mode 0 value 67608 (now 67602) (lr 0x80664f90)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 213, 651.2 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 585 so far
[hos host] vk: mirror syncs (600 submits): 97 dirty pages from their submits' captures, 67739 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56616 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 1065 runs, 195.9 MB compared; 142 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 25254 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 180 pages written, 180 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (100 MB), 9 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: presenter thread: the copy's fence 4.24 ms, the present 0.51 ms a present
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 213, 606.5 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 746 so far
[hos host] vk: mirror syncs (600 submits): 101 dirty pages from their submits' captures, 69054 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56724 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 1065 runs, 240.5 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 25966 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: memory: working set 15087 MB (peak 15259), private 21148 MB, staging 48 MB in 3 arenas, 38 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 37 allocations 694 MB; used 3020 of 15579 MB; heap 1 (device): 162 blocks 5155 MB, 13226 allocations 4246 MB; used 7150 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 16 blocks 512 MB, 93 images 259 MB; 52 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1259 images 1618 MB; 228 free ranges, the largest 19 MB
[hos host] vk: render target images: 5676 reused from the pool, 3152 made new, 3056 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1035 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2864 idle
[hos host] vk: VRAM: 7150 MB used of 11474 MB allowed (909 MB of it free in the blocks); textures 2878 MB (budget 7158 MB), 1361 textures
[hos host] vk: budget: VRAM 5155 of 11474 MB (7150 MB by the driver, 1994 MB unaccounted): textures 2878, blocks free 909, mirrors 696, shadows 300, other buffers 211, other images 111, staging 48; RAM 2977 MB (the process's private memory 21148 MB): imports 2236, capture spares 576, captures 122, capture table 31, retired imports 12, others 0
[hos host] vk: textures: decoded ASTC 1626 MB, rendered 856 MB (97 drawn in one frame, kept: 19496 KB), other 355 MB; used in the last 8 presents 311 MB
[hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 7 listed; 1046 MB of blocks hashed in 175 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 2653 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 2236 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 2236 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 2236 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 2236 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 2236 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 349 presents ago
[hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 13 (8585 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 4 (6459 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 210 waited for it, 0.59 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 100.8 ms; VRAM 7.0 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5946 queued, 5946 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14150 ms and 315.2 MB read in all); 8042 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11151 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2008 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8568 (5946 ready, 0 loading), 15676 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 12590 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.10, present 0.12, draw 0.80, descriptors 0.42, texture lookups 0.04, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 237 (696 MB): 3.80 MB synced in 795.9 copies, 3.27 MB written back
[hos host] vk: 7800 presents, 60.0 fps; a frame: 202 draws (0 skipped), 8 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.89 ms a frame (fences 3.89, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 222.2 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 51.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.0 ms a frame (the longest submission 5.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 16.77 ms a frame): its jobs 9.32 ms a frame (submits 1.76, presents 6.91, others 0.65), idle 7.45; the game waited for room at 5 presents, 0.96 ms each (0.01 a frame; 5.66 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 250 captures at a CPU signal; the submits' waits read in 0.009 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 8.86 ms after its hand-over (108.66 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 9.46 + 0.17 (the window's copy is in the idle), after it 0.12 + 0.36; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 94.0% of them took nothing that came with the signal; of the others, 0.12 a frame took pages of the signal's capture into a mirror (0.7 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 222 of 250 frames, its first three in 222
[hos host] vk: frame ahead, what came with the signal: 250 signals with a capture, each of 4.1 pages in 2.2 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.06+0.34, 0.06+0.02, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: render targets drawn over from another start, let go: 9616 so far (45730 MB), 8605 of them (41009 MB) between the passes
[hos Presentation Thread] vi: 7810 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 1, 15.5-17.8 299, 17.8-22 1, 22-30 0, 30-40 0, over 40 0
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 181 pages written, 181 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 10 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: readback pages refreshed from guest memory: 2631140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 216, 569.3 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 910 so far
[hos host] vk: mirror syncs (600 submits): 136 dirty pages from their submits' captures, 70729 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57531 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 1080 runs, 289.7 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27019 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos rmbl::VibrationThread] hid: 24576 vibration values, strongest 0.00 since the last count
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 190 pages written, 190 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (104 MB), 10 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.06, present 0.11, draw 0.82, descriptors 0.43, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 797.9 copies, 4.29 MB written back
[hos host] vk: 7920 presents, 60.1 fps; a frame: 201 draws (0 skipped), 8 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.98 ms a frame (fences 3.98, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 221.4 full barriers recorded, 101.3 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 50.5, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.1 ms a frame (the longest submission 13.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos BaseProcExecutor1] vcores: 95156 takes, 17469 waits (mean 28 us, longest 4.7 ms, 0 over HOS_VCORE_LOG_MS), 8980 kicks, 8803 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: write-backs left out (600 submits): 216, 523.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 116 dirty pages from their submits' captures, 71710 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1016 dispatches, 57898 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 975 runs, 327.5 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27610 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 3300000 (2104954 pages from the captures)
[hos ModuleSystemWorker5] barrier: 126566 looks: 44646 found the masks moved already, 79787 waited until they moved, 2133 for the whole 1 ms (a wait 264 us on average); 134098 wakes
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 222, 518.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 144 dirty pages from their submits' captures, 73128 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1028 dispatches, 59112 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 733 runs, 341.7 MB compared; 147 pages read into the shadows; 300 MB of shadows (43 MB of pages known, in 91 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 28382 in captures, 0 only in the collection, 73 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 190 pages written, 190 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 10 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.04, present 0.10, draw 0.82, descriptors 0.42, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.22, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 796.0 copies, 4.68 MB written back
[hos host] vk: 8040 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.97 ms a frame (fences 3.97, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 218.8 full barriers recorded, 101.3 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.1 ms a frame (the longest submission 5.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 121 dirty pages from their submits' captures, 71309 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57606 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 727 runs, 341.7 MB compared; 145 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27634 in captures, 0 only in the collection, 73 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 188 pages written, 188 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (112 MB), 11 spare buffers, 19 texture pages copied (0 MB)
[hos Presentation Thread] vi: 8110 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 300, 17.8-22 0, 22-30 0, 30-40 0, over 40 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 142 dirty pages from their submits' captures, 71278 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57613 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27672 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.03, present 0.11, draw 0.79, descriptors 0.41, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.20, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 237 (696 MB): 3.88 MB synced in 796.2 copies, 4.68 MB written back
[hos host] vk: 8160 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.95 ms a frame (fences 3.95, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 218.8 full barriers recorded, 101.3 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.3 ms a frame (the longest submission 6.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos sead::MainThread] nvn: GPU waits in 5 s: 6 sync waits (16 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 188 pages written, 188 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (100 MB), 9 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: recorder: 989574 packets run in 10.0 s (busy 0.49 s, spun idle 0.09 s, slept 8543 times, 3247 of them under 200 us), 0 calls made directly (no room); 10602 batches over 600 frames (17.7 a frame, 93 packets a batch), 5001 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 0 times, 0.00 ms (0.0 times, 0.00 ms a frame); every thread 559 times, 57.15 ms in all, the longest 0.38 ms; vkWaitForFences 550/56.63 ms (max 0.38), vmaDestroyImage 1/0.19 ms (max 0.19), vkDestroyImageView 2/0.17 ms (max 0.14), vmaDestroyBuffer 6/0.16 ms (max 0.04)
[hos host] vk: recorder: of its busy 0.49 s, 0.20 s in parts' ends and submits (the longest 0.35 ms), 0.00 s of that blocked on the queue's lock (1 times, the longest 0.31 ms), so 0.20 s in the ends, submits and holds; 0.30 s running the other packets; 1649 packets a frame
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: readback pages refreshed from guest memory: 2643140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 108 dirty pages from their submits' captures, 70605 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 57022 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 710 runs, 332.3 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27244 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vcores: 95239 takes, 17177 waits (mean 28 us, longest 3.8 ms, 0 over HOS_VCORE_LOG_MS), 8939 kicks, 8753 yields, 0 starved grants, 0 escapes in 5 s
[hos sead::MainThread] nvn: 900000 recordings made, 71656 live copies (18.4 MB), 5837 texture copies; 2 command buffers' segments (4.3 MB)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.11, present 0.11, draw 0.82, descriptors 0.44, texture lookups 0.04, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 237 (696 MB): 3.88 MB synced in 797.8 copies, 4.68 MB written back
[hos host] vk: 8280 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 4.06 ms a frame (fences 4.05, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 219.4 full barriers recorded, 101.5 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.4 ms a frame (the longest submission 13.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 184 pages written, 184 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 93 dirty pages from their submits' captures, 70338 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56794 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 710 runs, 332.3 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27276 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos SceneLowPriority] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 2989158 lookups, 7159 missed (0.24 %)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: presenter thread: the copy's fence 3.95 ms, the present 0.49 ms a present
[hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 97 dirty pages from their submits' captures, 70318 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56797 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 710 runs, 332.3 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27216 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 186 pages written, 186 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (110 MB), 11 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: memory: working set 15087 MB (peak 15259), private 21148 MB, staging 48 MB in 3 arenas, 38 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 36 allocations 680 MB; used 3020 of 15579 MB; heap 1 (device): 162 blocks 5155 MB, 13226 allocations 4246 MB; used 7150 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 16 blocks 512 MB, 93 images 259 MB; 52 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1259 images 1618 MB; 228 free ranges, the largest 19 MB
[hos host] vk: render target images: 5676 reused from the pool, 3152 made new, 3056 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1035 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2864 idle
[hos host] vk: VRAM: 7150 MB used of 11474 MB allowed (909 MB of it free in the blocks); textures 2878 MB (budget 7158 MB), 1361 textures
[hos host] vk: budget: VRAM 5155 of 11474 MB (7150 MB by the driver, 1994 MB unaccounted): textures 2878, blocks free 909, mirrors 696, shadows 300, other buffers 211, other images 111, staging 48; RAM 2977 MB (the process's private memory 21148 MB): imports 2236, capture spares 586, captures 112, capture table 31, retired imports 12, others 0
[hos host] vk: textures: decoded ASTC 1626 MB, rendered 856 MB (97 drawn in one frame, kept: 19496 KB), other 355 MB; used in the last 8 presents 311 MB
[hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 7 listed; 1046 MB of blocks hashed in 175 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 3253 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 2836 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 2836 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 2836 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 2836 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 2836 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 949 presents ago
[hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 13 (8585 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 4 (6459 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 151 waited for it, 0.07 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 17.9 ms; VRAM 7.0 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5946 queued, 5946 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14150 ms and 315.2 MB read in all); 8042 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11151 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2008 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8568 (5946 ready, 0 loading), 15676 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 12590 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.07, present 0.11, draw 0.81, descriptors 0.42, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.37, queue submit 0.00, slot wait 0.21
[hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 796.1 copies, 4.68 MB written back
[hos host] vk: 8400 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.90 ms a frame (fences 3.90, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 219.1 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.3 ms a frame (the longest submission 6.3 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 9.69 ms a frame (submits 1.72, presents 7.42, others 0.54), idle 6.98; the game waited for room at 0 presents, 0.00 ms each (0.00 a frame; 5.04 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 201 captures at a CPU signal; the submits' waits read in 0.010 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 9.22 ms after its hand-over (16.26 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 9.01 + 0.15 (the window's copy is in the idle), after it 0.07 + 0.19; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 97.3% of them took nothing that came with the signal; of the others, 0.05 a frame took pages of the signal's capture into a mirror (0.3 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 190 of 201 frames, its first three in 190
[hos host] vk: frame ahead, what came with the signal: 201 signals with a capture, each of 3.9 pages in 2.1 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.03+0.18, 0.04+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: render targets drawn over from another start, let go: 9616 so far (45730 MB), 8605 of them (41009 MB) between the passes
[hos Presentation Thread] vi: 8411 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 301, 17.8-22 0, 22-30 0, 30-40 0, over 40 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 219, 511.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 143 dirty pages from their submits' captures, 72283 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1022 dispatches, 58405 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 723 runs, 337.0 MB compared; 145 pages read into the shadows; 300 MB of shadows (43 MB of pages known, in 91 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 28050 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos host] vk: commit: 2539 MB left under Windows' limit as it is now, and the page files can grow by 79851 MB more: counted as room
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 184 pages written, 184 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: program picks: 4521066 kept (1356870 by the last picks' textures, a uniform buffer bound since), 26034 new, 2252899 with other textures (of 6800000)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 210, 490.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 125 dirty pages from their submits' captures, 69723 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 988 dispatches, 56306 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 707 runs, 332.3 MB compared; 141 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 26817 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #70000: event 0x100d6c8000 mode 0 value 69999 (lr 0x80664f14)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.05, present 0.12, draw 0.77, descriptors 0.40, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.20, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 237 (696 MB): 3.88 MB synced in 797.7 copies, 4.68 MB written back
[hos host] vk: 8520 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.95 ms a frame (fences 3.95, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 219.3 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.5 ms a frame (the longest submission 13.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos BaseProcExecutor1] vcores: 96102 takes, 17314 waits (mean 27 us, longest 3.8 ms, 0 over HOS_VCORE_LOG_MS), 8907 kicks, 8707 yields, 0 starved grants, 0 escapes in 5 s
[hos sead::MainThread] nvn: CommandBufferWaitEvent #105000: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444)
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 187 pages written, 187 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (120 MB), 9 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: readback pages refreshed from guest memory: 2655140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 71190 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1000 dispatches, 57504 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27635 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.03, present 0.10, draw 0.83, descriptors 0.43, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.22, dispatch 0.13, memory sync 0.35, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 796.1 copies, 4.68 MB written back
[hos host] vk: 8640 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.73 ms a frame (fences 3.73, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 219.1 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.5 ms a frame (the longest submission 6.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 115 dirty pages from their submits' captures, 71313 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57607 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27673 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 188 pages written, 188 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 10 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos Presentation Thread] vi: 8711 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 300, 17.8-22 0, 22-30 0, 30-40 0, over 40 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 121 dirty pages from their submits' captures, 70397 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1002 dispatches, 56882 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 710 runs, 332.3 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27166 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.02, present 0.11, draw 0.78, descriptors 0.41, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.12, memory sync 0.36, queue submit 0.00, slot wait 0.21
[hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 796.0 copies, 4.68 MB written back
[hos host] vk: 8760 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.75 ms a frame (fences 3.75, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 219.1 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.6 ms a frame (the longest submission 6.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 186 pages written, 186 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 9 spare buffers, 19 texture pages copied (0 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 1 sync waits (5 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 144 dirty pages from their submits' captures, 71573 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57827 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27686 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: recorder: 990859 packets run in 10.0 s (busy 0.49 s, spun idle 0.09 s, slept 8600 times, 3294 of them under 200 us), 0 calls made directly (no room); 10670 batches over 600 frames (17.8 a frame, 93 packets a batch), 5035 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 0 times, 0.00 ms (0.0 times, 0.00 ms a frame); every thread 522 times, 53.87 ms in all, the longest 0.49 ms; vkWaitForFences 519/53.74 ms (max 0.49), vmaDestroyBuffer 3/0.12 ms (max 0.07)
[hos host] vk: recorder: of its busy 0.49 s, 0.19 s in parts' ends and submits (the longest 0.58 ms), 0.00 s of that blocked on the queue's lock (1 times, the longest 0.55 ms), so 0.19 s in the ends, submits and holds; 0.29 s running the other packets; 1651 packets a frame
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos tera::FileResourceMgr] vcores: 95158 takes, 17161 waits (mean 26 us, longest 2.6 ms, 0 over HOS_VCORE_LOG_MS), 8788 kicks, 8595 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 186 pages written, 186 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (114 MB), 11 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 97 dirty pages from their submits' captures, 71185 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1000 dispatches, 57504 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27672 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.01, present 0.11, draw 0.80, descriptors 0.42, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.35, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 237 (696 MB): 3.88 MB synced in 797.7 copies, 4.68 MB written back
[hos host] vk: 8880 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.83 ms a frame (fences 3.74, timelines 0.09, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 219.1 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.6 ms a frame (the longest submission 9.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: queues ordered by their events: 48000 submissions stopped at a wait, 47992 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 107040, not met 0; captures at a CPU event's wait 5664, 0 of them at a wait met on arrival
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos ModuleSystemWorker2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 2857641 lookups, 6724 missed (0.24 %)
[hos sead::MainThread] vk: readbacks of GPU results: 0 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 0 kept their virtual core (nothing there to wait for), 3072 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: readback pages refreshed from guest memory: 2667140 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 101 dirty pages from their submits' captures, 71315 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57582 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 713 runs, 332.3 MB compared; 143 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 91 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27582 in captures, 0 only in the collection, 71 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 186 pages written, 186 copied (0.7 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: presenter thread: the copy's fence 3.77 ms, the present 0.48 ms a present
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: memory: working set 15087 MB (peak 15259), private 21148 MB, staging 48 MB in 3 arenas, 38 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 36 allocations 680 MB; used 3020 of 15579 MB; heap 1 (device): 162 blocks 5155 MB, 13226 allocations 4246 MB; used 7150 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 16 blocks 512 MB, 93 images 259 MB; 52 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1259 images 1618 MB; 228 free ranges, the largest 19 MB
[hos host] vk: render target images: 5676 reused from the pool, 3152 made new, 3056 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1035 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2864 idle
[hos host] vk: VRAM: 7150 MB used of 11474 MB allowed (909 MB of it free in the blocks); textures 2878 MB (budget 7158 MB), 1361 textures
[hos host] vk: budget: VRAM 5155 of 11474 MB (7150 MB by the driver, 1994 MB unaccounted): textures 2878, blocks free 909, mirrors 696, shadows 300, other buffers 211, other images 111, staging 48; RAM 2959 MB (the process's private memory 21148 MB): imports 2236, capture spares 572, captures 108, capture table 31, retired imports 12, others 0
[hos host] vk: textures: decoded ASTC 1626 MB, rendered 856 MB (96 drawn in one frame, kept: 19240 KB), other 355 MB; used in the last 8 presents 311 MB
[hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 7 listed; 1046 MB of blocks hashed in 175 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 3853 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 3436 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 3436 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 3436 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 3436 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 3436 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 1549 presents ago
[hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 13 (8585 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 4 (6459 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 179 waited for it, 0.13 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 22.2 ms; VRAM 7.0 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5946 queued, 5946 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14150 ms and 315.2 MB read in all); 8042 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11151 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2008 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8568 (5946 ready, 0 loading), 15676 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 12590 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 2.03, present 0.11, draw 0.77, descriptors 0.40, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.20, dispatch 0.13, memory sync 0.35, queue submit 0.00, slot wait 0.20
[hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 795.9 copies, 4.68 MB written back
[hos host] vk: 9000 presents, 59.9 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB)
[hos host] vk: host time in GPU resource waits: 3.52 ms a frame (fences 3.52, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 219.2 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none
[hos host] vk: GPU time 7.7 ms a frame (the longest submission 7.6 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 10.01 ms a frame (submits 1.63, presents 7.78, others 0.60), idle 6.66; the game waited for room at 2 presents, 0.30 ms each (0.00 a frame; 5.15 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 228 captures at a CPU signal; the submits' waits read in 0.009 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 9.55 ms after its hand-over (22.07 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 8.67 + 0.16 (the window's copy is in the idle), after it 0.08 + 0.19; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 95.8% of them took nothing that came with the signal; of the others, 0.08 a frame took pages of the signal's capture into a mirror (0.6 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 209 of 228 frames, its first three in 209
[hos host] vk: frame ahead, what came with the signal: 228 signals with a capture, each of 4.2 pages in 2.1 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.03+0.18, 0.05+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos Presentation Thread] vi: real-time step on: ratio 0.996, game's frame rate 60
[hos host] vk: render targets drawn over from another start, let go: 9617 so far (45730 MB), 8605 of them (41009 MB) between the passes
[hos host] vk: 1 textures retired (1361 left, 2878 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 128 dirty pages from their submits' captures, 70308 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56802 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 717 runs, 337.0 MB compared; 143 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27312 in captures, 0 only in the collection, 72 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB)
[hos host] vk: pool 0x1010eb8000+0xf00000 imported in 18.8 ms as 0x1010800000..0x1012000000 (213 imports, 672 ms in all)
[hos host] vk: slow upload 18.8 ms: #15853 0x100032a1000 2560x1440x1 x1 levels 1 NVN 0xa, 15360 of 15360 KB
[hos host] vk: texture #14767 0x10004b01a00 2560x1440 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564)
[hos host] vk: texture #14768 0x10004fb1e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564)
[hos host] vk: texture #14769 0x100050a2000 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564)
[hos host] vk: texture #14770 0x10004ec1c00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564)
[hos host] vk: texture #14801 0x10004921600 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564)
[hos host] vk: reaper: 28 images (78 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 5 (39 MB)
[hos Presentation Thread] vi: 9011 frames presented (58.8 fps over the last 5.1 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 1, 15.5-17.8 297, 17.8-22 0, 22-30 1, 30-40 0, over 40 1
[hos host] vk: mirror of pool 0x1010800000 (20 MB) given back: retired, 28968 submissions since its sync, GPU-written
[hos host] vk: replay reads of guest memory as submitted: 3320000 (2115936 pages from the captures)
[hos host] vk: 1 textures retired (1413 left, 2963 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: replay reads of guest memory as submitted: 3340000 (2128850 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 594 ranges, 2913 MB), 203 pages written, 203 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (98 MB), 12 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: write-backs left out (600 submits): 852, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3152 dirty pages from their submits' captures, 46163 from earlier ones', 112 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 721 dispatches, 37048 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6777 runs, 608.8 MB compared; 436 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15678 in captures, 1 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x1163400000 (8 MB, captured, state 2): 90; 0x1159c00000 (36 MB, captured, state 2): 20; 0x7ffe000000 (24 MB, captured, state 2): 1; 0x1051800000 (4 MB, captured, state 2): 1;
[hos host] vk: program picks: 4648785 kept (1390759 by the last picks' textures, a uniform buffer bound since), 26034 new, 2325180 with other textures (of 7000000)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 3360000 (2141601 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 51 sync waits (305 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: texture #15688 0x10001f54a00 80x45 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15683 0x10001f5b600 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15685 0x100020b6800 320x180 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15766 0x10005661000 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15765 0x10005570e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15682 0x10001e00400 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15684 0x10001ef0600 320x180 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15686 0x10001f40800 160x90 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15664 0x10000000000 128x128 x6 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15768 0x10005570e00 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: texture #15661 0x10000000000 256x256 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009)
[hos host] vk: replay reads of guest memory as submitted: 3380000 (2155760 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 11 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 117, 1180 MB with memory of 1592 MB (411 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 496 binds (146.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 44 spares now, 996 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3110 dirty pages from their submits' captures, 46514 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 727 dispatches, 37153 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6808 runs, 605.5 MB compared; 35 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15591 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 177 live (2240 MB), 3 retired still imported (32 MB), 35 retired given back (820 MB)
[hos host] vk: 1 textures retired (1434 left, 2989 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: replay reads of guest memory as submitted: 3400000 (2169447 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #110000: event 0x100d6c802c mode 0 value 73182 (now 73163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos sead::MainThread] nvn: 950000 recordings made, 71999 live copies (24.9 MB), 5844 texture copies; 2 command buffers' segments (4.1 MB)
[hos rmbl::VibrationThread] hid: 28672 vibration values, strongest 0.00 since the last count
[hos host] vk: replay reads of guest memory as submitted: 3420000 (2183145 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 849, 3772.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3366 dirty pages from their submits' captures, 46459 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 753 dispatches, 37257 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6649 runs, 588.1 MB compared; 38 pages read into the shadows; 300 MB of shadows (46 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15648 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos NavMeshCAStepDepWorker1] vcores: 109247 takes, 17514 waits (mean 51 us, longest 8.0 ms, 0 over HOS_VCORE_LOG_MS), 8525 kicks, 8536 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 1 textures retired (1465 left, 3069 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: replay reads of guest memory as submitted: 3440000 (2196415 pages from the captures)
[hos host] vk: a frame: 274 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.72, present 0.19, draw 7.60, descriptors 3.66, texture lookups 0.78, uploads 0.40, pipelines 0.28, programs 1.16, commands 1.95, dispatch 0.94, memory sync 0.36, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 299 (1183 MB): 9.79 MB synced in 1068.0 copies, 16.86 MB written back
[hos host] vk: 9120 presents, 52.8 fps; a frame: 2080 draws (0 skipped), 155 dispatches, 203 passes, 151 barriers, 6 submits, 0.8 uploads (0.9 MB), 112 ub updates (69 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1465 textures (3069 MB)
[hos host] vk: host time in GPU resource waits: 9.40 ms a frame (fences 4.37, timelines 5.04, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 756.8 full barriers recorded, 326.5 left out (nothing recorded since the one before), 21.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.7; 0x8: 2.8; 0x9: 0.9; 0x10: 5.0; 0x11: 38.5; 0x12: 72.2; 0x13: 1.9; 0x21: 3.9; 0x28: 14.8; 0x92: 3.0; 0x97: 0.9;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 42.6 from 27.8 batches at the game's barriers (45.3 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 142.0, an op 1513.1, the end 6.6, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.9; 5: 24.1; 48: 7.4;
[hos host] vk: GPU time 13.7 ms a frame (the longest submission 7.1 ms); 1.2 textures made a frame: colour targets 0.4 (0.6 MB), depth 0.2 (0.4 MB), ASTC or BC 0.6 (1.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.59 MB unswizzled on the CPU in 0.09 ms
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 600 ranges, 2963 MB), 237 pages written, 237 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 19 texture pages copied (0 MB)
[hos host] vk: 8 textures retired (1461 left, 3072 MB): budget 0, drawn over 8, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 3460000 (2209996 pages from the captures)
[hos host] vk: program picks: 4801422 kept (1459419 by the last picks' textures, a uniform buffer bound since), 26049 new, 2372528 with other textures (of 7200000)
[hos host] vk: 1 textures retired (1460 left, 3072 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 2722646 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 118, 1183 MB with memory of 1596 MB (412 MB spared); blocks given memory later 6212, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 498 binds (147.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 138 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 43 spares now, 992 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 826, 3665.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3380 dirty pages from their submits' captures, 45518 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 741 dispatches, 36338 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6800 runs, 605.4 MB compared; 40 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15307 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 178 live (2244 MB), 3 retired still imported (32 MB), 35 retired given back (820 MB)
[hos host] vk: replay reads of guest memory as submitted: 3480000 (2223551 pages from the captures)
[hos host] vk: (its written part: 1 of 1 KB readable, 654 bytes not zero; levels L0)
[hos host] vk: texture 0x10e52b7600 1224x1528x1 x3 levels 1 NVN 0x50 (5544 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 1 KB of it
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1016 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1064 kept their virtual core (nothing there to wait for), 2776 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 3500000 (2237223 pages from the captures)
[hos sead::MainThread] barrier: 152194 looks: 53890 found the masks moved already, 96156 waited until they moved, 2148 for the whole 1 ms (a wait 236 us on average); 167746 wakes
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3342 dirty pages from their submits' captures, 47306 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 769 dispatches, 37511 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6785 runs, 605.1 MB compared; 36 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15719 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 3520000 (2250854 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #75000: event 0x100d6c8014 mode 0 value 75000 (lr 0x80660bcc)
[hos host] vk: replay reads of guest memory as submitted: 3540000 (2264417 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 604 ranges, 2965 MB), 241 pages written, 241 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 148 texture pages copied (0 MB)
[hos host] vk: sparse mirrors: 118, 1183 MB with memory of 1596 MB (412 MB spared); blocks given memory later 6213, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 499 binds (147.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 138 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 43 spares now, 992 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3203 dirty pages from their submits' captures, 46519 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 730 dispatches, 37232 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6768 runs, 604.9 MB compared; 39 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15659 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 178 live (2244 MB), 3 retired still imported (32 MB), 35 retired given back (820 MB)
[hos host] vk: replay reads of guest memory as submitted: 3560000 (2277882 pages from the captures)
[hos host] vk: program picks: 4955167 kept (1530953 by the last picks' textures, a uniform buffer bound since), 26050 new, 2418782 with other textures (of 7400000)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 129 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 3580000 (2291609 pages from the captures)
[hos host] vk: texture 0x10d221bc00 768x768x1 x1 levels 10 NVN 0x7b -> vk 37 (ASTC decoded) (1500 textures, 3114 MB)
[hos host] vk: a frame: 277 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.69, present 0.17, draw 7.38, descriptors 3.21, texture lookups 0.18, uploads 0.04, pipelines 0.30, programs 1.24, commands 2.04, dispatch 0.90, memory sync 0.41, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 301 (1203 MB): 6.06 MB synced in 1099.8 copies, 17.79 MB written back
[hos host] vk: 9240 presents, 53.5 fps; a frame: 2288 draws (0 skipped), 166 dispatches, 211 passes, 150 barriers, 6 submits, 0.5 uploads (0.1 MB), 81 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12988 pipelines (3203 ms), 0 programs (10 failed), 1501 textures (3114 MB)
[hos host] vk: host time in GPU resource waits: 9.84 ms a frame (fences 4.52, timelines 5.33, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 816.5 full barriers recorded, 344.7 left out (nothing recorded since the one before), 17.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.2; 0x8: 3.0; 0x9: 1.0; 0x10: 3.2; 0x11: 41.3; 0x12: 70.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 140.9, an op 1635.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 14.9 ms a frame (the longest submission 6.1 ms); 0.5 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.4 (0.5 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 3600000 (2305892 pages from the captures)
[hos host] vk: 2 of 2 small textures the GPU drew written into their memory in 1.2 ms, 1 frames after their copies were recorded, the first #16055 0x116cf4d000 260x260 levels 9 NVN 0xd (26 so far, 5765 KB, 7 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 6006 dirty pages from their submits' captures, 51991 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 831 dispatches, 39056 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6653 runs, 617.7 MB compared; 54 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16092 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 140000 so far (x1 0x1179846000 x2 0 x3 0x1179846000 x4 0x654040)
[hos host] vk: replay reads of guest memory as submitted: 3620000 (2319919 pages from the captures)
[hos host] vk: 2 textures retired (1537 left, 3176 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: late pages' fix (600 recordings with one): of 1113 captures with late pages 1012 fixed (73.7 pages, 74.1 fixes a capture), waited as before: 0 no pool, 101 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 448 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 1 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 39, of an earlier recording 408, not recorded yet 0, of this part 0; pages fixed from stamps 73651, from snapshots 939; the shadows: 1288561 pages synced from captures, 1259005 made known by it, 27346 left unknown because the replay or the collection wrote them since
[hos Presentation Thread] vi: 9278 frames presented (53.3 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 101, 17.8-22 153, 22-30 9, 30-40 4, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 3640000 (2334660 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 121, 1209 MB with memory of 1620 MB (411 MB spared); blocks given memory later 6391, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 508 binds (147.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 141 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 40 spares now, 968 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2639 dirty pages from their submits' captures, 49944 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 803 dispatches, 39053 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6245 runs, 612.7 MB compared; 68 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16280 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd6400000 (48 MB, captured, state 2): 1;
[hos host] vk: imports: 179 live (2256 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB)
[hos host] vk: replay reads of guest memory as submitted: 3660000 (2349463 pages from the captures)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #115000: event 0x100d6c802c mode 0 value 76982 (now 76963) (lr 0x8028c554)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 606 ranges, 3192 MB), 265 pages written, 265 copied (1.0 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (100 MB), 12 spare buffers, 254 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 3680000 (2364162 pages from the captures)
[hos host] vk: program picks: 5106856 kept (1600648 by the last picks' textures, a uniform buffer bound since), 26064 new, 2467079 with other textures (of 7600000)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1092 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1380 kept their virtual core (nothing there to wait for), 2972 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: readback pages refreshed from guest memory: 2789369 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3321 dirty pages from their submits' captures, 44782 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 721 dispatches, 35120 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6692 runs, 588.7 MB compared; 36 pages read into the shadows; 300 MB of shadows (46 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15191 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 3700000 (2379282 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 260 sync waits (1458 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 3720000 (2393868 pages from the captures)
[hos host] vk: 3 textures retired (1534 left, 3164 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: texture slots registered again at another size: 400000 (of 509712 registrations; slot 16221 640x360 -> 256x256)
[hos host] vk: replay reads of guest memory as submitted: 3740000 (2408580 pages from the captures)
[hos host] vk: recorder: 4643494 packets run in 10.0 s (busy 1.23 s, spun idle 0.12 s, slept 10974 times, 4093 of them under 200 us), 0 calls made directly (no room); 15604 batches over 547 frames (28.5 a frame, 298 packets a batch), 7753 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 86 times, 22.90 ms (0.2 times, 0.04 ms a frame); every thread 644 times, 64.01 ms in all, the longest 10.86 ms; vkWaitForFences 474/33.76 ms (max 0.51), the queue's lock 35/24.53 ms (max 10.86), vkQueueBindSparse 75/2.54 ms (max 0.23), vkDestroyImageView 28/2.32 ms (max 0.28), vmaDestroyBuffer 26/0.37 ms (max 0.04), vkEndCommandBuffer 2/0.34 ms (max 0.19)
[hos host] vk: recorder: of its busy 1.23 s, 0.27 s in parts' ends and submits (the longest 1.15 ms), 0.01 s of that blocked on the queue's lock (25 times, the longest 1.11 ms), so 0.26 s in the ends, submits and holds; 0.96 s running the other packets; 8489 packets a frame
[hos host] vk: sparse mirrors: 121, 1210 MB with memory of 1620 MB (409 MB spared); blocks given memory later 6411, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 509 binds (147.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 141 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 40 spares now, 968 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3391 dirty pages from their submits' captures, 45525 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 730 dispatches, 36188 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6743 runs, 588.9 MB compared; 35 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14716 in captures, 0 only in the collection, 26 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 179 live (2256 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB)
[hos host] vk: mirror of pool 0x7fa3000000 (4 MB) given back: idle, 6036 submissions since its sync
[hos host] vk: mirror of pool 0x7fa3400000 (4 MB) given back: idle, 6036 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 3760000 (2423160 pages from the captures)
[hos host] vk: a frame: 310 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.75, present 0.20, draw 7.75, descriptors 3.52, texture lookups 0.30, uploads 0.14, pipelines 0.32, programs 1.24, commands 2.07, dispatch 0.89, memory sync 0.39, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 303 (1211 MB): 5.83 MB synced in 1087.1 copies, 17.72 MB written back
[hos host] vk: 9360 presents, 50.1 fps; a frame: 2282 draws (0 skipped), 163 dispatches, 206 passes, 148 barriers, 6 submits, 0.3 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12989 pipelines (3204 ms), 0 programs (10 failed), 1534 textures (3164 MB)
[hos host] vk: host time in GPU resource waits: 10.88 ms a frame (fences 4.13, timelines 6.75, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 803.1 full barriers recorded, 339.4 left out (nothing recorded since the one before), 16.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 40.4; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 138.6, an op 1580.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 15.9 ms a frame (the longest submission 11.8 ms); 0.8 textures made a frame: colour targets 0.3 (1.7 MB), depth 0.2 (1.1 MB), ASTC or BC 0.2 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.06 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: replay reads of guest memory as submitted: 3780000 (2437742 pages from the captures)
[hos nn.npln.Callback] vcores: 132370 takes, 23179 waits (mean 77 us, longest 13.2 ms, 0 over HOS_VCORE_LOG_MS), 10669 kicks, 10863 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2965 dirty pages from their submits' captures, 45280 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 724 dispatches, 36038 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6575 runs, 571.7 MB compared; 60 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14889 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 1;
[hos host] vk: replay reads of guest memory as submitted: 3800000 (2452435 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 614 ranges, 3204 MB), 245 pages written, 245 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (96 MB), 12 spare buffers, 254 texture pages copied (0 MB)
[hos host] vk: program picks: 5260085 kept (1672308 by the last picks' textures, a uniform buffer bound since), 26066 new, 2513848 with other textures (of 7800000)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 106 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 3820000 (2466846 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 3840000 (2480075 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 122, 1213 MB with memory of 1624 MB (410 MB spared); blocks given memory later 6426, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 516 binds (148.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 144 taken from the spares, 161 made new; 183 kept as spares, 0 destroyed; 39 spares now, 964 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 851, 3776.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 4073 dirty pages from their submits' captures, 47471 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 759 dispatches, 37784 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7303 runs, 628.3 MB compared; 37 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15443 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 180 live (2260 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB)
[hos host] vk: replay reads of guest memory as submitted: 3860000 (2493453 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 3880000 (2506184 pages from the captures)
[hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 41712489 lookups, 97484 missed (0.23 %)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3068 dirty pages from their submits' captures, 47063 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 740 dispatches, 37373 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7167 runs, 611.6 MB compared; 38 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15518 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 3900000 (2519678 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #80000: event 0x100d6c8020 mode 0 value 80000 (lr 0x80664f14)
[hos ModuleSystemWorker1] nvn: 1000000 recordings made, 73669 live copies (23.7 MB), 5862 texture copies; 5 command buffers' segments (4.3 MB)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 611 ranges, 3212 MB), 250 pages written, 250 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (120 MB), 12 spare buffers, 492 texture pages copied (1 MB)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1108 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1420 kept their virtual core (nothing there to wait for), 3188 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 3920000 (2532367 pages from the captures)
[hos host] vk: program picks: 5411383 kept (1739339 by the last picks' textures, a uniform buffer bound since), 26070 new, 2562546 with other textures (of 8000000)
[hos host] vk: a frame: 288 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.09, present 0.18, draw 7.93, descriptors 3.44, texture lookups 0.09, uploads 0.01, pipelines 0.32, programs 1.33, commands 2.22, dispatch 0.95, memory sync 0.45, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 305 (1214 MB): 5.83 MB synced in 1099.3 copies, 17.96 MB written back
[hos host] vk: 9480 presents, 55.4 fps; a frame: 2247 draws (0 skipped), 171 dispatches, 201 passes, 138 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12989 pipelines (3204 ms), 0 programs (10 failed), 1550 textures (3178 MB)
[hos host] vk: host time in GPU resource waits: 8.89 ms a frame (fences 4.72, timelines 4.17, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 782.6 full barriers recorded, 323.0 left out (nothing recorded since the one before), 17.2 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 30.9; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 128.9, an op 1591.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 14.1 ms a frame (the longest submission 5.1 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: readback pages refreshed from guest memory: 2856733 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 122, 1214 MB with memory of 1624 MB (409 MB spared); blocks given memory later 6442, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 518 binds (148.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 144 taken from the spares, 161 made new; 183 kept as spares, 0 destroyed; 39 spares now, 964 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 4165 dirty pages from their submits' captures, 47755 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 773 dispatches, 38061 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7170 runs, 611.5 MB compared; 38 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15885 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 180 live (2260 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB)
[hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.7 ms, 1 frames after their copies were recorded, the first #16156 0x116d1f9000 260x260 levels 9 NVN 0xd (27 so far, 5975 KB, 8 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: replay reads of guest memory as submitted: 3940000 (2546318 pages from the captures)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #120000: event 0x100d6c802c mode 0 value 80782 (now 80763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 3960000 (2560246 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 3980000 (2574962 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 874, 3883.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3551 dirty pages from their submits' captures, 48737 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 757 dispatches, 38112 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7371 runs, 629.6 MB compared; 37 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15384 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 4000000 (2588790 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 4020000 (2602822 pages from the captures)
[hos Presentation Thread] vi: 9546 frames presented (53.5 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 70, 17.8-22 193, 22-30 5, 30-40 0, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 4040000 (2617337 pages from the captures)
[hos host] vk: sparse mirrors: 124, 1225 MB with memory of 1632 MB (406 MB spared); blocks given memory later 6566, given back 8041; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 528 binds (148.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 146 taken from the spares, 161 made new; 183 kept as spares, 0 destroyed; 37 spares now, 956 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3976 dirty pages from their submits' captures, 46467 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 738 dispatches, 36830 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7147 runs, 596.0 MB compared; 40 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15253 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2264 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 591 ranges, 3223 MB), 267 pages written, 267 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 886 texture pages copied (3 MB)
[hos host] vk: replay reads of guest memory as submitted: 4060000 (2632214 pages from the captures)
[hos host] vk: presenter thread: the copy's fence 4.33 ms, the present 0.78 ms a present
[hos host] vk: program picks: 5563234 kept (1807667 by the last picks' textures, a uniform buffer bound since), 26080 new, 2610685 with other textures (of 8200000)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 685 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 4080000 (2646804 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3308 dirty pages from their submits' captures, 46109 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 731 dispatches, 36625 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7044 runs, 579.0 MB compared; 35 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15292 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 4100000 (2661154 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] nvn: GPU waits in 5 s: 271 sync waits (1270 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 4120000 (2675815 pages from the captures)
[hos host] vk: memory: working set 15103 MB (peak 15259), private 21420 MB, staging 80 MB in 5 arenas, 40 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 12 blocks 736 MB, 36 allocations 660 MB; used 3040 of 15579 MB; heap 1 (device): 169 blocks 5378 MB, 21892 allocations 5110 MB; used 7374 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 15 blocks 480 MB, 104 images 296 MB; 60 free ranges, the largest 17 MB
[hos host] vk: image blocks (textures, compacted): 61 blocks 1952 MB, 1451 images 1917 MB; 251 free ranges, the largest 30 MB
[hos host] vk: render target images: 5719 reused from the pool, 3222 made new, 3116 destroyed from it, 989 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1096 with the pool empty (0 of them with targets still in the dead queue), 286 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 501 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2924 idle
[hos host] vk: VRAM: 7374 MB used of 11474 MB allowed (268 MB of it free in the blocks); textures 3181 MB (budget 7158 MB), 1563 textures
[hos host] vk: budget: VRAM 5378 of 11474 MB (7374 MB by the driver, 1995 MB unaccounted): textures 3181, mirrors 1225, shadows 300, blocks free 268, other buffers 211, other images 111, staging 80; RAM 2975 MB (the process's private memory 21420 MB): imports 2264, capture spares 552, captures 98, capture table 31, retired imports 20, other buffers 10, others 0
[hos host] vk: textures: decoded ASTC 1837 MB, rendered 873 MB (106 drawn in one frame, kept: 47145 KB), other 426 MB; used in the last 8 presents 1965 MB
[hos host] vk: decoded ASTC used again: 25 textures (86 MB) copied from ones the game finalized, not decoded; 25 of those retired, 2 listed; 1089 MB of blocks hashed in 184 ms
[hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 4453 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 2149 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 4453 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 15 (6972 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); 1466x282x1 NVN 0x25: 4 (6459 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 593 waited for it, 5.19 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 114.6 ms; VRAM 7.2 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 5964 queued, 5964 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14177 ms and 315.6 MB read in all); 8062 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11165 graphics pipelines (1 not in a cache), 711 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2056 pipelines made at create taken by draws; 2627 create calls with nothing kept for their stages, 0 before the index; units 8591 (5964 ready, 0 loading), 15729 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 12618 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 305 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.94, present 0.20, draw 7.74, descriptors 3.32, texture lookups 0.10, uploads 0.01, pipelines 0.31, programs 1.32, commands 2.19, dispatch 0.92, memory sync 0.46, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 307 (1225 MB): 5.93 MB synced in 1105.6 copies, 18.06 MB written back
[hos host] vk: 9600 presents, 54.1 fps; a frame: 2253 draws (0 skipped), 175 dispatches, 200 passes, 134 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12999 pipelines (3209 ms), 0 programs (10 failed), 1563 textures (3181 MB)
[hos host] vk: host time in GPU resource waits: 9.92 ms a frame (fences 4.91, timelines 5.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 780.5 full barriers recorded, 317.5 left out (nothing recorded since the one before), 17.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.9 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1587.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 14.7 ms a frame (the longest submission 11.1 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 18.82 ms a frame): its jobs 15.94 ms a frame (submits 1.09, presents 3.85, others 11.00), idle 2.87; the game waited for room at 5 presents, 0.44 ms each (0.00 a frame; 2.51 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 594 captures at a CPU signal; the submits' waits read in 0.043 ms a frame; readbacks: 2945 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.01 ms after its hand-over (99.66 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.36 + 0.15 (the window's copy is in the idle), after it 2.04 + 0.67; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 10.69 parts a frame recorded in the signal's job, 71.6% of them took nothing that came with the signal; of the others, 3.03 a frame took pages of the signal's capture into a mirror (103.5 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 6 of 594 frames, its first three in 6
[hos host] vk: frame ahead, what came with the signal: 594 signals with a capture, each of 136.2 pages in 12.6 pools; between a frame's last submit and that signal 0.02 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (4 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.06+0.19, 0.16+0.05, 0.10+0.10, 0.27+0.07, 0.03+0.02, 0.09+0.04, 0.14+0.05, 0.10+0.06, 0.06+0.05, 0.02+0.03, 0.01+0.01, 0.00+0.00
[hos host] vk: mirror of pool 0x7fa1800000 (4 MB) given back: idle, 6034 submissions since its sync
[hos host] vk: render targets drawn over from another start, let go: 9721 so far (46153 MB), 8706 of them (41432 MB) between the passes
[hos host] vk: replay reads of guest memory as submitted: 4140000 (2690453 pages from the captures)
[hos host] vk: mirror of pool 0x7f81800000 (12 MB) given back: idle, 6056 submissions since its sync
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 122, 1216 MB with memory of 1616 MB (399 MB spared); blocks given memory later 6574, given back 8041; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 532 binds (149.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 146 taken from the spares, 161 made new; 185 kept as spares, 0 destroyed; 39 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3124 dirty pages from their submits' captures, 45270 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 715 dispatches, 35446 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7252 runs, 597.5 MB compared; 32 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14916 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2264 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB)
[hos host] vk: replay reads of guest memory as submitted: 4160000 (2704458 pages from the captures)
[hos sead::MainThread] vk: readbacks of GPU results: 1104 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1289 kept their virtual core (nothing there to wait for), 3063 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 4180000 (2718671 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 591 ranges, 3223 MB), 245 pages written, 245 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (96 MB), 12 spare buffers, 886 texture pages copied (3 MB)
[hos host] vk: readback pages refreshed from guest memory: 2923824 pages so far (0 held: may be written ahead in their submission)
[hos NavMeshCAStepDepWorker1] vcores: 139238 takes, 25434 waits (mean 84 us, longest 19.4 ms, 0 over HOS_VCORE_LOG_MS), 11792 kicks, 12005 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3018 dirty pages from their submits' captures, 45946 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 707 dispatches, 36389 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7235 runs, 597.2 MB compared; 34 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15336 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: program picks: 5714423 kept (1874161 by the last picks' textures, a uniform buffer bound since), 26081 new, 2659495 with other textures (of 8400000)
[hos host] vk: replay reads of guest memory as submitted: 4200000 (2732867 pages from the captures)
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 160000 so far (x1 0x117da08010 x2 0 x3 0x117da08000 x4 0x9b040)
[hos host] vk: replay reads of guest memory as submitted: 4220000 (2747574 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 4240000 (2761945 pages from the captures)
[hos host] vk: sparse mirrors: 123, 1224 MB with memory of 1620 MB (396 MB spared); blocks given memory later 6644, given back 8041; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 536 binds (149.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 147 taken from the spares, 161 made new; 185 kept as spares, 0 destroyed; 38 spares now, 968 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3561 dirty pages from their submits' captures, 47909 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 753 dispatches, 37736 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7382 runs, 615.3 MB compared; 34 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15700 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2264 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB)
[hos host] vk: replay reads of guest memory as submitted: 4260000 (2776112 pages from the captures)
[hos host] vk: queues ordered by their events: 54000 submissions stopped at a wait, 53993 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 124852, not met 0; captures at a CPU event's wait 6392, 0 of them at a wait met on arrival
[hos host] vk: mirror of pool 0x7f8a400000 (4 MB) given back: idle, 6025 submissions since its sync
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #125000: event 0x100d6c802c mode 0 value 84582 (now 84563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: mirror of pool 0x7fa9000000 (4 MB) given back: idle, 6028 submissions since its sync
[hos host] vk: replay reads of guest memory as submitted: 4280000 (2790517 pages from the captures)
[hos host] vk: upload over a drawn image: #1212 0x10001fe0800 640x360x1 x1 NVN 0x14 at present 9706 (texture frame 9707), finalized, 842 KB dirty in 1 ranges
[hos host] vk: reaper: 79 images (265 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 13 (80 MB)
[hos host] vk: pool 0x100fcfca00+0xf00000 imported in 3.2 ms as 0x100f800000..0x1012000000 (219 imports, 680 ms in all)
[hos host] vk: pool 0x100f800000..0x1012000000 took in 2 pools
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: texture #1928 0x1001efcd000 192x192x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4630 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #1925 0x1001efd9200 128x128x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4630 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14058 0x10004470e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4563 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14052 0x1001e071200 2560x1440x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 4563 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #506 0x1147c16000 12288x12288x1 x1 NVN 0x32 the GPU drew goes with no copy of its texels (live, unused for 4563 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 70 textures retired (1524 left, 2868 MB): budget 68, drawn over 2, dead targets 0; GPU-drawn with no copy kept: 1 live, 4 finalized, 316 MB
[hos host] vk: mirror of pool 0x1010800000 (24 MB) given back: retired, 111 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x100f800000 (8 MB) given back: retired, 79 submissions since its sync, GPU-written
[hos host] vk: mirror of pool 0x7fa2000000 (8 MB) given back: idle, 6053 submissions since its sync
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9712 (texture frame 9713), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9714 (texture frame 9715), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9716 (texture frame 9717), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9718 (texture frame 9719), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3006 dirty pages from their submits' captures, 50999 from earlier ones', 224 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 39758 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7679 runs, 650.7 MB compared; 150 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16523 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9ac00000 (20 MB, captured, state 2): 224;
[hos sead::MainThread] nvn: CommandBufferSignalEvent #85000: event 0x100d6c8028 mode 0 value 84999 (lr 0x80667410)
[hos host] vk: a frame: 307 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.86, present 0.30, draw 7.52, descriptors 3.27, texture lookups 0.27, uploads 0.14, pipelines 0.30, programs 1.32, commands 2.04, dispatch 0.98, memory sync 0.43, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 311 (1227 MB): 6.30 MB synced in 1101.7 copies, 18.10 MB written back
[hos host] vk: 9720 presents, 53.7 fps; a frame: 2138 draws (0 skipped), 176 dispatches, 192 passes, 134 barriers, 6 submits, 0.8 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12999 pipelines (3209 ms), 0 programs (10 failed), 1552 textures (2914 MB)
[hos host] vk: host time in GPU resource waits: 8.97 ms a frame (fences 4.28, timelines 4.69, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 752.6 full barriers recorded, 309.7 left out (nothing recorded since the one before), 17.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1443.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 14.4 ms a frame (the longest submission 6.8 ms); 0.8 textures made a frame: colour targets 0.3 (1.1 MB), depth 0.1 (0.6 MB), ASTC or BC 0.4 (0.5 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.54 MB unswizzled on the CPU in 0.08 ms
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9722 (texture frame 9723), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4300000 (2803053 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 598 ranges, 3301 MB), 250 pages written, 250 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (96 MB), 11 spare buffers, 886 texture pages copied (3 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9726 (texture frame 9727), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9728 (texture frame 9729), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9730 (texture frame 9731), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9732 (texture frame 9733), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9734 (texture frame 9735), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9736 (texture frame 9737), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9738 (texture frame 9739), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9740 (texture frame 9741), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4320000 (2813616 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 5863466 kept (1937369 by the last picks' textures, a uniform buffer bound since), 26133 new, 2710400 with other textures (of 8600000)
[hos host] vk: mirror of pool 0x7f81400000 (4 MB) given back: idle, 6043 submissions since its sync
[hos host] vk: sparse mirrors: 122, 1232 MB with memory of 1628 MB (395 MB spared); blocks given memory later 6644, given back 8043; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 546 binds (152.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 151 taken from the spares, 162 made new; 190 kept as spares, 0 destroyed; 39 spares now, 996 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 974, 4327.6 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1491 dirty pages from their submits' captures, 53486 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 865 dispatches, 41809 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8244 runs, 686.2 MB compared; 92 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17476 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: replay reads of guest memory as submitted: 4340000 (2823321 pages from the captures)
[hos host] vk: texture #15867 0x10004ce1e00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705)
[hos host] vk: texture #15871 0x10004fb1e00 160x90 x1 NVN 0x3 the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705)
[hos host] vk: texture #15883 0x100023a0e00 1280x720 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705)
[hos host] vk: texture #15890 0x100021aea00 120x67 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705)
[hos host] vk: texture #15889 0x100020d0c00 120x67 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705)
[hos host] vk: texture #16132 0x10009cf3200 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9776 (texture frame 9777), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9778 (texture frame 9779), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9780 (texture frame 9781), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: slow upload 16.3 ms: #16359 0x11142cb200 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB
[hos host] vk: replay reads of guest memory as submitted: 4360000 (2834863 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9782 (texture frame 9783), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9784 (texture frame 9785), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9786 (texture frame 9787), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1100 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1111 kept their virtual core (nothing there to wait for), 2729 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9788 (texture frame 9789), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9790 (texture frame 9791), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9792 (texture frame 9793), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9794 (texture frame 9795), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3704 dirty pages from their submits' captures, 54448 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 884 dispatches, 41290 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8182 runs, 700.3 MB compared; 73 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17023 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9796 (texture frame 9797), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 4380000 (2846523 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 608 ranges, 3413 MB), 241 pages written, 241 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (98 MB), 11 spare buffers, 1182 texture pages copied (4 MB)
[hos Presentation Thread] vi: 9815 frames presented (53.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 2, 15.5-17.8 110, 17.8-22 141, 22-30 12, 30-40 3, over 40 1
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9818 (texture frame 9819), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4400000 (2856617 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9820 (texture frame 9821), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9822 (texture frame 9823), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9824 (texture frame 9825), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: (its written part: 3 of 3 KB readable, 1728 bytes not zero; levels L0)
[hos host] vk: texture 0x111a589c00 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 3 KB of it
[hos host] vk: (its written part: 3 of 3 KB readable, 1728 bytes not zero; levels L0)
[hos host] vk: texture 0x111a589c00 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 3 KB of it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9826 (texture frame 9827), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9828 (texture frame 9829), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: 2 textures retired (1690 left, 3187 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 2991319 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 123, 1232 MB with memory of 1632 MB (399 MB spared); blocks given memory later 6647, given back 8044; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 551 binds (153.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 152 taken from the spares, 162 made new; 191 kept as spares, 0 destroyed; 39 spares now, 996 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2133 dirty pages from their submits' captures, 54838 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 887 dispatches, 43207 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8127 runs, 713.7 MB compared; 89 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17927 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: replay reads of guest memory as submitted: 4420000 (2868569 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 211 textures retired (1478 left, 2862 MB): budget 211, drawn over 0, dead targets 0
[hos host] vk: a frame: 177 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.24, present 0.30, draw 7.52, descriptors 3.80, texture lookups 0.81, uploads 0.83, pipelines 0.27, programs 1.23, commands 1.74, dispatch 0.88, memory sync 0.55, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 314 (1232 MB): 5.72 MB synced in 1075.4 copies, 17.95 MB written back
[hos host] vk: 9840 presents, 51.0 fps; a frame: 1752 draws (0 skipped), 166 dispatches, 169 passes, 131 barriers, 6 submits, 3.3 uploads (3.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12999 pipelines (3209 ms), 0 programs (10 failed), 1478 textures (2862 MB)
[hos host] vk: host time in GPU resource waits: 8.90 ms a frame (fences 1.64, timelines 7.25, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 670.6 full barriers recorded, 286.0 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.1; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.5, an op 1064.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 14.2 ms a frame (the longest submission 9.5 ms); 1.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 1.2 (2.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 3.11 MB unswizzled on the CPU in 0.45 ms
[hos host] vk: late pages' fix (600 recordings with one): of 673 captures with late pages 673 fixed (134.0 pages, 134.9 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 524 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 1 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 29, of an earlier recording 494, not recorded yet 0, of this part 0; pages fixed from stamps 89153, from snapshots 1027; the shadows: 264414 pages synced from captures, 262355 made known by it, 2059 left unknown because the replay or the collection wrote them since
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9850 (texture frame 9851), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker2] nvn: 1050000 recordings made, 74295 live copies (22.8 MB), 5872 texture copies; 6 command buffers' segments (4.3 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 259 sync waits (1546 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: program picks: 6005421 kept (1996383 by the last picks' textures, a uniform buffer bound since), 26166 new, 2768412 with other textures (of 8800000)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9854 (texture frame 9855), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4440000 (2880769 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9856 (texture frame 9857), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9858 (texture frame 9859), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9860 (texture frame 9861), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9865 (texture frame 9866), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9867 (texture frame 9868), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9869 (texture frame 9870), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9871 (texture frame 9872), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4460000 (2892704 pages from the captures)
[hos host] vk: (its written part: 3 of 3 KB readable, 1728 bytes not zero; levels L0)
[hos host] vk: texture 0x111a589c00 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 3 KB of it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9873 (texture frame 9874), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0)
[hos host] vk: texture 0x111b189e00 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it
[hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0)
[hos host] vk: texture 0x111bd8a000 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9875 (texture frame 9876), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1997 dirty pages from their submits' captures, 55258 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 912 dispatches, 43503 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8190 runs, 714.9 MB compared; 92 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17975 in captures, 44 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: recorder: 6128515 packets run in 10.0 s (busy 1.64 s, spun idle 0.12 s, slept 11677 times, 4214 of them under 200 us), 0 calls made directly (no room); 17871 batches over 529 frames (33.8 a frame, 343 packets a batch), 9004 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 61 times, 5.51 ms (0.1 times, 0.01 ms a frame); every thread 637 times, 43.11 ms in all, the longest 0.66 ms; vkWaitForFences 465/29.57 ms (max 0.62), the queue's lock 43/5.29 ms (max 0.57), vkDestroyImageView 29/2.22 ms (max 0.35), vkQueueBindSparse 42/2.18 ms (max 0.33), vmaFreeMemory 11/2.16 ms (max 0.66), vmaDestroyBuffer 40/1.34 ms (max 0.48)
[hos host] vk: recorder: of its busy 1.64 s, 0.36 s in parts' ends and submits (the longest 1.17 ms), 0.07 s of that blocked on the queue's lock (111 times, the longest 1.13 ms), so 0.30 s in the ends, submits and holds; 1.27 s running the other packets; 11585 packets a frame
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 250 textures retired (1237 left, 2512 MB): budget 250, drawn over 0, dead targets 0
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9882 (texture frame 9883), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9884 (texture frame 9885), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9886 (texture frame 9887), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4480000 (2905378 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9888 (texture frame 9889), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 46 textures retired (1191 left, 2441 MB): budget 46, drawn over 0, dead targets 0
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9890 (texture frame 9891), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9892 (texture frame 9893), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #130000: event 0x100d6c802c mode 0 value 88382 (now 88363) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.07, writes collected 0.05
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9897 (texture frame 9898), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 619 ranges, 3448 MB), 306 pages written, 306 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.072; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.052, bytes copied 0.000; 25 kept (112 MB), 11 spare buffers, 14339 texture pages copied (56 MB)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9899 (texture frame 9900), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9901 (texture frame 9902), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos BaseProcExecutor2] vcores: 132444 takes, 21912 waits (mean 74 us, longest 9.1 ms, 0 over HOS_VCORE_LOG_MS), 10361 kicks, 10542 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9903 (texture frame 9904), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9905 (texture frame 9906), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9907 (texture frame 9908), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4500000 (2915517 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 43619 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 123, 1232 MB with memory of 1632 MB (399 MB spared); blocks given memory later 6647, given back 8044; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 551 binds (153.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 152 taken from the spares, 162 made new; 191 kept as spares, 0 destroyed; 39 spares now, 996 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1715 dirty pages from their submits' captures, 55158 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 881 dispatches, 43356 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8264 runs, 716.2 MB compared; 82 pages read into the shadows; 292 MB of shadows (46 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18616 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9916 (texture frame 9917), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: VRAM compaction 4 begins: the textures' blocks hold 2208 MB, 706 MB of it free space
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9918 (texture frame 9919), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0)
[hos host] vk: texture 0x11331b0400 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it
[hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0)
[hos host] vk: texture 0x11321b0200 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9920 (texture frame 9921), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9922 (texture frame 9923), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0)
[hos host] vk: texture 0x111c18a000 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it
[hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0)
[hos host] vk: texture 0x111b189e00 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9924 (texture frame 9925), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4520000 (2925313 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9929 (texture frame 9930), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9931 (texture frame 9932), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9933 (texture frame 9934), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9935 (texture frame 9936), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9937 (texture frame 9938), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9939 (texture frame 9940), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] vk: readbacks of GPU results: 1073 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1419 kept their virtual core (nothing there to wait for), 3189 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 4540000 (2938009 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9948 (texture frame 9949), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9950 (texture frame 9951), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9952 (texture frame 9953), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: program picks: 6149747 kept (2056700 by the last picks' textures, a uniform buffer bound since), 26189 new, 2824063 with other textures (of 9000000)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9954 (texture frame 9955), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2049 dirty pages from their submits' captures, 56810 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 43788 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8259 runs, 715.7 MB compared; 95 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17528 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9956 (texture frame 9957), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4560000 (2950381 pages from the captures)
[hos host] vk: a frame: 206 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.37, present 0.37, draw 7.94, descriptors 3.79, texture lookups 0.71, uploads 0.75, pipelines 0.30, programs 1.40, commands 1.91, dispatch 0.89, memory sync 0.74, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 315 (1233 MB): 5.64 MB synced in 1086.0 copies, 17.90 MB written back
[hos host] vk: 9960 presents, 48.3 fps; a frame: 1917 draws (0 skipped), 162 dispatches, 179 passes, 131 barriers, 6 submits, 1.5 uploads (2.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13003 pipelines (3211 ms), 0 programs (10 failed), 1193 textures (2441 MB)
[hos host] vk: host time in GPU resource waits: 8.82 ms a frame (fences 1.46, timelines 7.36, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 695.5 full barriers recorded, 295.4 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1116.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 15.0 ms a frame (the longest submission 12.6 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 2.80 MB unswizzled on the CPU in 0.38 ms
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9961 (texture frame 9962), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9963 (texture frame 9964), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9965 (texture frame 9966), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0)
[hos host] vk: texture 0x11321b0200 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it
[hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0)
[hos host] vk: texture 0x1132db0400 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9967 (texture frame 9968), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos BaseProcExecutor4] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 13624206 lookups, 115902 missed (0.85 %)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9969 (texture frame 9970), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0)
[hos host] vk: texture 0x111b189e00 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it
[hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0)
[hos host] vk: texture 0x111bd8a000 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9971 (texture frame 9972), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9978 (texture frame 9979), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4580000 (2960759 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9980 (texture frame 9981), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: CommandBufferSignalEvent #90000: event 0x100d6c8034 mode 0 value 89999 (lr 0x806616b0)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9982 (texture frame 9983), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] vk: capture: a submit 0.21 ms (lock 0.00, scan 0.04 over 620 ranges, 3444 MB), 334 pages written, 334 copied (1.3 MB)
[hos host] vk: a capture taken (ms): all 0.068; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.050, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 14576 texture pages copied (56 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9984 (texture frame 9985), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: VRAM compaction 4 over: 368 textures moved (282 MB) in 34 passes, 0 left where they were; 20 blocks (640 MB) given back to the driver, 2848 MB in all so far
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9986 (texture frame 9987), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9993 (texture frame 9994), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9995 (texture frame 9996), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4600000 (2972360 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 125, 1230 MB with memory of 1640 MB (409 MB spared); blocks given memory later 6649, given back 8107; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 555 binds (153.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 154 taken from the spares, 162 made new; 191 kept as spares, 0 destroyed; 37 spares now, 988 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1582 dirty pages from their submits' captures, 55596 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 43459 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8160 runs, 714.9 MB compared; 112 pages read into the shadows; 292 MB of shadows (46 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18099 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos Presentation Thread] nvn: QueueSubmitCommands #60000: queue 000010100D640020 (its #9998), 1 handles (lr 0x805282f4)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.08, writes collected 0.05
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9997 (texture frame 9998), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: EventSignal (CPU) #20000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9999 (texture frame 10000), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10001 (texture frame 10002), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10003 (texture frame 10004), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10008 (texture frame 10009), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10010 (texture frame 10011), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10012 (texture frame 10013), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10014 (texture frame 10015), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4620000 (2982233 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10016 (texture frame 10017), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10018 (texture frame 10019), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10025 (texture frame 10026), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10027 (texture frame 10028), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 1.0 ms, 1 frames after their copies were recorded, the first #16432 0x116d470000 260x260 levels 9 NVN 0xd (28 so far, 6184 KB, 9 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10029 (texture frame 10030), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10031 (texture frame 10032), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4640000 (2994205 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10033 (texture frame 10034), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: readback pages refreshed from guest memory: 3058429 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10035 (texture frame 10036), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1516 dirty pages from their submits' captures, 55840 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 883 dispatches, 43404 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8120 runs, 714.5 MB compared; 91 pages read into the shadows; 292 MB of shadows (46 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17382 in captures, 0 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10042 (texture frame 10043), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10044 (texture frame 10045), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: program picks: 6299013 kept (2122582 by the last picks' textures, a uniform buffer bound since), 26189 new, 2874797 with other textures (of 9200000)
[hos host] vk: mirror of pool 0x7fa1c00000 (4 MB) given back: idle, 6062 submissions since its sync
[hos host] vk: 192 mirrors given back (1612 MB in all); 1227 MB of mirrors now, 160 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10046 (texture frame 10047), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4660000 (3006962 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10048 (texture frame 10049), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10050 (texture frame 10051), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker1] nvn: DrawMeshTasksIndirect: 180000 so far (x1 0x117db40010 x2 0 x3 0x117db40000 x4 0x44b040)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10052 (texture frame 10053), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10057 (texture frame 10058), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos Presentation Thread] vi: 10060 frames presented (48.9 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 23, 17.8-22 196, 22-30 17, 30-40 2, over 40 7
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10059 (texture frame 10060), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10061 (texture frame 10062), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10063 (texture frame 10064), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10065 (texture frame 10066), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4680000 (3017318 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10067 (texture frame 10068), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0)
[hos host] vk: texture 0x11331b0400 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it
[hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0)
[hos host] vk: texture 0x11321b0200 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 623 ranges, 3444 MB), 264 pages written, 264 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.041; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.026, bytes copied 0.000; 24 kept (182 MB), 11 spare buffers, 8291 texture pages copied (32 MB)
[hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0)
[hos host] vk: texture 0x111c18a000 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 90536 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10074 (texture frame 10075), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: sparse mirrors: 125, 1226 MB with memory of 1640 MB (413 MB spared); blocks given memory later 6649, given back 8124; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 559 binds (155.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 155 taken from the spares, 162 made new; 192 kept as spares, 0 destroyed; 37 spares now, 988 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1720 dirty pages from their submits' captures, 54714 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 43164 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8121 runs, 714.5 MB compared; 103 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16926 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10076 (texture frame 10077), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10078 (texture frame 10079), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a frame: 232 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.61, present 0.29, draw 8.06, descriptors 3.90, texture lookups 0.65, uploads 0.74, pipelines 0.30, programs 1.31, commands 1.99, dispatch 0.89, memory sync 0.71, queue submit 0.00, slot wait 0.27
[hos host] vk: a frame: mirrors 317 (1226 MB): 5.58 MB synced in 1081.4 copies, 17.87 MB written back
[hos host] vk: 10080 presents, 49.4 fps; a frame: 2111 draws (0 skipped), 161 dispatches, 182 passes, 131 barriers, 6 submits, 1.1 uploads (2.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13005 pipelines (3212 ms), 0 programs (10 failed), 1205 textures (2444 MB)
[hos host] vk: host time in GPU resource waits: 8.53 ms a frame (fences 1.28, timelines 7.25, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 704.6 full barriers recorded, 298.4 left out (nothing recorded since the one before), 13.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.1, an op 1217.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 14.7 ms a frame (the longest submission 9.6 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 2.54 MB unswizzled on the CPU in 0.37 ms
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10080 (texture frame 10081), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10082 (texture frame 10083), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10084 (texture frame 10085), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4700000 (3026965 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10089 (texture frame 10090), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10091 (texture frame 10092), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10093 (texture frame 10094), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a 192 MB capture buffer made ahead in 74.1 ms, off the game's thread
[hos host] vk: (its written part: 3 of 3 KB readable, 0 bytes not zero; levels L0)
[hos host] vk: texture 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6 (39444 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it
[hos BaseProcExecutor0] vcores: BaseProcExecutor0 (priority 16) waited 126.5 ms for a core: given by sead::MainThread (priority 16), which held it 126.6 ms and let it go 123.1 ms after it was asked; the cores it may use now: core 0: it;
[hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 59.9 ms for a core: given by sead::MainThread (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 0: it;
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10095 (texture frame 10096), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a 128 MB capture buffer made ahead in 40.9 ms, off the game's thread
[hos sead::MainThread] nvn: GPU waits in 5 s: 240 sync waits (1722 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #135000: event 0x100d6c802c mode 0 value 92182 (now 92163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.09
[hos host] vk: mirror of pool 0x7f9f000000 (8 MB) given back: idle, 6009 submissions since its sync
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10097 (texture frame 10098), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] vk: readbacks of GPU results: 1080 of 3000 waited for their write-back, 0.12 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1600 kept their virtual core (nothing there to wait for), 3008 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: mirror of pool 0x7f85800000 (8 MB) given back: idle, 6022 submissions since its sync
[hos host] vk: mirror of pool 0x7f96800000 (4 MB) given back: idle, 6001 submissions since its sync
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10102 (texture frame 10103), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4720000 (3039150 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10104 (texture frame 10105), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: mirror of pool 0x7f84400000 (4 MB) given back: idle, 6051 submissions since its sync
[hos host] vk: mirror of pool 0x7faf400000 (4 MB) given back: idle, 6049 submissions since its sync
[hos host] vk: mirror of pool 0x7fd6400000 (48 MB) given back: idle, 6049 submissions since its sync
[hos host] vk: mirror of pool 0x7f7ec00000 (4 MB) given back: idle, 6048 submissions since its sync
[hos host] vk: mirror of pool 0x7fa0400000 (4 MB) given back: idle, 6048 submissions since its sync
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10106 (texture frame 10107), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10108 (texture frame 10109), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.5 ms, 1 frames after their copies were recorded, the first #16437 0x116d72b000 260x260 levels 9 NVN 0xd (29 so far, 6394 KB, 10 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: texture #14280 0x116c3ac000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 10110 (drawn at 5293)
[hos host] vk: 208 mirrors given back (1687 MB in all); 1153 MB of mirrors now, 160 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10110 (texture frame 10111), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 7802 dirty pages from their submits' captures, 123514 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1084 dispatches, 49057 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8151 runs, 714.8 MB compared; 179 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16950 in captures, 125 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10117 (texture frame 10118), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4740000 (3051287 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10119 (texture frame 10120), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10121 (texture frame 10122), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10123 (texture frame 10124), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos BaseProcExecutor5] vcores: 122136 takes, 22285 waits (mean 89 us, longest 126.5 ms, 2 over HOS_VCORE_LOG_MS), 10814 kicks, 10908 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10125 (texture frame 10126), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10127 (texture frame 10128), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10132 (texture frame 10133), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos rmbl::VibrationThread] hid: 32768 vibration values, strongest 0.00 since the last count
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10134 (texture frame 10135), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker4] nvn: 1100000 recordings made, 66797 live copies (22.6 MB), 5874 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10136 (texture frame 10137), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4760000 (3061605 pages from the captures)
[hos host] vk: program picks: 6447819 kept (2190306 by the last picks' textures, a uniform buffer bound since), 26189 new, 2925991 with other textures (of 9400000)
[hos host] vk: (its written part: 3 of 3 KB readable, 1630 bytes not zero; levels L0)
[hos host] vk: texture 0x11286a9400 3072x2096x1 x1 levels 1 NVN 0x4b (6336 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10138 (texture frame 10139), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10140 (texture frame 10141), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10142 (texture frame 10143), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10151 (texture frame 10152), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10153 (texture frame 10154), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10155 (texture frame 10156), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos Presentation Thread] vk: capture: a submit 0.85 ms (lock 0.08, scan 0.09 over 604 ranges, 3428 MB), 1043 pages written, 1043 copied (4.1 MB)
[hos host] vk: a capture taken (ms): all 0.076; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.047, bytes copied 0.000; 25 kept (188 MB), 11 spare buffers, 3818 texture pages copied (14 MB)
[hos host] vk: sparse mirrors: 107, 1065 MB with memory of 1484 MB (419 MB spared); blocks given memory later 7748, given back 10466; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 600 binds (184.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 155 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 55 spares now, 1144 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 10327 dirty pages from their submits' captures, 66398 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1002 dispatches, 44795 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8136 runs, 714.6 MB compared; 448 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16205 in captures, 118 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: replay reads of guest memory as submitted: 4780000 (3071388 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10157 (texture frame 10158), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10159 (texture frame 10160), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: presenter thread: the copy's fence 2.35 ms, the present 0.83 ms a present
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10164 (texture frame 10165), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10166 (texture frame 10167), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10168 (texture frame 10169), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10170 (texture frame 10171), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10172 (texture frame 10173), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10174 (texture frame 10175), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4800000 (3081823 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10176 (texture frame 10177), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10178 (texture frame 10179), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10180 (texture frame 10181), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10182 (texture frame 10183), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10184 (texture frame 10185), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10186 (texture frame 10187), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4820000 (3094061 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10193 (texture frame 10194), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10195 (texture frame 10196), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1403 dirty pages from their submits' captures, 52279 from earlier ones', 1172 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 853 dispatches, 41256 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8210 runs, 714.5 MB compared; 60 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16717 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 887; 0x7f9ac00000 (20 MB, captured, state 2): 285;
[hos host] vk: reaper: 648 images (1388 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 221 (376 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.07, writes collected 0.18
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10197 (texture frame 10198), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: memory: working set 15561 MB (peak 15827), private 21495 MB, staging 80 MB in 5 arenas, 40 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 844 MB; used 3280 of 15579 MB; heap 1 (device): 159 blocks 4823 MB, 19030 allocations 4239 MB; used 6818 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 100 images 304 MB; 61 free ranges, the largest 24 MB
[hos host] vk: image blocks (textures, compacted): 50 blocks 1600 MB, 1166 images 1540 MB; 267 free ranges, the largest 19 MB
[hos host] vk: render target images: 5719 reused from the pool, 3265 made new, 3163 destroyed from it, 989 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1139 with the pool empty (0 of them with targets still in the dead queue), 286 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 501 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2971 idle
[hos host] vk: VRAM: 6818 MB used of 11474 MB allowed (583 MB of it free in the blocks); textures 2460 MB (budget 7158 MB), 1264 textures
[hos host] vk: budget: VRAM 4823 of 11474 MB (6818 MB by the driver, 1995 MB unaccounted): textures 2460, mirrors 1065, blocks free 583, shadows 292, other buffers 211, other images 111, staging 80, dead images 19; RAM 3208 MB (the process's private memory 21495 MB): imports 2280, capture spares 672, captures 152, retired imports 52, capture table 32, other buffers 20, others 0
[hos host] vk: textures: decoded ASTC 1391 MB, rendered 527 MB (94 drawn in one frame, kept: 42654 KB), other 508 MB; used in the last 8 presents 1692 MB
[hos host] vk: decoded ASTC used again: 66 textures (184 MB) copied from ones the game finalized, not decoded; 66 of those retired, 0 listed; 1154 MB of blocks hashed in 198 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 5053 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x111b189e00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 22 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 18 (7500 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 4 (703 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); 1466x282x1 NVN 0x25: 4 (6459 KB);
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 22376 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos Presentation Thread] vk: presents after the CPU's signal (600): 590 waited for it, 6.84 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10199 (texture frame 10200), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: status: 600 frames, the slowest 146.0 ms; VRAM 6.7 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6151 queued, 6151 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15240 ms and 330.9 MB read in all); 8346 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11699 graphics pipelines (1 not in a cache), 711 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2147 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 8985 (6151 ready, 0 loading), 16724 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13162 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 180 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 14.10, present 0.18, draw 7.86, descriptors 3.46, texture lookups 0.21, uploads 0.32, pipelines 0.32, programs 1.37, commands 2.12, dispatch 1.01, memory sync 1.51, queue submit 0.00, slot wait 0.27
[hos host] vk: a frame: mirrors 317 (1065 MB): 8.55 MB synced in 1078.2 copies, 17.87 MB written back
[hos host] vk: 10200 presents, 42.7 fps; a frame: 2042 draws (0 skipped), 162 dispatches, 177 passes, 131 barriers, 6 submits, 1.3 uploads (1.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13553 pipelines (3509 ms), 0 programs (10 failed), 1264 textures (2460 MB)
[hos host] vk: host time in GPU resource waits: 9.27 ms a frame (fences 2.34, timelines 6.93, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 691.3 full barriers recorded, 294.7 left out (nothing recorded since the one before), 13.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.4, an op 1217.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 13.8 ms a frame (the longest submission 15.3 ms); 0.8 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.8 (0.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.47 MB unswizzled on the CPU in 0.25 ms
[hos host] vk: the render thread (600 presents, 20.52 ms a frame): its jobs 19.46 ms a frame (submits 1.69, presents 6.76, others 11.00), idle 1.07; the game waited for room at 36 presents, 0.60 ms each (0.04 a frame; 14.40 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 591 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 2985 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 18.66 ms after its hand-over (55.41 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.99 + 0.28 (the window's copy is in the idle), after it 2.03 + 0.63; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.62 parts a frame recorded in the signal's job, 70.5% of them took nothing that came with the signal; of the others, 2.54 a frame took pages of the signal's capture into a mirror (80.5 pages a frame) and 0.00 pages read as guest memory had them (2.4 pages a frame); a frame's first part took nothing in 0 of 589 frames, its first three in 0
[hos host] vk: frame ahead, what came with the signal: 590 signals with a capture, each of 367.2 pages in 12.8 pools; between a frame's last submit and that signal 0.54 texture and sampler registrations, 0.64 releases and 0.00 other CPU signals a signal (55 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.40+0.11, 0.17+0.12, 0.05+0.03, 0.10+0.03, 0.95+0.11, 0.12+0.18, 0.10+0.03, 0.10+0.02, 0.01+0.01, 0.01+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10201 (texture frame 10202), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10203 (texture frame 10204), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10208 (texture frame 10209), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: render targets drawn over from another start, let go: 9765 so far (46318 MB), 8750 of them (41596 MB) between the passes
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10210 (texture frame 10211), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4840000 (3103953 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10212 (texture frame 10213), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: late pages found at the replay (600 takes): 449 pages in 5 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10217 (texture frame 10218), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10219 (texture frame 10220), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10221 (texture frame 10222), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10226 (texture frame 10227), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10228 (texture frame 10229), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10230 (texture frame 10231), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4860000 (3112665 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 3125944 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: program picks: 6596448 kept (2259090 by the last picks' textures, a uniform buffer bound since), 26189 new, 2977362 with other textures (of 9600000)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10235 (texture frame 10236), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: sparse mirrors: 107, 1065 MB with memory of 1484 MB (419 MB spared); blocks given memory later 7748, given back 10466; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 600 binds (184.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 155 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 55 spares now, 1144 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 461 dirty pages from their submits' captures, 55761 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 43219 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8438 runs, 732.5 MB compared; 52 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17750 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10242 (texture frame 10243), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10244 (texture frame 10245), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: CommandBufferSignalEvent #95000: event 0x100d6c8040 mode 0 value 94999 (lr 0x806621e8)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10246 (texture frame 10247), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 606 ranges, 3438 MB), 244 pages written, 244 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (82 MB), 11 spare buffers, 4231 texture pages copied (16 MB)
[hos host] vk: replay reads of guest memory as submitted: 4880000 (3121599 pages from the captures)
[hos sead::MainThread] vk: readbacks of GPU results: 1042 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1476 kept their virtual core (nothing there to wait for), 2876 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10260 (texture frame 10261), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10264 (texture frame 10265), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10266 (texture frame 10267), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4900000 (3130480 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10278 (texture frame 10279), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 244 dirty pages from their submits' captures, 56958 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 893 dispatches, 44246 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8880 runs, 768.9 MB compared; 45 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17826 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10280 (texture frame 10281), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10282 (texture frame 10283), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10294 (texture frame 10295), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10296 (texture frame 10297), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: CommandBufferWaitEvent #140000: event 0x100d6c802c mode 0 value 95982 (now 95963) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 4920000 (3139328 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10298 (texture frame 10299), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10300 (texture frame 10301), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5568 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos Presentation Thread] vi: 10305 frames presented (48.9 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 19, 15.5-17.8 122, 17.8-22 58, 22-30 34, 30-40 6, over 40 6
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 4940000 (3149711 pages from the captures)
[hos host] vk: a frame: 180 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.84, present 0.16, draw 6.30, descriptors 2.71, texture lookups 0.10, uploads 0.14, pipelines 0.26, programs 1.12, commands 1.71, dispatch 0.85, memory sync 0.42, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 318 (1065 MB): 5.29 MB synced in 1042.7 copies, 17.88 MB written back
[hos host] vk: 10320 presents, 60.4 fps; a frame: 1878 draws (0 skipped), 164 dispatches, 158 passes, 131 barriers, 6 submits, 0.5 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13553 pipelines (3509 ms), 0 programs (10 failed), 1282 textures (2462 MB)
[hos host] vk: host time in GPU resource waits: 1.82 ms a frame (fences 1.39, timelines 0.43, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 610.4 full barriers recorded, 273.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 1140.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.9 ms a frame (the longest submission 4.4 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.76 MB unswizzled on the CPU in 0.12 ms
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10320 (texture frame 10321), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: sparse mirrors: 108, 1065 MB with memory of 1488 MB (422 MB spared); blocks given memory later 7748, given back 10466; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 601 binds (185.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 156 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 54 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 261 dirty pages from their submits' captures, 55661 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 884 dispatches, 43008 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8527 runs, 733.6 MB compared; 47 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17427 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10322 (texture frame 10323), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10324 (texture frame 10325), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10326 (texture frame 10327), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4960000 (3161977 pages from the captures)
[hos host] vk: program picks: 6745057 kept (2326487 by the last picks' textures, a uniform buffer bound since), 26189 new, 3028753 with other textures (of 9800000)
[hos Presentation Thread] vk: capture: a submit 0.29 ms (lock 0.02, scan 0.05 over 596 ranges, 3440 MB), 658 pages written, 485 copied (1.9 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (282 MB), 12 spare buffers, 4627 texture pages copied (18 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10342 (texture frame 10343), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos OptimizeChangeBanc] nvn: 1150000 recordings made, 68335 live copies (24.1 MB), 5874 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10344 (texture frame 10345), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10346 (texture frame 10347), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10348 (texture frame 10349), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10350 (texture frame 10351), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 4980000 (3170716 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10352 (texture frame 10353), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: GPU waits in 5 s: 137 sync waits (724 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 16245 dirty pages from their submits' captures, 108939 from earlier ones', 40350 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1094 dispatches, 47348 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8529 runs, 733.2 MB compared; 260 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16632 in captures, 90 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool: 0x7fab000000 (68 MB, captured, state 2): 16287; 0x7fb5c00000 (44 MB, captured, state 2): 10659; 0x7fb8800000 (52 MB, captured, state 2): 8665; 0x7fd3c00000 (16 MB, captured, state 2): 3252; 0x1145400000 (40 MB, captured, state 2): 533; 0x104e000000 (24 MB, captured, state 2): 226; 0x103cc00000 (12 MB, captured, state 2): 198; 0x1038400000 (4 MB, captured, state 2): 72;
[hos host] vk: recorder: 5246293 packets run in 10.0 s (busy 1.53 s, spun idle 0.11 s, slept 10088 times, 3308 of them under 200 us), 0 calls made directly (no room); 15467 batches over 494 frames (31.3 a frame, 339 packets a batch), 7332 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 124 times, 8.49 ms (0.3 times, 0.02 ms a frame); every thread 662 times, 42.83 ms in all, the longest 2.42 ms; vkWaitForFences 422/25.22 ms (max 2.42), the queue's lock 46/5.37 ms (max 0.80), vkDestroyImageView 45/3.85 ms (max 0.84), vmaDestroyBuffer 56/3.21 ms (max 0.72), vkQueueBindSparse 72/2.33 ms (max 0.37), vmaFreeMemory 15/2.10 ms (max 0.45)
[hos host] vk: recorder: of its busy 1.53 s, 0.39 s in parts' ends and submits (the longest 1.48 ms), 0.12 s of that blocked on the queue's lock (199 times, the longest 1.44 ms), so 0.27 s in the ends, submits and holds; 1.14 s running the other packets; 10620 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 5000000 (3179749 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10374 (texture frame 10375), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a submission's parts ended: 33785 where the game had ended its render pass, 6215 at 512 draws
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10376 (texture frame 10377), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10378 (texture frame 10379), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10380 (texture frame 10381), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10382 (texture frame 10383), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10384 (texture frame 10385), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5020000 (3188818 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.04, writes collected 0.19
[hos nn.npln.Worker] vcores: 150978 takes, 30278 waits (mean 67 us, longest 16.3 ms, 0 over HOS_VCORE_LOG_MS), 15559 kicks, 15572 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 108, 1135 MB with memory of 1488 MB (352 MB spared); blocks given memory later 8950, given back 10537; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 630 binds (188.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 156 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 54 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1778 dirty pages from their submits' captures, 52602 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 847 dispatches, 41823 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8251 runs, 714.2 MB compared; 71 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16766 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10406 (texture frame 10407), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5040000 (3199876 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10408 (texture frame 10409), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10410 (texture frame 10411), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10412 (texture frame 10413), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1026 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 5 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1659 kept their virtual core (nothing there to wait for), 3205 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10414 (texture frame 10415), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10416 (texture frame 10417), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5060000 (3208884 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.22 ms (lock 0.03, scan 0.05 over 630 ranges, 3538 MB), 350 pages written, 350 copied (1.4 MB)
[hos host] vk: a capture taken (ms): all 0.032; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (130 MB), 11 spare buffers, 4627 texture pages copied (18 MB)
[hos host] vk: program picks: 6892419 kept (2392272 by the last picks' textures, a uniform buffer bound since), 26190 new, 3081390 with other textures (of 10000000)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15662 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10438 (texture frame 10439), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: readback pages refreshed from guest memory: 3193053 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: a frame: 196 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.21, present 0.16, draw 7.12, descriptors 3.02, texture lookups 0.07, uploads 0.12, pipelines 0.30, programs 1.28, commands 1.97, dispatch 0.94, memory sync 1.21, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 318 (1135 MB): 8.88 MB synced in 1056.6 copies, 17.87 MB written back
[hos host] vk: 10440 presents, 52.6 fps; a frame: 1965 draws (0 skipped), 163 dispatches, 170 passes, 131 barriers, 6 submits, 0.2 uploads (0.7 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13553 pipelines (3509 ms), 0 programs (10 failed), 1282 textures (2462 MB)
[hos host] vk: host time in GPU resource waits: 6.76 ms a frame (fences 1.93, timelines 4.83, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 660.5 full barriers recorded, 287.2 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1173.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 13.0 ms a frame (the longest submission 43.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.67 MB unswizzled on the CPU in 0.12 ms
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10440 (texture frame 10441), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1464 dirty pages from their submits' captures, 52649 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 845 dispatches, 41747 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8254 runs, 714.2 MB compared; 71 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16621 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10442 (texture frame 10443), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 200000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10444 (texture frame 10445), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10446 (texture frame 10447), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5080000 (3217827 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10448 (texture frame 10449), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: queues ordered by their events: 60000 submissions stopped at a wait, 59980 went on once it was reached, 18 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 2 (stopped 50.00 ms at most); waits met 143922, not met 0; captures at a CPU event's wait 7043, 0 of them at a wait met on arrival
[hos host] vk: late pages found at the replay (600 takes): 15 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 5100000 (3226850 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10468 (texture frame 10469), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10470 (texture frame 10471), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10472 (texture frame 10473), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10474 (texture frame 10475), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10476 (texture frame 10477), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker5] nvn: 1200000 recordings made, 74180 live copies (23.6 MB), 5875 texture copies; 7 command buffers' segments (4.4 MB)
[hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.5 ms, 1 frames after their copies were recorded, the first #16568 0x116dad4000 260x260 levels 9 NVN 0xd (30 so far, 6603 KB, 10 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10478 (texture frame 10479), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: sparse mirrors: 109, 1137 MB with memory of 1492 MB (354 MB spared); blocks given memory later 8950, given back 10537; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 631 binds (188.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 157 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 53 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1407 dirty pages from their submits' captures, 53531 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 861 dispatches, 41847 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8273 runs, 715.0 MB compared; 69 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16850 in captures, 1 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 182 live (2284 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos ModuleSystemWorker5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 21838561 lookups, 145583 missed (0.67 %)
[hos host] vk: replay reads of guest memory as submitted: 5120000 (3236169 pages from the captures)
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #145000: event 0x100d6c802c mode 0 value 99782 (now 99763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10500 (texture frame 10501), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10502 (texture frame 10503), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10504 (texture frame 10505), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5140000 (3245271 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10506 (texture frame 10507), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #100000: event 0x100d6c8000 mode 0 value 99999 (lr 0x80291440)
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 623 ranges, 3545 MB), 233 pages written, 233 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (118 MB), 10 spare buffers, 5326 texture pages copied (20 MB)
[hos host] vk: replay reads of guest memory as submitted: 5160000 (3257834 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10522 (texture frame 10523), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1055 dirty pages from their submits' captures, 55587 from earlier ones', 8 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 898 dispatches, 43358 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8589 runs, 735.8 MB compared; 60 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17549 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9ac00000 (20 MB, captured, state 2): 8;
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10524 (texture frame 10525), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10526 (texture frame 10527), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: program picks: 7040380 kept (2458709 by the last picks' textures, a uniform buffer bound since), 26192 new, 3133427 with other textures (of 10200000)
[hos host] vk: replay reads of guest memory as submitted: 5180000 (3268601 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10540 (texture frame 10541), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10542 (texture frame 10543), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10544 (texture frame 10545), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5200000 (3278373 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10558 (texture frame 10559), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a frame: 184 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.61, present 0.15, draw 6.86, descriptors 2.96, texture lookups 0.08, uploads 0.12, pipelines 0.29, programs 1.22, commands 1.85, dispatch 0.88, memory sync 0.46, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 320 (1142 MB): 5.44 MB synced in 1052.5 copies, 17.94 MB written back
[hos host] vk: 10560 presents, 60.9 fps; a frame: 1953 draws (0 skipped), 167 dispatches, 169 passes, 131 barriers, 6 submits, 0.3 uploads (0.7 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13553 pipelines (3509 ms), 0 programs (10 failed), 1296 textures (2474 MB)
[hos host] vk: host time in GPU resource waits: 3.14 ms a frame (fences 1.34, timelines 1.80, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 658.6 full barriers recorded, 285.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1170.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.5 ms a frame (the longest submission 4.7 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.66 MB unswizzled on the CPU in 0.10 ms
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10560 (texture frame 10561), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10562 (texture frame 10563), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: sparse mirrors: 110, 1142 MB with memory of 1500 MB (357 MB spared); blocks given memory later 8950, given back 10543; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 634 binds (189.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 52 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 160 dirty pages from their submits' captures, 57765 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 929 dispatches, 44582 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8872 runs, 755.9 MB compared; 52 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18037 in captures, 1 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool: 0x7f8a800000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 182 live (2284 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 8055 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1112 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1495 kept their virtual core (nothing there to wait for), 2857 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10574 (texture frame 10575), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10576 (texture frame 10577), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5220000 (3288336 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10578 (texture frame 10579), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10580 (texture frame 10581), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos BaseProcExecutor0] vcores: BaseProcExecutor0 (priority 16) waited 80.5 ms for a core: given by sead::MainThread (priority 16), which held it 80.6 ms and let it go 77.6 ms after it was asked; the cores it may use now: core 0: it;
[hos Presentation Thread] vi: 10583 frames presented (54.7 fps over the last 5.1 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 44, 15.5-17.8 175, 17.8-22 33, 22-30 19, 30-40 3, over 40 4
[hos host] vk: late pages found at the replay (600 takes): 35 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 5240000 (3300109 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.14
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10598 (texture frame 10599), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10600 (texture frame 10601), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10602 (texture frame 10603), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10604 (texture frame 10605), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 4664 dirty pages from their submits' captures, 135697 from earlier ones', 6 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1104 dispatches, 49783 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8709 runs, 739.0 MB compared; 774 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16630 in captures, 534 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9ac00000 (20 MB, captured, state 2): 6;
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10606 (texture frame 10607), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos Presentation Thread] vk: capture: a submit 0.44 ms (lock 0.02, scan 0.06 over 598 ranges, 3461 MB), 652 pages written, 652 copied (2.5 MB)
[hos host] vk: a capture taken (ms): all 0.033; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 24 kept (90 MB), 12 spare buffers, 6126 texture pages copied (23 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10608 (texture frame 10609), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5260000 (3311953 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 5280000 (3324974 pages from the captures)
[hos host] vk: program picks: 7189455 kept (2526672 by the last picks' textures, a uniform buffer bound since), 26201 new, 3184343 with other textures (of 10400000)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10628 (texture frame 10629), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10630 (texture frame 10631), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10632 (texture frame 10633), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10634 (texture frame 10635), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: GPU waits in 5 s: 208 sync waits (879 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10636 (texture frame 10637), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5300000 (3335137 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 3260610 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 110, 1070 MB with memory of 1500 MB (429 MB spared); blocks given memory later 9032, given back 11766; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 645 binds (191.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 52 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1562 dirty pages from their submits' captures, 55625 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 915 dispatches, 43658 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8552 runs, 721.6 MB compared; 72 pages read into the shadows; 292 MB of shadows (47 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17829 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 182 live (2284 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10652 (texture frame 10653), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10654 (texture frame 10655), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10656 (texture frame 10657), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: 1250000 recordings made, 69744 live copies (21.4 MB), 5875 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: replay reads of guest memory as submitted: 5320000 (3344520 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10670 (texture frame 10671), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10672 (texture frame 10673), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10674 (texture frame 10675), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a frame: 183 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.83, present 0.16, draw 6.77, descriptors 2.93, texture lookups 0.10, uploads 0.15, pipelines 0.28, programs 1.21, commands 1.84, dispatch 0.92, memory sync 0.99, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 320 (1070 MB): 8.21 MB synced in 1073.5 copies, 18.04 MB written back
[hos host] vk: 10680 presents, 51.8 fps; a frame: 1959 draws (0 skipped), 173 dispatches, 168 passes, 131 barriers, 6 submits, 0.4 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13554 pipelines (3509 ms), 0 programs (10 failed), 1322 textures (2516 MB)
[hos host] vk: host time in GPU resource waits: 5.12 ms a frame (fences 2.24, timelines 2.88, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 659.6 full barriers recorded, 283.9 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.3, an op 1165.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.1 ms a frame (the longest submission 16.4 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.70 MB unswizzled on the CPU in 0.12 ms
[hos AudioFrameworkThread] vcores: 148870 takes, 24741 waits (mean 75 us, longest 80.5 ms, 1 over HOS_VCORE_LOG_MS), 12169 kicks, 12197 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 5340000 (3355741 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10686 (texture frame 10687), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10688 (texture frame 10689), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 118 dirty pages from their submits' captures, 60247 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 948 dispatches, 44887 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9198 runs, 776.8 MB compared; 70 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18292 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10690 (texture frame 10691), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12723 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10692 (texture frame 10693), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10694 (texture frame 10695), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #150000: event 0x100d6c802c mode 0 value 103582 (now 103563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 595 ranges, 3471 MB), 254 pages written, 254 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 6126 texture pages copied (23 MB)
[hos host] vk: replay reads of guest memory as submitted: 5360000 (3365347 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 345 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10708 (texture frame 10709), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10710 (texture frame 10711), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10712 (texture frame 10713), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5380000 (3376705 pages from the captures)
[hos host] vk: program picks: 7341009 kept (2598217 by the last picks' textures, a uniform buffer bound since), 26203 new, 3232787 with other textures (of 10600000)
[hos host] vk: sparse mirrors: 109, 1059 MB with memory of 1492 MB (432 MB spared); blocks given memory later 9032, given back 11888; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 648 binds (191.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 211 kept as spares, 0 destroyed; 53 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 852 dirty pages from their submits' captures, 56060 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 908 dispatches, 43573 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8884 runs, 742.3 MB compared; 62 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17774 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 182 live (2284 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10734 (texture frame 10735), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1067 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1606 kept their virtual core (nothing there to wait for), 3002 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10736 (texture frame 10737), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5400000 (3387557 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10750 (texture frame 10751), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10752 (texture frame 10753), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10754 (texture frame 10755), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5420000 (3396904 pages from the captures)
[hos host] vk: presenter thread: the copy's fence 1.64 ms, the present 0.76 ms a present
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10766 (texture frame 10767), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10768 (texture frame 10769), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10770 (texture frame 10771), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #105000: event 0x100d6c8014 mode 0 value 105001 (lr 0x80660c94)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 454 dirty pages from their submits' captures, 57406 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 925 dispatches, 44504 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9079 runs, 760.1 MB compared; 46 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18312 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10772 (texture frame 10773), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5440000 (3406288 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10788 (texture frame 10789), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.6 ms, 1 frames after their copies were recorded, the first #16605 0x116dd8f000 260x260 levels 9 NVN 0xd (31 so far, 6813 KB, 11 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10790 (texture frame 10791), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10792 (texture frame 10793), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.05 over 596 ranges, 3471 MB), 293 pages written, 293 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (176 MB), 11 spare buffers, 7161 texture pages copied (27 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10794 (texture frame 10795), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5460000 (3417732 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: memory: working set 15275 MB (peak 15952), private 21269 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 896 MB, 36 allocations 822 MB; used 3232 of 15579 MB; heap 1 (device): 159 blocks 4823 MB, 19000 allocations 4256 MB; used 6818 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 102 images 304 MB; 61 free ranges, the largest 24 MB
[hos host] vk: image blocks (textures, compacted): 50 blocks 1600 MB, 1220 images 1577 MB; 271 free ranges, the largest 4 MB
[hos host] vk: render target images: 5719 reused from the pool, 3267 made new, 3163 destroyed from it, 989 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1141 with the pool empty (0 of them with targets still in the dead queue), 286 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 501 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2971 idle
[hos host] vk: VRAM: 6818 MB used of 11474 MB allowed (566 MB of it free in the blocks); textures 2518 MB (budget 7158 MB), 1328 textures
[hos host] vk: budget: VRAM 4823 of 11474 MB (6818 MB by the driver, 1995 MB unaccounted): textures 2518, mirrors 1059, blocks free 566, shadows 292, other buffers 211, other images 111, staging 64; RAM 3180 MB (the process's private memory 21269 MB): imports 2284, capture spares 722, captures 110, retired imports 32, capture table 32, others 0
[hos host] vk: textures: decoded ASTC 1439 MB, rendered 527 MB (95 drawn in one frame, kept: 42890 KB), other 518 MB; used in the last 8 presents 1696 MB
[hos host] vk: decoded ASTC used again: 79 textures (195 MB) copied from ones the game finalized, not decoded; 79 of those retired, 0 listed; 1165 MB of blocks hashed in 201 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 5653 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x111bd8a000 4096x4096x1 x1 levels 1 NVN 0x4b, last used 733 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 18 (7500 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 4 (703 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); 1466x282x1 NVN 0x25: 4 (6459 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 325 waited for it, 3.92 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 124.9 ms; VRAM 6.7 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6153 queued, 6153 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15243 ms and 331.0 MB read in all); 8348 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11700 graphics pipelines (1 not in a cache), 712 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2158 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 8987 (6153 ready, 0 loading), 16669 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13164 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 760 known, 3 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 178 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.41, present 0.16, draw 6.80, descriptors 2.94, texture lookups 0.07, uploads 0.10, pipelines 0.29, programs 1.20, commands 1.85, dispatch 0.91, memory sync 0.48, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 320 (1059 MB): 5.52 MB synced in 1060.0 copies, 18.11 MB written back
[hos host] vk: 10800 presents, 60.0 fps; a frame: 1991 draws (0 skipped), 176 dispatches, 167 passes, 131 barriers, 6 submits, 0.2 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13554 pipelines (3509 ms), 0 programs (10 failed), 1328 textures (2518 MB)
[hos host] vk: host time in GPU resource waits: 2.13 ms a frame (fences 1.51, timelines 0.62, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 654.8 full barriers recorded, 282.3 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.3, an op 1168.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.4 ms a frame (the longest submission 18.6 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.63 MB unswizzled on the CPU in 0.10 ms
[hos host] vk: queues ordered by their events: 62591 submissions stopped at a wait, 62559 went on once it was reached, 32 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 3 (stopped 50.30 ms at most); waits met 152524, not met 0; captures at a CPU event's wait 7218, 0 of them at a wait met on arrival
[hos host] vk: the render thread (600 presents, 17.59 ms a frame): its jobs 16.01 ms a frame (submits 5.57, presents 4.90, others 5.54), idle 1.58; the game waited for room at 19 presents, 0.92 ms each (0.03 a frame; 7.80 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 327 captures at a CPU signal; the submits' waits read in 0.044 ms a frame; readbacks: 1717 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.19 ms after its hand-over (48.10 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.79 + 0.33 (the window's copy is in the idle), after it 1.30 + 0.38; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.01 parts a frame recorded in the signal's job, 73.6% of them took nothing that came with the signal; of the others, 2.11 a frame took pages of the signal's capture into a mirror (91.8 pages a frame) and 0.03 pages read as guest memory had them (123.4 pages a frame); a frame's first part took nothing in 22 of 327 frames, its first three in 21
[hos host] vk: frame ahead, what came with the signal: 328 signals with a capture, each of 509.0 pages in 12.2 pools; between a frame's last submit and that signal 0.09 texture and sampler registrations, 0.01 releases and 0.00 other CPU signals a signal (14 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.26+0.11, 0.30+0.13, 0.05+0.02, 0.08+0.02, 0.45+0.05, 0.05+0.03, 0.09+0.02, 0.02+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9554 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10806 (texture frame 10807), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10808 (texture frame 10809), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: render targets drawn over from another start, let go: 9765 so far (46318 MB), 8750 of them (41596 MB) between the passes
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10810 (texture frame 10811), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5480000 (3430038 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 109, 1059 MB with memory of 1492 MB (432 MB spared); blocks given memory later 9032, given back 11890; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 649 binds (191.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 211 kept as spares, 0 destroyed; 53 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 312 dirty pages from their submits' captures, 61374 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 939 dispatches, 44958 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9116 runs, 761.2 MB compared; 57 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18517 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 182 live (2284 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 7493383 kept (2672394 by the last picks' textures, a uniform buffer bound since), 26208 new, 3280408 with other textures (of 10800000)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10824 (texture frame 10825), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10826 (texture frame 10827), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10828 (texture frame 10829), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10830 (texture frame 10831), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5500000 (3440862 pages from the captures)
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 220000 so far (x1 0x117a4f2040 x2 0x6001d40 x3 0x117a4f2000 x4 0xf33080)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10842 (texture frame 10843), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10844 (texture frame 10845), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10846 (texture frame 10847), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5520000 (3449941 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 3327894 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 187 dirty pages from their submits' captures, 57793 from earlier ones', 12 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 948 dispatches, 44415 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9058 runs, 759.7 MB compared; 66 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18165 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7f83c00000 (4 MB, captured, state 2): 12;
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10860 (texture frame 10861), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10862 (texture frame 10863), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10864 (texture frame 10865), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5540000 (3461658 pages from the captures)
[hos Presentation Thread] vi: 10876 frames presented (58.5 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 35, 15.5-17.8 231, 17.8-22 16, 22-30 6, 30-40 4, over 40 1
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10878 (texture frame 10879), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10880 (texture frame 10881), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5560000 (3472007 pages from the captures)
[hos host] vk: reaper: 29 images (48 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 8 (19 MB)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 589 ranges, 3175 MB), 275 pages written, 275 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 7513 texture pages copied (29 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10896 (texture frame 10897), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #155000: event 0x100d6c802c mode 0 value 107382 (now 107363) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: sparse mirrors: 109, 1059 MB with memory of 1492 MB (432 MB spared); blocks given memory later 9034, given back 11899; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 652 binds (192.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 211 kept as spares, 0 destroyed; 53 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 143 dirty pages from their submits' captures, 57900 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 941 dispatches, 44420 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9046 runs, 759.5 MB compared; 56 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18235 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 182 live (2284 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10898 (texture frame 10899), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1155 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1658 kept their virtual core (nothing there to wait for), 2950 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10900 (texture frame 10901), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10902 (texture frame 10903), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10904 (texture frame 10905), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5580000 (3483571 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10906 (texture frame 10907), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11560 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 7645334 kept (2746963 by the last picks' textures, a uniform buffer bound since), 26208 new, 3328457 with other textures (of 11000000)
[hos host] vk: a frame: 173 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.02, present 0.16, draw 6.33, descriptors 2.76, texture lookups 0.07, uploads 0.09, pipelines 0.27, programs 1.09, commands 1.73, dispatch 0.88, memory sync 0.49, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 320 (1059 MB): 5.42 MB synced in 1059.0 copies, 18.09 MB written back
[hos host] vk: 10920 presents, 60.0 fps; a frame: 1968 draws (0 skipped), 175 dispatches, 162 passes, 131 barriers, 6 submits, 0.2 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13554 pipelines (3509 ms), 0 programs (10 failed), 1328 textures (2518 MB)
[hos host] vk: host time in GPU resource waits: 1.42 ms a frame (fences 1.39, timelines 0.02, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 634.1 full barriers recorded, 277.1 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1118.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.0 ms a frame (the longest submission 4.3 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.62 MB unswizzled on the CPU in 0.09 ms
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 5600000 (3495708 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10922 (texture frame 10923), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10924 (texture frame 10925), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10926 (texture frame 10927), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10928 (texture frame 10929), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10935 (texture frame 10936), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10937 (texture frame 10938), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 192 dirty pages from their submits' captures, 56914 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 920 dispatches, 43682 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8878 runs, 740.6 MB compared; 51 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17668 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10939 (texture frame 10940), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5620000 (3508225 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10941 (texture frame 10942), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: GPU waits in 5 s: 45 sync waits (92 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10946 (texture frame 10947), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10948 (texture frame 10949), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: recorder: 5972872 packets run in 10.0 s (busy 1.72 s, spun idle 0.11 s, slept 10383 times, 3466 of them under 200 us), 0 calls made directly (no room); 17526 batches over 580 frames (30.2 a frame, 341 packets a batch), 8435 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 41 times, 3.25 ms (0.1 times, 0.01 ms a frame); every thread 602 times, 44.73 ms in all, the longest 0.76 ms; vkWaitForFences 502/34.97 ms (max 0.59), vmaDestroyBuffer 41/4.42 ms (max 0.76), vmaFreeMemory 14/2.60 ms (max 0.67), the queue's lock 19/1.64 ms (max 0.22), vkQueueBindSparse 23/0.60 ms (max 0.16), a part's submit 1/0.39 ms (max 0.39)
[hos host] vk: recorder: of its busy 1.72 s, 0.43 s in parts' ends and submits (the longest 1.17 ms), 0.12 s of that blocked on the queue's lock (259 times, the longest 1.12 ms), so 0.31 s in the ends, submits and holds; 1.29 s running the other packets; 10298 packets a frame
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10950 (texture frame 10951), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10952 (texture frame 10953), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10957 (texture frame 10958), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5640000 (3520786 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10959 (texture frame 10960), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10961 (texture frame 10962), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10964 (texture frame 10965), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10966 (texture frame 10967), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10968 (texture frame 10969), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10975 (texture frame 10976), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10977 (texture frame 10978), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5660000 (3532521 pages from the captures)
[hos host] vk: sparse mirrors: 111, 1060 MB with memory of 1500 MB (439 MB spared); blocks given memory later 9034, given back 11900; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 655 binds (193.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 160 taken from the spares, 162 made new; 211 kept as spares, 0 destroyed; 51 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 860 dirty pages from their submits' captures, 60135 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 927 dispatches, 44836 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9092 runs, 757.5 MB compared; 57 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18009 in captures, 3 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7f83c00000 (4 MB, captured, state 2): 2; 0x7fe2800000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 183 live (2288 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos BaseProcExecutor3] vcores: 152285 takes, 23508 waits (mean 68 us, longest 6.5 ms, 0 over HOS_VCORE_LOG_MS), 11165 kicks, 10966 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10984 (texture frame 10985), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10986 (texture frame 10987), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.04 over 587 ranges, 3167 MB), 299 pages written, 299 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (134 MB), 11 spare buffers, 7675 texture pages copied (29 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10988 (texture frame 10989), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10993 (texture frame 10994), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10995 (texture frame 10996), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10997 (texture frame 10998), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5680000 (3544826 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11002 (texture frame 11003), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11004 (texture frame 11005), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11006 (texture frame 11007), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11011 (texture frame 11012), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11013 (texture frame 11014), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11015 (texture frame 11016), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5700000 (3556983 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 23394 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11017 (texture frame 11018), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11019 (texture frame 11020), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: program picks: 7795370 kept (2818288 by the last picks' textures, a uniform buffer bound since), 26212 new, 3378417 with other textures (of 11200000)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11021 (texture frame 11022), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 425 dirty pages from their submits' captures, 56320 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 916 dispatches, 43524 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8656 runs, 736.9 MB compared; 54 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17700 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 1;
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11023 (texture frame 11024), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11025 (texture frame 11026), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11027 (texture frame 11028), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 5720000 (3569823 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11032 (texture frame 11033), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #110000: event 0x100d6c801c mode 0 value 110000 (lr 0x802915c8)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11034 (texture frame 11035), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.6 ms, 1 frames after their copies were recorded, the first #16647 0x116e006000 260x260 levels 9 NVN 0xd (32 so far, 7022 KB, 11 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11036 (texture frame 11037), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11038 (texture frame 11039), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a frame: 159 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.23, present 0.16, draw 6.51, descriptors 2.91, texture lookups 0.13, uploads 0.20, pipelines 0.26, programs 1.11, commands 1.73, dispatch 0.91, memory sync 0.55, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 323 (1063 MB): 5.55 MB synced in 1060.1 copies, 17.98 MB written back
[hos host] vk: 11040 presents, 59.8 fps; a frame: 1816 draws (0 skipped), 170 dispatches, 170 passes, 132 barriers, 6 submits, 0.6 uploads (1.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1364 textures (2555 MB)
[hos host] vk: host time in GPU resource waits: 1.92 ms a frame (fences 1.24, timelines 0.67, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 629.9 full barriers recorded, 278.4 left out (nothing recorded since the one before), 23.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.9; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.9, an op 1049.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.1 ms a frame (the longest submission 4.8 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.25 MB unswizzled on the CPU in 0.18 ms
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11040 (texture frame 11041), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11045 (texture frame 11046), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11047 (texture frame 11048), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11049 (texture frame 11050), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5740000 (3583051 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11051 (texture frame 11052), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11053 (texture frame 11054), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11055 (texture frame 11056), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11057 (texture frame 11058), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1117 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1713 kept their virtual core (nothing there to wait for), 3151 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: readback pages refreshed from guest memory: 3395639 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 111, 1049 MB with memory of 1488 MB (438 MB spared); blocks given memory later 9034, given back 11900; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 657 binds (193.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 161 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 51 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1284 dirty pages from their submits' captures, 54301 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 42410 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7723 runs, 708.1 MB compared; 74 pages read into the shadows; 292 MB of shadows (47 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17607 in captures, 1 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 183 live (2288 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11062 (texture frame 11063), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 18964628 lookups, 140714 missed (0.74 %)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11064 (texture frame 11065), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11066 (texture frame 11067), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11068 (texture frame 11069), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5760000 (3595251 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11070 (texture frame 11071), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11072 (texture frame 11073), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11074 (texture frame 11075), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11076 (texture frame 11077), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: pool 0x1163f6e000+0x7e9000 imported in 2.2 ms as 0x1163c00000..0x1164800000 (222 imports, 685 ms in all)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 579 ranges, 3131 MB), 279 pages written, 279 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.037; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 25 kept (132 MB), 12 spare buffers, 8301 texture pages copied (32 MB)
[hos host] vk: late pages' fix (600 recordings with one): of 1013 captures with late pages 1013 fixed (95.7 pages, 96.5 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 442 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 162, of an earlier recording 280, not recorded yet 0, of this part 0; pages fixed from stamps 95932, from snapshots 1017; the shadows: 546305 pages synced from captures, 528079 made known by it, 18226 left unknown because the replay or the collection wrote them since
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11078 (texture frame 11079), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: 1 textures retired (1367 left, 2567 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11080 (texture frame 11081), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11082 (texture frame 11083), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11084 (texture frame 11085), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11086 (texture frame 11087), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11088 (texture frame 11089), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11091 (texture frame 11092), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5780000 (3606260 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11093 (texture frame 11094), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11095 (texture frame 11096), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #160000: event 0x100d6c802c mode 0 value 111182 (now 111163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11097 (texture frame 11098), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11099 (texture frame 11100), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1856 dirty pages from their submits' captures, 55141 from earlier ones', 1013 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 866 dispatches, 41469 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7471 runs, 705.7 MB compared; 2096 pages read into the shadows; 304 MB of shadows (54 MB of pages known, in 99 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18046 in captures, 40 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool: 0x1163c00000 (12 MB, captured, state 2): 1013;
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11101 (texture frame 11102), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11103 (texture frame 11104), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11105 (texture frame 11106), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11110 (texture frame 11111), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11112 (texture frame 11113), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5800000 (3618601 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11114 (texture frame 11115), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11116 (texture frame 11117), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11118 (texture frame 11119), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: program picks: 7948036 kept (2891733 by the last picks' textures, a uniform buffer bound since), 26216 new, 3425747 with other textures (of 11400000)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11120 (texture frame 11121), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11122 (texture frame 11123), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos sead::MainThread] nvn: 1300000 recordings made, 70235 live copies (23.0 MB), 5878 texture copies; 4 command buffers' segments (4.4 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11124 (texture frame 11125), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11126 (texture frame 11127), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11128 (texture frame 11129), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11130 (texture frame 11131), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11132 (texture frame 11133), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11134 (texture frame 11135), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11136 (texture frame 11137), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5820000 (3628447 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11138 (texture frame 11139), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: sparse mirrors: 112, 1063 MB with memory of 1500 MB (436 MB spared); blocks given memory later 9115, given back 11938; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 660 binds (193.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 162 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 50 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 974, 4327.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1368 dirty pages from their submits' captures, 52931 from earlier ones', 1013 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 883 dispatches, 41064 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7454 runs, 698.0 MB compared; 79 pages read into the shadows; 304 MB of shadows (51 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18416 in captures, 40 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool: 0x1163c00000 (12 MB, captured, state 2): 1013;
[hos host] vk: imports: 184 live (2300 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11140 (texture frame 11141), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11142 (texture frame 11143), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11144 (texture frame 11145), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11146 (texture frame 11147), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11148 (texture frame 11149), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: a frame: 147 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.57, present 0.16, draw 7.51, descriptors 3.28, texture lookups 0.08, uploads 0.24, pipelines 0.31, programs 1.34, commands 2.01, dispatch 0.95, memory sync 0.58, queue submit 0.00, slot wait 0.29
[hos host] vk: a frame: mirrors 325 (1079 MB): 5.80 MB synced in 1053.5 copies, 17.89 MB written back
[hos host] vk: 11160 presents, 57.9 fps; a frame: 1946 draws (0 skipped), 149 dispatches, 187 passes, 131 barriers, 6 submits, 0.4 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1367 textures (2567 MB)
[hos host] vk: host time in GPU resource waits: 4.96 ms a frame (fences 1.21, timelines 3.75, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 683.4 full barriers recorded, 296.6 left out (nothing recorded since the one before), 21.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1048.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.1 ms a frame (the longest submission 5.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.1 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.47 MB unswizzled on the CPU in 0.24 ms
[hos ModuleSystemWorker4] barrier: 180943 looks: 66255 found the masks moved already, 112380 waited until they moved, 2308 for the whole 1 ms (a wait 273 us on average); 188127 wakes
[hos host] vk: replay reads of guest memory as submitted: 5840000 (3638638 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 569 ranges, 3120 MB), 261 pages written, 261 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (118 MB), 12 spare buffers, 2386 texture pages copied (9 MB)
[hos Presentation Thread] vi: 11172 frames presented (59.2 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 33, 15.5-17.8 210, 17.8-22 51, 22-30 2, 30-40 0, over 40 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6542 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11175 (texture frame 11176), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1371 dirty pages from their submits' captures, 51150 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 852 dispatches, 39992 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7346 runs, 672.9 MB compared; 67 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16232 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11177 (texture frame 11178), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11179 (texture frame 11180), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: replay reads of guest memory as submitted: 5860000 (3650630 pages from the captures)
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11181 (texture frame 11182), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11183 (texture frame 11184), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11185 (texture frame 11186), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it
[hos host] vk: late pages found at the replay (600 takes): 12 pages in 12 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: replay reads of guest memory as submitted: 5880000 (3662281 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 113, 1061 MB with memory of 1548 MB (486 MB spared); blocks given memory later 9115, given back 12228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 663 binds (194.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 163 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 49 spares now, 1080 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 370 dirty pages from their submits' captures, 54923 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 909 dispatches, 42084 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7710 runs, 708.6 MB compared; 64 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17364 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 184 live (2300 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos host] vk: program picks: 8100063 kept (2963733 by the last picks' textures, a uniform buffer bound since), 26228 new, 3473708 with other textures (of 11600000)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1111 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1511 kept their virtual core (nothing there to wait for), 2841 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 5900000 (3674176 pages from the captures)
[hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 240000 so far (x1 0x117a4f2040 x2 0x6001d40 x3 0x117a4f2000 x4 0xf33080)
[hos sead::MainThread] nvn: GPU waits in 5 s: 167 sync waits (528 ms), 0 polls timed out (0 ms), 3 queue finishes (38 ms)
[hos host] vk: replay reads of guest memory as submitted: 5920000 (3683773 pages from the captures)
[hos host] vk: queues ordered by their events: 66000 submissions stopped at a wait, 65966 went on once it was reached, 32 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 3 (stopped 50.30 ms at most); waits met 163918, not met 0; captures at a CPU event's wait 7435, 0 of them at a wait met on arrival
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 562 ranges, 3108 MB), 258 pages written, 258 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (130 MB), 11 spare buffers, 3653 texture pages copied (14 MB)
[hos host] vk: readback pages refreshed from guest memory: 3462809 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 324 dirty pages from their submits' captures, 56080 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 926 dispatches, 43180 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7755 runs, 724.1 MB compared; 61 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17643 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 5940000 (3695021 pages from the captures)
[hos ModuleSystemWorker5] vcores: 157095 takes, 33222 waits (mean 61 us, longest 11.3 ms, 0 over HOS_VCORE_LOG_MS), 15990 kicks, 15991 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a frame: 144 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.73, present 0.14, draw 6.31, descriptors 2.72, texture lookups 0.06, uploads 0.18, pipelines 0.26, programs 1.14, commands 1.71, dispatch 0.81, memory sync 0.47, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 325 (1060 MB): 5.38 MB synced in 1045.7 copies, 17.70 MB written back
[hos host] vk: 11280 presents, 60.2 fps; a frame: 1925 draws (0 skipped), 150 dispatches, 186 passes, 131 barriers, 6 submits, 0.4 uploads (1.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1367 textures (2567 MB)
[hos host] vk: host time in GPU resource waits: 1.98 ms a frame (fences 1.47, timelines 0.50, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 680.6 full barriers recorded, 296.1 left out (nothing recorded since the one before), 20.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.4, an op 1024.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.2 ms a frame (the longest submission 5.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.29 MB unswizzled on the CPU in 0.18 ms
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10428 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.6 ms, 1 frames after their copies were recorded, the first #16653 0x116e25b000 260x260 levels 9 NVN 0xd (33 so far, 7232 KB, 12 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed)
[hos host] vk: replay reads of guest memory as submitted: 5960000 (3705722 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #165000: event 0x100d6c802c mode 0 value 114982 (now 114963) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #115000: event 0x100d6c801c mode 0 value 114997 (lr 0x802911b8)
[hos host] vk: sparse mirrors: 113, 1060 MB with memory of 1548 MB (487 MB spared); blocks given memory later 9115, given back 12242; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 665 binds (194.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 163 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 49 spares now, 1080 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 480 dirty pages from their submits' captures, 54561 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 41610 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7658 runs, 707.5 MB compared; 66 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17243 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb5c00000 (44 MB, captured, state 2): 1;
[hos host] vk: imports: 184 live (2300 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB)
[hos host] vk: replay reads of guest memory as submitted: 5980000 (3718162 pages from the captures)
[hos rmbl::VibrationThread] hid: 36864 vibration values, strongest 0.00 since the last count
[hos host] vk: program picks: 8251640 kept (3036090 by the last picks' textures, a uniform buffer bound since), 26228 new, 3522131 with other textures (of 11800000)
[hos host] vk: texture #15310 0x100089d2a00 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 11334 (drawn at 9705)
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 254 dirty pages from their submits' captures, 55429 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 877 dispatches, 42515 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7589 runs, 691.5 MB compared; 48 pages read into the shadows; 304 MB of shadows (51 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17467 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 6000000 (3726749 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 552 ranges, 3106 MB), 286 pages written, 286 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.015; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 4250 texture pages copied (16 MB)
[hos host] vk: texture #14274 0x116c30d000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 11362 (drawn at 5293)
[hos host] vk: presenter thread: the copy's fence 1.34 ms, the present 0.76 ms a present
[hos host] vk: replay reads of guest memory as submitted: 6020000 (3735336 pages from the captures)
[hos host] vk: sparse mirrors: 113, 1059 MB with memory of 1548 MB (488 MB spared); blocks given memory later 9115, given back 12263; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 668 binds (194.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 163 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 49 spares now, 1080 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 191 dirty pages from their submits' captures, 54686 from earlier ones', 4 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 915 dispatches, 41612 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7761 runs, 710.0 MB compared; 56 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17013 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 4;
[hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1129 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1730 kept their virtual core (nothing there to wait for), 3390 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 7930 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 6040000 (3744362 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: memory: working set 15182 MB (peak 15952), private 21247 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 16 blocks 928 MB, 37 allocations 836 MB; used 3248 of 15579 MB; heap 1 (device): 161 blocks 4887 MB, 19045 allocations 4466 MB; used 6882 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 149 images 465 MB; 101 free ranges, the largest 21 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1267 images 1617 MB; 278 free ranges, the largest 29 MB
[hos host] vk: render target images: 5725 reused from the pool, 3332 made new, 3181 destroyed from it, 989 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1194 with the pool empty (0 of them with targets still in the dead queue), 287 with none in it a present yet, 25 with none of a kind they can use, 466 with all smaller (the largest 51% short on average), 504 with all bigger (over 1.5 times), 855 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2989 idle
[hos host] vk: VRAM: 6882 MB used of 11474 MB allowed (420 MB of it free in the blocks); textures 2719 MB (budget 7158 MB), 1422 textures
[hos host] vk: budget: VRAM 4887 of 11474 MB (6882 MB by the driver, 1995 MB unaccounted): textures 2719, mirrors 1056, blocks free 420, shadows 304, other buffers 211, other images 111, staging 64; RAM 3168 MB (the process's private memory 21247 MB): imports 2300, capture spares 732, captures 104, capture table 32, others 0
[hos host] vk: textures: decoded ASTC 1475 MB, rendered 669 MB (130 drawn in one frame, kept: 152242 KB), other 522 MB; used in the last 8 presents 1717 MB
[hos host] vk: decoded ASTC used again: 91 textures (212 MB) copied from ones the game finalized, not decoded; 91 of those retired, 18 listed; 1176 MB of blocks hashed in 203 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 6253 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 421 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 20 (7697 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 204x106x1 NVN 0x1: 1 (21 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 5 (879 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); 1466x282x1 NVN 0x25: 4 (6459 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 253 waited for it, 2.35 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 23.4 ms; VRAM 6.7 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6153 queued, 6153 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15243 ms and 331.0 MB read in all); 8348 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11700 graphics pipelines (1 not in a cache), 712 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2172 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 8987 (6153 ready, 0 loading), 16189 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13164 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 760 known, 3 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 152 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.21, present 0.15, draw 6.70, descriptors 2.98, texture lookups 0.15, uploads 0.22, pipelines 0.27, programs 1.15, commands 1.79, dispatch 0.83, memory sync 0.46, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 325 (1056 MB): 5.45 MB synced in 1060.5 copies, 17.75 MB written back
[hos host] vk: 11400 presents, 60.0 fps; a frame: 2010 draws (0 skipped), 152 dispatches, 180 passes, 131 barriers, 6 submits, 0.6 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1422 textures (2719 MB)
[hos host] vk: host time in GPU resource waits: 1.76 ms a frame (fences 1.56, timelines 0.20, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 683.1 full barriers recorded, 295.3 left out (nothing recorded since the one before), 15.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1078.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.8 ms a frame (the longest submission 5.0 ms); 0.7 textures made a frame: colour targets 0.2 (0.9 MB), depth 0.4 (1.5 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.51 MB unswizzled on the CPU in 0.21 ms
[hos host] vk: the render thread (600 presents, 16.78 ms a frame): its jobs 15.43 ms a frame (submits 6.19, presents 4.88, others 4.35), idle 1.36; the game waited for room at 30 presents, 0.10 ms each (0.01 a frame; 10.78 jobs queued at a present); 3 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 11.09 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 261 captures at a CPU signal; the submits' waits read in 0.041 ms a frame; readbacks: 1411 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.73 ms after its hand-over (21.02 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.43 + 0.27 (the window's copy is in the idle), after it 1.11 + 0.33; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.48 parts a frame recorded in the signal's job, 79.2% of them took nothing that came with the signal; of the others, 1.75 a frame took pages of the signal's capture into a mirror (33.0 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 37 of 261 frames, its first three in 36
[hos host] vk: frame ahead, what came with the signal: 261 signals with a capture, each of 77.0 pages in 9.9 pools; between a frame's last submit and that signal 0.01 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.21+0.06, 0.34+0.13, 0.03+0.02, 0.03+0.02, 0.20+0.02, 0.18+0.03, 0.06+0.02, 0.05+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: render targets drawn over from another start, let go: 9788 so far (46474 MB), 8773 of them (41752 MB) between the passes
[hos host] vk: program picks: 8403939 kept (3109804 by the last picks' textures, a uniform buffer bound since), 26228 new, 3569832 with other textures (of 12000000)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 157 dirty pages from their submits' captures, 54358 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 871 dispatches, 41925 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7664 runs, 694.0 MB compared; 52 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17257 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 6060000 (3753504 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: texture #15694 0x1001e7cf000 1120x1333x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 5845 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14767 0x10004b01a00 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1735 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14761 0x10005552400 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 1735 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14769 0x100050a2000 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1708 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14768 0x10004fb1e00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1696 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #14770 0x10004ec1c00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1645 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 341 textures retired (1046 left, 2104 MB): budget 337, drawn over 4, dead targets 0; GPU-drawn with no copy kept: 0 live, 6 finalized, 26 MB
[hos host] vk: replay reads of guest memory as submitted: 6080000 (3764635 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 558 ranges, 3108 MB), 279 pages written, 279 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (138 MB), 11 spare buffers, 4250 texture pages copied (16 MB)
[hos host] vk: texture #16733 0x10008612200 957x957 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 11449 (drawn at 11388)
[hos host] vk: readback pages refreshed from guest memory: 3530428 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 107, 1046 MB with memory of 1516 MB (469 MB spared); blocks given memory later 9121, given back 12304; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 678 binds (196.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 163 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 55 spares now, 1112 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 285 dirty pages from their submits' captures, 54312 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 878 dispatches, 41932 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7877 runs, 711.5 MB compared; 58 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17670 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 6100000 (3774630 pages from the captures)
[hos Presentation Thread] vi: 11472 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 9, 15.5-17.8 280, 17.8-22 11, 22-30 0, 30-40 0, over 40 0
[hos host] vk: texture #16647 0x116e006000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 11481 (drawn at 11034)
[hos host] vk: replay reads of guest memory as submitted: 6120000 (3786066 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 385 dirty pages from their submits' captures, 53582 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 870 dispatches, 41418 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7561 runs, 677.3 MB compared; 70 pages read into the shadows; 304 MB of shadows (51 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17298 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12254 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #170000: event 0x100d6c802c mode 0 value 118782 (now 118763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 6140000 (3795198 pages from the captures)
[hos host] vk: program picks: 8556862 kept (3182304 by the last picks' textures, a uniform buffer bound since), 26228 new, 3616909 with other textures (of 12200000)
[hos host] vk: 3 textures retired (1045 left, 2092 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: a frame: 161 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.91, present 0.21, draw 6.37, descriptors 2.80, texture lookups 0.18, uploads 0.23, pipelines 0.27, programs 1.11, commands 1.71, dispatch 0.81, memory sync 0.45, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 325 (1046 MB): 5.39 MB synced in 1058.4 copies, 17.80 MB written back
[hos host] vk: 11520 presents, 60.0 fps; a frame: 1957 draws (0 skipped), 153 dispatches, 178 passes, 131 barriers, 6 submits, 0.4 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1045 textures (2092 MB)
[hos host] vk: host time in GPU resource waits: 1.38 ms a frame (fences 1.22, timelines 0.17, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 685.3 full barriers recorded, 296.1 left out (nothing recorded since the one before), 13.3 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 1068.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.1 ms a frame (the longest submission 5.1 ms); 0.5 textures made a frame: colour targets 0.2 (1.6 MB), depth 0.2 (1.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.53 MB unswizzled on the CPU in 0.22 ms
[hos host] vk: texture #14272 0x116c2d8000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 11522 (drawn at 5293)
[hos host] vk: reaper: 431 images (863 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 282 (412 MB)
[hos host] vk: sparse mirrors: 109, 1056 MB with memory of 1532 MB (475 MB spared); blocks given memory later 9121, given back 12304; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 680 binds (196.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 165 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 53 spares now, 1096 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 337 dirty pages from their submits' captures, 53514 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 867 dispatches, 41464 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7792 runs, 711.0 MB compared; 64 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17180 in captures, 2 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 6160000 (3806084 pages from the captures)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1112 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1302 kept their virtual core (nothing there to wait for), 2794 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 554 ranges, 3060 MB), 286 pages written, 286 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (128 MB), 11 spare buffers, 4488 texture pages copied (17 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 52 sync waits (73 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: recorder: 6121759 packets run in 10.0 s (busy 1.79 s, spun idle 0.11 s, slept 10352 times, 3143 of them under 200 us), 0 calls made directly (no room); 18332 batches over 596 frames (30.8 a frame, 334 packets a batch), 9002 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 59 times, 6.79 ms (0.1 times, 0.01 ms a frame); every thread 631 times, 45.11 ms in all, the longest 1.28 ms; vkWaitForFences 522/31.67 ms (max 0.45), vmaFreeMemory 18/3.46 ms (max 0.60), vmaDestroyBuffer 14/3.16 ms (max 1.28), the queue's lock 21/2.53 ms (max 0.39), vkQueueBindSparse 29/2.05 ms (max 0.52), vkDestroyImageView 18/1.75 ms (max 0.31)
[hos host] vk: recorder: of its busy 1.79 s, 0.52 s in parts' ends and submits (the longest 2.66 ms), 0.20 s of that blocked on the queue's lock (349 times, the longest 2.61 ms), so 0.32 s in the ends, submits and holds; 1.28 s running the other packets; 10271 packets a frame
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 2 textures retired (1060 left, 2103 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6180000 (3818659 pages from the captures)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #120000: event 0x100d6c8030 mode 0 value 119999 (lr 0x806612a8)
[hos host] vk: replay reads of guest memory as submitted: 6200000 (3831028 pages from the captures)
[hos ModuleSystemWorker3] nvn: 1350000 recordings made, 70785 live copies (23.2 MB), 5950 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 623 dirty pages from their submits' captures, 55304 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 937 dispatches, 42082 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8286 runs, 703.0 MB compared; 65 pages read into the shadows; 304 MB of shadows (51 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17249 in captures, 2 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos aal::LowPrioWorkerThread] vcores: 152095 takes, 22663 waits (mean 68 us, longest 7.1 ms, 0 over HOS_VCORE_LOG_MS), 10699 kicks, 10553 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 6220000 (3842972 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: texture #16057 0x116cf82000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 11598 (drawn at 9242)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 4548 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 6240000 (3854371 pages from the captures)
[hos host] vk: program picks: 8709400 kept (3255707 by the last picks' textures, a uniform buffer bound since), 26230 new, 3664369 with other textures (of 12400000)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 112, 1065 MB with memory of 1552 MB (487 MB spared); blocks given memory later 9122, given back 12350; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 688 binds (198.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 168 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 50 spares now, 1076 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 207 dirty pages from their submits' captures, 56984 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 920 dispatches, 43988 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9104 runs, 758.4 MB compared; 62 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17803 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 17 textures retired (1061 left, 2094 MB): budget 17, drawn over 0, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 35 textures retired (1099 left, 2193 MB): budget 35, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6260000 (3867177 pages from the captures)
[hos ModuleSystemWorker4] nvn: DrawMeshTasksIndirect: 260000 so far (x1 0x117a4f2040 x2 0x6001d40 x3 0x117a4f2000 x4 0xf33080)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.04 over 564 ranges, 3150 MB), 263 pages written, 263 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (122 MB), 11 spare buffers, 5309 texture pages copied (20 MB)
[hos host] vk: texture 0x10001e00400 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 25 textures retired (1260 left, 2497 MB): budget 25, drawn over 0, dead targets 0
[hos host] vk: a frame: 179 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.93, present 0.15, draw 7.26, descriptors 3.81, texture lookups 1.20, uploads 1.24, pipelines 0.26, programs 1.04, commands 1.68, dispatch 0.86, memory sync 0.48, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 344 (1112 MB): 6.00 MB synced in 1059.3 copies, 18.01 MB written back
[hos host] vk: 11640 presents, 56.5 fps; a frame: 1859 draws (0 skipped), 169 dispatches, 163 passes, 132 barriers, 6 submits, 2.8 uploads (2.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1272 textures (2510 MB)
[hos host] vk: host time in GPU resource waits: 4.36 ms a frame (fences 2.13, timelines 2.22, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 657.8 full barriers recorded, 280.5 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.6; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.9 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.7, an op 1090.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.0 ms a frame (the longest submission 5.0 ms); 2.8 textures made a frame: colour targets 0.4 (1.1 MB), depth 0.1 (0.6 MB), ASTC or BC 2.3 (3.5 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.92 MB unswizzled on the CPU in 0.27 ms
[hos host] vk: replay reads of guest memory as submitted: 6280000 (3879327 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 3597538 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1724 dirty pages from their submits' captures, 50481 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 847 dispatches, 40189 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8064 runs, 683.5 MB compared; 72 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16595 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 2;
[hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 14961078 lookups, 126778 missed (0.85 %)
[hos host] vk: replay reads of guest memory as submitted: 6300000 (3890708 pages from the captures)
[hos host] vk: texture #17190 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (7 so far)
[hos host] vk: sparse mirrors: 126, 1112 MB with memory of 1636 MB (523 MB spared); blocks given memory later 9122, given back 12350; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 702 binds (201.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 182 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 36 spares now, 992 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 365 dirty pages from their submits' captures, 55198 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 875 dispatches, 42641 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8558 runs, 720.3 MB compared; 58 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17566 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: texture #16214 0x10004bf1600 160x90 x1 NVN 0x3 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635)
[hos host] vk: texture #16190 0x10004651000 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635)
[hos host] vk: texture #16194 0x100053ea200 640x360 x1 NVN 0xa the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635)
[hos host] vk: texture #16193 0x10005372000 640x360 x1 NVN 0x5 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635)
[hos host] vk: texture #16195 0x100054da400 640x360 x1 NVN 0x11 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635)
[hos host] vk: texture #16197 0x10005191c00 1600x1600 x3 NVN 0x32 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635)
[hos host] vk: texture #16227 0x10004380c00 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #175000: event 0x100d6c802c mode 0 value 122582 (now 122563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 6320000 (3902172 pages from the captures)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1099 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 19 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1464 kept their virtual core (nothing there to wait for), 2888 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 5 textures retired (1334 left, 2623 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6340000 (3915189 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 16 textures retired (1328 left, 2642 MB): budget 16, drawn over 0, dead targets 0
[hos host] vk: program picks: 8859341 kept (3318222 by the last picks' textures, a uniform buffer bound since), 26233 new, 3714425 with other textures (of 12600000)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.04 over 572 ranges, 3284 MB), 257 pages written, 257 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (138 MB), 8 spare buffers, 6127 texture pages copied (23 MB)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 49 textures retired (1355 left, 2651 MB): budget 49, drawn over 0, dead targets 0
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1624 dirty pages from their submits' captures, 48847 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 827 dispatches, 38231 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7201 runs, 627.7 MB compared; 77 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15701 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos ModuleSystemWorker4] tex: the game's texture heap: 189 compaction rounds asked for in 10800 frames without one running
[hos host] vk: replay reads of guest memory as submitted: 6360000 (3928180 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13321 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 6 textures retired (1397 left, 2696 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: texture #16204 0x10004741200 2560x1440 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11742 (drawn at 11635)
[hos host] vk: texture #16205 0x10004bf1600 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11742 (drawn at 11635)
[hos host] vk: texture #16206 0x10004b01400 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11742 (drawn at 11635)
[hos host] vk: texture 0x100023a0e00 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: texture #16228 0x10001e00400 640x360 x1 NVN 0x14 the GPU drew, finalized by the game, sampled again at frame 11742 (drawn at 11635)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 25 textures retired (1378 left, 2697 MB): budget 25, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6380000 (3941778 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 4 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: a frame: 265 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.64, present 0.17, draw 6.87, descriptors 3.24, texture lookups 0.49, uploads 0.28, pipelines 0.27, programs 1.11, commands 1.76, dispatch 0.85, memory sync 0.46, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 348 (1119 MB): 5.53 MB synced in 1066.4 copies, 17.88 MB written back
[hos host] vk: 11760 presents, 59.0 fps; a frame: 1887 draws (0 skipped), 164 dispatches, 187 passes, 134 barriers, 6 submits, 2.1 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1378 textures (2697 MB)
[hos host] vk: host time in GPU resource waits: 4.83 ms a frame (fences 1.80, timelines 3.03, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 729.4 full barriers recorded, 304.2 left out (nothing recorded since the one before), 16.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 26.5; 0x12: 69.3; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 124.8, an op 1174.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.2 ms a frame (the longest submission 4.2 ms); 2.3 textures made a frame: colour targets 0.2 (0.7 MB), depth 0.1 (0.4 MB), ASTC or BC 2.0 (3.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.29 MB unswizzled on the CPU in 0.05 ms
[hos Presentation Thread] vi: 11763 frames presented (58.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 50, 15.5-17.8 197, 17.8-22 28, 22-30 10, 30-40 6, over 40 0
[hos host] vk: sparse mirrors: 130, 1119 MB with memory of 1656 MB (536 MB spared); blocks given memory later 9126, given back 12351; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 708 binds (202.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 186 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 32 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1836 dirty pages from their submits' captures, 52988 from earlier ones', 4 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 873 dispatches, 41954 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7344 runs, 702.9 MB compared; 74 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16993 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool: 0x7f86400000 (8 MB, captured, state 2): 4;
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 4 textures retired (1374 left, 2696 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6400000 (3954130 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 6 textures retired (1368 left, 2690 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 20 textures retired (1352 left, 2663 MB): budget 20, drawn over 0, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 17 textures retired (1340 left, 2647 MB): budget 17, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6420000 (3965802 pages from the captures)
[hos host] vk: texture #17008 0x10004bf1c00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741)
[hos host] vk: texture #17026 0x100021aea00 120x67 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741)
[hos host] vk: texture #17031 0x100020e5a00 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741)
[hos host] vk: texture #16990 0x10005552400 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741)
[hos host] vk: texture #17003 0x100050a2000 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741)
[hos host] vk: texture #17002 0x10004fb1e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741)
[hos host] vk: texture #17033 0x10004921600 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 11803 (drawn at 11741)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 9 textures retired (1337 left, 2640 MB): budget 8, drawn over 1, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 182 dirty pages from their submits' captures, 74789 from earlier ones', 41178 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 987 dispatches, 45762 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7659 runs, 736.7 MB compared; 62 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18460 in captures, 2 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool: 0x7fab000000 (68 MB, captured, state 2): 16936; 0x7fb5c00000 (44 MB, captured, state 2): 10657; 0x7fb8800000 (52 MB, captured, state 2): 8785; 0x7fd3c00000 (16 MB, captured, state 2): 3830; 0x7fa9000000 (4 MB, captured, state 2): 644; 0x7faf400000 (4 MB, captured, state 2): 155; 0x113b400000 (24 MB, captured, state 2): 29; 0x7f94000000 (16 MB, captured, state 2): 26;
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 38 textures retired (1317 left, 2607 MB): budget 38, drawn over 0, dead targets 0
[hos sead::MainThread] vk: capture: a submit 0.28 ms (lock 0.02, scan 0.05 over 567 ranges, 3340 MB), 704 pages written, 533 copied (2.1 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (318 MB), 11 spare buffers, 6348 texture pages copied (24 MB)
[hos OptimizeChangeBanc] nvn: 1400000 recordings made, 69330 live copies (23.3 MB), 5950 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: program picks: 9006207 kept (3367113 by the last picks' textures, a uniform buffer bound since), 26274 new, 3767518 with other textures (of 12800000)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #125000: event 0x100d6c803c mode 0 value 124999 (lr 0x80661bfc)
[hos host] vk: replay reads of guest memory as submitted: 6440000 (3977550 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 118 sync waits (649 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 9 textures retired (1308 left, 2590 MB): budget 9, drawn over 0, dead targets 0
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 7 textures retired (1303 left, 2586 MB): budget 7, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6460000 (3990729 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 12 textures retired (1296 left, 2586 MB): budget 12, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 3665028 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10276 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 130, 1195 MB with memory of 1656 MB (460 MB spared); blocks given memory later 10338, given back 12351; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 743 binds (211.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 186 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 32 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 7944 dirty pages from their submits' captures, 87919 from earlier ones', 3774 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1051 dispatches, 46111 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7136 runs, 699.4 MB compared; 359 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16523 in captures, 236 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 3774;
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos sead::MainThread] vk: readbacks of GPU results: 1019 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1452 kept their virtual core (nothing there to wait for), 3156 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos BaseProcExecutor2] vcores: 151035 takes, 28215 waits (mean 68 us, longest 17.1 ms, 0 over HOS_VCORE_LOG_MS), 14212 kicks, 14221 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: late pages found at the replay (600 takes): 371 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 2 textures retired (1294 left, 2586 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6480000 (4001677 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 10 textures retired (1289 left, 2569 MB): budget 10, drawn over 0, dead targets 0
[hos host] vk: a frame: 266 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.63, present 0.18, draw 6.79, descriptors 2.90, texture lookups 0.16, uploads 0.03, pipelines 0.29, programs 1.22, commands 1.86, dispatch 0.88, memory sync 1.12, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 348 (1187 MB): 8.67 MB synced in 1065.1 copies, 17.50 MB written back
[hos host] vk: 11880 presents, 55.2 fps; a frame: 1809 draws (0 skipped), 138 dispatches, 167 passes, 131 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1289 textures (2569 MB)
[hos host] vk: host time in GPU resource waits: 3.72 ms a frame (fences 1.69, timelines 2.03, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 630.0 full barriers recorded, 286.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 995.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.2 ms a frame (the longest submission 5.7 ms); 0.4 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.4 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 5 textures retired (1285 left, 2562 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 871 dirty pages from their submits' captures, 55953 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 933 dispatches, 43815 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7446 runs, 733.7 MB compared; 78 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17502 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 6500000 (4013890 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #180000: event 0x100d6c802c mode 0 value 126382 (now 126363) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.18
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 21 textures retired (1272 left, 2524 MB): budget 21, drawn over 0, dead targets 0
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos Presentation Thread] vk: capture: a submit 0.21 ms (lock 0.03, scan 0.06 over 613 ranges, 3388 MB), 324 pages written, 324 copied (1.3 MB)
[hos host] vk: a capture taken (ms): all 0.029; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (114 MB), 11 spare buffers, 8463 texture pages copied (33 MB)
[hos host] vk: replay reads of guest memory as submitted: 6520000 (4026453 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: program picks: 9151396 kept (3411432 by the last picks' textures, a uniform buffer bound since), 26280 new, 3822323 with other textures (of 13000000)
[hos host] vk: sparse mirrors: 130, 1189 MB with memory of 1656 MB (466 MB spared); blocks given memory later 10360, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 749 binds (212.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 186 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 32 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 396 dirty pages from their submits' captures, 57256 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 987 dispatches, 44313 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7548 runs, 734.7 MB compared; 68 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18161 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7f86000000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: replay reads of guest memory as submitted: 6540000 (4038464 pages from the captures)
[hos ModuleSystemWorker2] nvn: 1450000 recordings made, 75348 live copies (23.7 MB), 5950 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: 1 textures retired (1277 left, 2512 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: presenter thread: the copy's fence 1.44 ms, the present 0.75 ms a present
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17717 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 6560000 (4050721 pages from the captures)
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 434 dirty pages from their submits' captures, 58607 from earlier ones', 910 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 45389 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7719 runs, 751.5 MB compared; 72 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18459 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 910;
[hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back
[hos host] vk: replay reads of guest memory as submitted: 6580000 (4062584 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.07
[hos host] vk: memory: working set 15171 MB (peak 15952), private 21363 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 17 blocks 784 MB, 36 allocations 692 MB; used 3108 of 15579 MB; heap 1 (device): 166 blocks 5047 MB, 20797 allocations 4379 MB; used 7042 of 8467 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 111 images 372 MB; 66 free ranges, the largest 28 MB
[hos host] vk: image blocks (textures, compacted): 57 blocks 1824 MB, 1176 images 1506 MB; 413 free ranges, the largest 32 MB
[hos host] vk: render target images: 5753 reused from the pool, 3435 made new, 3322 destroyed from it, 993 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1294 with the pool empty (0 of them with targets still in the dead queue), 287 with none in it a present yet, 25 with none of a kind they can use, 467 with all smaller (the largest 51% short on average), 506 with all bigger (over 1.5 times), 855 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3130 idle
[hos host] vk: VRAM: 7042 MB used of 8467 MB allowed (667 MB of it free in the blocks); textures 2514 MB (budget 2140 MB), 1293 textures
[hos host] vk: budget: VRAM 5047 of 8467 MB (7042 MB by the driver, 1995 MB unaccounted): textures 2514, mirrors 1174, blocks free 667, shadows 304, other buffers 211, other images 111, staging 64; RAM 3034 MB (the process's private memory 21363 MB): imports 2304, capture spares 564, captures 134, capture table 32, others 0
[hos host] vk: textures: decoded ASTC 1544 MB, rendered 589 MB (88 drawn in one frame, kept: 45085 KB), other 345 MB; used in the last 8 presents 1899 MB
[hos host] vk: decoded ASTC used again: 185 textures (477 MB) copied from ones the game finalized, not decoded; 185 of those retired, 1 listed; 1360 MB of blocks hashed in 231 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 6853 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 28 (8783 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 102x30x1 NVN 0x25: 1 (11 KB); 137x134x1 NVN 0x1: 1 (17 KB); 202x38x1 NVN 0x25: 1 (29 KB); 204x106x1 NVN 0x1: 1 (21 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 8 (1407 KB); 266x102x1 NVN 0x1: 1 (26 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 261 waited for it, 3.48 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 48.6 ms; VRAM 6.9 of 8.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6168 queued, 6168 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15261 ms and 331.4 MB read in all); 8363 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11712 graphics pipelines (1 not in a cache), 713 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2212 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9002 (6168 ready, 0 loading), 16426 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13188 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 763 known, 6 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 221 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.57, present 0.16, draw 6.00, descriptors 2.58, texture lookups 0.19, uploads 0.09, pipelines 0.26, programs 1.07, commands 1.63, dispatch 0.79, memory sync 0.57, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 348 (1174 MB): 5.41 MB synced in 1055.5 copies, 17.49 MB written back
[hos host] vk: 12000 presents, 60.0 fps; a frame: 1691 draws (0 skipped), 138 dispatches, 168 passes, 131 barriers, 6 submits, 0.8 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1293 textures (2514 MB)
[hos host] vk: host time in GPU resource waits: 2.16 ms a frame (fences 1.47, timelines 0.70, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 634.9 full barriers recorded, 287.5 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 923.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.8 ms a frame (the longest submission 4.7 ms); 0.8 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.7 (1.7 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 17.22 ms a frame): its jobs 15.76 ms a frame (submits 6.17, presents 5.01, others 4.57), idle 1.46; the game waited for room at 21 presents, 1.17 ms each (0.04 a frame; 8.24 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 267 captures at a CPU signal; the submits' waits read in 0.040 ms a frame; readbacks: 1488 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.96 ms after its hand-over (37.83 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.60 + 0.27 (the window's copy is in the idle), after it 1.26 + 0.81; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.67 parts a frame recorded in the signal's job, 73.2% of them took nothing that came with the signal; of the others, 2.30 a frame took pages of the signal's capture into a mirror (61.0 pages a frame) and 0.03 pages read as guest memory had them (168.4 pages a frame); a frame's first part took nothing in 31 of 267 frames, its first three in 31
[hos host] vk: frame ahead, what came with the signal: 267 signals with a capture, each of 560.8 pages in 11.1 pools; between a frame's last submit and that signal 0.05 texture and sampler registrations, 0.14 releases and 0.00 other CPU signals a signal (7 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.36+0.09, 0.41+0.13, 0.05+0.04, 0.03+0.02, 0.10+0.05, 0.16+0.27, 0.09+0.18, 0.05+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 616 ranges, 3342 MB), 268 pages written, 268 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (128 MB), 10 spare buffers, 4776 texture pages copied (18 MB)
[hos Presentation Thread] vi: real-time step on: ratio 0.997, game's frame rate 60
[hos host] vk: render targets drawn over from another start, let go: 9953 so far (47123 MB), 8938 of them (42401 MB) between the passes
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1086 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1406 kept their virtual core (nothing there to wait for), 2946 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 6600000 (4073949 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 126, 1160 MB with memory of 1588 MB (427 MB spared); blocks given memory later 10383, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 757 binds (212.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 187 taken from the spares, 162 made new; 222 kept as spares, 0 destroyed; 35 spares now, 1036 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 322 dirty pages from their submits' captures, 56737 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 958 dispatches, 44051 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7550 runs, 733.8 MB compared; 83 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18036 in captures, 1 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: 224 mirrors given back (1757 MB in all); 1158 MB of mirrors now, 172 MB of them idle and kept for what the GPU wrote there alone
[hos sead::MainThread] nvn: DrawMeshTasks: 280000 so far (x1 0 x2 0x20 x3 0x10201414 x4 0x6001ba0)
[hos host] vk: commit: 2481 MB left under Windows' limit as it is now, and the page files can grow by 79851 MB more: counted as room
[hos host] vk: replay reads of guest memory as submitted: 6620000 (4085301 pages from the captures)
[hos host] vk: program picks: 9295234 kept (3454419 by the last picks' textures, a uniform buffer bound since), 26280 new, 3878485 with other textures (of 13200000)
[hos Presentation Thread] vi: 12053 frames presented (57.9 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 35, 15.5-17.8 210, 17.8-22 31, 22-30 10, 30-40 3, over 40 1
[hos host] vk: readback pages refreshed from guest memory: 3732143 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1099, 4882.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 99 dirty pages from their submits' captures, 60368 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 970 dispatches, 46189 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7908 runs, 768.4 MB compared; 77 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18965 in captures, 1 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: queues ordered by their events: 72000 submissions stopped at a wait, 71966 went on once it was reached, 32 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 3 (stopped 50.30 ms at most); waits met 184199, not met 0; captures at a CPU event's wait 7756, 0 of them at a wait met on arrival
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9456 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 6640000 (4096845 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #130000: event 0x100d6c8048 mode 0 value 129999 (lr 0x80666598)
[hos host] vk: 2 textures retired (1293 left, 2500 MB): budget 0, drawn over 0, dead targets 2
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #185000: event 0x100d6c802c mode 0 value 130182 (now 130163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: 1 textures retired (1292 left, 2486 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6660000 (4107675 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 615 ranges, 3342 MB), 278 pages written, 278 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 25 kept (140 MB), 11 spare buffers, 7124 texture pages copied (27 MB)
[hos host] vk: sparse mirrors: 123, 1178 MB with memory of 1596 MB (417 MB spared); blocks given memory later 10423, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 763 binds (213.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 188 taken from the spares, 162 made new; 227 kept as spares, 0 destroyed; 39 spares now, 1032 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 221 dirty pages from their submits' captures, 57591 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 928 dispatches, 44835 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7694 runs, 750.4 MB compared; 97 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18535 in captures, 1 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: a frame: 256 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.77, present 0.18, draw 5.42, descriptors 2.27, texture lookups 0.08, uploads 0.00, pipelines 0.24, programs 1.00, commands 1.49, dispatch 0.77, memory sync 0.47, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 350 (1178 MB): 5.52 MB synced in 1044.4 copies, 17.45 MB written back
[hos host] vk: 12120 presents, 60.0 fps; a frame: 1617 draws (0 skipped), 136 dispatches, 166 passes, 131 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1292 textures (2486 MB)
[hos host] vk: host time in GPU resource waits: 1.69 ms a frame (fences 1.47, timelines 0.22, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 622.7 full barriers recorded, 285.6 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 831.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.5 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 6680000 (4119607 pages from the captures)
[hos host] vk: recorder: 5707589 packets run in 10.0 s (busy 1.61 s, spun idle 0.11 s, slept 10472 times, 2900 of them under 200 us), 0 calls made directly (no room); 17238 batches over 580 frames (29.7 a frame, 331 packets a batch), 8506 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 110 times, 8.90 ms (0.2 times, 0.02 ms a frame); every thread 704 times, 43.13 ms in all, the longest 0.44 ms; vkWaitForFences 495/28.15 ms (max 0.41), vkQueueBindSparse 83/4.77 ms (max 0.36), vkDestroyImageView 49/3.43 ms (max 0.29), vmaFreeMemory 16/2.32 ms (max 0.39), vkEndCommandBuffer 9/1.68 ms (max 0.44), the queue's lock 17/1.55 ms (max 0.22)
[hos host] vk: recorder: of its busy 1.61 s, 0.40 s in parts' ends and submits (the longest 1.14 ms), 0.09 s of that blocked on the queue's lock (164 times, the longest 1.10 ms), so 0.30 s in the ends, submits and holds; 1.21 s running the other packets; 9841 packets a frame
[hos sead::MainThread] nvn: GPU waits in 5 s: 67 sync waits (175 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 405 dirty pages from their submits' captures, 64795 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 958 dispatches, 45046 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7280 runs, 729.1 MB compared; 93 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18653 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 6700000 (4131296 pages from the captures)
[hos BaseProcExecutor0] vcores: 158125 takes, 28237 waits (mean 70 us, longest 21.3 ms, 0 over HOS_VCORE_LOG_MS), 13413 kicks, 13178 yields, 0 starved grants, 0 escapes in 5 s
[hos PlayReporter] vcores: PlayReporter (priority 29) waited 71.5 ms for a core: given by PreActorModelCompose (priority 26), which held it 0.0 ms and was never asked; the cores it may use now: core 0: it;
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1081 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1473 kept their virtual core (nothing there to wait for), 2879 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: program picks: 9440102 kept (3495015 by the last picks' textures, a uniform buffer bound since), 26280 new, 3933617 with other textures (of 13400000)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21010 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 6720000 (4142673 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 119, 1181 MB with memory of 1580 MB (398 MB spared); blocks given memory later 10517, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 769 binds (214.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 188 taken from the spares, 162 made new; 231 kept as spares, 0 destroyed; 43 spares now, 1048 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 274 dirty pages from their submits' captures, 59406 from earlier ones', 5 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 957 dispatches, 44754 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7425 runs, 746.3 MB compared; 99 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18121 in captures, 1 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 3; 0x7fd3c00000 (16 MB, captured, state 2): 1; 0x1042000000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 622 ranges, 3347 MB), 304 pages written, 304 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 25 kept (122 MB), 10 spare buffers, 7360 texture pages copied (28 MB)
[hos host] vk: replay reads of guest memory as submitted: 6740000 (4154135 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 6760000 (4166426 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 295 dirty pages from their submits' captures, 57883 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 950 dispatches, 45054 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7429 runs, 746.3 MB compared; 85 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18581 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a frame: 249 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.07, present 0.15, draw 5.69, descriptors 2.45, texture lookups 0.08, uploads 0.18, pipelines 0.24, programs 1.02, commands 1.54, dispatch 0.76, memory sync 0.50, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 350 (1171 MB): 5.61 MB synced in 1048.1 copies, 17.36 MB written back
[hos host] vk: 12240 presents, 60.0 fps; a frame: 1602 draws (0 skipped), 130 dispatches, 169 passes, 132 barriers, 6 submits, 0.5 uploads (1.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1308 textures (2503 MB)
[hos host] vk: host time in GPU resource waits: 1.76 ms a frame (fences 1.59, timelines 0.17, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 628.4 full barriers recorded, 288.8 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.5, an op 823.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.6 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.15 MB unswizzled on the CPU in 0.17 ms
[hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 39708874 lookups, 142783 missed (0.36 %)
[hos host] vk: replay reads of guest memory as submitted: 6780000 (4179928 pages from the captures)
[hos host] vk: 1 textures retired (1319 left, 2506 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: reaper: 515 images (1015 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 48 (86 MB)
[hos host] vk: replay reads of guest memory as submitted: 6800000 (4193927 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 3799676 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 119, 1177 MB with memory of 1576 MB (398 MB spared); blocks given memory later 10517, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 775 binds (214.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 191 taken from the spares, 162 made new; 234 kept as spares, 0 destroyed; 43 spares now, 1052 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 538 dirty pages from their submits' captures, 56964 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 943 dispatches, 44224 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7359 runs, 730.2 MB compared; 85 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18097 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9c000000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 24304 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 9584186 kept (3539728 by the last picks' textures, a uniform buffer bound since), 26288 new, 3989525 with other textures (of 13600000)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 622 ranges, 3369 MB), 275 pages written, 275 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.021, bytes copied 0.000; 25 kept (126 MB), 11 spare buffers, 8785 texture pages copied (34 MB)
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #190000: event 0x100d6c802c mode 0 value 133982 (now 133963) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 6820000 (4205116 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1100, 4884.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 59240 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 974 dispatches, 45919 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7893 runs, 768.2 MB compared; 91 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19136 in captures, 0 only in the collection, 44 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 1;
[hos ModuleSystemWorker5] vk: readbacks of GPU results: 1094 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1641 kept their virtual core (nothing there to wait for), 3223 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 21 textures retired (1305 left, 2498 MB): budget 21, drawn over 0, dead targets 0
[hos host] vk: 156 textures retired (1166 left, 2270 MB): budget 156, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6840000 (4216898 pages from the captures)
[hos host] vk: texture #17032 0x10001fe0800 640x360x1 x1 NVN 0x14 the GPU drew goes with no copy of its texels (finalized, unused for 609 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #17004 0x10004ec1c00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 609 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #17001 0x10004b01a00 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 609 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #16995 0x10005552400 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 609 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 39 textures retired (1185 left, 2291 MB): budget 39, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 19 MB
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #135000: event 0x100d6c8010 mode 0 value 135001 (lr 0x80660bcc)
[hos Presentation Thread] vi: 12352 frames presented (59.8 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 22, 15.5-17.8 254, 17.8-22 21, 22-30 2, 30-40 0, over 40 0
[hos host] vk: a frame: 187 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.76, present 0.15, draw 5.45, descriptors 2.42, texture lookups 0.20, uploads 0.30, pipelines 0.23, programs 0.94, commands 1.45, dispatch 0.76, memory sync 0.53, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 358 (1192 MB): 5.50 MB synced in 1047.5 copies, 17.43 MB written back
[hos host] vk: 12360 presents, 59.3 fps; a frame: 1547 draws (0 skipped), 134 dispatches, 163 passes, 131 barriers, 6 submits, 1.3 uploads (1.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1205 textures (2306 MB)
[hos host] vk: host time in GPU resource waits: 3.03 ms a frame (fences 1.73, timelines 1.31, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 611.1 full barriers recorded, 281.9 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 731.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.1 ms a frame (the longest submission 4.9 ms); 0.9 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.9 (1.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.75 MB unswizzled on the CPU in 0.26 ms
[hos host] vk: replay reads of guest memory as submitted: 6860000 (4230340 pages from the captures)
[hos host] vk: sparse mirrors: 125, 1194 MB with memory of 1608 MB (413 MB spared); blocks given memory later 10517, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 781 binds (216.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 197 taken from the spares, 162 made new; 234 kept as spares, 0 destroyed; 37 spares now, 1020 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1055 dirty pages from their submits' captures, 56413 from earlier ones', 17 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 940 dispatches, 44141 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7371 runs, 732.3 MB compared; 78 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18433 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 17;
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: 1 textures retired (1208 left, 2307 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: (its written part: 3 of 3 KB readable, 2016 bytes not zero; levels L0)
[hos host] vk: texture 0x10faa26e00 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 3 KB of it
[hos host] vk: replay reads of guest memory as submitted: 6880000 (4241792 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.04 over 620 ranges, 3411 MB), 253 pages written, 253 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (118 MB), 12 spare buffers, 2716 texture pages copied (10 MB)
[hos host] vk: 3 textures retired (1214 left, 2326 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos ModuleSystemWorker2] nvn: 1500000 recordings made, 73611 live copies (23.4 MB), 5950 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: replay reads of guest memory as submitted: 6900000 (4254905 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1491 dirty pages from their submits' captures, 51816 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 875 dispatches, 40991 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7640 runs, 692.7 MB compared; 68 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17105 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: program picks: 9732966 kept (3602077 by the last picks' textures, a uniform buffer bound since), 26288 new, 4040745 with other textures (of 13800000)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9881 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 6920000 (4268132 pages from the captures)
[hos host] vk: 20 textures retired (1240 left, 2412 MB): budget 20, drawn over 0, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 3 textures retired (1237 left, 2412 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 6940000 (4280516 pages from the captures)
[hos ModuleSystemWorker4] nvn: DrawMeshTasksIndirect: 300000 so far (x1 0x1179e9c000 x2 0 x3 0x1179e9c000 x4 0x654040)
[hos host] vk: 6 textures retired (1237 left, 2412 MB): budget 6, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: GPU waits in 5 s: 110 sync waits (431 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 128, 1207 MB with memory of 1628 MB (420 MB spared); blocks given memory later 10517, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 786 binds (217.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 201 taken from the spares, 162 made new; 235 kept as spares, 0 destroyed; 34 spares now, 1000 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1220 dirty pages from their submits' captures, 53354 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 895 dispatches, 41556 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8363 runs, 703.9 MB compared; 52 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17129 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 6960000 (4291812 pages from the captures)
[hos aal::LowPrioWorkerThread] vcores: 151761 takes, 22324 waits (mean 66 us, longest 71.5 ms, 1 over HOS_VCORE_LOG_MS), 10711 kicks, 10619 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 6980000 (4303613 pages from the captures)
[hos host] vk: 3 textures retired (1241 left, 2409 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: late pages' fix: stamps for 0x113c9f0000 (pool 0x113b400000; 576 blocks, 36864 KB)
[hos Presentation Thread] vk: capture: a submit 0.33 ms (lock 0.02, scan 0.05 over 589 ranges, 3355 MB), 591 pages written, 591 copied (2.3 MB)
[hos host] vk: a capture taken (ms): all 0.048; GPU wait for late pages 0.019 (1 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (752 MB), 3 spare buffers, 3622 texture pages copied (14 MB)
[hos host] vk: a frame: 171 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.31, present 0.16, draw 6.96, descriptors 3.12, texture lookups 0.21, uploads 0.31, pipelines 0.28, programs 1.21, commands 1.83, dispatch 0.89, memory sync 0.82, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 363 (1209 MB): 7.22 MB synced in 1054.7 copies, 17.97 MB written back
[hos host] vk: 12480 presents, 55.2 fps; a frame: 1979 draws (0 skipped), 166 dispatches, 174 passes, 131 barriers, 6 submits, 1.1 uploads (2.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1241 textures (2409 MB)
[hos host] vk: host time in GPU resource waits: 4.98 ms a frame (fences 1.49, timelines 3.48, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 682.3 full barriers recorded, 291.3 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 1104.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.1 ms a frame (the longest submission 11.6 ms); 0.7 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.6 (1.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.86 MB unswizzled on the CPU in 0.27 ms
[hos host] vk: replay reads of guest memory as submitted: 7000000 (4314559 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 3866900 pages so far (0 held: may be written ahead in their submission)
[hos sead::MainThread] vk: readbacks of GPU results: 1059 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 19 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1680 kept their virtual core (nothing there to wait for), 3184 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 21296 dirty pages from their submits' captures, 88983 from earlier ones', 4 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1068 dispatches, 47535 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9168 runs, 760.4 MB compared; 254 pages read into the shadows; 316 MB of shadows (51 MB of pages known, in 101 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18908 in captures, 118 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7f81400000 (4 MB, captured, state 2): 4;
[hos sead::MainThread] nvn: CommandBufferWaitEvent #195000: event 0x100d6c802c mode 0 value 137782 (now 137763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.16
[hos host] vk: replay reads of guest memory as submitted: 7020000 (4326198 pages from the captures)
[hos host] vk: program picks: 9883916 kept (3671438 by the last picks' textures, a uniform buffer bound since), 26288 new, 4089795 with other textures (of 14000000)
[hos host] vk: 8 textures retired (1233 left, 2408 MB): budget 8, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: EventSignal (CPU) #25000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos rmbl::VibrationThread] hid: 40960 vibration values, strongest 0.00 since the last count
[hos host] vk: 8 textures retired (1225 left, 2390 MB): budget 8, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 7040000 (4335886 pages from the captures)
[hos host] vk: 1 textures retired (1224 left, 2390 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: sparse mirrors: 128, 1139 MB with memory of 1628 MB (488 MB spared); blocks given memory later 10636, given back 13668; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 806 binds (222.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 201 taken from the spares, 162 made new; 235 kept as spares, 0 destroyed; 34 spares now, 1000 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1999 dirty pages from their submits' captures, 50765 from earlier ones', 7 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 873 dispatches, 39956 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8487 runs, 687.8 MB compared; 78 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18029 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9ac00000 (20 MB, captured, state 2): 6; 0x7f81400000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 7060000 (4347328 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21374 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 9 textures retired (1217 left, 2384 MB): budget 9, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 7080000 (4358466 pages from the captures)
[hos host] vk: 15 textures retired (1202 left, 2383 MB): budget 15, drawn over 0, dead targets 0
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1364 dirty pages from their submits' captures, 52022 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 877 dispatches, 40811 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8716 runs, 707.0 MB compared; 63 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19120 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: late pages found at the replay (600 takes): 38 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: presenter thread: the copy's fence 1.54 ms, the present 0.73 ms a present
[hos Presentation Thread] vk: capture: a submit 0.24 ms (lock 0.03, scan 0.04 over 594 ranges, 3363 MB), 309 pages written, 309 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.042; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 24 kept (104 MB), 11 spare buffers, 6879 texture pages copied (26 MB)
[hos host] vk: 3 textures retired (1199 left, 2382 MB): budget 3, drawn over 0, dead targets 0
[hos ModuleSystemWorker2] nvn: 1550000 recordings made, 69412 live copies (22.9 MB), 5950 texture copies; 5 command buffers' segments (4.3 MB)
[hos host] vk: replay reads of guest memory as submitted: 7100000 (4368349 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 7120000 (4380519 pages from the captures)
[hos host] vk: pool 0x7fab000000..0x7faf400000 (68 MB) trimmed to its 2 live part(s), 12 MB: the rest outside every game pool for 1856 submissions (2 trims, 92 MB in all)
[hos host] vk: 365 pools mirrored in VRAM (1137 MB)
[hos host] vk: program picks: 10034788 kept (3741539 by the last picks' textures, a uniform buffer bound since), 26288 new, 4138923 with other textures (of 14200000)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: memory: working set 15334 MB (peak 15952), private 21569 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 11 blocks 864 MB, 35 allocations 836 MB; used 3200 of 15579 MB; heap 1 (device): 166 blocks 5047 MB, 20136 allocations 4235 MB; used 7042 of 8466 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 107 images 339 MB; 65 free ranges, the largest 28 MB
[hos host] vk: image blocks (textures, compacted): 57 blocks 1824 MB, 1091 images 1418 MB; 410 free ranges, the largest 32 MB
[hos host] vk: render target images: 5753 reused from the pool, 3435 made new, 3326 destroyed from it, 993 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1294 with the pool empty (0 of them with targets still in the dead queue), 287 with none in it a present yet, 25 with none of a kind they can use, 467 with all smaller (the largest 51% short on average), 506 with all bigger (over 1.5 times), 855 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3134 idle
[hos host] vk: VRAM: 7042 MB used of 8466 MB allowed (811 MB of it free in the blocks); textures 2394 MB (budget 2107 MB), 1204 textures
[hos host] vk: budget: VRAM 5047 of 8466 MB (7042 MB by the driver, 1995 MB unaccounted): textures 2394, mirrors 1136, blocks free 811, shadows 316, other buffers 213, other images 111, staging 64; RAM 3184 MB (the process's private memory 21569 MB): imports 2248, capture spares 750, captures 86, retired imports 68, capture table 32, others 0
[hos host] vk: textures: decoded ASTC 1417 MB, rendered 540 MB (88 drawn in one frame, kept: 45085 KB), other 405 MB; used in the last 8 presents 1729 MB
[hos host] vk: decoded ASTC used again: 227 textures (552 MB) copied from ones the game finalized, not decoded; 227 of those retired, 6 listed; 1412 MB of blocks hashed in 239 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 7453 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 28 (8783 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 102x30x1 NVN 0x25: 1 (11 KB); 137x134x1 NVN 0x1: 1 (17 KB); 202x38x1 NVN 0x25: 1 (29 KB); 204x106x1 NVN 0x1: 1 (21 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 8 (1407 KB); 266x102x1 NVN 0x1: 1 (26 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB);
[hos host] vk: sparse mirrors: 128, 1136 MB with memory of 1568 MB (431 MB spared); blocks given memory later 10636, given back 13668; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 810 binds (224.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 203 taken from the spares, 162 made new; 237 kept as spares, 0 destroyed; 34 spares now, 1060 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 881 dirty pages from their submits' captures, 53011 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 854 dispatches, 41092 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8773 runs, 707.0 MB compared; 68 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19032 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos Presentation Thread] vk: presents after the CPU's signal (600): 267 waited for it, 3.40 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 85.5 ms; VRAM 6.9 of 8.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6168 queued, 6168 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15261 ms and 331.4 MB read in all); 8363 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11712 graphics pipelines (1 not in a cache), 713 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2220 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9002 (6168 ready, 0 loading), 16669 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13188 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 763 known, 6 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 184 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.04, present 0.18, draw 7.37, descriptors 3.17, texture lookups 0.14, uploads 0.13, pipelines 0.30, programs 1.27, commands 2.02, dispatch 1.01, memory sync 0.78, queue submit 0.00, slot wait 0.28
[hos host] vk: a frame: mirrors 365 (1136 MB): 5.41 MB synced in 1049.3 copies, 18.12 MB written back
[hos host] vk: 12600 presents, 57.4 fps; a frame: 1967 draws (0 skipped), 176 dispatches, 168 passes, 131 barriers, 6 submits, 0.5 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1204 textures (2394 MB)
[hos host] vk: host time in GPU resource waits: 4.08 ms a frame (fences 1.49, timelines 2.59, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 670.2 full barriers recorded, 285.0 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1100.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.8 ms a frame (the longest submission 8.2 ms); 0.4 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.4 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.59 MB unswizzled on the CPU in 0.10 ms
[hos host] vk: the render thread (600 presents, 17.14 ms a frame): its jobs 15.24 ms a frame (submits 5.71, presents 5.06, others 4.47), idle 1.90; the game waited for room at 33 presents, 0.42 ms each (0.02 a frame; 10.37 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 268 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 1481 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.50 ms after its hand-over (34.54 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.04 + 0.28 (the window's copy is in the idle), after it 1.19 + 0.44; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.31 parts a frame recorded in the signal's job, 74.5% of them took nothing that came with the signal; of the others, 2.12 a frame took pages of the signal's capture into a mirror (114.0 pages a frame) and 0.01 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 14 of 267 frames, its first three in 14
[hos host] vk: frame ahead, what came with the signal: 268 signals with a capture, each of 516.6 pages in 11.9 pools; between a frame's last submit and that signal 0.07 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (5 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.16+0.07, 0.49+0.14, 0.03+0.02, 0.04+0.02, 0.16+0.08, 0.20+0.07, 0.07+0.02, 0.03+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: replay reads of guest memory as submitted: 7140000 (4391213 pages from the captures)
[hos host] vk: render targets drawn over from another start, let go: 9956 so far (47138 MB), 8939 of them (42402 MB) between the passes
[hos host] vk: 3 textures retired (1210 left, 2396 MB): budget 3, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: CommandBufferSignalEvent #140000: event 0x100d6c8018 mode 0 value 140000 (lr 0x80660c94)
[hos host] vk: replay reads of guest memory as submitted: 7160000 (4401260 pages from the captures)
[hos Presentation Thread] vi: 12637 frames presented (56.8 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 48, 15.5-17.8 183, 17.8-22 32, 22-30 16, 30-40 4, over 40 2
[hos host] vk: 1 textures retired (1209 left, 2380 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 273 dirty pages from their submits' captures, 60488 from earlier ones', 10770 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 934 dispatches, 44325 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9328 runs, 760.6 MB compared; 69 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20555 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb5c00000 (44 MB, captured, state 2): 10770;
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1108 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1516 kept their virtual core (nothing there to wait for), 2836 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 7180000 (4413672 pages from the captures)
[hos host] vk: 240 mirrors given back (1810 MB in all); 1108 MB of mirrors now, 172 MB of them idle and kept for what the GPU wrote there alone
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 597 ranges, 3321 MB), 282 pages written, 282 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (74 MB), 9 spare buffers, 6880 texture pages copied (26 MB)
[hos host] vk: 4 textures retired (1205 left, 2379 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5118 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 7200000 (4423823 pages from the captures)
[hos host] vk: 1 textures retired (1204 left, 2379 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: readback pages refreshed from guest memory: 3934517 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 7220000 (4434612 pages from the captures)
[hos host] vk: sparse mirrors: 116, 1079 MB with memory of 1480 MB (400 MB spared); blocks given memory later 10636, given back 13668; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 817 binds (224.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 203 taken from the spares, 162 made new; 249 kept as spares, 0 destroyed; 46 spares now, 1148 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 654 dirty pages from their submits' captures, 57345 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 928 dispatches, 43261 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9149 runs, 743.6 MB compared; 239 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20050 in captures, 129 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos host] vk: program picks: 10185445 kept (3810041 by the last picks' textures, a uniform buffer bound since), 26290 new, 4188264 with other textures (of 14400000)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #200000: event 0x100d6c802c mode 0 value 141582 (now 141563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 7240000 (4444192 pages from the captures)
[hos host] vk: texture #17002 0x10004fb1e00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 600 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 1 textures retired (1203 left, 2378 MB): budget 1, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: recorder: 5668541 packets run in 10.0 s (busy 1.64 s, spun idle 0.11 s, slept 10394 times, 3169 of them under 200 us), 0 calls made directly (no room); 17297 batches over 583 frames (29.7 a frame, 328 packets a batch), 8506 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 73 times, 5.97 ms (0.1 times, 0.01 ms a frame); every thread 803 times, 57.59 ms in all, the longest 1.42 ms; vkWaitForFences 508/35.80 ms (max 0.50), vmaDestroyBuffer 164/11.77 ms (max 1.42), vmaFreeMemory 16/2.99 ms (max 0.75), vkQueueBindSparse 52/2.70 ms (max 0.40), the queue's lock 21/2.05 ms (max 0.25), vkDestroyImageView 24/1.62 ms (max 0.31)
[hos host] vk: recorder: of its busy 1.64 s, 0.41 s in parts' ends and submits (the longest 1.74 ms), 0.10 s of that blocked on the queue's lock (198 times, the longest 1.68 ms), so 0.31 s in the ends, submits and holds; 1.23 s running the other packets; 9723 packets a frame
[hos host] vk: a frame: 180 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.02, present 0.20, draw 6.25, descriptors 2.65, texture lookups 0.11, uploads 0.01, pipelines 0.27, programs 1.13, commands 1.71, dispatch 0.89, memory sync 0.44, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 365 (1079 MB): 5.86 MB synced in 1049.4 copies, 18.12 MB written back
[hos host] vk: 12720 presents, 60.1 fps; a frame: 1914 draws (0 skipped), 177 dispatches, 164 passes, 131 barriers, 6 submits, 0.2 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1203 textures (2378 MB)
[hos host] vk: host time in GPU resource waits: 1.88 ms a frame (fences 1.44, timelines 0.44, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 642.5 full barriers recorded, 279.0 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 1063.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.8 ms a frame (the longest submission 5.2 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 7260000 (4454760 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 162 dirty pages from their submits' captures, 59005 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 944 dispatches, 45270 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9490 runs, 778.0 MB compared; 55 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21156 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: texture #17003 0x100050a2000 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 606 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 1 textures retired (1202 left, 2377 MB): budget 1, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: replay reads of guest memory as submitted: 7280000 (4466694 pages from the captures)
[hos BaseProcExecutor2] vcores: 151691 takes, 26919 waits (mean 71 us, longest 40.2 ms, 0 over HOS_VCORE_LOG_MS), 13432 kicks, 13332 yields, 0 starved grants, 0 escapes in 5 s
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 590 ranges, 3308 MB), 270 pages written, 270 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (74 MB), 11 spare buffers, 8474 texture pages copied (33 MB)
[hos host] vk: replay reads of guest memory as submitted: 7300000 (4476711 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9824 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 113, 1067 MB with memory of 1464 MB (396 MB spared); blocks given memory later 10636, given back 13762; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 822 binds (226.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 203 taken from the spares, 162 made new; 252 kept as spares, 0 destroyed; 49 spares now, 1164 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 124 dirty pages from their submits' captures, 57618 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 940 dispatches, 44339 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9304 runs, 760.3 MB compared; 58 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20621 in captures, 1 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 7320000 (4488442 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] nvn: GPU waits in 5 s: 124 sync waits (408 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 6 textures retired (1196 left, 2357 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 7340000 (4497832 pages from the captures)
[hos host] vk: program picks: 10335912 kept (3879975 by the last picks' textures, a uniform buffer bound since), 26290 new, 4237797 with other textures (of 14600000)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1121 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1696 kept their virtual core (nothing there to wait for), 3168 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 239 dirty pages from their submits' captures, 59099 from earlier ones', 2478 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 912 dispatches, 43329 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8994 runs, 740.2 MB compared; 57 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19815 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fbc000000 (12 MB, captured, state 2): 2478;
[hos host] vk: replay reads of guest memory as submitted: 7360000 (4509281 pages from the captures)
[hos host] vk: 1 textures retired (1203 left, 2361 MB): budget 1, drawn over 0, dead targets 0
[hos BaseProcExecutor1] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 19190882 lookups, 141020 missed (0.73 %)
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 320000 so far (x1 0x117c952020 x2 0 x3 0x117c952000 x4 0x168040)
[hos host] vk: replay reads of guest memory as submitted: 7380000 (4518259 pages from the captures)
[hos host] vk: 5 textures retired (1198 left, 2346 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: a frame: 159 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.83, present 0.15, draw 6.25, descriptors 2.64, texture lookups 0.08, uploads 0.00, pipelines 0.26, programs 1.13, commands 1.73, dispatch 0.90, memory sync 0.43, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 366 (1067 MB): 5.54 MB synced in 1051.7 copies, 18.03 MB written back
[hos host] vk: 12840 presents, 59.9 fps; a frame: 1889 draws (0 skipped), 171 dispatches, 165 passes, 131 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1198 textures (2346 MB)
[hos host] vk: host time in GPU resource waits: 1.46 ms a frame (fences 1.34, timelines 0.11, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 644.9 full barriers recorded, 281.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 1024.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.0 ms a frame (the longest submission 4.2 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 115, 1069 MB with memory of 1476 MB (406 MB spared); blocks given memory later 10637, given back 13762; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 827 binds (227.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 253 kept as spares, 0 destroyed; 47 spares now, 1152 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 252 dirty pages from their submits' captures, 54746 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 42145 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8393 runs, 715.2 MB compared; 52 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19455 in captures, 1 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos host] vk: 55 textures retired (1154 left, 2237 MB): budget 55, drawn over 0, dead targets 0
[hos sead::MainThread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.04 over 590 ranges, 3308 MB), 303 pages written, 303 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (124 MB), 12 spare buffers, 6950 texture pages copied (27 MB)
[hos host] vk: replay reads of guest memory as submitted: 7400000 (4530623 pages from the captures)
[hos host] vk: 21 textures retired (1135 left, 2193 MB): budget 21, drawn over 0, dead targets 0
[hos host] vk: reaper: 447 images (671 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 136 (191 MB)
[hos host] vk: 22 textures retired (1113 left, 2132 MB): budget 22, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 7420000 (4542497 pages from the captures)
[hos host] vk: queues ordered by their events: 78000 submissions stopped at a wait, 77951 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 204385, not met 0; captures at a CPU event's wait 8093, 0 of them at a wait met on arrival
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14323 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos ModuleSystemWorker2] nvn: CommandBufferSignalEvent #145000: event 0x100d6c8024 mode 0 value 145000 (lr 0x80665068)
[hos host] vk: readback pages refreshed from guest memory: 4002129 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 923 dirty pages from their submits' captures, 51522 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 850 dispatches, 40169 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8067 runs, 679.3 MB compared; 85 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18273 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: program picks: 10487821 kept (3952924 by the last picks' textures, a uniform buffer bound since), 26290 new, 4285888 with other textures (of 14800000)
[hos host] vk: replay reads of guest memory as submitted: 7440000 (4553997 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #205000: event 0x100d6c802c mode 0 value 145382 (now 145363) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: 24 textures retired (1093 left, 2102 MB): budget 24, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 7460000 (4562497 pages from the captures)
[hos host] vk: sparse mirrors: 115, 1069 MB with memory of 1476 MB (406 MB spared); blocks given memory later 10637, given back 13762; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 827 binds (227.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 253 kept as spares, 0 destroyed; 47 spares now, 1152 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 730 dirty pages from their submits' captures, 54150 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 874 dispatches, 40206 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8121 runs, 680.4 MB compared; 65 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18672 in captures, 1 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos Presentation Thread] vi: 12938 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 25, 15.5-17.8 254, 17.8-22 21, 22-30 1, 30-40 0, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 7480000 (4573281 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.01, scan 0.05 over 576 ranges, 3303 MB), 290 pages written, 290 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.029; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (112 MB), 10 spare buffers, 10983 texture pages copied (42 MB)
[hos host] vk: a frame: 163 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.73, present 0.15, draw 6.84, descriptors 2.86, texture lookups 0.11, uploads 0.02, pipelines 0.29, programs 1.24, commands 1.90, dispatch 0.87, memory sync 0.60, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 368 (1054 MB): 5.54 MB synced in 1055.8 copies, 17.90 MB written back
[hos host] vk: 12960 presents, 60.0 fps; a frame: 2105 draws (0 skipped), 163 dispatches, 178 passes, 131 barriers, 6 submits, 0.3 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1096 textures (2103 MB)
[hos host] vk: host time in GPU resource waits: 2.64 ms a frame (fences 1.37, timelines 1.27, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 700.1 full barriers recorded, 297.3 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.9 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1099.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.9 ms a frame (the longest submission 3.8 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 7500000 (4583982 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 303 dirty pages from their submits' captures, 55149 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 950 dispatches, 41871 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8359 runs, 698.7 MB compared; 74 pages read into the shadows; 316 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19438 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb5c00000 (44 MB, captured, state 2): 1;
[hos sead::MainThread] vk: readbacks of GPU results: 1118 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 10 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1482 kept their virtual core (nothing there to wait for), 2614 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: (its written part: 3 of 3 KB readable, 1560 bytes not zero; levels L0)
[hos host] vk: texture 0x10d787b400 2480x2480x1 x1 levels 1 NVN 0x4b (6045 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it
[hos host] vk: (its written part: 3 of 3 KB readable, 1560 bytes not zero; levels L0)
[hos host] vk: texture 0x10d787b400 2480x2480x1 x1 levels 1 NVN 0x4b (6045 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it
[hos host] vk: program picks: 10641908 kept (4026683 by the last picks' textures, a uniform buffer bound since), 26290 new, 4331801 with other textures (of 15000000)
[hos host] vk: replay reads of guest memory as submitted: 7520000 (4595958 pages from the captures)
[hos host] vk: VRAM compaction 5 begins: the textures' blocks hold 1696 MB, 552 MB of it free space
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21770 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06
[hos host] vk: sparse mirrors: 113, 1053 MB with memory of 1452 MB (398 MB spared); blocks given memory later 10638, given back 13785; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 833 binds (228.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 255 kept as spares, 0 destroyed; 49 spares now, 1176 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 345 dirty pages from their submits' captures, 54029 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 891 dispatches, 41518 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8488 runs, 700.9 MB compared; 67 pages read into the shadows; 316 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19270 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 7540000 (4605855 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture slots registered again at another size: 600000 (of 759767 registrations; slot 7164 1280x720 -> 2560x1440)
[hos ModuleSystemWorker3] nvn: 1600000 recordings made, 70455 live copies (23.0 MB), 5950 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: replay reads of guest memory as submitted: 7560000 (4614180 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.05 over 574 ranges, 3309 MB), 282 pages written, 282 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.032; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.024, bytes copied 0.000; 24 kept (126 MB), 12 spare buffers, 9513 texture pages copied (37 MB)
[hos BaseProcExecutor0] vcores: 154595 takes, 24209 waits (mean 67 us, longest 11.5 ms, 0 over HOS_VCORE_LOG_MS), 11597 kicks, 11466 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 759 dirty pages from their submits' captures, 52291 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 40800 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8563 runs, 701.9 MB compared; 64 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18666 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: VRAM compaction 5 over: 386 textures moved (408 MB) in 33 passes, 0 left where they were; 15 blocks (480 MB) given back to the driver, 3328 MB in all so far
[hos host] vk: replay reads of guest memory as submitted: 7580000 (4625021 pages from the captures)
[hos host] vk: program picks: 10798108 kept (4104657 by the last picks' textures, a uniform buffer bound since), 26290 new, 4375601 with other textures (of 15200000)
[hos host] vk: replay reads of guest memory as submitted: 7600000 (4635916 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 4069375 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: a frame: 142 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.14, present 0.37, draw 6.41, descriptors 2.78, texture lookups 0.19, uploads 0.02, pipelines 0.27, programs 1.11, commands 1.75, dispatch 0.89, memory sync 0.52, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 368 (1053 MB): 5.42 MB synced in 1059.9 copies, 17.98 MB written back
[hos host] vk: 13080 presents, 60.0 fps; a frame: 2044 draws (0 skipped), 168 dispatches, 178 passes, 131 barriers, 6 submits, 0.2 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1106 textures (2106 MB)
[hos host] vk: host time in GPU resource waits: 1.83 ms a frame (fences 1.32, timelines 0.50, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 701.1 full barriers recorded, 297.0 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1042.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.9 ms a frame (the longest submission 4.8 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: texture #16653 0x116e25b000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 13081 (drawn at 11285)
[hos host] vk: sparse mirrors: 113, 1053 MB with memory of 1452 MB (398 MB spared); blocks given memory later 10638, given back 13786; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 834 binds (229.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 255 kept as spares, 0 destroyed; 49 spares now, 1176 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 161 dirty pages from their submits' captures, 55374 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 42447 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8687 runs, 718.1 MB compared; 72 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19743 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 83 sync waits (209 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #210000: event 0x100d6c802c mode 0 value 149182 (now 149163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: a submission's parts ended: 51879 where the game had ended its render pass, 8121 at 512 draws
[hos host] vk: replay reads of guest memory as submitted: 7620000 (4644358 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17923 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 577 dirty pages from their submits' captures, 54482 from earlier ones', 32 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 875 dispatches, 41865 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8421 runs, 700.7 MB compared; 54 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19328 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool: 0x7fe2800000 (4 MB, captured, state 2): 32;
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1118 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1451 kept their virtual core (nothing there to wait for), 3157 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 7640000 (4655028 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.04 over 573 ranges, 3319 MB), 285 pages written, 285 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 6150 texture pages copied (24 MB)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #150000: event 0x100d6c802c mode 0 value 149999 (lr 0x80667688)
[hos host] vk: replay reads of guest memory as submitted: 7660000 (4664788 pages from the captures)
[hos host] vk: sparse mirrors: 113, 1052 MB with memory of 1452 MB (399 MB spared); blocks given memory later 10638, given back 13799; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 835 binds (229.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 255 kept as spares, 0 destroyed; 49 spares now, 1176 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 388 dirty pages from their submits' captures, 54538 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 881 dispatches, 41999 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8435 runs, 701.6 MB compared; 54 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19361 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos host] vk: presenter thread: the copy's fence 1.35 ms, the present 0.74 ms a present
[hos host] vk: program picks: 10954590 kept (4186845 by the last picks' textures, a uniform buffer bound since), 26292 new, 4419117 with other textures (of 15400000)
[hos ModuleSystemWorker2] barrier: 209606 looks: 78534 found the masks moved already, 128670 waited until they moved, 2402 for the whole 1 ms (a wait 298 us on average); 208977 wakes
[hos host] vk: replay reads of guest memory as submitted: 7680000 (4675370 pages from the captures)
[hos host] vk: 256 mirrors given back (1863 MB in all); 1070 MB of mirrors now, 172 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: replay reads of guest memory as submitted: 7700000 (4686225 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 517 dirty pages from their submits' captures, 57933 from earlier ones', 24 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 902 dispatches, 42184 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8429 runs, 701.5 MB compared; 63 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19236 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fbc000000 (12 MB, captured, state 2): 19; 0x1043000000 (4 MB, captured, state 2): 5;
[hos host] vk: memory: working set 15224 MB (peak 15952), private 20919 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms)
[hos host] vk: memory heaps: heap 0 (system): 12 blocks 864 MB, 35 allocations 840 MB; used 3200 of 15579 MB; heap 1 (device): 148 blocks 4471 MB, 19027 allocations 3928 MB; used 6466 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 17 blocks 544 MB, 112 images 355 MB; 72 free ranges, the largest 17 MB
[hos host] vk: image blocks (textures, compacted): 39 blocks 1248 MB, 1030 images 1162 MB; 274 free ranges, the largest 19 MB
[hos host] vk: render target images: 5753 reused from the pool, 3478 made new, 3364 destroyed from it, 993 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1325 with the pool empty (0 of them with targets still in the dead queue), 288 with none in it a present yet, 25 with none of a kind they can use, 470 with all smaller (the largest 51% short on average), 509 with all bigger (over 1.5 times), 860 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3172 idle
[hos host] vk: VRAM: 6466 MB used of 11474 MB allowed (542 MB of it free in the blocks); textures 2153 MB (budget 7158 MB), 1148 textures
[hos host] vk: budget: VRAM 4471 of 11474 MB (6466 MB by the driver, 1995 MB unaccounted): textures 2153, mirrors 1070, blocks free 542, shadows 316, other buffers 213, other images 111, staging 64; RAM 3189 MB (the process's private memory 20919 MB): imports 2248, capture spares 754, captures 86, retired imports 68, capture table 33, others 0
[hos host] vk: textures: decoded ASTC 1218 MB, rendered 563 MB (93 drawn in one frame, kept: 70981 KB), other 332 MB; used in the last 8 presents 1737 MB
[hos host] vk: decoded ASTC used again: 333 textures (771 MB) copied from ones the game finalized, not decoded; 333 of those retired, 8 listed; 1456 MB of blocks hashed in 249 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 8053 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 29 (8959 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 102x30x1 NVN 0x25: 1 (11 KB); 137x134x1 NVN 0x1: 1 (17 KB); 202x38x1 NVN 0x25: 1 (29 KB); 204x106x1 NVN 0x1: 1 (21 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 9 (1583 KB); 266x102x1 NVN 0x1: 1 (26 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 203 waited for it, 1.83 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 27.0 ms; VRAM 6.3 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6168 queued, 6168 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15261 ms and 331.4 MB read in all); 8363 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11712 graphics pipelines (1 not in a cache), 713 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2226 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9002 (6168 ready, 0 loading), 16176 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13188 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 763 known, 6 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 143 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.26, present 0.16, draw 6.55, descriptors 2.94, texture lookups 0.26, uploads 0.20, pipelines 0.27, programs 1.08, commands 1.76, dispatch 0.85, memory sync 0.50, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 371 (1070 MB): 5.78 MB synced in 1073.2 copies, 17.99 MB written back
[hos host] vk: 13200 presents, 60.1 fps; a frame: 2094 draws (0 skipped), 168 dispatches, 173 passes, 131 barriers, 6 submits, 1.1 uploads (0.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1148 textures (2153 MB)
[hos host] vk: host time in GPU resource waits: 2.56 ms a frame (fences 1.56, timelines 1.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 679.7 full barriers recorded, 290.4 left out (nothing recorded since the one before), 13.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 1131.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.5 ms a frame (the longest submission 4.6 ms); 1.2 textures made a frame: colour targets 0.1 (0.4 MB), depth 0.3 (1.1 MB), ASTC or BC 0.8 (1.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.86 MB unswizzled on the CPU in 0.14 ms
[hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 15.01 ms a frame (submits 6.99, presents 4.63, others 3.39), idle 1.66; the game waited for room at 11 presents, 0.10 ms each (0.00 a frame; 5.74 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 209 captures at a CPU signal; the submits' waits read in 0.041 ms a frame; readbacks: 1193 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.24 ms after its hand-over (25.73 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.64 + 0.28 (the window's copy is in the idle), after it 1.00 + 0.31; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.78 parts a frame recorded in the signal's job, 80.8% of them took nothing that came with the signal; of the others, 1.68 a frame took pages of the signal's capture into a mirror (26.2 pages a frame) and 0.01 pages read as guest memory had them (12.0 pages a frame); a frame's first part took nothing in 40 of 209 frames, its first three in 40
[hos host] vk: frame ahead, what came with the signal: 209 signals with a capture, each of 66.9 pages in 9.0 pools; between a frame's last submit and that signal 0.01 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.18+0.06, 0.24+0.13, 0.03+0.02, 0.03+0.01, 0.09+0.02, 0.30+0.02, 0.05+0.03, 0.06+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: render targets drawn over from another start, let go: 9992 so far (47293 MB), 8975 of them (42557 MB) between the passes
[hos host] vk: replay reads of guest memory as submitted: 7720000 (4697297 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17572 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] vk: capture: a submit 0.20 ms (lock 0.01, scan 0.05 over 565 ranges, 3326 MB), 341 pages written, 341 copied (1.3 MB)
[hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 8826 texture pages copied (34 MB)
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 340000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 7740000 (4706447 pages from the captures)
[hos Presentation Thread] vi: 13239 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 26, 15.5-17.8 254, 17.8-22 19, 22-30 2, 30-40 0, over 40 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 115, 1065 MB with memory of 1476 MB (410 MB spared); blocks given memory later 10638, given back 13853; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 844 binds (230.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 210 taken from the spares, 162 made new; 257 kept as spares, 0 destroyed; 47 spares now, 1152 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 231 dirty pages from their submits' captures, 57408 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 945 dispatches, 44300 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9496 runs, 760.4 MB compared; 87 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20778 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos host] vk: 2 textures retired (1157 left, 2161 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 7760000 (4717419 pages from the captures)
[hos host] vk: 6 textures retired (1151 left, 2136 MB): budget 0, drawn over 6, dead targets 0
[hos host] vk: program picks: 11104857 kept (4257044 by the last picks' textures, a uniform buffer bound since), 26302 new, 4468840 with other textures (of 15600000)
[hos host] vk: readback pages refreshed from guest memory: 4137034 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 7780000 (4728206 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 481 dirty pages from their submits' captures, 58992 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 977 dispatches, 45238 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9686 runs, 764.4 MB compared; 88 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21183 in captures, 2 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1107 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 10 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1334 kept their virtual core (nothing there to wait for), 2762 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #215000: event 0x100d6c802c mode 0 value 152982 (now 152963) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.07
[hos host] vk: replay reads of guest memory as submitted: 7800000 (4738134 pages from the captures)
[hos host] vk: recorder: 6141123 packets run in 10.0 s (busy 1.68 s, spun idle 0.11 s, slept 10520 times, 3228 of them under 200 us), 0 calls made directly (no room); 18321 batches over 600 frames (30.5 a frame, 335 packets a batch), 9010 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 86 times, 5.67 ms (0.1 times, 0.01 ms a frame); every thread 669 times, 38.33 ms in all, the longest 0.67 ms; vkWaitForFences 509/27.36 ms (max 0.48), vmaDestroyBuffer 33/3.13 ms (max 0.67), vkDestroyImageView 61/2.56 ms (max 0.38), vmaFreeMemory 21/2.23 ms (max 0.49), vkQueueBindSparse 31/1.58 ms (max 0.30), the queue's lock 6/1.08 ms (max 0.64)
[hos host] vk: recorder: of its busy 1.68 s, 0.43 s in parts' ends and submits (the longest 1.56 ms), 0.11 s of that blocked on the queue's lock (243 times, the longest 1.46 ms), so 0.32 s in the ends, submits and holds; 1.25 s running the other packets; 10235 packets a frame
[hos host] vk: a frame: 168 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.28, present 0.18, draw 5.67, descriptors 2.45, texture lookups 0.09, uploads 0.09, pipelines 0.24, programs 0.99, commands 1.55, dispatch 0.89, memory sync 0.52, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 374 (1065 MB): 5.46 MB synced in 1063.7 copies, 18.16 MB written back
[hos host] vk: 13320 presents, 60.0 fps; a frame: 1699 draws (0 skipped), 179 dispatches, 155 passes, 131 barriers, 6 submits, 0.3 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1153 textures (2137 MB)
[hos host] vk: host time in GPU resource waits: 1.39 ms a frame (fences 1.21, timelines 0.18, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 613.6 full barriers recorded, 272.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.7 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 937.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.9 ms a frame (the longest submission 4.4 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.56 MB unswizzled on the CPU in 0.08 ms
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 549 ranges, 3336 MB), 271 pages written, 271 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (86 MB), 10 spare buffers, 3822 texture pages copied (14 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5813 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 117, 1065 MB with memory of 1484 MB (418 MB spared); blocks given memory later 10640, given back 13858; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 851 binds (230.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 212 taken from the spares, 162 made new; 257 kept as spares, 0 destroyed; 45 spares now, 1144 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 58496 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 928 dispatches, 44688 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9585 runs, 762.8 MB compared; 60 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21106 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2256 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 7820000 (4750100 pages from the captures)
[hos Presentation Thread] nvn: QueueSubmitCommands #80000: queue 000010100D5EB020 (its #13331), 8 handles (lr 0x805282f4)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos BaseProcExecutor0] vcores: 153669 takes, 22243 waits (mean 71 us, longest 12.5 ms, 0 over HOS_VCORE_LOG_MS), 10624 kicks, 10535 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 7840000 (4760843 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 737 dirty pages from their submits' captures, 57067 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 954 dispatches, 44283 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9284 runs, 759.3 MB compared; 60 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20990 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x105dc00000 (4 MB, captured, state 2): 2;
[hos host] vk: replay reads of guest memory as submitted: 7860000 (4773197 pages from the captures)
[hos host] vk: program picks: 11254306 kept (4325002 by the last picks' textures, a uniform buffer bound since), 26312 new, 4519381 with other textures (of 15800000)
[hos host] vk: replay reads of guest memory as submitted: 7880000 (4785337 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 57 sync waits (145 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #155000: event 0x100d6c8038 mode 0 value 154999 (lr 0x8066189c)
[hos host] vk: sparse mirrors: 119, 1080 MB with memory of 1536 MB (455 MB spared); blocks given memory later 10642, given back 13858; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 854 binds (231.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 214 taken from the spares, 162 made new; 257 kept as spares, 0 destroyed; 43 spares now, 1092 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 989 dirty pages from their submits' captures, 56617 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 948 dispatches, 44155 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8673 runs, 748.1 MB compared; 78 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20758 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 189 live (2260 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 7900000 (4796852 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.18 ms (lock 0.00, scan 0.05 over 545 ranges, 3414 MB), 297 pages written, 297 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.039; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.028, bytes copied 0.000; 24 kept (90 MB), 11 spare buffers, 11044 texture pages copied (43 MB)
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 15368935 lookups, 134707 missed (0.88 %)
[hos host] vk: texture 0x10001e00400 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: replay reads of guest memory as submitted: 7920000 (4809590 pages from the captures)
[hos host] vk: pool 0x1011f97c00+0xf00000 imported in 4.9 ms as 0x100f800000..0x1014000000 (227 imports, 693 ms in all)
[hos host] vk: pool 0x100f800000..0x1014000000 took in 2 pools
[hos host] vk: texture 0x10001e00400 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since)
[hos host] vk: 1 textures retired (1366 left, 2599 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: a frame: 195 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.69, present 0.18, draw 6.81, descriptors 3.61, texture lookups 1.23, uploads 0.94, pipelines 0.24, programs 0.97, commands 1.54, dispatch 0.92, memory sync 0.60, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 389 (1107 MB): 6.34 MB synced in 1053.6 copies, 17.94 MB written back
[hos host] vk: 13440 presents, 56.1 fps; a frame: 1688 draws (0 skipped), 166 dispatches, 156 passes, 132 barriers, 6 submits, 2.3 uploads (2.7 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1409 textures (2642 MB)
[hos host] vk: host time in GPU resource waits: 4.18 ms a frame (fences 1.76, timelines 2.43, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 618.5 full barriers recorded, 273.0 left out (nothing recorded since the one before), 14.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 26.3; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.4, an op 955.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.5 ms a frame (the longest submission 5.6 ms); 3.2 textures made a frame: colour targets 0.9 (2.7 MB), depth 0.2 (1.2 MB), ASTC or BC 2.1 (3.7 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.81 MB unswizzled on the CPU in 0.30 ms
[hos host] [debug] vk: a staging arena of 22 MB took 5.8 ms to make
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 27955 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] vk: readbacks of GPU results: 1109 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 23 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1589 kept their virtual core (nothing there to wait for), 3275 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: reaper: 214 images (574 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 23 (74 MB)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1828 dirty pages from their submits' captures, 52671 from earlier ones', 76 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 904 dispatches, 41559 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8237 runs, 697.7 MB compared; 71 pages read into the shadows; 284 MB of shadows (51 MB of pages known, in 101 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19990 in captures, 1 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x7f98800000 (8 MB, captured, state 2): 55; 0x7f9ec00000 (4 MB, captured, state 2): 14; 0x7f8e000000 (4 MB, captured, state 2): 4; 0x7fa9000000 (4 MB, captured, state 2): 3;
[hos host] vk: texture #15756 0x1001edcae00 122x98x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 5900 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15754 0x1001edcf000 58x34x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 5900 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15733 0x10000032200 800x62x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 4441 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 280 textures retired (1218 left, 2513 MB): budget 280, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 3 finalized, 0 MB
[hos host] vk: replay reads of guest memory as submitted: 7940000 (4821890 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 45 textures retired (1193 left, 2473 MB): budget 45, drawn over 0, dead targets 0
[hos ModuleSystemWorker5] nvn: 1650000 recordings made, 70958 live copies (23.7 MB), 5974 texture copies; 4 command buffers' segments (4.2 MB)
[hos host] vk: replay reads of guest memory as submitted: 7960000 (4832405 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 4204161 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 132, 1113 MB with memory of 1604 MB (490 MB spared); blocks given memory later 10646, given back 13865; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 874 binds (235.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 228 taken from the spares, 163 made new; 259 kept as spares, 0 destroyed; 31 spares now, 1096 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1161 dirty pages from their submits' captures, 50591 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 841 dispatches, 39635 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8031 runs, 665.4 MB compared; 54 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18152 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: 41 textures retired (1181 left, 2430 MB): budget 0, drawn over 0, dead targets 41
[hos host] vk: program picks: 11405083 kept (4392933 by the last picks' textures, a uniform buffer bound since), 26312 new, 4568604 with other textures (of 16000000)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #220000: event 0x100d6c802c mode 0 value 156782 (now 156763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 7980000 (4844384 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 547 ranges, 3442 MB), 276 pages written, 276 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (86 MB), 10 spare buffers, 2513 texture pages copied (9 MB)
[hos host] vk: write-backs left out (600 submits): 1001, 4442.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 208 dirty pages from their submits' captures, 55305 from earlier ones', 103 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 921 dispatches, 42577 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8734 runs, 720.0 MB compared; 53 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19795 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fe1000000 (4 MB, captured, state 2): 45; 0x1043400000 (4 MB, captured, state 2): 35; 0x7fe2800000 (4 MB, captured, state 2): 11; 0x105dc00000 (4 MB, captured, state 2): 4; 0x7f9ac00000 (20 MB, captured, state 2): 3; 0x7fa3400000 (4 MB, captured, state 2): 2; 0x1044400000 (4 MB, captured, state 2): 2; 0x7fbc000000 (12 MB, captured, state 2): 1;
[hos Presentation Thread] vi: 13526 frames presented (57.3 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 49, 15.5-17.8 199, 17.8-22 20, 22-30 10, 30-40 6, over 40 3
[hos host] vk: replay reads of guest memory as submitted: 8000000 (4857068 pages from the captures)
[hos host] vk: texture #18378 0x100020e5a00 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 13549 (drawn at 13488)
[hos host] vk: 2 textures retired (1205 left, 2437 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8020000 (4868864 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 4098 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a frame: 207 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.20, present 0.25, draw 6.47, descriptors 3.16, texture lookups 0.79, uploads 0.55, pipelines 0.25, programs 0.97, commands 1.61, dispatch 0.92, memory sync 0.46, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 391 (1110 MB): 5.46 MB synced in 1056.9 copies, 18.01 MB written back
[hos host] vk: 13560 presents, 57.5 fps; a frame: 1681 draws (0 skipped), 172 dispatches, 165 passes, 133 barriers, 6 submits, 1.2 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1205 textures (2437 MB)
[hos host] vk: host time in GPU resource waits: 3.75 ms a frame (fences 1.76, timelines 1.99, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 655.2 full barriers recorded, 282.5 left out (nothing recorded since the one before), 15.2 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 26.2; 0x12: 69.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.8 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.6, an op 961.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.3 ms a frame (the longest submission 4.1 ms); 1.4 textures made a frame: colour targets 0.2 (0.5 MB), depth 0.1 (0.4 MB), ASTC or BC 1.1 (1.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.44 MB unswizzled on the CPU in 0.07 ms
[hos host] vk: sparse mirrors: 131, 1110 MB with memory of 1600 MB (490 MB spared); blocks given memory later 10647, given back 13865; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 876 binds (235.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 228 taken from the spares, 163 made new; 260 kept as spares, 0 destroyed; 32 spares now, 1100 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 154 dirty pages from their submits' captures, 57517 from earlier ones', 78 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 976 dispatches, 43965 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9363 runs, 758.8 MB compared; 60 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21542 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 78;
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 8040000 (4880458 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 8060000 (4891665 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 542 ranges, 3372 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (92 MB), 11 spare buffers, 4245 texture pages copied (16 MB)
[hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 125 dirty pages from their submits' captures, 57528 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 909 dispatches, 43912 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9323 runs, 743.8 MB compared; 54 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20829 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: program picks: 11555015 kept (4458513 by the last picks' textures, a uniform buffer bound since), 26315 new, 4618669 with other textures (of 16200000)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1120 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1337 kept their virtual core (nothing there to wait for), 2759 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 8080000 (4902253 pages from the captures)
[hos ModuleSystemWorker1] nvn: DrawMeshTasksIndirect: 360000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040)
[hos PreActorModelCompose] vcores: 148170 takes, 22496 waits (mean 68 us, longest 7.4 ms, 0 over HOS_VCORE_LOG_MS), 10579 kicks, 10536 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 1 textures retired (1210 left, 2436 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 132, 1113 MB with memory of 1604 MB (490 MB spared); blocks given memory later 10651, given back 13865; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 880 binds (235.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 230 taken from the spares, 163 made new; 261 kept as spares, 0 destroyed; 31 spares now, 1096 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: 76 textures retired (1203 left, 2449 MB): budget 76, drawn over 0, dead targets 0
[hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 308 dirty pages from their submits' captures, 57390 from earlier ones', 23 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 967 dispatches, 43720 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9305 runs, 760.0 MB compared; 53 pages read into the shadows; 284 MB of shadows (51 MB of pages known, in 102 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21313 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa6800000 (4 MB, captured, state 2): 6; 0x7fa2000000 (8 MB, captured, state 2): 6; 0x7f9ec00000 (4 MB, captured, state 2): 6; 0x7fb5c00000 (44 MB, captured, state 2): 5;
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 8100000 (4913313 pages from the captures)
[hos host] vk: 20 textures retired (1217 left, 2474 MB): budget 20, drawn over 0, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10246 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #160000: event 0x100d6c8044 mode 0 value 159999 (lr 0x802742f0)
[hos host] vk: late pages' fix (600 recordings with one): of 1230 captures with late pages 1229 fixed (74.3 pages, 74.9 fixes a capture), waited as before: 0 no pool, 1 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 470 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 576 blocks (36864 KB); of the reads on the CPU, 1 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 171, of an earlier recording 298, not recorded yet 0, of this part 0; pages fixed from stamps 90276, from snapshots 1076; the shadows: 1228694 pages synced from captures, 1197451 made known by it, 31243 left unknown because the replay or the collection wrote them since
[hos sead::MainThread] nvn: GPU waits in 5 s: 91 sync waits (564 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 8120000 (4924719 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #15753 0x1001ede6200 122x98x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4671 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15750 0x1001edea400 58x34x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4671 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 3 textures retired (1270 left, 2583 MB): budget 3, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 2 finalized, 0 MB
[hos OptimizeChangeBanc] nvn: 1700000 recordings made, 69995 live copies (23.2 MB), 6003 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: a frame: 167 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.77, present 0.25, draw 6.35, descriptors 2.97, texture lookups 0.60, uploads 0.41, pipelines 0.25, programs 1.03, commands 1.64, dispatch 0.94, memory sync 0.93, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 394 (1128 MB): 7.65 MB synced in 1070.1 copies, 18.09 MB written back
[hos host] vk: 13680 presents, 54.3 fps; a frame: 1706 draws (0 skipped), 175 dispatches, 160 passes, 131 barriers, 6 submits, 1.2 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1270 textures (2583 MB)
[hos host] vk: host time in GPU resource waits: 3.46 ms a frame (fences 1.77, timelines 1.69, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 645.4 full barriers recorded, 277.8 left out (nothing recorded since the one before), 13.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 919.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.8 ms a frame (the longest submission 5.5 ms); 2.0 textures made a frame: colour targets 0.3 (2.1 MB), depth 0.4 (2.0 MB), ASTC or BC 1.2 (1.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.60 MB unswizzled on the CPU in 0.11 ms
[hos host] vk: readback pages refreshed from guest memory: 4271702 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 949, 4216.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 31530 dirty pages from their submits' captures, 89124 from earlier ones', 20 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1085 dispatches, 45848 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8127 runs, 682.7 MB compared; 108 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18515 in captures, 2 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool: 0x7fe2800000 (4 MB, captured, state 2): 16; 0x7f98800000 (8 MB, captured, state 2): 2; 0x105dc00000 (4 MB, captured, state 2): 2;
[hos host] vk: queues ordered by their events: 84000 submissions stopped at a wait, 83952 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 224729, not met 0; captures at a CPU event's wait 8398, 0 of them at a wait met on arrival
[hos host] vk: texture #15760 0x1001edbea00 146x105x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4681 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15757 0x1001edc4c00 82x41x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4681 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 7 textures retired (1287 left, 2615 MB): budget 7, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 2 finalized, 0 MB
[hos host] vk: replay reads of guest memory as submitted: 8140000 (4934601 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.40 ms (lock 0.04, scan 0.06 over 572 ranges, 3534 MB), 772 pages written, 772 copied (3.0 MB)
[hos host] vk: a capture taken (ms): all 0.039; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.020, bytes copied 0.000; 25 kept (110 MB), 10 spare buffers, 7928 texture pages copied (30 MB)
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #225000: event 0x100d6c802c mode 0 value 160582 (now 160563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.04, writes collected 0.16
[hos host] vk: texture #15739 0x1001ee02800 42x41x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4691 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 1 textures retired (1293 left, 2643 MB): budget 1, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: texture #17983 0x10008942600 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13708 (drawn at 13647)
[hos host] vk: texture #18670 0x1000a0b2c00 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 13708 (drawn at 13647)
[hos host] vk: texture #18528 0x10008612200 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13709 (drawn at 13647)
[hos host] vk: texture #15592 0x1001ef3d200 640x360x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 8148 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15743 0x1001edfbc00 74x32x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15742 0x1001edfe600 106x105x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15747 0x1001edf2800 42x41x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15749 0x1001edee600 106x105x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #15746 0x1001edf5a00 138x96x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 8 textures retired (1293 left, 2668 MB): budget 6, drawn over 2, dead targets 0; GPU-drawn with no copy kept: 0 live, 6 finalized, 0 MB
[hos host] vk: program picks: 11706106 kept (4531198 by the last picks' textures, a uniform buffer bound since), 26319 new, 4667574 with other textures (of 16400000)
[hos rmbl::VibrationThread] hid: 45056 vibration values, strongest 0.00 since the last count
[hos host] vk: replay reads of guest memory as submitted: 8160000 (4946228 pages from the captures)
[hos host] vk: texture #18778 0x10008612200 839x839 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13719 (drawn at 13657)
[hos host] vk: texture #15337 0x1001f05fc00 637x133x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4573 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 15 textures retired (1283 left, 2660 MB): budget 10, drawn over 5, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: texture #18779 0x10008612200 853x853 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13720 (drawn at 13658)
[hos host] vk: texture #18781 0x10008612200 872x872 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13721 (drawn at 13659)
[hos host] vk: texture #18782 0x10008612200 890x890 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13722 (drawn at 13660)
[hos host] vk: sparse mirrors: 133, 1132 MB with memory of 1612 MB (479 MB spared); blocks given memory later 11018, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 907 binds (242.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 231 taken from the spares, 163 made new; 261 kept as spares, 0 destroyed; 30 spares now, 1088 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1311 dirty pages from their submits' captures, 52085 from earlier ones', 22 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 875 dispatches, 40518 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8033 runs, 680.4 MB compared; 100 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18756 in captures, 2 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool: 0x1043400000 (4 MB, captured, state 2): 14; 0x1044400000 (4 MB, captured, state 2): 8;
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: 51 textures retired (1241 left, 2619 MB): budget 43, drawn over 8, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8180000 (4957463 pages from the captures)
[hos host] vk: 39 textures retired (1246 left, 2633 MB): budget 34, drawn over 5, dead targets 0
[hos host] vk: 8 new ASTC textures left for a later frame (this frame's uploads took their budget; 12000 so far)
[hos host] vk: 23 textures retired (1229 left, 2605 MB): budget 14, drawn over 9, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8200000 (4969343 pages from the captures)
[hos host] vk: 9 textures retired (1228 left, 2604 MB): budget 3, drawn over 6, dead targets 0
[hos host] vk: texture #16156 0x116d1f9000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 13762 (drawn at 9483)
[hos host] vk: presenter thread: the copy's fence 1.41 ms, the present 0.71 ms a present
[hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1346 dirty pages from their submits' captures, 52793 from earlier ones', 822 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 869 dispatches, 40816 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8070 runs, 680.5 MB compared; 108 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18388 in captures, 1 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa5c00000 (4 MB, captured, state 2): 796; 0x7fa3800000 (8 MB, captured, state 2): 26;
[hos sead::MainThread] vk: readbacks of GPU results: 1059 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 15 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1512 kept their virtual core (nothing there to wait for), 3096 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 6 textures retired (1229 left, 2601 MB): budget 0, drawn over 6, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8220000 (4980198 pages from the captures)
[hos host] vk: 12 textures retired (1226 left, 2610 MB): budget 8, drawn over 4, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10077 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 585 ranges, 3555 MB), 247 pages written, 247 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (86 MB), 12 spare buffers, 8040 texture pages copied (31 MB)
[hos host] vk: 7 textures retired (1227 left, 2609 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 321 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: memory: working set 15409 MB (peak 15952), private 21735 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 16 blocks 912 MB, 37 allocations 816 MB; used 3332 of 15579 MB; heap 1 (device): 170 blocks 5175 MB, 20210 allocations 4452 MB; used 7170 of 8421 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 19 blocks 608 MB, 165 images 531 MB; 111 free ranges, the largest 17 MB
[hos host] vk: image blocks (textures, compacted): 59 blocks 1888 MB, 1065 images 1473 MB; 390 free ranges, the largest 21 MB
[hos host] vk: render target images: 5853 reused from the pool, 3676 made new, 3509 destroyed from it, 1019 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1471 with the pool empty (0 of them with targets still in the dead queue), 296 with none in it a present yet, 34 with none of a kind they can use, 485 with all smaller (the largest 50% short on average), 528 with all bigger (over 1.5 times), 861 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3317 idle
[hos host] vk: VRAM: 7170 MB used of 8421 MB allowed (723 MB of it free in the blocks); textures 2640 MB (budget 2200 MB), 1236 textures
[hos host] vk: budget: VRAM 5175 of 8421 MB (7170 MB by the driver, 1995 MB unaccounted): textures 2640, mirrors 1138, blocks free 723, shadows 284, other buffers 213, other images 111, staging 64; RAM 3249 MB (the process's private memory 21735 MB): imports 2260, capture spares 722, retired imports 140, captures 92, capture table 33, other buffers 2, others 0
[hos host] vk: textures: decoded ASTC 1413 MB, rendered 765 MB (111 drawn in one frame, kept: 198416 KB), other 401 MB; used in the last 8 presents 1752 MB
[hos host] vk: decoded ASTC used again: 359 textures (814 MB) copied from ones the game finalized, not decoded; 359 of those retired, 4 listed; 1607 MB of blocks hashed in 281 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 8653 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34, last used 2 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 52 (11349 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB);
[hos host] vk: 10 textures retired (1226 left, 2621 MB): budget 8, drawn over 2, dead targets 0
[hos Presentation Thread] vk: presents after the CPU's signal (600): 237 waited for it, 4.24 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 53.1 ms; VRAM 7.0 of 8.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6171 queued, 6171 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15263 ms and 331.5 MB read in all); 8366 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11714 graphics pipelines (1 not in a cache), 714 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2241 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9005 (6171 ready, 0 loading), 16283 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13192 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 149 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.26, present 0.18, draw 6.99, descriptors 3.12, texture lookups 0.53, uploads 0.30, pipelines 0.30, programs 1.17, commands 1.87, dispatch 0.89, memory sync 0.66, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 395 (1138 MB): 5.68 MB synced in 1069.0 copies, 17.92 MB written back
[hos host] vk: 13800 presents, 59.6 fps; a frame: 2048 draws (0 skipped), 165 dispatches, 177 passes, 132 barriers, 6 submits, 0.5 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1227 textures (2624 MB)
[hos host] vk: host time in GPU resource waits: 4.24 ms a frame (fences 1.74, timelines 2.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 705.4 full barriers recorded, 296.6 left out (nothing recorded since the one before), 12.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.7, an op 1122.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.4 ms a frame (the longest submission 4.4 ms); 1.2 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.7 (2.5 MB), ASTC or BC 0.5 (0.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.18 MB unswizzled on the CPU in 0.03 ms
[hos host] vk: the render thread (600 presents, 17.41 ms a frame): its jobs 15.39 ms a frame (submits 6.31, presents 4.74, others 4.34), idle 2.02; the game waited for room at 44 presents, 0.55 ms each (0.04 a frame; 13.31 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 244 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 1443 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.60 ms after its hand-over (50.32 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.24 + 0.34 (the window's copy is in the idle), after it 1.21 + 0.83; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.73 parts a frame recorded in the signal's job, 72.5% of them took nothing that came with the signal; of the others, 2.36 a frame took pages of the signal's capture into a mirror (164.2 pages a frame) and 0.04 pages read as guest memory had them (3.7 pages a frame); a frame's first part took nothing in 19 of 244 frames, its first three in 19
[hos host] vk: frame ahead, what came with the signal: 244 signals with a capture, each of 680.2 pages in 12.0 pools; between a frame's last submit and that signal 0.10 texture and sampler registrations, 0.09 releases and 0.00 other CPU signals a signal (15 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.22+0.10, 0.30+0.13, 0.06+0.02, 0.03+0.02, 0.14+0.12, 0.23+0.27, 0.14+0.15, 0.07+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: sparse mirrors: 134, 1138 MB with memory of 1620 MB (481 MB spared); blocks given memory later 11037, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 910 binds (243.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 232 taken from the spares, 163 made new; 261 kept as spares, 0 destroyed; 29 spares now, 1080 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 539 dirty pages from their submits' captures, 53984 from earlier ones', 867 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 887 dispatches, 41613 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8385 runs, 700.2 MB compared; 59 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19490 in captures, 2 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa6400000 (4 MB, captured, state 2): 864; 0x7f81400000 (4 MB, captured, state 2): 3;
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: program picks: 11858053 kept (4605563 by the last picks' textures, a uniform buffer bound since), 26319 new, 4715627 with other textures (of 16600000)
[hos host] vk: replay reads of guest memory as submitted: 8240000 (4988378 pages from the captures)
[hos host] vk: render targets drawn over from another start, let go: 10207 so far (48225 MB), 9148 of them (43361 MB) between the passes
[hos host] vk: 8 textures retired (1227 left, 2622 MB): budget 0, drawn over 8, dead targets 0
[hos Presentation Thread] vi: 13813 frames presented (57.3 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 50, 15.5-17.8 195, 17.8-22 24, 22-30 12, 30-40 4, over 40 2
[hos host] vk: 10 textures retired (1223 left, 2601 MB): budget 0, drawn over 10, dead targets 0
[hos host] vk: 10 textures retired (1219 left, 2586 MB): budget 0, drawn over 10, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8260000 (4995757 pages from the captures)
[hos host] vk: 8 textures retired (1213 left, 2552 MB): budget 0, drawn over 8, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 172 dirty pages from their submits' captures, 54173 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 845 dispatches, 41442 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8406 runs, 700.7 MB compared; 54 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19533 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 5 textures retired (1217 left, 2560 MB): budget 0, drawn over 5, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8280000 (5005118 pages from the captures)
[hos host] vk: 7 textures retired (1216 left, 2559 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: 9 textures retired (1216 left, 2556 MB): budget 0, drawn over 9, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8300000 (5015846 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 4338812 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: 3 textures retired (1221 left, 2576 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: sparse mirrors: 129, 1130 MB with memory of 1596 MB (466 MB spared); blocks given memory later 11094, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 917 binds (245.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 232 taken from the spares, 163 made new; 266 kept as spares, 0 destroyed; 34 spares now, 1104 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 425 dirty pages from their submits' captures, 55953 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 911 dispatches, 42436 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8365 runs, 700.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19808 in captures, 1 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 590 ranges, 3562 MB), 275 pages written, 275 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (106 MB), 12 spare buffers, 3159 texture pages copied (12 MB)
[hos host] vk: recorder: 5597598 packets run in 10.0 s (busy 1.67 s, spun idle 0.11 s, slept 10270 times, 3164 of them under 200 us), 0 calls made directly (no room); 17345 batches over 573 frames (30.3 a frame, 323 packets a batch), 8640 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 161 times, 15.00 ms (0.3 times, 0.03 ms a frame); every thread 836 times, 73.03 ms in all, the longest 9.46 ms; vkWaitForFences 487/40.17 ms (max 9.46), vmaDestroyBuffer 101/11.51 ms (max 1.04), vkDestroyImageView 124/10.30 ms (max 0.34), vkQueueBindSparse 66/4.57 ms (max 0.44), vmaFreeMemory 12/1.74 ms (max 0.39), vkEndCommandBuffer 12/1.63 ms (max 0.42)
[hos host] vk: recorder: of its busy 1.67 s, 0.42 s in parts' ends and submits (the longest 1.36 ms), 0.10 s of that blocked on the queue's lock (185 times, the longest 1.23 ms), so 0.32 s in the ends, submits and holds; 1.25 s running the other packets; 9769 packets a frame
[hos host] vk: 273 mirrors given back (1973 MB in all); 1112 MB of mirrors now, 140 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: 5 textures retired (1223 left, 2585 MB): budget 0, drawn over 5, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9736 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 12012421 kept (4683370 by the last picks' textures, a uniform buffer bound since), 26321 new, 4761257 with other textures (of 16800000)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #230000: event 0x100d6c802c mode 0 value 164382 (now 164363) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: 7 textures retired (1224 left, 2587 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8320000 (5024694 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 6 textures retired (1225 left, 2589 MB): budget 0, drawn over 6, dead targets 0
[hos sead::MainThread] nvn: 1750000 recordings made, 75461 live copies (25.3 MB), 6098 texture copies; 2 command buffers' segments (4.1 MB)
[hos ModuleSystemWorker2] vcores: 155732 takes, 30602 waits (mean 69 us, longest 20.0 ms, 0 over HOS_VCORE_LOG_MS), 15200 kicks, 15142 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 8 textures retired (1225 left, 2590 MB): budget 0, drawn over 8, dead targets 0
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 418 dirty pages from their submits' captures, 55117 from earlier ones', 761 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 904 dispatches, 42503 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8422 runs, 701.8 MB compared; 69 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19538 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 761;
[hos host] vk: a frame: 166 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.51, present 0.17, draw 6.71, descriptors 2.90, texture lookups 0.21, uploads 0.00, pipelines 0.29, programs 1.15, commands 1.84, dispatch 0.87, memory sync 0.47, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 395 (1088 MB): 5.53 MB synced in 1069.2 copies, 17.98 MB written back
[hos host] vk: 13920 presents, 60.0 fps; a frame: 2129 draws (0 skipped), 168 dispatches, 178 passes, 131 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1226 textures (2593 MB)
[hos host] vk: host time in GPU resource waits: 1.70 ms a frame (fences 1.45, timelines 0.25, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 699.5 full barriers recorded, 295.2 left out (nothing recorded since the one before), 13.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1111.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.6 ms a frame (the longest submission 3.4 ms); 0.7 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.7 (2.2 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 8340000 (5034316 pages from the captures)
[hos host] vk: 9 textures retired (1229 left, 2595 MB): budget 0, drawn over 9, dead targets 0
[hos sead::MainThread] nvn: CommandBufferSignalEvent #165000: event 0x100d6c800c mode 0 value 165001 (lr 0x80660adc)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1127 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1549 kept their virtual core (nothing there to wait for), 3059 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 10 textures retired (1225 left, 2577 MB): budget 0, drawn over 10, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8360000 (5045015 pages from the captures)
[hos host] vk: 9 textures retired (1221 left, 2560 MB): budget 0, drawn over 9, dead targets 0
[hos host] vk: sparse mirrors: 117, 1076 MB with memory of 1484 MB (407 MB spared); blocks given memory later 11094, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 923 binds (246.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 232 taken from the spares, 163 made new; 278 kept as spares, 0 destroyed; 46 spares now, 1216 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: 7 textures retired (1222 left, 2564 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 216 dirty pages from their submits' captures, 55003 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 870 dispatches, 42107 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8455 runs, 717.3 MB compared; 60 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19471 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: replay reads of guest memory as submitted: 8380000 (5055799 pages from the captures)
[hos host] vk: 4 textures retired (1220 left, 2555 MB): budget 0, drawn over 4, dead targets 0
[hos sead::MainThread] nvn: GPU waits in 5 s: 103 sync waits (310 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 592 ranges, 3563 MB), 268 pages written, 268 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (128 MB), 11 spare buffers, 3778 texture pages copied (14 MB)
[hos host] vk: 6 textures retired (1222 left, 2560 MB): budget 0, drawn over 6, dead targets 0
[hos host] vk: program picks: 12165641 kept (4759416 by the last picks' textures, a uniform buffer bound since), 26321 new, 4808037 with other textures (of 17000000)
[hos host] vk: replay reads of guest memory as submitted: 8400000 (5066247 pages from the captures)
[hos host] vk: 30 textures retired (1225 left, 2582 MB): budget 22, drawn over 8, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 945 dirty pages from their submits' captures, 54051 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 870 dispatches, 41941 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7910 runs, 694.1 MB compared; 58 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19289 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 14 textures retired (1223 left, 2578 MB): budget 8, drawn over 6, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8420000 (5079184 pages from the captures)
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 29086233 lookups, 139017 missed (0.48 %)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6955 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 5 textures retired (1230 left, 2596 MB): budget 0, drawn over 5, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 11 textures retired (1227 left, 2587 MB): budget 0, drawn over 11, dead targets 0
[hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 380000 so far (x1 0x117daa3070 x2 0 x3 0x117daa3040 x4 0x9b040)
[hos host] vk: replay reads of guest memory as submitted: 8440000 (5091500 pages from the captures)
[hos host] vk: 7 textures retired (1232 left, 2578 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: reaper: 791 images (1605 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 246 (279 MB)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 117, 1078 MB with memory of 1484 MB (405 MB spared); blocks given memory later 11097, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 927 binds (248.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 233 taken from the spares, 163 made new; 279 kept as spares, 0 destroyed; 46 spares now, 1216 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 593 dirty pages from their submits' captures, 55819 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 42960 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7776 runs, 707.6 MB compared; 66 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19823 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: 7 textures retired (1231 left, 2571 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: a frame: 187 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.76, present 0.17, draw 6.81, descriptors 2.89, texture lookups 0.24, uploads 0.03, pipelines 0.29, programs 1.21, commands 1.87, dispatch 0.84, memory sync 0.52, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 396 (1078 MB): 5.53 MB synced in 1073.4 copies, 17.81 MB written back
[hos host] vk: 14040 presents, 59.9 fps; a frame: 2147 draws (0 skipped), 157 dispatches, 179 passes, 131 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1232 textures (2575 MB)
[hos host] vk: host time in GPU resource waits: 2.05 ms a frame (fences 1.28, timelines 0.76, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 693.9 full barriers recorded, 296.3 left out (nothing recorded since the one before), 13.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1200.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.4 ms a frame (the longest submission 4.3 ms); 1.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.7 (2.0 MB), ASTC or BC 0.3 (0.5 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: replay reads of guest memory as submitted: 8460000 (5104000 pages from the captures)
[hos host] vk: 9 textures retired (1235 left, 2577 MB): budget 0, drawn over 9, dead targets 0
[hos host] vk: texture #19093 0x10008612200 799x799 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14051 (drawn at 13967)
[hos host] vk: texture #19102 0x10008612200 794x794 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14052 (drawn at 13966)
[hos host] vk: texture #19152 0x10008612200 900x900 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14056 (drawn at 13994)
[hos host] vk: 56 textures retired (1191 left, 2538 MB): budget 53, drawn over 3, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8480000 (5114089 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 595 ranges, 3565 MB), 264 pages written, 264 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 25 kept (110 MB), 12 spare buffers, 6911 texture pages copied (26 MB)
[hos host] vk: texture #18190 0x100050a2000 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #18230 0x10001fe0800 640x360x1 x1 NVN 0x14 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #18188 0x10004b01a00 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #18183 0x10005552400 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #18191 0x10004ec1c00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #18189 0x10004fb1e00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 51 textures retired (1152 left, 2526 MB): budget 41, drawn over 10, dead targets 0; GPU-drawn with no copy kept: 0 live, 6 finalized, 21 MB
[hos host] vk: program picks: 12319093 kept (4828623 by the last picks' textures, a uniform buffer bound since), 26321 new, 4854585 with other textures (of 17200000)
[hos host] vk: readback pages refreshed from guest memory: 4406327 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1263 dirty pages from their submits' captures, 53227 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 854 dispatches, 41673 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7580 runs, 689.8 MB compared; 59 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19525 in captures, 2 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 87 textures retired (1098 left, 2448 MB): budget 79, drawn over 8, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8500000 (5127226 pages from the captures)
[hos host] vk: 6 textures retired (1107 left, 2513 MB): budget 0, drawn over 6, dead targets 0
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1120 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1382 kept their virtual core (nothing there to wait for), 2970 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #19151 0x10008612200 904x904 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14095 (drawn at 14033)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #235000: event 0x100d6c802c mode 0 value 168182 (now 168163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06
[hos host] vk: 32 textures retired (1092 left, 2505 MB): budget 23, drawn over 9, dead targets 0
[hos host] vk: 33 textures retired (1083 left, 2503 MB): budget 23, drawn over 10, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8520000 (5139525 pages from the captures)
[hos Presentation Thread] vi: 14112 frames presented (59.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 30, 15.5-17.8 241, 17.8-22 25, 22-30 3, 30-40 0, over 40 0
[hos host] vk: sparse mirrors: 121, 1087 MB with memory of 1500 MB (412 MB spared); blocks given memory later 11098, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 932 binds (248.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 237 taken from the spares, 163 made new; 279 kept as spares, 0 destroyed; 42 spares now, 1200 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2125 dirty pages from their submits' captures, 54911 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 934 dispatches, 43397 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7575 runs, 704.9 MB compared; 203 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19513 in captures, 127 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: 26 textures retired (1034 left, 2175 MB): budget 21, drawn over 5, dead targets 0
[hos host] vk: VRAM compaction 6 begins: the textures' blocks hold 1664 MB, 549 MB of it free space
[hos host] vk: replay reads of guest memory as submitted: 8540000 (5150345 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 29470 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 3 textures retired (1059 left, 2258 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #19310 0x10008612200 1421x1421 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14154 (drawn at 14093)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 595 ranges, 3609 MB), 290 pages written, 290 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.039; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.029, bytes copied 0.000; 25 kept (206 MB), 9 spare buffers, 9831 texture pages copied (38 MB)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 468 dirty pages from their submits' captures, 56934 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 920 dispatches, 44107 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8136 runs, 741.4 MB compared; 64 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20718 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa9000000 (4 MB, captured, state 2): 2;
[hos host] vk: a frame: 194 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.88, present 0.32, draw 6.91, descriptors 3.07, texture lookups 0.64, uploads 0.20, pipelines 0.29, programs 1.17, commands 1.86, dispatch 0.83, memory sync 0.56, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 401 (1089 MB): 5.55 MB synced in 1065.8 copies, 17.66 MB written back
[hos host] vk: 14160 presents, 59.8 fps; a frame: 1960 draws (0 skipped), 147 dispatches, 177 passes, 131 barriers, 6 submits, 0.9 uploads (1.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1097 textures (2360 MB)
[hos host] vk: host time in GPU resource waits: 3.46 ms a frame (fences 1.40, timelines 2.06, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 668.2 full barriers recorded, 292.9 left out (nothing recorded since the one before), 14.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.8 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1060.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.3 ms a frame (the longest submission 4.2 ms); 1.9 textures made a frame: colour targets 0.3 (1.2 MB), depth 1.0 (5.7 MB), ASTC or BC 0.7 (1.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.99 MB unswizzled on the CPU in 0.16 ms
[hos host] vk: replay reads of guest memory as submitted: 8560000 (5161528 pages from the captures)
[hos host] vk: program picks: 12468951 kept (4891719 by the last picks' textures, a uniform buffer bound since), 26321 new, 4904727 with other textures (of 17400000)
[hos host] vk: VRAM compaction 6 over: 300 textures moved (270 MB) in 27 passes, 0 left where they were; 15 blocks (480 MB) given back to the driver, 3808 MB in all so far
[hos host] vk: replay reads of guest memory as submitted: 8580000 (5172744 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #170000: event 0x100d6c8014 mode 0 value 170000 (lr 0x802915c8)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: sparse mirrors: 121, 1092 MB with memory of 1512 MB (420 MB spared); blocks given memory later 11102, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 937 binds (249.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 239 taken from the spares, 163 made new; 281 kept as spares, 0 destroyed; 42 spares now, 1188 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 647 dirty pages from their submits' captures, 58914 from earlier ones', 129 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 975 dispatches, 45765 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8293 runs, 741.9 MB compared; 64 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21013 in captures, 2 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 129;
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: 1 textures retired (1148 left, 2524 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8600000 (5183875 pages from the captures)
[hos BaseProcExecutor1] vcores: 154795 takes, 24921 waits (mean 69 us, longest 22.4 ms, 0 over HOS_VCORE_LOG_MS), 11644 kicks, 11676 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 3 textures retired (1158 left, 2525 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: pool 0x7f96c00000 (8 MB): mirror made at its first binding (401 so)
[hos host] vk: 8 textures retired (1195 left, 2545 MB): budget 0, drawn over 8, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8620000 (5196577 pages from the captures)
[hos host] vk: 9 textures retired (1202 left, 2515 MB): budget 0, drawn over 9, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1100 dirty pages from their submits' captures, 56327 from earlier ones', 1771 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 953 dispatches, 43867 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7933 runs, 736.4 MB compared; 85 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19954 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9c400000 (32 MB, captured, state 2): 1756; 0x7f9ec00000 (4 MB, captured, state 2): 14; 0x7fd3c00000 (16 MB, captured, state 2): 1;
[hos host] vk: 10 textures retired (1199 left, 2491 MB): budget 0, drawn over 10, dead targets 0
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1072 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 25 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1470 kept their virtual core (nothing there to wait for), 3138 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 605 ranges, 3601 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 7656 texture pages copied (29 MB)
[hos host] vk: 1 textures retired (1120 left, 2216 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 23158 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 8640000 (5208228 pages from the captures)
[hos host] vk: 2 textures retired (1119 left, 2211 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] nvn: GPU waits in 5 s: 146 sync waits (422 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 5 textures retired (1115 left, 2183 MB): budget 0, drawn over 5, dead targets 0
[hos host] vk: a frame: 171 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.99, present 0.25, draw 6.33, descriptors 2.89, texture lookups 0.49, uploads 0.27, pipelines 0.25, programs 1.11, commands 1.62, dispatch 0.79, memory sync 0.53, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 407 (1127 MB): 5.77 MB synced in 1066.7 copies, 17.54 MB written back
[hos host] vk: 14280 presents, 60.1 fps; a frame: 1750 draws (0 skipped), 142 dispatches, 171 passes, 132 barriers, 6 submits, 0.8 uploads (1.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1115 textures (2183 MB)
[hos host] vk: host time in GPU resource waits: 2.87 ms a frame (fences 1.41, timelines 1.47, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 636.0 full barriers recorded, 286.6 left out (nothing recorded since the one before), 15.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.5, an op 1011.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.9 ms a frame (the longest submission 4.6 ms); 1.4 textures made a frame: colour targets 0.1 (0.9 MB), depth 0.6 (1.7 MB), ASTC or BC 0.7 (1.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.03 MB unswizzled on the CPU in 0.15 ms
[hos host] vk: readback pages refreshed from guest memory: 4473439 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 8660000 (5219767 pages from the captures)
[hos host] vk: sparse mirrors: 126, 1127 MB with memory of 1584 MB (456 MB spared); blocks given memory later 11218, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 946 binds (251.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 244 taken from the spares, 163 made new; 281 kept as spares, 0 destroyed; 37 spares now, 1116 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 618 dirty pages from their submits' captures, 59978 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 975 dispatches, 46357 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7806 runs, 748.7 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20867 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 2;
[hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB)
[hos host] vk: program picks: 12609757 kept (4946761 by the last picks' textures, a uniform buffer bound since), 26324 new, 4963918 with other textures (of 17600000)
[hos host] vk: 3 textures retired (1112 left, 2171 MB): budget 0, drawn over 3, dead targets 0
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #240000: event 0x100d6c802c mode 0 value 171982 (now 171963) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.07
[hos host] vk: replay reads of guest memory as submitted: 8680000 (5231520 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 121 dirty pages from their submits' captures, 61295 from earlier ones', 78 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 964 dispatches, 46593 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7891 runs, 750.7 MB compared; 68 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21229 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 77; 0x7fd3c00000 (16 MB, captured, state 2): 1;
[hos host] vk: replay reads of guest memory as submitted: 8700000 (5243354 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 612 ranges, 3614 MB), 272 pages written, 272 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 24 kept (102 MB), 12 spare buffers, 7701 texture pages copied (30 MB)
[hos ModuleSystemWorker2] nvn: 1800000 recordings made, 77844 live copies (25.1 MB), 6219 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: pool 0x7fbf733900+0x20d6e0 imported in 7.4 ms as 0x7fbd800000..0x7fc1400000 (228 imports, 700 ms in all)
[hos host] vk: pool 0x7fbd800000..0x7fc1400000 took in 2 pools
[hos host] vk: 288 mirrors given back (2071 MB in all); 1176 MB of mirrors now, 140 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: replay reads of guest memory as submitted: 8720000 (5255523 pages from the captures)
[hos host] vk: presenter thread: the copy's fence 1.37 ms, the present 0.73 ms a present
[hos host] vk: sparse mirrors: 126, 1180 MB with memory of 1640 MB (459 MB spared); blocks given memory later 11222, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 960 binds (254.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 250 taken from the spares, 164 made new; 288 kept as spares, 0 destroyed; 38 spares now, 1120 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1126 dirty pages from their submits' captures, 56337 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 904 dispatches, 43766 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7133 runs, 682.2 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19659 in captures, 1 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 187 live (2260 MB), 4 retired still imported (132 MB), 41 retired given back (952 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 7445 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 8740000 (5267640 pages from the captures)
[hos host] vk: program picks: 12756202 kept (5010234 by the last picks' textures, a uniform buffer bound since), 26325 new, 5017472 with other textures (of 17800000)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: memory: working set 15190 MB (peak 15952), private 21222 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 864 MB, 37 allocations 778 MB; used 3276 of 15579 MB; heap 1 (device): 160 blocks 4855 MB, 20726 allocations 4225 MB; used 6850 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 21 blocks 672 MB, 105 images 259 MB; 69 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 47 blocks 1504 MB, 1099 images 1484 MB; 268 free ranges, the largest 11 MB
[hos host] vk: render target images: 5883 reused from the pool, 3929 made new, 3822 destroyed from it, 1146 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 148 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1574 with the pool empty (0 of them with targets still in the dead queue), 331 with none in it a present yet, 35 with none of a kind they can use, 548 with all smaller (the largest 49% short on average), 564 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3630 idle
[hos host] vk: VRAM: 6850 MB used of 11474 MB allowed (629 MB of it free in the blocks); textures 2380 MB (budget 7158 MB), 1210 textures
[hos host] vk: budget: VRAM 4855 of 11474 MB (6850 MB by the driver, 1995 MB unaccounted): textures 2380, mirrors 1172, blocks free 629, shadows 284, other buffers 213, other images 111, staging 64; RAM 3203 MB (the process's private memory 21222 MB): imports 2260, capture spares 646, retired imports 132, captures 116, capture table 33, other buffers 16, others 0
[hos host] vk: textures: decoded ASTC 1427 MB, rendered 514 MB (69 drawn in one frame, kept: 42597 KB), other 405 MB; used in the last 8 presents 1885 MB
[hos host] vk: decoded ASTC used again: 359 textures (814 MB) copied from ones the game finalized, not decoded; 359 of those retired, 3 listed; 1684 MB of blocks hashed in 294 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 9253 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 52 (11349 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 277 waited for it, 2.77 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 37.0 ms; VRAM 6.7 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6174 queued, 6174 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15269 ms and 331.5 MB read in all); 8369 programs (1621 of them compute variants without their code, 0 graphics programs no note names), 11716 graphics pipelines (1 not in a cache), 715 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2247 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9008 (6174 ready, 0 loading), 16392 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13196 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 710 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 190 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.44, present 0.14, draw 8.18, descriptors 3.57, texture lookups 0.40, uploads 0.35, pipelines 0.30, programs 1.21, commands 2.56, dispatch 0.79, memory sync 0.47, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 414 (1172 MB): 6.55 MB synced in 1088.6 copies, 17.50 MB written back
[hos host] vk: 14400 presents, 57.9 fps; a frame: 2027 draws (0 skipped), 138 dispatches, 182 passes, 131 barriers, 6 submits, 1.1 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13571 pipelines (3514 ms), 0 programs (10 failed), 1210 textures (2380 MB)
[hos host] vk: host time in GPU resource waits: 4.30 ms a frame (fences 1.66, timelines 2.64, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 670.1 full barriers recorded, 294.8 left out (nothing recorded since the one before), 19.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 1210.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.2 ms a frame (the longest submission 5.3 ms); 1.0 textures made a frame: colour targets 0.1 (0.6 MB), depth 0.1 (0.3 MB), ASTC or BC 0.8 (1.7 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.25 MB unswizzled on the CPU in 0.19 ms
[hos host] vk: queues ordered by their events: 89326 submissions stopped at a wait, 89280 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 242510, not met 0; captures at a CPU event's wait 8747, 0 of them at a wait met on arrival
[hos host] vk: the render thread (600 presents, 16.80 ms a frame): its jobs 15.56 ms a frame (submits 6.15, presents 4.64, others 4.77), idle 1.23; the game waited for room at 17 presents, 0.77 ms each (0.02 a frame; 7.03 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 280 captures at a CPU signal; the submits' waits read in 0.042 ms a frame; readbacks: 1536 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.83 ms after its hand-over (34.54 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.26 + 0.29 (the window's copy is in the idle), after it 1.42 + 0.56; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.38 parts a frame recorded in the signal's job, 76.5% of them took nothing that came with the signal; of the others, 1.96 a frame took pages of the signal's capture into a mirror (37.9 pages a frame) and 0.01 pages read as guest memory had them (9.0 pages a frame); a frame's first part took nothing in 18 of 280 frames, its first three in 18
[hos host] vk: frame ahead, what came with the signal: 280 signals with a capture, each of 65.5 pages in 10.0 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.05 releases and 0.00 other CPU signals a signal (1 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.26+0.06, 0.42+0.14, 0.11+0.04, 0.04+0.02, 0.19+0.17, 0.29+0.09, 0.06+0.02, 0.05+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: replay reads of guest memory as submitted: 8760000 (5279301 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 926, 4109.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1428 dirty pages from their submits' captures, 51502 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 805 dispatches, 40522 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7027 runs, 666.9 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17852 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: readbacks of GPU results: 1069 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 22 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1244 kept their virtual core (nothing there to wait for), 2596 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: render targets drawn over from another start, let go: 10396 so far (49226 MB), 9336 of them (44360 MB) between the passes
[hos Presentation Thread] vi: 14410 frames presented (59.4 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 62, 15.5-17.8 198, 17.8-22 27, 22-30 9, 30-40 1, over 40 1
[hos host] vk: replay reads of guest memory as submitted: 8780000 (5291840 pages from the captures)
[hos host] vk: pool 0x1017549a00+0xf00000 imported in 3.7 ms as 0x1017400000..0x101a000000 (229 imports, 704 ms in all)
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 400000 so far (x1 0x117a4f2040 x2 0x6001d40 x3 0x117a4f2000 x4 0xf33080)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 611 ranges, 3618 MB), 250 pages written, 250 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 24 kept (78 MB), 12 spare buffers, 8400 texture pages copied (32 MB)
[hos host] vk: replay reads of guest memory as submitted: 8800000 (5304316 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 123, 1174 MB with memory of 1624 MB (449 MB spared); blocks given memory later 11236, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 966 binds (255.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 251 taken from the spares, 164 made new; 292 kept as spares, 0 destroyed; 41 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1280 dirty pages from their submits' captures, 54409 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 850 dispatches, 42221 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7054 runs, 671.7 MB compared; 56 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19149 in captures, 0 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 187 live (2276 MB), 5 retired still imported (160 MB), 41 retired given back (952 MB)
[hos host] vk: texture storage read as submitted (2001 times; 111350 pages from the copies, 0 of them rewritten in guest memory since their submit; 3787 pages kept, 271 spans, 957 MB)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #175000: event 0x100d6c8020 mode 0 value 175000 (lr 0x80664f14)
[hos host] vk: replay reads of guest memory as submitted: 8820000 (5316626 pages from the captures)
[hos host] vk: 87 textures retired (1242 left, 2440 MB): budget 87, drawn over 0, dead targets 0
[hos host] vk: program picks: 12906179 kept (5073449 by the last picks' textures, a uniform buffer bound since), 26332 new, 5067488 with other textures (of 18000000)
[hos host] vk: recorder: 6355175 packets run in 10.0 s (busy 1.76 s, spun idle 0.12 s, slept 11223 times, 3543 of them under 200 us), 0 calls made directly (no room); 18316 batches over 593 frames (30.9 a frame, 347 packets a batch), 8835 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 191 times, 19.08 ms (0.3 times, 0.03 ms a frame); every thread 913 times, 62.23 ms in all, the longest 1.33 ms; vkWaitForFences 519/26.28 ms (max 0.45), vkDestroyImageView 193/15.25 ms (max 0.44), vmaDestroyBuffer 96/10.39 ms (max 1.33), vmaFreeMemory 18/4.22 ms (max 0.82), vkQueueBindSparse 50/3.41 ms (max 0.38), the queue's lock 26/2.00 ms (max 0.19)
[hos host] vk: recorder: of its busy 1.76 s, 0.44 s in parts' ends and submits (the longest 1.34 ms), 0.13 s of that blocked on the queue's lock (255 times, the longest 1.30 ms), so 0.31 s in the ends, submits and holds; 1.32 s running the other packets; 10717 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 8840000 (5330624 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 4540955 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3423 dirty pages from their submits' captures, 63001 from earlier ones', 5 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 906 dispatches, 44031 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7348 runs, 701.3 MB compared; 74 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18879 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7f86000000 (4 MB, captured, state 2): 5;
[hos host] vk: queues ordered by their events: 90000 submissions stopped at a wait, 89953 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 244685, not met 0; captures at a CPU event's wait 8811, 0 of them at a wait met on arrival
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #245000: event 0x100d6c802c mode 0 value 175782 (now 175763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: replay reads of guest memory as submitted: 8860000 (5343224 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11526 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos BaseProcExecutor1] vcores: 152064 takes, 25000 waits (mean 68 us, longest 7.4 ms, 0 over HOS_VCORE_LOG_MS), 11876 kicks, 11879 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: texture #19745 0x10004bf1c00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454)
[hos host] vk: texture #19746 0x10004b01a00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454)
[hos host] vk: texture #19729 0x1000589ac00 640x360 x1 NVN 0x11 the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454)
[hos host] vk: texture #19744 0x10004ce1e00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454)
[hos host] vk: texture #19743 0x10009cf3200 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454)
[hos host] vk: a frame: 209 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.90, present 0.16, draw 7.00, descriptors 3.10, texture lookups 0.40, uploads 0.18, pipelines 0.29, programs 1.21, commands 1.87, dispatch 0.80, memory sync 0.51, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 415 (1168 MB): 5.95 MB synced in 1092.7 copies, 17.59 MB written back
[hos host] vk: 14520 presents, 58.8 fps; a frame: 2080 draws (0 skipped), 142 dispatches, 185 passes, 132 barriers, 6 submits, 1.4 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13571 pipelines (3514 ms), 0 programs (10 failed), 1273 textures (2487 MB)
[hos host] vk: host time in GPU resource waits: 3.66 ms a frame (fences 1.34, timelines 2.32, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 695.9 full barriers recorded, 300.1 left out (nothing recorded since the one before), 17.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.2; 0x12: 68.9; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.6, an op 1211.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.2 ms a frame (the longest submission 3.5 ms); 2.3 textures made a frame: colour targets 0.7 (2.5 MB), depth 0.3 (1.8 MB), ASTC or BC 1.2 (1.7 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.38 MB unswizzled on the CPU in 0.06 ms
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 8880000 (5354701 pages from the captures)
[hos host] vk: sparse mirrors: 121, 1168 MB with memory of 1612 MB (443 MB spared); blocks given memory later 11245, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 969 binds (255.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 251 taken from the spares, 164 made new; 294 kept as spares, 0 destroyed; 43 spares now, 1148 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 296 dirty pages from their submits' captures, 57809 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 914 dispatches, 44544 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7647 runs, 720.4 MB compared; 47 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19993 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 187 live (2276 MB), 5 retired still imported (160 MB), 41 retired given back (952 MB)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 615 ranges, 3642 MB), 293 pages written, 293 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 25 kept (92 MB), 11 spare buffers, 5309 texture pages copied (20 MB)
[hos host] vk: 2 textures retired (1278 left, 2494 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8900000 (5366261 pages from the captures)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1071 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 34 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1601 kept their virtual core (nothing there to wait for), 3263 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 598 dirty pages from their submits' captures, 56367 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 43457 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7540 runs, 703.9 MB compared; 63 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19363 in captures, 2 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 8920000 (5378568 pages from the captures)
[hos host] vk: program picks: 13055698 kept (5140932 by the last picks' textures, a uniform buffer bound since), 26336 new, 5117965 with other textures (of 18200000)
[hos sead::MainThread] nvn: GPU waits in 5 s: 155 sync waits (606 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: texture #19891 0x10008612200 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14571 (drawn at 14510)
[hos host] vk: texture #19606 0x10009932a00 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 14571 (drawn at 14510)
[hos host] vk: replay reads of guest memory as submitted: 8940000 (5389411 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos ModuleSystemWorker3] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 16216970 lookups, 124086 missed (0.77 %)
[hos host] vk: sparse mirrors: 132, 1194 MB with memory of 1668 MB (473 MB spared); blocks given memory later 11247, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 982 binds (257.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 262 taken from the spares, 164 made new; 294 kept as spares, 0 destroyed; 32 spares now, 1092 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1091 dirty pages from their submits' captures, 52761 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 849 dispatches, 41002 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8037 runs, 680.9 MB compared; 59 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18434 in captures, 1 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2280 MB), 5 retired still imported (160 MB), 41 retired given back (952 MB)
[hos host] vk: replay reads of guest memory as submitted: 8960000 (5397798 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 611 ranges, 3661 MB), 254 pages written, 254 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (104 MB), 11 spare buffers, 6800 texture pages copied (26 MB)
[hos host] vk: 36 textures retired (1371 left, 2669 MB): budget 0, drawn over 36, dead targets 0
[hos host] vk: reaper: 714 images (2416 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 86 (178 MB)
[hos host] vk: 74 textures retired (1305 left, 2602 MB): budget 73, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8980000 (5409110 pages from the captures)
[hos host] vk: texture #19734 0x100089d2a00 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14631 (drawn at 14454)
[hos host] vk: texture #19739 0x10004fb1e00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #19741 0x10004ec1c00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #19740 0x100050a2000 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #19765 0x10001fe0800 640x360x1 x1 NVN 0x14 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #19731 0x10005552400 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #19738 0x10004b01a00 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 146 textures retired (1203 left, 2418 MB): budget 127, drawn over 5, dead targets 14; GPU-drawn with no copy kept: 0 live, 6 finalized, 21 MB
[hos host] vk: a frame: 159 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.72, present 0.18, draw 6.88, descriptors 3.03, texture lookups 0.49, uploads 0.19, pipelines 0.29, programs 1.16, commands 1.89, dispatch 0.84, memory sync 0.50, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 428 (1206 MB): 5.86 MB synced in 1079.8 copies, 17.81 MB written back
[hos host] vk: 14640 presents, 59.1 fps; a frame: 2103 draws (0 skipped), 157 dispatches, 177 passes, 131 barriers, 6 submits, 1.3 uploads (0.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13571 pipelines (3514 ms), 0 programs (10 failed), 1203 textures (2417 MB)
[hos host] vk: host time in GPU resource waits: 4.30 ms a frame (fences 1.32, timelines 2.98, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 690.1 full barriers recorded, 293.5 left out (nothing recorded since the one before), 12.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1067.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.6 ms a frame (the longest submission 3.8 ms); 2.1 textures made a frame: colour targets 0.1 (0.8 MB), depth 0.7 (2.3 MB), ASTC or BC 1.3 (2.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.59 MB unswizzled on the CPU in 0.11 ms
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1317 dirty pages from their submits' captures, 52333 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 857 dispatches, 40906 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8102 runs, 681.5 MB compared; 63 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18412 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 9000000 (5421917 pages from the captures)
[hos host] vk: 21 textures retired (1185 left, 2368 MB): budget 19, drawn over 2, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 24059 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 13207653 kept (5218825 by the last picks' textures, a uniform buffer bound since), 26336 new, 5166010 with other textures (of 18400000)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 9020000 (5432813 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 4608064 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 134, 1206 MB with memory of 1684 MB (477 MB spared); blocks given memory later 11255, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 986 binds (257.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 264 taken from the spares, 164 made new; 294 kept as spares, 0 destroyed; 30 spares now, 1076 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1189 dirty pages from their submits' captures, 56191 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 915 dispatches, 43976 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8792 runs, 736.4 MB compared; 64 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19800 in captures, 4 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2280 MB), 5 retired still imported (160 MB), 41 retired given back (952 MB)
[hos host] vk: replay reads of guest memory as submitted: 9040000 (5443107 pages from the captures)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #250000: event 0x100d6c802c mode 0 value 179582 (now 179563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.07
[hos host] vk: 3 textures retired (1194 left, 2368 MB): budget 3, drawn over 0, dead targets 0
[hos Presentation Thread] vi: 14706 frames presented (59.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 68, 15.5-17.8 178, 17.8-22 37, 22-30 13, 30-40 0, over 40 0
[hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.01, scan 0.04 over 613 ranges, 3726 MB), 291 pages written, 291 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.041; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.027, bytes copied 0.000; 24 kept (92 MB), 12 spare buffers, 8352 texture pages copied (32 MB)
[hos host] vk: replay reads of guest memory as submitted: 9060000 (5454577 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #180000: event 0x100d6c8028 mode 0 value 179999 (lr 0x80667410)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2425 dirty pages from their submits' captures, 59734 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 946 dispatches, 43820 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8821 runs, 737.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20077 in captures, 4 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1121 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 28 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1511 kept their virtual core (nothing there to wait for), 2841 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 9080000 (5464908 pages from the captures)
[hos host] vk: pool 0x7f9ab25200+0x1c3b60 imported in 2.3 ms as 0x7f9a800000..0x7f9c000000 (231 imports, 707 ms in all)
[hos ModuleSystemWorker4] nvn: 1850000 recordings made, 79245 live copies (25.0 MB), 6228 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: 1 textures retired (1200 left, 2365 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9100000 (5473232 pages from the captures)
[hos host] vk: a frame: 170 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.27, present 0.17, draw 6.51, descriptors 2.71, texture lookups 0.13, uploads 0.03, pipelines 0.26, programs 1.15, commands 1.89, dispatch 0.92, memory sync 0.61, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 429 (1220 MB): 6.08 MB synced in 1063.7 copies, 17.99 MB written back
[hos host] vk: 14760 presents, 60.7 fps; a frame: 1863 draws (0 skipped), 170 dispatches, 156 passes, 132 barriers, 6 submits, 0.2 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13573 pipelines (3515 ms), 0 programs (10 failed), 1202 textures (2365 MB)
[hos host] vk: host time in GPU resource waits: 2.68 ms a frame (fences 1.45, timelines 1.23, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 613.0 full barriers recorded, 274.5 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.6, an op 1030.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.9 ms a frame (the longest submission 5.5 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17059 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 13357451 kept (5288733 by the last picks' textures, a uniform buffer bound since), 26336 new, 5216212 with other textures (of 18600000)
[hos host] vk: sparse mirrors: 134, 1220 MB with memory of 1688 MB (467 MB spared); blocks given memory later 11429, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 995 binds (258.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 265 taken from the spares, 164 made new; 295 kept as spares, 0 destroyed; 30 spares now, 1072 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 325 dirty pages from their submits' captures, 68915 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 949 dispatches, 46384 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9290 runs, 774.5 MB compared; 64 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21027 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 188 live (2284 MB), 6 retired still imported (180 MB), 41 retired given back (952 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 9120000 (5483935 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 612 ranges, 3725 MB), 284 pages written, 284 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.015; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (94 MB), 12 spare buffers, 6868 texture pages copied (26 MB)
[hos host] vk: replay reads of guest memory as submitted: 9140000 (5494062 pages from the captures)
[hos AudioFrameworkThread] vcores: 154519 takes, 26002 waits (mean 72 us, longest 31.5 ms, 0 over HOS_VCORE_LOG_MS), 12383 kicks, 12325 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 57739 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 927 dispatches, 44502 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9057 runs, 756.1 MB compared; 59 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20359 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 3 textures retired (1204 left, 2371 MB): budget 3, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 420000 so far (x1 0x117db40010 x2 0 x3 0x117db40000 x4 0x44b040)
[hos host] vk: replay reads of guest memory as submitted: 9160000 (5506589 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 9180000 (5516498 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 133, 1218 MB with memory of 1684 MB (465 MB spared); blocks given memory later 11429, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 996 binds (258.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 265 taken from the spares, 164 made new; 296 kept as spares, 0 destroyed; 31 spares now, 1076 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1308 dirty pages from their submits' captures, 69224 from earlier ones', 734 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 931 dispatches, 42830 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8461 runs, 701.3 MB compared; 139 pages read into the shadows; 284 MB of shadows (51 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19164 in captures, 94 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 534; 0x7fa0800000 (8 MB, captured, state 2): 200;
[hos host] vk: imports: 188 live (2284 MB), 6 retired still imported (180 MB), 41 retired given back (952 MB)
[hos host] vk: 2 textures retired (1203 left, 2349 MB): budget 2, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: GPU waits in 5 s: 129 sync waits (420 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: program picks: 13506894 kept (5360236 by the last picks' textures, a uniform buffer bound since), 26336 new, 5266769 with other textures (of 18800000)
[hos host] vk: replay reads of guest memory as submitted: 9200000 (5524998 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 40989 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 4 textures retired (1209 left, 2348 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: a frame: 157 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.42, present 0.22, draw 6.55, descriptors 2.80, texture lookups 0.12, uploads 0.03, pipelines 0.27, programs 1.17, commands 1.80, dispatch 1.02, memory sync 1.07, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 429 (1158 MB): 8.56 MB synced in 1073.4 copies, 17.99 MB written back
[hos host] vk: 14880 presents, 53.7 fps; a frame: 1847 draws (0 skipped), 170 dispatches, 157 passes, 131 barriers, 6 submits, 0.2 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13742 pipelines (3589 ms), 0 programs (10 failed), 1209 textures (2348 MB)
[hos host] vk: host time in GPU resource waits: 2.26 ms a frame (fences 1.46, timelines 0.80, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 618.5 full barriers recorded, 275.3 left out (nothing recorded since the one before), 13.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.1, an op 1007.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.2 ms a frame (the longest submission 6.6 ms); 0.2 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: texture #16605 0x116dd8f000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 14881 (drawn at 10787)
[hos host] vk: late pages found at the replay (600 takes): 57 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: readback pages refreshed from guest memory: 4675763 pages so far (0 held: may be written ahead in their submission)
[hos ModuleSystemWorker5] vk: readbacks of GPU results: 1105 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1550 kept their virtual core (nothing there to wait for), 3058 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 12355 dirty pages from their submits' captures, 148738 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1043 dispatches, 50097 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8627 runs, 719.3 MB compared; 278 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18693 in captures, 141 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 9220000 (5536415 pages from the captures)
[hos host] vk: 304 mirrors given back (2177 MB in all); 1149 MB of mirrors now, 140 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: (its written part: 3 of 3 KB readable, 0 bytes not zero; levels L0)
[hos host] vk: texture 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6 (39444 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it
[hos Presentation Thread] vk: capture: a submit 0.49 ms (lock 0.02, scan 0.08 over 582 ranges, 3590 MB), 922 pages written, 922 copied (3.6 MB)
[hos host] vk: a capture taken (ms): all 0.098; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.071, bytes copied 0.000; 24 kept (188 MB), 12 spare buffers, 23211 texture pages copied (90 MB)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #255000: event 0x100d6c802c mode 0 value 183382 (now 183363) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.10, writes collected 0.14
[hos host] vk: (its written part: 2 of 2 KB readable, 2531 bytes not zero; levels L0)
[hos host] vk: texture 0x10debdc600 1240x1240x1 x3 levels 1 NVN 0x4d (4563 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 2 KB of it
[hos host] vk: 89 textures retired (1123 left, 2224 MB): budget 89, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9240000 (5549029 pages from the captures)
[hos rmbl::VibrationThread] hid: 49152 vibration values, strongest 0.00 since the last count
[hos host] vk: 38 textures retired (1099 left, 2206 MB): budget 38, drawn over 0, dead targets 0
[hos host] vk: sparse mirrors: 126, 1055 MB with memory of 1580 MB (524 MB spared); blocks given memory later 11507, given back 15533; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1034 binds (269.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 267 taken from the spares, 164 made new; 305 kept as spares, 0 destroyed; 38 spares now, 1180 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 15361 dirty pages from their submits' captures, 88940 from earlier ones', 6577 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1033 dispatches, 46454 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8142 runs, 711.7 MB compared; 114 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18963 in captures, 67 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd6400000 (48 MB, captured, state 2): 4081; 0x7fbc000000 (12 MB, captured, state 2): 2489; 0x7fa9000000 (4 MB, captured, state 2): 6; 0x7fff800000 (8 MB, captured, state 2): 1;
[hos host] vk: imports: 188 live (2284 MB), 6 retired still imported (180 MB), 41 retired given back (952 MB)
[hos host] vk: replay reads of guest memory as submitted: 9260000 (5561653 pages from the captures)
[hos host] vk: 3 textures retired (1119 left, 2219 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: 39 textures retired (1117 left, 2213 MB): budget 39, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9280000 (5573557 pages from the captures)
[hos host] vk: 143 textures retired (1100 left, 2297 MB): budget 143, drawn over 0, dead targets 0
[hos host] vk: presenter thread: the copy's fence 1.35 ms, the present 0.79 ms a present
[hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1717 dirty pages from their submits' captures, 51266 from earlier ones', 55 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 944 dispatches, 40635 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7655 runs, 675.0 MB compared; 119 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17958 in captures, 8 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool: 0x7f84000000 (4 MB, captured, state 2): 48; 0x7fa2000000 (8 MB, captured, state 2): 7;
[hos host] vk: program picks: 13658400 kept (5433737 by the last picks' textures, a uniform buffer bound since), 26345 new, 5315254 with other textures (of 19000000)
[hos host] vk: pool 0x7fb5c00000..0x7fb8800000 (44 MB) trimmed to its 2 live part(s), 16 MB: the rest outside every game pool for 1856 submissions (3 trims, 120 MB in all)
[hos host] vk: pool 0x7fb6000000..0x7fb6400000: 344 KB the CPU wrote since retired pool 0x7fb5c00000's last sync kept from guest memory, not its mirror
[hos host] vk: pool 0x7fb7c00000..0x7fb8800000: 536 KB the CPU wrote since retired pool 0x7fb5c00000's last sync kept from guest memory, not its mirror
[hos host] vk: replay reads of guest memory as submitted: 9300000 (5584189 pages from the captures)
[hos Presentation Thread] vi: 14978 frames presented (54.4 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 26, 15.5-17.8 164, 17.8-22 49, 22-30 22, 30-40 9, over 40 2
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #185000: event 0x100d6c8034 mode 0 value 184999 (lr 0x806616b0)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15346 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] vk: capture: a submit 0.37 ms (lock 0.07, scan 0.06 over 591 ranges, 3606 MB), 565 pages written, 565 copied (2.2 MB)
[hos host] vk: a capture taken (ms): all 0.052; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.030, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 6896 texture pages copied (26 MB)
[hos sead::MainThread] nvn: 1900000 recordings made, 72864 live copies (24.3 MB), 6231 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.05, writes collected 0.19
[hos host] vk: replay reads of guest memory as submitted: 9320000 (5596500 pages from the captures)
[hos host] vk: memory: working set 15472 MB (peak 15952), private 21787 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 912 MB, 35 allocations 832 MB; used 3412 of 15579 MB; heap 1 (device): 163 blocks 4951 MB, 19498 allocations 4009 MB; used 7074 of 8421 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 19 blocks 608 MB, 115 images 302 MB; 69 free ranges, the largest 28 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 966 images 1294 MB; 381 free ranges, the largest 26 MB
[hos host] vk: render target images: 5911 reused from the pool, 4185 made new, 4068 destroyed from it, 1159 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 150 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1792 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 576 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3876 idle
[hos host] vk: VRAM: 7074 MB used of 8421 MB allowed (941 MB of it free in the blocks); textures 2232 MB (budget 2116 MB), 1087 textures
[hos host] vk: budget: VRAM 4951 of 8421 MB (7074 MB by the driver, 2123 MB unaccounted): textures 2232, mirrors 1103, blocks free 941, shadows 284, other buffers 213, other images 111, staging 64; RAM 3346 MB (the process's private memory 21787 MB): imports 2256, capture spares 738, retired imports 224, captures 94, capture table 34, others 0
[hos host] vk: textures: decoded ASTC 1279 MB, rendered 548 MB (70 drawn in one frame, kept: 42773 KB), other 370 MB; used in the last 8 presents 1570 MB
[hos host] vk: decoded ASTC used again: 402 textures (932 MB) copied from ones the game finalized, not decoded; 402 of those retired, 2 listed; 1823 MB of blocks hashed in 316 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 9853 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10e42b7400 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 56 (12053 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB);
[hos sead::MainThread] nvn: EventSignal (CPU) #30000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos Presentation Thread] vk: presents after the CPU's signal (600): 337 waited for it, 3.62 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 51.6 ms; VRAM 6.9 of 8.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6457 queued, 6457 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16325 ms and 347.6 MB read in all); 8761 programs (1626 of them compute variants without their code, 0 graphics programs no note names), 12277 graphics pipelines (1 not in a cache), 716 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2268 pipelines made at create taken by draws; 3153 create calls with nothing kept for their stages, 0 before the index; units 9610 (6457 ready, 0 loading), 17770 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13948 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 711 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: 9452 programs from the program cache
[hos host] vk: a frame: 172 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.36, present 0.20, draw 6.86, descriptors 3.22, texture lookups 0.71, uploads 0.54, pipelines 0.26, programs 1.12, commands 1.75, dispatch 1.00, memory sync 1.55, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 437 (1103 MB): 8.24 MB synced in 1078.5 copies, 17.78 MB written back
[hos host] vk: 15000 presents, 53.4 fps; a frame: 1851 draws (0 skipped), 157 dispatches, 154 passes, 132 barriers, 6 submits, 1.6 uploads (1.7 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14144 pipelines (3721 ms), 0 programs (10 failed), 1087 textures (2232 MB)
[hos host] vk: host time in GPU resource waits: 4.98 ms a frame (fences 1.60, timelines 3.38, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 606.8 full barriers recorded, 273.5 left out (nothing recorded since the one before), 14.3 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.7; 0x12: 68.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.0, an op 986.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.6 ms a frame (the longest submission 5.7 ms); 2.2 textures made a frame: colour targets 0.6 (1.5 MB), depth 0.2 (1.1 MB), ASTC or BC 1.5 (2.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.81 MB unswizzled on the CPU in 0.17 ms
[hos host] vk: the render thread (600 presents, 17.55 ms a frame): its jobs 16.26 ms a frame (submits 5.51, presents 4.89, others 5.86), idle 1.29; the game waited for room at 34 presents, 0.83 ms each (0.05 a frame; 14.74 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 339 captures at a CPU signal; the submits' waits read in 0.040 ms a frame; readbacks: 1875 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.42 ms after its hand-over (38.22 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.96 + 0.30 (the window's copy is in the idle), after it 1.42 + 0.76; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.55 parts a frame recorded in the signal's job, 71.1% of them took nothing that came with the signal; of the others, 2.45 a frame took pages of the signal's capture into a mirror (126.2 pages a frame) and 0.02 pages read as guest memory had them (14.4 pages a frame); a frame's first part took nothing in 9 of 339 frames, its first three in 8
[hos host] vk: frame ahead, what came with the signal: 339 signals with a capture, each of 403.4 pages in 13.3 pools; between a frame's last submit and that signal 0.68 texture and sampler registrations, 0.88 releases and 0.00 other CPU signals a signal (65 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.21+0.12, 0.40+0.13, 0.08+0.03, 0.06+0.02, 0.18+0.05, 0.33+0.35, 0.08+0.04, 0.06+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: pipeline cache: play file 9 saved (36 KB): 1 pipelines, 1 from the driver's cache, 0 compiled; 1189 more from the start's cache
[hos host] vk: sparse mirrors: 127, 1103 MB with memory of 1552 MB (448 MB spared); blocks given memory later 12262, given back 15533; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1080 binds (275.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 273 taken from the spares, 164 made new; 310 kept as spares, 0 destroyed; 37 spares now, 1208 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 671 dirty pages from their submits' captures, 56345 from earlier ones', 5 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 966 dispatches, 43306 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8033 runs, 709.3 MB compared; 65 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19569 in captures, 5 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa5c00000 (4 MB, captured, state 2): 2; 0x7f86000000 (4 MB, captured, state 2): 2; 0x7f84000000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 189 live (2256 MB), 7 retired still imported (224 MB), 41 retired given back (952 MB)
[hos Presentation Thread] vi: real-time step on: ratio 0.995, game's frame rate 60
[hos host] vk: render targets drawn over from another start, let go: 10647 so far (50241 MB), 9573 of them (45308 MB) between the passes
[hos host] vk: texture #20270 0x116f87b000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 15010 (drawn at 14830)
[hos host] vk: replay reads of guest memory as submitted: 9340000 (5608736 pages from the captures)
[hos host] vk: 2 textures retired (1086 left, 2219 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: texture #20481 0x10004bf1c00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 15041 (drawn at 14980)
[hos host] vk: replay reads of guest memory as submitted: 9360000 (5618773 pages from the captures)
[hos sead::MainThread] vk: readbacks of GPU results: 1053 of 3000 waited for their write-back, 0.09 ms each; 0 looked at the stopped queues first, 37 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1662 kept their virtual core (nothing there to wait for), 3202 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: recorder: 5640656 packets run in 10.0 s (busy 1.63 s, spun idle 0.11 s, slept 10428 times, 3313 of them under 200 us), 0 calls made directly (no room); 17330 batches over 571 frames (30.4 a frame, 325 packets a batch), 8508 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 157 times, 9.69 ms (0.3 times, 0.02 ms a frame); every thread 782 times, 47.42 ms in all, the longest 1.04 ms; vkWaitForFences 490/27.34 ms (max 0.59), vmaDestroyBuffer 59/4.94 ms (max 1.04), vkDestroyImageView 65/4.90 ms (max 0.31), vkQueueBindSparse 114/4.80 ms (max 0.67), vmaFreeMemory 20/2.63 ms (max 0.42), vkEndCommandBuffer 3/0.90 ms (max 0.36)
[hos host] vk: recorder: of its busy 1.63 s, 0.40 s in parts' ends and submits (the longest 1.45 ms), 0.08 s of that blocked on the queue's lock (154 times, the longest 1.38 ms), so 0.32 s in the ends, submits and holds; 1.23 s running the other packets; 9879 packets a frame
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 340 dirty pages from their submits' captures, 58162 from earlier ones', 249 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 933 dispatches, 44865 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8394 runs, 759.5 MB compared; 68 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20042 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 249;
[hos host] vk: replay reads of guest memory as submitted: 9380000 (5629559 pages from the captures)
[hos host] vk: program picks: 13811946 kept (5508826 by the last picks' textures, a uniform buffer bound since), 26347 new, 5361706 with other textures (of 19200000)
[hos ModuleSystemWorker1] vcores: 158529 takes, 37801 waits (mean 84 us, longest 35.8 ms, 0 over HOS_VCORE_LOG_MS), 18511 kicks, 18361 yields, 0 starved grants, 0 escapes in 5 s
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 591 ranges, 3606 MB), 273 pages written, 273 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 6443 texture pages copied (25 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 18067 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: readback pages refreshed from guest memory: 4743016 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 9400000 (5642192 pages from the captures)
[hos host] vk: sparse mirrors: 120, 1042 MB with memory of 1472 MB (429 MB spared); blocks given memory later 12262, given back 15535; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1085 binds (276.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 273 taken from the spares, 164 made new; 317 kept as spares, 0 destroyed; 44 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 142 dirty pages from their submits' captures, 60322 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 951 dispatches, 45059 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8194 runs, 741.8 MB compared; 64 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20707 in captures, 3 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 189 live (2256 MB), 7 retired still imported (224 MB), 41 retired given back (952 MB)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #260000: event 0x100d6c802c mode 0 value 187182 (now 187163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #20338 0x116fbd1000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 15115 (drawn at 14946)
[hos host] vk: replay reads of guest memory as submitted: 9420000 (5650967 pages from the captures)
[hos host] vk: a frame: 152 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.04, present 0.17, draw 5.55, descriptors 2.34, texture lookups 0.08, uploads 0.01, pipelines 0.23, programs 1.02, commands 1.51, dispatch 0.82, memory sync 0.48, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 437 (1037 MB): 5.44 MB synced in 1053.8 copies, 17.67 MB written back
[hos host] vk: 15120 presents, 60.1 fps; a frame: 1837 draws (0 skipped), 150 dispatches, 150 passes, 131 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14144 pipelines (3721 ms), 0 programs (10 failed), 1095 textures (2224 MB)
[hos host] vk: host time in GPU resource waits: 1.53 ms a frame (fences 1.40, timelines 0.13, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 569.3 full barriers recorded, 268.8 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1020.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.1 ms a frame (the longest submission 4.3 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 103 dirty pages from their submits' captures, 58666 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 950 dispatches, 45239 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8383 runs, 759.4 MB compared; 57 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20701 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 9440000 (5661198 pages from the captures)
[hos sead::MainThread] barrier: 238370 looks: 90914 found the masks moved already, 144864 waited until they moved, 2592 for the whole 1 ms (a wait 320 us on average); 229561 wakes
[hos host] vk: replay reads of guest memory as submitted: 9460000 (5671851 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 94 sync waits (343 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 38 made): 28202372 lookups, 147255 missed (0.52 %)
[hos host] vk: program picks: 13964327 kept (5583623 by the last picks' textures, a uniform buffer bound since), 26354 new, 5409318 with other textures (of 19400000)
[hos host] vk: texture #18847 0x116f0e1000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 15177 (drawn at 13683)
[hos host] vk: sparse mirrors: 121, 1039 MB with memory of 1496 MB (456 MB spared); blocks given memory later 12262, given back 15621; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1087 binds (276.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 274 taken from the spares, 164 made new; 317 kept as spares, 0 destroyed; 43 spares now, 1264 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1099, 4882.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 208 dirty pages from their submits' captures, 59409 from earlier ones', 599 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 982 dispatches, 45780 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8607 runs, 777.3 MB compared; 86 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20637 in captures, 1 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool: 0x7fbc000000 (12 MB, captured, state 2): 563; 0x7fd6400000 (48 MB, captured, state 2): 36;
[hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 593 ranges, 3596 MB), 288 pages written, 288 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.024, bytes copied 0.000; 25 kept (112 MB), 10 spare buffers, 5350 texture pages copied (20 MB)
[hos host] vk: replay reads of guest memory as submitted: 9480000 (5684149 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 19335 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #20474 0x10004b01a00 2560x1440 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980)
[hos host] vk: texture #20475 0x10004fb1e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980)
[hos host] vk: texture #20476 0x100050a2000 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980)
[hos host] vk: texture #20477 0x10004ec1c00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980)
[hos host] vk: texture #20511 0x10001fe0800 640x360 x1 NVN 0x14 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1120 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1189 kept their virtual core (nothing there to wait for), 2907 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 440000 so far (x1 0x1178b9a000 x2 0 x3 0x1178b9a000 x4 0x654040)
[hos host] vk: reaper: 711 images (1493 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 142 (223 MB)
[hos host] vk: replay reads of guest memory as submitted: 9500000 (5696363 pages from the captures)
[hos host] vk: 53 textures retired (1126 left, 2334 MB): budget 53, drawn over 0, dead targets 0
[hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 817 dirty pages from their submits' captures, 56731 from earlier ones', 15 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 971 dispatches, 43865 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7957 runs, 723.0 MB compared; 90 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19501 in captures, 2 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 10; 0x1042000000 (4 MB, captured, state 2): 4; 0x7fa2000000 (8 MB, captured, state 2): 1;
[hos host] vk: 60 textures retired (1115 left, 2297 MB): budget 60, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9520000 (5708354 pages from the captures)
[hos host] vk: 3 textures retired (1114 left, 2293 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: a frame: 188 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.73, present 0.15, draw 6.20, descriptors 2.84, texture lookups 0.51, uploads 0.38, pipelines 0.24, programs 1.06, commands 1.58, dispatch 0.81, memory sync 0.48, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 438 (1039 MB): 5.47 MB synced in 1060.8 copies, 17.68 MB written back
[hos host] vk: 15240 presents, 58.3 fps; a frame: 1896 draws (0 skipped), 151 dispatches, 156 passes, 132 barriers, 6 submits, 1.1 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14144 pipelines (3721 ms), 0 programs (10 failed), 1114 textures (2293 MB)
[hos host] vk: host time in GPU resource waits: 3.04 ms a frame (fences 1.52, timelines 1.53, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 601.1 full barriers recorded, 274.2 left out (nothing recorded since the one before), 14.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.6; 0x12: 68.9; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.7 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.1, an op 1043.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.5 ms a frame (the longest submission 4.6 ms); 1.4 textures made a frame: colour targets 0.3 (1.0 MB), depth 0.1 (0.6 MB), ASTC or BC 1.0 (1.7 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.06 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: replay reads of guest memory as submitted: 9540000 (5720276 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #190000: event 0x100d6c8040 mode 0 value 189999 (lr 0x806621e8)
[hos host] vk: texture #18104 0x116e859000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 15246 (drawn at 13394)
[hos host] vk: slow upload 22.5 ms: #20779 0x10f72be800 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB
[hos host] vk: 27 textures retired (1147 left, 2379 MB): budget 27, drawn over 0, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 126, 1099 MB with memory of 1576 MB (476 MB spared); blocks given memory later 12270, given back 15631; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1096 binds (278.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 279 taken from the spares, 164 made new; 317 kept as spares, 0 destroyed; 38 spares now, 1184 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1870 dirty pages from their submits' captures, 50352 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 869 dispatches, 39815 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7450 runs, 658.7 MB compared; 55 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17980 in captures, 2 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB)
[hos host] vk: replay reads of guest memory as submitted: 9560000 (5733596 pages from the captures)
[hos host] vk: 5 textures retired (1173 left, 2405 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: texture #19901 0x100021c0600 960x540 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 15267 (drawn at 15206)
[hos host] vk: texture #20521 0x10008942600 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 15267 (drawn at 15206)
[hos host] vk: texture #20534 0x10008942600 640x360 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 15267 (drawn at 15206)
[hos host] vk: texture #20526 0x10004651000 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 15267 (drawn at 15206)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 591 ranges, 3642 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (116 MB), 11 spare buffers, 6358 texture pages copied (24 MB)
[hos host] vk: program picks: 14117171 kept (5657953 by the last picks' textures, a uniform buffer bound since), 26357 new, 5456471 with other textures (of 19600000)
[hos host] vk: replay reads of guest memory as submitted: 9580000 (5748010 pages from the captures)
[hos host] vk: 1 textures retired (1172 left, 2404 MB): budget 0, drawn over 1, dead targets 0
[hos Presentation Thread] vi: 15270 frames presented (58.3 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 27, 15.5-17.8 229, 17.8-22 26, 22-30 7, 30-40 1, over 40 2
[hos host] vk: replay reads of guest memory as submitted: 9600000 (5761879 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 4810773 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1980 dirty pages from their submits' captures, 44027 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 763 dispatches, 34822 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6371 runs, 566.7 MB compared; 55 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15530 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 2 textures retired (1171 left, 2402 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: queues ordered by their events: 96000 submissions stopped at a wait, 95951 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 264735, not met 0; captures at a CPU event's wait 9195, 0 of them at a wait met on arrival
[hos host] vk: 320 mirrors given back (2262 MB in all); 1093 MB of mirrors now, 140 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: replay reads of guest memory as submitted: 9620000 (5776128 pages from the captures)
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #265000: event 0x100d6c802c mode 0 value 190982 (now 190963) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05
[hos host] vk: 1 textures retired (1172 left, 2401 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9640000 (5789140 pages from the captures)
[hos host] vk: late pages' fix (600 recordings with one): of 1048 captures with late pages 1048 fixed (81.2 pages, 81.7 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 451 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 576 blocks (36864 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 153, of an earlier recording 298, not recorded yet 0, of this part 0; pages fixed from stamps 84022, from snapshots 1075; the shadows: 820081 pages synced from captures, 799886 made known by it, 20195 left unknown because the replay or the collection wrote them since
[hos host] vk: replay reads of guest memory as submitted: 9660000 (5803710 pages from the captures)
[hos host] vk: sparse mirrors: 122, 1091 MB with memory of 1556 MB (464 MB spared); blocks given memory later 12282, given back 15631; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1101 binds (278.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 279 taken from the spares, 164 made new; 321 kept as spares, 0 destroyed; 42 spares now, 1204 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2247 dirty pages from their submits' captures, 45284 from earlier ones', 6 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 810 dispatches, 36026 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6580 runs, 584.8 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15995 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 6;
[hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB)
[hos host] vk: 1 textures retired (1180 left, 2402 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9680000 (5817441 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21888 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 2 textures retired (1183 left, 2405 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9700000 (5830796 pages from the captures)
[hos host] vk: program picks: 14272498 kept (5737562 by the last picks' textures, a uniform buffer bound since), 26357 new, 5501144 with other textures (of 19800000)
[hos host] vk: 21 textures retired (1162 left, 2373 MB): budget 21, drawn over 0, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1841 dirty pages from their submits' captures, 43820 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 780 dispatches, 34883 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6430 runs, 568.1 MB compared; 77 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15598 in captures, 0 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 590 ranges, 3640 MB), 249 pages written, 249 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (96 MB), 11 spare buffers, 6958 texture pages copied (27 MB)
[hos host] vk: replay reads of guest memory as submitted: 9720000 (5844432 pages from the captures)
[hos host] vk: 2 textures retired (1160 left, 2370 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: a frame: 306 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.30, present 0.17, draw 7.87, descriptors 3.42, texture lookups 0.41, uploads 0.28, pipelines 0.31, programs 1.31, commands 2.20, dispatch 0.86, memory sync 0.51, queue submit 0.00, slot wait 0.29
[hos host] vk: a frame: mirrors 444 (1096 MB): 6.14 MB synced in 1087.6 copies, 17.74 MB written back
[hos host] vk: 15360 presents, 56.4 fps; a frame: 2476 draws (0 skipped), 155 dispatches, 189 passes, 135 barriers, 6 submits, 0.9 uploads (0.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14144 pipelines (3721 ms), 0 programs (10 failed), 1160 textures (2370 MB)
[hos host] vk: host time in GPU resource waits: 7.08 ms a frame (fences 2.35, timelines 4.73, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 721.5 full barriers recorded, 307.5 left out (nothing recorded since the one before), 17.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.7, an op 1613.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.2 ms a frame (the longest submission 3.7 ms); 0.9 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.8 (1.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.27 MB unswizzled on the CPU in 0.04 ms
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1101 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 11 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1343 kept their virtual core (nothing there to wait for), 3009 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos nn.npln.Worker] vcores: 143257 takes, 24474 waits (mean 71 us, longest 17.1 ms, 0 over HOS_VCORE_LOG_MS), 11744 kicks, 11708 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 9740000 (5858675 pages from the captures)
[hos host] vk: 15 textures retired (1145 left, 2347 MB): budget 15, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9760000 (5871928 pages from the captures)
[hos host] vk: sparse mirrors: 122, 1097 MB with memory of 1564 MB (466 MB spared); blocks given memory later 12321, given back 15631; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1108 binds (279.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 280 taken from the spares, 164 made new; 322 kept as spares, 0 destroyed; 42 spares now, 1196 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 8944 dirty pages from their submits' captures, 61324 from earlier ones', 18757 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 921 dispatches, 38108 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6592 runs, 585.2 MB compared; 90 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15902 in captures, 78 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool: 0x7fbd800000 (60 MB, captured, state 2): 12488; 0x7fd6400000 (48 MB, captured, state 2): 3302; 0x7fab000000 (8 MB, captured, state 2): 1636; 0x7f9c400000 (32 MB, captured, state 2): 594; 0x7faf400000 (4 MB, captured, state 2): 295; 0x7f8e000000 (4 MB, captured, state 2): 278; 0x7fa9000000 (4 MB, captured, state 2): 98; 0x7fdfc00000 (12 MB, captured, state 2): 28;
[hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB)
[hos OptimizeChangeBanc] nvn: 1950000 recordings made, 72629 live copies (25.3 MB), 6235 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: 27 textures retired (1118 left, 2293 MB): budget 27, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9780000 (5886425 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.12
[hos host] vk: replay reads of guest memory as submitted: 9800000 (5899561 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 826, 3665.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 7661 dirty pages from their submits' captures, 62643 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 938 dispatches, 38010 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6423 runs, 583.2 MB compared; 89 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15938 in captures, 1 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: texture #13893 0x10005280e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 214 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 10 textures retired (1132 left, 2286 MB): budget 10, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB
[hos host] vk: replay reads of guest memory as submitted: 9820000 (5913333 pages from the captures)
[hos host] vk: program picks: 14428169 kept (5817159 by the last picks' textures, a uniform buffer bound since), 26357 new, 5545473 with other textures (of 20000000)
[hos host] vk: replay reads of guest memory as submitted: 9840000 (5927465 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 219 sync waits (1031 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos Presentation Thread] vk: capture: a submit 0.43 ms (lock 0.04, scan 0.08 over 578 ranges, 3642 MB), 797 pages written, 681 copied (2.7 MB)
[hos host] vk: a capture taken (ms): all 0.053; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.027, bytes copied 0.000; 24 kept (218 MB), 12 spare buffers, 11599 texture pages copied (45 MB)
[hos host] vk: replay reads of guest memory as submitted: 9860000 (5940847 pages from the captures)
[hos host] vk: 26 textures retired (1122 left, 2291 MB): budget 26, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 4877348 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16531 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 122, 1112 MB with memory of 1564 MB (451 MB spared); blocks given memory later 13381, given back 16453; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1141 binds (286.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 280 taken from the spares, 164 made new; 322 kept as spares, 0 destroyed; 42 spares now, 1196 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 13005 dirty pages from their submits' captures, 73557 from earlier ones', 2567 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 915 dispatches, 38592 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6248 runs, 565.5 MB compared; 87 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15949 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb7c00000 (12 MB, captured, state 2): 2567;
[hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB)
[hos host] vk: replay reads of guest memory as submitted: 9880000 (5954870 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 6 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 1 textures retired (1129 left, 2292 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9900000 (5968038 pages from the captures)
[hos OptimizeChangeBanc] nvn: 2000000 recordings made, 77008 live copies (29.1 MB), 6235 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: a frame: 340 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 15.67, present 0.18, draw 8.81, descriptors 3.71, texture lookups 0.29, uploads 0.17, pipelines 0.35, programs 1.52, commands 2.53, dispatch 1.03, memory sync 1.75, queue submit 0.00, slot wait 0.30
[hos host] vk: a frame: mirrors 445 (1129 MB): 10.25 MB synced in 1115.1 copies, 17.68 MB written back
[hos host] vk: 15480 presents, 47.4 fps; a frame: 2568 draws (0 skipped), 151 dispatches, 195 passes, 134 barriers, 6 submits, 0.6 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14146 pipelines (3722 ms), 0 programs (10 failed), 1141 textures (2302 MB)
[hos host] vk: host time in GPU resource waits: 6.82 ms a frame (fences 1.93, timelines 4.89, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 722.9 full barriers recorded, 308.2 left out (nothing recorded since the one before), 22.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1644.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.6 ms a frame (the longest submission 5.9 ms); 0.5 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.5 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 9920000 (5981625 pages from the captures)
[hos ParallelDecompThread] zstd: 4 stream(s) decoding on the host at once
[hos BaseProcExecutor0] vcores: BaseProcExecutor0 (priority 16) waited 70.6 ms for a core: given by sead::MainThread (priority 16), which held it 70.7 ms and let it go 68.0 ms after it was asked; the cores it may use now: core 0: it;
[hos host] vk: late pages' fix: stamps for 0x1051bf0000 (pool 0x1051800000; 640 blocks, 40960 KB)
[hos host] vk: late pages' fix: stamps for 0x11373f0000 (pool 0x1137000000; 704 blocks, 45056 KB)
[hos host] vk: late pages' fix: stamps for 0x11377f0000 (pool 0x1137000000; 768 blocks, 49152 KB)
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 14881 dirty pages from their submits' captures, 90635 from earlier ones', 84 from guest memory as it was at the replay; 12 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1006 dispatches, 39888 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6353 runs, 581.7 MB compared; 92 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15461 in captures, 5 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool: 0x7fdfc00000 (12 MB, captured, state 2): 84;
[hos host] vk: texture #16205 0x10004bf1600 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 253 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 19 textures retired (1140 left, 2291 MB): budget 19, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: (its written part: 3 of 3 KB readable, 0 bytes not zero; levels L0)
[hos host] vk: texture 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6 (39444 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it
[hos host] vk: (its written part: 0 of 0 KB readable, 288 bytes not zero; levels L0)
[hos host] vk: texture 0x10dff20200 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 0 KB of it
[hos host] vk: replay reads of guest memory as submitted: 9940000 (5995005 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #270000: event 0x100d6c802c mode 0 value 194782 (now 194763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.10, writes collected 0.25
[hos host] vk: program picks: 14583155 kept (5895985 by the last picks' textures, a uniform buffer bound since), 26358 new, 5590486 with other textures (of 20200000)
[hos host] vk: 1 textures retired (1145 left, 2287 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: slow upload 15.5 ms: #20943 0x10dae87a00 768x768x1 x1 levels 10 NVN 0x7b ASTC, 528 of 528 KB
[hos host] vk: replay reads of guest memory as submitted: 9960000 (6008064 pages from the captures)
[hos Presentation Thread] vi: 15509 frames presented (47.6 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 5, 15.5-17.8 74, 17.8-22 104, 22-30 42, 30-40 10, over 40 4
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #195000: event 0x100d6c800c mode 0 value 195002 (lr 0x80660adc)
[hos host] vk: replay reads of guest memory as submitted: 9980000 (6021927 pages from the captures)
[hos host] vk: sparse mirrors: 122, 1107 MB with memory of 1536 MB (428 MB spared); blocks given memory later 13728, given back 16473; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1186 binds (289.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 282 taken from the spares, 164 made new; 324 kept as spares, 0 destroyed; 42 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 8793 dirty pages from their submits' captures, 80196 from earlier ones', 2 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 934 dispatches, 38828 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6178 runs, 564.3 MB compared; 387 pages read into the shadows; 284 MB of shadows (51 MB of pages known, in 102 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15149 in captures, 465 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool: 0x7f98400000 (4 MB, captured, state 2): 2;
[hos host] vk: imports: 190 live (2260 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 949 of 3000 waited for their write-back, 0.15 ms each; 0 looked at the stopped queues first, 8 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1108 kept their virtual core (nothing there to wait for), 2988 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 10000000 (6034958 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.61 ms (lock 0.06, scan 0.06 over 590 ranges, 3610 MB), 931 pages written, 931 copied (3.6 MB)
[hos host] vk: a capture taken (ms): all 0.109; GPU wait for late pages 0.014 (1 of 600 takes), pools lock 0.000, retires and stamps 0.066, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 22828 texture pages copied (89 MB)
[hos host] vk: replay reads of guest memory as submitted: 10020000 (6047343 pages from the captures)
[hos host] vk: texture #20946 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (8 so far)
[hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2158 dirty pages from their submits' captures, 45066 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 788 dispatches, 36174 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6356 runs, 581.8 MB compared; 75 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15464 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 10040000 (6060549 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 34538 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 10060000 (6073721 pages from the captures)
[hos host] vk: presenter thread: the copy's fence 1.77 ms, the present 0.75 ms a present
[hos host] vk: recorder: 6159760 packets run in 10.0 s (busy 1.65 s, spun idle 0.12 s, slept 11017 times, 3419 of them under 200 us), 0 calls made directly (no room); 16947 batches over 521 frames (32.5 a frame, 363 packets a batch), 8771 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 144 times, 10.07 ms (0.3 times, 0.02 ms a frame); every thread 806 times, 66.68 ms in all, the longest 1.07 ms; vkWaitForFences 467/37.13 ms (max 1.07), the queue's lock 84/11.12 ms (max 0.75), vmaFreeMemory 21/5.22 ms (max 0.89), vkDestroyImageView 40/3.61 ms (max 0.96), vkQueueBindSparse 108/3.01 ms (max 0.50), vmaDestroyBuffer 57/2.79 ms (max 0.73)
[hos host] vk: recorder: of its busy 1.65 s, 0.44 s in parts' ends and submits (the longest 1.26 ms), 0.12 s of that blocked on the queue's lock (210 times, the longest 1.22 ms), so 0.32 s in the ends, submits and holds; 1.22 s running the other packets; 11823 packets a frame
[hos host] vk: program picks: 14736938 kept (5972729 by the last picks' textures, a uniform buffer bound since), 26361 new, 5636700 with other textures (of 20400000)
[hos host] vk: replay reads of guest memory as submitted: 10080000 (6085698 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 376 pages in 5 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 112, 1070 MB with memory of 1496 MB (425 MB spared); blocks given memory later 13728, given back 16889; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1196 binds (290.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 282 taken from the spares, 164 made new; 334 kept as spares, 0 destroyed; 52 spares now, 1264 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 826, 3665.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1785 dirty pages from their submits' captures, 45841 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 785 dispatches, 36556 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6322 runs, 581.9 MB compared; 79 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16297 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 190 live (2260 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB)
[hos host] vk: replay reads of guest memory as submitted: 10100000 (6097447 pages from the captures)
[hos host] vk: late pages' fix: stamps for 0x10383f0000 (pool 0x1038000000; 832 blocks, 53248 KB)
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 460000 so far (x1 0x117da08020 x2 0 x3 0x117da08000 x4 0x9b040)
[hos BaseProcExecutor1] vcores: 147891 takes, 42524 waits (mean 92 us, longest 70.6 ms, 1 over HOS_VCORE_LOG_MS), 20751 kicks, 20682 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 10120000 (6109020 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.12
[hos host] vk: memory: working set 15681 MB (peak 16021), private 21990 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 12 blocks 928 MB, 36 allocations 850 MB; used 3360 of 15579 MB; heap 1 (device): 163 blocks 4951 MB, 19364 allocations 4078 MB; used 7074 of 8536 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 18 blocks 576 MB, 125 images 322 MB; 71 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 53 blocks 1696 MB, 1037 images 1356 MB; 377 free ranges, the largest 32 MB
[hos host] vk: render target images: 5911 reused from the pool, 4234 made new, 4107 destroyed from it, 1160 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1836 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 581 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3915 idle
[hos host] vk: VRAM: 7074 MB used of 8536 MB allowed (873 MB of it free in the blocks); textures 2314 MB (budget 2134 MB), 1168 textures
[hos host] vk: budget: VRAM 4951 of 8536 MB (7074 MB by the driver, 2123 MB unaccounted): textures 2314, mirrors 1074, blocks free 873, shadows 284, other buffers 229, other images 111, staging 64; RAM 3297 MB (the process's private memory 21990 MB): imports 2260, capture spares 748, retired imports 152, captures 88, capture table 35, other buffers 14, others 0
[hos host] vk: textures: decoded ASTC 1409 MB, rendered 535 MB (82 drawn in one frame, kept: 47264 KB), other 337 MB; used in the last 8 presents 2098 MB
[hos host] vk: decoded ASTC used again: 404 textures (933 MB) copied from ones the game finalized, not decoded; 404 of those retired, 0 listed; 1912 MB of blocks hashed in 328 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 10453 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10e42b7400 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 61 (12932 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 407 waited for it, 5.40 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 126.5 ms; VRAM 6.9 of 8.3 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6459 queued, 6459 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16328 ms and 347.7 MB read in all); 8763 programs (1626 of them compute variants without their code, 0 graphics programs no note names), 12279 graphics pipelines (1 not in a cache), 716 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2282 pipelines made at create taken by draws; 3153 create calls with nothing kept for their stages, 0 before the index; units 9612 (6459 ready, 0 loading), 17669 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13952 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 711 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 273 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 15.55, present 0.16, draw 9.17, descriptors 3.85, texture lookups 0.29, uploads 0.18, pipelines 0.36, programs 1.62, commands 2.62, dispatch 0.95, memory sync 1.41, queue submit 0.00, slot wait 0.32
[hos host] vk: a frame: mirrors 447 (1074 MB): 7.81 MB synced in 1106.9 copies, 17.64 MB written back
[hos host] vk: 15600 presents, 45.2 fps; a frame: 2706 draws (0 skipped), 149 dispatches, 218 passes, 135 barriers, 6 submits, 0.5 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14146 pipelines (3722 ms), 0 programs (10 failed), 1168 textures (2314 MB)
[hos host] vk: host time in GPU resource waits: 9.93 ms a frame (fences 2.61, timelines 7.32, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 722.7 full barriers recorded, 309.8 left out (nothing recorded since the one before), 44.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.5, an op 1803.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 13.7 ms a frame (the longest submission 14.8 ms); 0.4 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: the render thread (600 presents, 18.96 ms a frame): its jobs 17.32 ms a frame (submits 4.45, presents 4.67, others 8.20), idle 1.64; the game waited for room at 29 presents, 0.37 ms each (0.02 a frame; 15.24 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 408 captures at a CPU signal; the submits' waits read in 0.041 ms a frame; readbacks: 2148 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 16.40 ms after its hand-over (45.09 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.73 + 0.21 (the window's copy is in the idle), after it 2.32 + 0.89; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 10.96 parts a frame recorded in the signal's job, 68.3% of them took nothing that came with the signal; of the others, 3.46 a frame took pages of the signal's capture into a mirror (166.6 pages a frame) and 0.01 pages read as guest memory had them (7.0 pages a frame); a frame's first part took nothing in 11 of 407 frames, its first three in 11
[hos host] vk: frame ahead, what came with the signal: 408 signals with a capture, each of 565.7 pages in 14.4 pools; between a frame's last submit and that signal 2.02 texture and sampler registrations, 0.09 releases and 0.00 other CPU signals a signal (116 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.71+0.15, 0.37+0.10, 0.25+0.07, 0.30+0.06, 0.05+0.02, 0.10+0.29, 0.22+0.05, 0.17+0.03, 0.05+0.06, 0.07+0.05, 0.02+0.01, 0.01+0.00
[hos host] vk: replay reads of guest memory as submitted: 10140000 (6122830 pages from the captures)
[hos host] vk: render targets drawn over from another start, let go: 10689 so far (50405 MB), 9614 of them (45466 MB) between the passes
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 597 ranges, 3583 MB), 260 pages written, 260 copied (1.0 MB)
[hos host] vk: readback pages refreshed from guest memory: 4944795 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: a capture taken (ms): all 0.037; GPU wait for late pages 0.002 (1 of 600 takes), pools lock 0.000, retires and stamps 0.019, bytes copied 0.000; 24 kept (144 MB), 11 spare buffers, 4432 texture pages copied (17 MB)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2578 dirty pages from their submits' captures, 43019 from earlier ones', 166 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 774 dispatches, 34707 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6291 runs, 566.0 MB compared; 72 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15594 in captures, 0 only in the collection, 22 only by the replay itself
[hos host] vk: read live, by pool: 0x1038000000 (4 MB, captured, state 2): 166;
[hos host] vk: replay reads of guest memory as submitted: 10160000 (6135351 pages from the captures)
[hos ModuleSystemWorker2] nvn: 2050000 recordings made, 73816 live copies (25.4 MB), 6237 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: replay reads of guest memory as submitted: 10180000 (6148284 pages from the captures)
[hos host] vk: a submission's parts ended: 69145 where the game had ended its render pass, 10855 at 512 draws
[hos host] vk: commit: 1770 MB left under Windows' limit as it is now, and the page files can grow by 79851 MB more: counted as room
[hos host] vk: program picks: 14889496 kept (6049570 by the last picks' textures, a uniform buffer bound since), 26361 new, 5684142 with other textures (of 20600000)
[hos host] vk: replay reads of guest memory as submitted: 10200000 (6160443 pages from the captures)
[hos host] vk: sparse mirrors: 113, 1074 MB with memory of 1500 MB (425 MB spared); blocks given memory later 13729, given back 16891; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1200 binds (291.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 283 taken from the spares, 164 made new; 334 kept as spares, 0 destroyed; 51 spares now, 1260 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3730 dirty pages from their submits' captures, 47063 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 803 dispatches, 36624 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6620 runs, 584.5 MB compared; 79 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15567 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 190 live (2260 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB)
[hos host] vk: replay reads of guest memory as submitted: 10220000 (6174128 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10240000 (6186790 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17791 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] vk: readbacks of GPU results: 1079 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1147 kept their virtual core (nothing there to wait for), 2949 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2253 dirty pages from their submits' captures, 43528 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 792 dispatches, 34917 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6271 runs, 549.1 MB compared; 75 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15621 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 10260000 (6197786 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 235 sync waits (1375 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 39 made): 22622743 lookups, 148417 missed (0.66 %)
[hos host] vk: late pages found at the replay (600 takes): 4 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 10280000 (6208637 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #275000: event 0x100d6c802c mode 0 value 198582 (now 198563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.04 over 611 ranges, 3597 MB), 297 pages written, 297 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.029; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 24 kept (136 MB), 12 spare buffers, 5526 texture pages copied (21 MB)
[hos host] vk: pool 0x7fb1c8f900+0x773a90 imported in 6.4 ms as 0x7fb0800000..0x7fb4400000 (233 imports, 714 ms in all)
[hos host] vk: replay reads of guest memory as submitted: 10300000 (6222169 pages from the captures)
[hos host] vk: 337 mirrors given back (2335 MB in all); 1113 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: sparse mirrors: 114, 1113 MB with memory of 1596 MB (482 MB spared); blocks given memory later 13731, given back 16891; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1207 binds (291.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 286 taken from the spares, 165 made new; 337 kept as spares, 0 destroyed; 51 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 799, 3550.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2459 dirty pages from their submits' captures, 43471 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 803 dispatches, 34964 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6441 runs, 566.7 MB compared; 83 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15019 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 189 live (2260 MB), 7 retired still imported (212 MB), 43 retired given back (1024 MB)
[hos host] vk: replay reads of guest memory as submitted: 10320000 (6235239 pages from the captures)
[hos host] vk: program picks: 15042135 kept (6127554 by the last picks' textures, a uniform buffer bound since), 26377 new, 5731487 with other textures (of 20800000)
[hos host] vk: a frame: 261 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 13.59, present 0.16, draw 8.64, descriptors 3.55, texture lookups 0.11, uploads 0.03, pipelines 0.34, programs 1.64, commands 2.43, dispatch 0.90, memory sync 0.64, queue submit 0.00, slot wait 0.33
[hos host] vk: a frame: mirrors 451 (1113 MB): 6.35 MB synced in 1103.6 copies, 17.72 MB written back
[hos host] vk: 15720 presents, 52.3 fps; a frame: 2671 draws (0 skipped), 153 dispatches, 220 passes, 134 barriers, 6 submits, 0.3 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14153 pipelines (3725 ms), 0 programs (10 failed), 1201 textures (2358 MB)
[hos host] vk: host time in GPU resource waits: 7.28 ms a frame (fences 1.57, timelines 5.70, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 732.9 full barriers recorded, 311.5 left out (nothing recorded since the one before), 45.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1764.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 13.2 ms a frame (the longest submission 3.8 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.15 MB unswizzled on the CPU in 0.02 ms
[hos host] vk: replay reads of guest memory as submitted: 10340000 (6247330 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10360000 (6260685 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2220 dirty pages from their submits' captures, 44012 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 801 dispatches, 35044 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6421 runs, 566.6 MB compared; 78 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15939 in captures, 1 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 10380000 (6274114 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10400000 (6288002 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10420000 (6301288 pages from the captures)
[hos Presentation Thread] vi: 15771 frames presented (52.4 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 61, 17.8-22 189, 22-30 11, 30-40 0, over 40 1
[hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #200000: event 0x100d6c8010 mode 0 value 200000 (lr 0x80660adc)
[hos host] vk: readback pages refreshed from guest memory: 5011880 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 115, 1116 MB with memory of 1600 MB (483 MB spared); blocks given memory later 13731, given back 16900; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1212 binds (292.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 288 taken from the spares, 165 made new; 338 kept as spares, 0 destroyed; 50 spares now, 1220 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2649 dirty pages from their submits' captures, 44301 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 807 dispatches, 35536 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6441 runs, 567.1 MB compared; 74 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15262 in captures, 0 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 189 live (2260 MB), 7 retired still imported (212 MB), 43 retired given back (1024 MB)
[hos host] vk: replay reads of guest memory as submitted: 10440000 (6314784 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 615 ranges, 3609 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.033; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (116 MB), 6 spare buffers, 6286 texture pages copied (24 MB)
[hos host] vk: replay reads of guest memory as submitted: 10460000 (6326851 pages from the captures)
[hos host] vk: program picks: 15192250 kept (6202680 by the last picks' textures, a uniform buffer bound since), 26383 new, 5781366 with other textures (of 21000000)
[hos host] vk: replay reads of guest memory as submitted: 10480000 (6338835 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16039 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: pool 0x7fb0800000..0x7fb4400000 (60 MB) trimmed to its 1 live part(s), 20 MB: the rest outside every game pool for 1856 submissions (4 trims, 160 MB in all)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2254 dirty pages from their submits' captures, 44641 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 775 dispatches, 35713 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6449 runs, 567.0 MB compared; 63 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16283 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 10500000 (6350926 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 10520000 (6364029 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10540000 (6377142 pages from the captures)
[hos sead::MainThread] vk: readbacks of GPU results: 1095 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 8 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1257 kept their virtual core (nothing there to wait for), 3095 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: sparse mirrors: 115, 1124 MB with memory of 1568 MB (443 MB spared); blocks given memory later 13732, given back 16900; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1217 binds (292.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 290 taken from the spares, 165 made new; 340 kept as spares, 0 destroyed; 50 spares now, 1252 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 849, 3772.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1927 dirty pages from their submits' captures, 46259 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 816 dispatches, 36915 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6647 runs, 601.0 MB compared; 87 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15866 in captures, 0 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 189 live (2220 MB), 8 retired still imported (272 MB), 43 retired given back (1024 MB)
[hos host] vk: replay reads of guest memory as submitted: 10560000 (6389602 pages from the captures)
[hos host] vk: a frame: 276 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 13.53, present 0.15, draw 8.73, descriptors 3.58, texture lookups 0.10, uploads 0.02, pipelines 0.35, programs 1.61, commands 2.50, dispatch 0.92, memory sync 0.51, queue submit 0.00, slot wait 0.30
[hos host] vk: a frame: mirrors 455 (1124 MB): 5.92 MB synced in 1100.7 copies, 17.71 MB written back
[hos host] vk: 15840 presents, 51.8 fps; a frame: 2651 draws (0 skipped), 152 dispatches, 216 passes, 134 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14153 pipelines (3725 ms), 0 programs (10 failed), 1247 textures (2385 MB)
[hos host] vk: host time in GPU resource waits: 7.75 ms a frame (fences 1.94, timelines 5.81, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 730.7 full barriers recorded, 312.2 left out (nothing recorded since the one before), 40.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.2, an op 1725.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 13.7 ms a frame (the longest submission 3.1 ms); 0.4 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: texture #14173 0x10004ce1800 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 608 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 13 textures retired (1235 left, 2363 MB): budget 13, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: reaper: 319 images (563 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 59 (86 MB)
[hos host] vk: replay reads of guest memory as submitted: 10580000 (6402191 pages from the captures)
[hos BaseProcExecutor1] vcores: 136465 takes, 23648 waits (mean 81 us, longest 22.6 ms, 0 over HOS_VCORE_LOG_MS), 11121 kicks, 11283 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 10600000 (6416067 pages from the captures)
[hos host] vk: program picks: 15342233 kept (6275489 by the last picks' textures, a uniform buffer bound since), 26386 new, 5831380 with other textures (of 21200000)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 620 ranges, 3569 MB), 258 pages written, 258 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 7568 texture pages copied (29 MB)
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2991 dirty pages from their submits' captures, 45131 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 36051 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6004 runs, 562.4 MB compared; 68 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16172 in captures, 0 only in the collection, 22 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 10620000 (6429191 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10640000 (6443738 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10660000 (6457791 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #280000: event 0x100d6c802c mode 0 value 202382 (now 202363) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 10680000 (6471437 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 117, 1129 MB with memory of 1576 MB (447 MB spared); blocks given memory later 13732, given back 16901; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1220 binds (293.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 292 taken from the spares, 165 made new; 340 kept as spares, 0 destroyed; 48 spares now, 1244 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3584 dirty pages from their submits' captures, 44288 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 783 dispatches, 35507 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5944 runs, 577.5 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15882 in captures, 0 only in the collection, 22 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 189 live (2220 MB), 8 retired still imported (272 MB), 43 retired given back (1024 MB)
[hos ModuleSystemWorker3] tex: the game's texture heap: 261 compaction rounds asked for in 14400 frames without one running
[hos host] vk: 19 textures retired (1245 left, 2408 MB): budget 19, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 10700000 (6484123 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10720000 (6497564 pages from the captures)
[hos host] vk: texture #16197 0x10005191c00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 493 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 3 textures retired (1251 left, 2391 MB): budget 3, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB
[hos host] vk: readback pages refreshed from guest memory: 5078691 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3508 dirty pages from their submits' captures, 44910 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 804 dispatches, 35904 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5825 runs, 560.9 MB compared; 68 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16430 in captures, 1 only in the collection, 24 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 10740000 (6511951 pages from the captures)
[hos host] vk: program picks: 15492266 kept (6345596 by the last picks' textures, a uniform buffer bound since), 26386 new, 5881347 with other textures (of 21400000)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 25678 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] nvn: GPU waits in 5 s: 264 sync waits (1403 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 10760000 (6525057 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 603 ranges, 3532 MB), 291 pages written, 291 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.036; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 24 kept (92 MB), 12 spare buffers, 8939 texture pages copied (34 MB)
[hos host] vk: a frame: 289 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.50, present 0.18, draw 8.25, descriptors 3.48, texture lookups 0.18, uploads 0.07, pipelines 0.33, programs 1.52, commands 2.28, dispatch 0.78, memory sync 0.47, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 460 (1137 MB): 5.87 MB synced in 1095.8 copies, 17.54 MB written back
[hos host] vk: 15960 presents, 55.6 fps; a frame: 2578 draws (0 skipped), 138 dispatches, 213 passes, 134 barriers, 6 submits, 0.5 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14153 pipelines (3725 ms), 0 programs (10 failed), 1252 textures (2392 MB)
[hos host] vk: host time in GPU resource waits: 8.05 ms a frame (fences 3.25, timelines 4.79, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 715.1 full barriers recorded, 311.2 left out (nothing recorded since the one before), 38.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1692.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 13.4 ms a frame (the longest submission 4.2 ms); 0.5 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.5 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.19 MB unswizzled on the CPU in 0.02 ms
[hos host] vk: replay reads of guest memory as submitted: 10780000 (6537532 pages from the captures)
[hos host] vk: late pages' fix (600 recordings with one): of 924 captures with late pages 922 fixed (77.4 pages, 77.8 fixes a capture), waited as before: 0 no pool, 2 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 314 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 6 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 77, of an earlier recording 231, not recorded yet 0, of this part 0; pages fixed from stamps 70535, from snapshots 785; the shadows: 321762 pages synced from captures, 314443 made known by it, 7319 left unknown because the replay or the collection wrote them since
[hos host] vk: late pages found at the replay (600 takes): 17 pages in 17 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 120, 1137 MB with memory of 1592 MB (454 MB spared); blocks given memory later 13732, given back 16901; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1223 binds (295.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 295 taken from the spares, 165 made new; 340 kept as spares, 0 destroyed; 45 spares now, 1228 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2883 dirty pages from their submits' captures, 44322 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 776 dispatches, 35436 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5958 runs, 578.3 MB compared; 66 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15899 in captures, 0 only in the collection, 26 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 191 live (2232 MB), 8 retired still imported (272 MB), 43 retired given back (1024 MB)
[hos host] vk: replay reads of guest memory as submitted: 10800000 (6550719 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 10820000 (6563737 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1099 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 21 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1275 kept their virtual core (nothing there to wait for), 3077 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 480000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040)
[hos BaseProcCreateAndDeleteThread] nvn: 2100000 recordings made, 74329 live copies (25.6 MB), 6242 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: replay reads of guest memory as submitted: 10840000 (6577437 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos rmbl::VibrationThread] hid: 53248 vibration values, strongest 0.00 since the last count
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3117 dirty pages from their submits' captures, 45226 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 857 dispatches, 35915 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5768 runs, 560.6 MB compared; 74 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16275 in captures, 0 only in the collection, 17 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9a400000 (4 MB, captured, state 2): 2;
[hos host] vk: replay reads of guest memory as submitted: 10860000 (6590486 pages from the captures)
[hos host] vk: program picks: 15642906 kept (6414825 by the last picks' textures, a uniform buffer bound since), 26386 new, 5930707 with other textures (of 21600000)
[hos host] vk: texture #16206 0x10004b01400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 544 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 1 textures retired (1255 left, 2392 MB): budget 1, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: replay reads of guest memory as submitted: 10880000 (6602954 pages from the captures)
[hos host] vk: texture #16204 0x10004741200 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 537 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 34 textures retired (1237 left, 2391 MB): budget 34, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 3 MB
[hos host] vk: replay reads of guest memory as submitted: 10900000 (6616535 pages from the captures)
[hos host] vk: sparse mirrors: 120, 1138 MB with memory of 1588 MB (449 MB spared); blocks given memory later 13734, given back 16905; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1229 binds (296.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 297 taken from the spares, 165 made new; 342 kept as spares, 0 destroyed; 45 spares now, 1232 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2945 dirty pages from their submits' captures, 44036 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 791 dispatches, 35397 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5889 runs, 577.0 MB compared; 65 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15882 in captures, 0 only in the collection, 20 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 6 retired still imported (212 MB), 45 retired given back (1084 MB)
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #205000: event 0x100d6c8008 mode 0 value 204995 (lr 0x80660f34)
[hos host] vk: queues ordered by their events: 102000 submissions stopped at a wait, 101937 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 283496, not met 0; captures at a CPU event's wait 9932, 0 of them at a wait met on arrival
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 595 ranges, 3524 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 24 kept (100 MB), 10 spare buffers, 6018 texture pages copied (23 MB)
[hos host] vk: replay reads of guest memory as submitted: 10920000 (6629971 pages from the captures)
[hos Presentation Thread] vi: 16046 frames presented (54.9 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 131, 17.8-22 139, 22-30 5, 30-40 0, over 40 0
[hos host] vk: 12 textures retired (1234 left, 2390 MB): budget 12, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 10940000 (6642903 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3205 dirty pages from their submits' captures, 44912 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 817 dispatches, 36189 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5878 runs, 576.6 MB compared; 65 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16390 in captures, 0 only in the collection, 18 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 10960000 (6656527 pages from the captures)
[hos host] vk: a frame: 297 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.29, present 0.16, draw 7.92, descriptors 3.31, texture lookups 0.12, uploads 0.03, pipelines 0.32, programs 1.47, commands 2.19, dispatch 0.79, memory sync 0.49, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 462 (1138 MB): 5.75 MB synced in 1091.3 copies, 17.49 MB written back
[hos host] vk: 16080 presents, 56.5 fps; a frame: 2535 draws (0 skipped), 136 dispatches, 214 passes, 134 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14153 pipelines (3725 ms), 0 programs (10 failed), 1258 textures (2410 MB)
[hos host] vk: host time in GPU resource waits: 7.94 ms a frame (fences 2.93, timelines 5.01, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 713.6 full barriers recorded, 311.1 left out (nothing recorded since the one before), 38.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.2, an op 1604.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 13.0 ms a frame (the longest submission 4.1 ms); 0.4 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.4 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 10980000 (6669004 pages from the captures)
[hos host] vk: 2 textures retired (1257 left, 2383 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11907 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 11000000 (6682688 pages from the captures)
[hos host] vk: program picks: 15794073 kept (6484367 by the last picks' textures, a uniform buffer bound since), 26397 new, 5979529 with other textures (of 21800000)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #285000: event 0x100d6c802c mode 0 value 206182 (now 206163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: readback pages refreshed from guest memory: 5145875 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 121, 1140 MB with memory of 1592 MB (451 MB spared); blocks given memory later 13741, given back 16905; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1234 binds (297.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 298 taken from the spares, 165 made new; 342 kept as spares, 0 destroyed; 44 spares now, 1228 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3393 dirty pages from their submits' captures, 43672 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 822 dispatches, 35275 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5737 runs, 560.2 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15890 in captures, 0 only in the collection, 17 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 5 retired still imported (184 MB), 46 retired given back (1112 MB)
[hos host] vk: replay reads of guest memory as submitted: 11020000 (6696672 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 29 pages in 29 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: recorder: 7366661 packets run in 10.0 s (busy 1.82 s, spun idle 0.13 s, slept 12599 times, 4172 of them under 200 us), 0 calls made directly (no room); 18766 batches over 540 frames (34.8 a frame, 393 packets a batch), 10026 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 65 times, 8.00 ms (0.1 times, 0.01 ms a frame); every thread 709 times, 70.58 ms in all, the longest 1.00 ms; vkWaitForFences 503/47.78 ms (max 0.79), the queue's lock 99/11.78 ms (max 0.65), vmaFreeMemory 14/4.40 ms (max 1.00), vkQueueBindSparse 39/2.58 ms (max 0.40), vmaDestroyBuffer 29/1.81 ms (max 0.98), vkDestroyImageView 13/1.69 ms (max 0.54)
[hos host] vk: recorder: of its busy 1.82 s, 0.49 s in parts' ends and submits (the longest 1.64 ms), 0.16 s of that blocked on the queue's lock (251 times, the longest 1.60 ms), so 0.33 s in the ends, submits and holds; 1.34 s running the other packets; 13642 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 11040000 (6710275 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 598 ranges, 3532 MB), 270 pages written, 270 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (102 MB), 12 spare buffers, 7648 texture pages copied (29 MB)
[hos host] vk: write-backs left out (600 submits): 775, 3441.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2740 dirty pages from their submits' captures, 42109 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 753 dispatches, 33779 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5632 runs, 547.5 MB compared; 62 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15429 in captures, 0 only in the collection, 17 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 11060000 (6723314 pages from the captures)
[hos host] vk: texture #21160 0x10002199c00 120x67 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (9 so far)
[hos host] vk: texture #21161 0x100021a9e00 60x33 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (10 so far)
[hos NavMeshCAStepDepWorker1] vcores: 146298 takes, 26205 waits (mean 80 us, longest 15.2 ms, 0 over HOS_VCORE_LOG_MS), 12215 kicks, 12413 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 11080000 (6735981 pages from the captures)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1095 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1370 kept their virtual core (nothing there to wait for), 2726 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 11100000 (6749191 pages from the captures)
[hos host] vk: sparse mirrors: 121, 1130 MB with memory of 1584 MB (453 MB spared); blocks given memory later 13741, given back 16905; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1236 binds (297.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 343 kept as spares, 0 destroyed; 44 spares now, 1236 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2886 dirty pages from their submits' captures, 42602 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 784 dispatches, 33940 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5636 runs, 547.6 MB compared; 65 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15944 in captures, 0 only in the collection, 27 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 5 retired still imported (184 MB), 46 retired given back (1112 MB)
[hos host] vk: presenter thread: the copy's fence 2.48 ms, the present 0.85 ms a present
[hos host] vk: replay reads of guest memory as submitted: 11120000 (6762453 pages from the captures)
[hos host] vk: program picks: 15947037 kept (6555303 by the last picks' textures, a uniform buffer bound since), 26401 new, 6026561 with other textures (of 22000000)
[hos host] vk: replay reads of guest memory as submitted: 11140000 (6775001 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2860 dirty pages from their submits' captures, 46195 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 820 dispatches, 37089 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6268 runs, 602.3 MB compared; 43 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16613 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 11160000 (6788503 pages from the captures)
[hos host] vk: memory: working set 15368 MB (peak 16021), private 21609 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 854 MB; used 3368 of 15579 MB; heap 1 (device): 162 blocks 4919 MB, 20364 allocations 4210 MB; used 7042 of 8536 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 18 blocks 576 MB, 132 images 306 MB; 70 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1134 images 1449 MB; 377 free ranges, the largest 32 MB
[hos host] vk: render target images: 5911 reused from the pool, 4242 made new, 4108 destroyed from it, 1160 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1844 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 581 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3916 idle
[hos host] vk: VRAM: 7042 MB used of 8536 MB allowed (708 MB of it free in the blocks); textures 2391 MB (budget 2134 MB), 1272 textures
[hos host] vk: budget: VRAM 4919 of 8536 MB (7042 MB by the driver, 2123 MB unaccounted): textures 2391, mirrors 1130, blocks free 708, shadows 284, other buffers 229, other images 111, staging 64; RAM 3309 MB (the process's private memory 21609 MB): imports 2236, capture spares 754, retired imports 184, captures 100, capture table 35, others 0
[hos host] vk: textures: decoded ASTC 1505 MB, rendered 516 MB (81 drawn in one frame, kept: 47260 KB), other 338 MB; used in the last 8 presents 2006 MB
[hos host] vk: decoded ASTC used again: 427 textures (948 MB) copied from ones the game finalized, not decoded; 427 of those retired, 2 listed; 1941 MB of blocks hashed in 333 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 11053 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 68 (14164 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 596 waited for it, 5.03 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 33.5 ms; VRAM 6.9 of 8.3 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2312 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 307 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.69, present 0.17, draw 7.63, descriptors 3.23, texture lookups 0.09, uploads 0.01, pipelines 0.31, programs 1.36, commands 2.12, dispatch 0.78, memory sync 0.50, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 464 (1130 MB): 5.71 MB synced in 1091.5 copies, 17.66 MB written back
[hos host] vk: 16200 presents, 58.7 fps; a frame: 2433 draws (0 skipped), 139 dispatches, 205 passes, 134 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1272 textures (2391 MB)
[hos host] vk: host time in GPU resource waits: 7.94 ms a frame (fences 3.54, timelines 4.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 718.4 full barriers recorded, 313.0 left out (nothing recorded since the one before), 28.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1479.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.7 ms a frame (the longest submission 3.7 ms); 0.1 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 18.22 ms a frame): its jobs 17.19 ms a frame (submits 1.19, presents 4.23, others 11.76), idle 1.04; the game waited for room at 28 presents, 0.23 ms each (0.01 a frame; 10.52 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 597 captures at a CPU signal; the submits' waits read in 0.042 ms a frame; readbacks: 2995 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 16.38 ms after its hand-over (47.75 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.55 + 0.14 (the window's copy is in the idle), after it 2.72 + 0.62; after a frame's last recording that reads or writes the game's memory come 1.04 recordings (0.00 ms of GPU time), after its last that reads it 1.04 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 11.51 parts a frame recorded in the signal's job, 70.7% of them took nothing that came with the signal; of the others, 3.38 a frame took pages of the signal's capture into a mirror (88.8 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 1 of 597 frames, its first three in 1
[hos host] vk: frame ahead, what came with the signal: 597 signals with a capture, each of 174.1 pages in 13.9 pools; between a frame's last submit and that signal 0.08 texture and sampler registrations, 0.08 releases and 0.00 other CPU signals a signal (24 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.99+0.22, 0.11+0.03, 0.34+0.11, 0.47+0.07, 0.02+0.01, 0.12+0.03, 0.27+0.03, 0.21+0.02, 0.07+0.02, 0.06+0.05, 0.03+0.02, 0.02+0.01
[hos host] vk: render targets drawn over from another start, let go: 10689 so far (50405 MB), 9614 of them (45466 MB) between the passes
[hos host] vk: replay reads of guest memory as submitted: 11180000 (6802494 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 602 ranges, 3504 MB), 264 pages written, 264 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (130 MB), 8 spare buffers, 3311 texture pages copied (12 MB)
[hos host] vk: replay reads of guest memory as submitted: 11200000 (6815978 pages from the captures)
[hos host] vk: 1 textures retired (1276 left, 2382 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 120, 1126 MB with memory of 1580 MB (453 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1238 binds (297.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 344 kept as spares, 0 destroyed; 45 spares now, 1240 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2768 dirty pages from their submits' captures, 47616 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 849 dispatches, 38127 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6488 runs, 623.1 MB compared; 66 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17552 in captures, 0 only in the collection, 26 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 5 retired still imported (184 MB), 46 retired given back (1112 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 288 sync waits (1269 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos BaseProcExecutor3] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 25648112 lookups, 124201 missed (0.48 %)
[hos host] vk: replay reads of guest memory as submitted: 11220000 (6828705 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16792 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 11240000 (6841408 pages from the captures)
[hos host] vk: program picks: 16099651 kept (6624340 by the last picks' textures, a uniform buffer bound since), 26401 new, 6073947 with other textures (of 22200000)
[hos host] vk: readback pages refreshed from guest memory: 5212970 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2756 dirty pages from their submits' captures, 45745 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 794 dispatches, 36722 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6216 runs, 602.0 MB compared; 68 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17085 in captures, 0 only in the collection, 27 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 11260000 (6854080 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 23 pages in 23 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 11280000 (6866943 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #290000: event 0x100d6c802c mode 0 value 209982 (now 209963) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #210000: event 0x100d6c801c mode 0 value 209997 (lr 0x802911b8)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 601 ranges, 3464 MB), 246 pages written, 246 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 24 kept (102 MB), 12 spare buffers, 5478 texture pages copied (21 MB)
[hos host] vk: sparse mirrors: 117, 1119 MB with memory of 1568 MB (448 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1240 binds (297.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 347 kept as spares, 0 destroyed; 48 spares now, 1252 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2715 dirty pages from their submits' captures, 47119 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 37797 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6288 runs, 617.8 MB compared; 70 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17507 in captures, 0 only in the collection, 25 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 5 retired still imported (184 MB), 46 retired given back (1112 MB)
[hos host] vk: replay reads of guest memory as submitted: 11300000 (6879774 pages from the captures)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1092 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1549 kept their virtual core (nothing there to wait for), 3059 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 11320000 (6892659 pages from the captures)
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.08, present 0.17, draw 7.16, descriptors 2.98, texture lookups 0.08, uploads 0.01, pipelines 0.31, programs 1.24, commands 2.06, dispatch 0.79, memory sync 0.43, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 464 (1119 MB): 5.58 MB synced in 1079.5 copies, 17.70 MB written back
[hos host] vk: 16320 presents, 59.5 fps; a frame: 2312 draws (0 skipped), 139 dispatches, 198 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1276 textures (2382 MB)
[hos host] vk: host time in GPU resource waits: 6.93 ms a frame (fences 2.36, timelines 4.56, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 705.4 full barriers recorded, 309.2 left out (nothing recorded since the one before), 23.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.2, an op 1377.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.7 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 11340000 (6905598 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 874, 3883.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2631 dirty pages from their submits' captures, 46906 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 812 dispatches, 37650 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6029 runs, 598.7 MB compared; 69 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17353 in captures, 0 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vi: 16339 frames presented (58.6 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 6, 15.5-17.8 230, 17.8-22 55, 22-30 2, 30-40 0, over 40 0
[hos host] vk: program picks: 16251970 kept (6693314 by the last picks' textures, a uniform buffer bound since), 26401 new, 6121628 with other textures (of 22400000)
[hos host] vk: replay reads of guest memory as submitted: 11360000 (6918426 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 11380000 (6931345 pages from the captures)
[hos ModuleSystemWorker3] nvn: 2150000 recordings made, 72002 live copies (24.6 MB), 6246 texture copies; 4 command buffers' segments (4.4 MB)
[hos host] vk: sparse mirrors: 117, 1119 MB with memory of 1568 MB (448 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1240 binds (297.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 347 kept as spares, 0 destroyed; 48 spares now, 1252 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2381 dirty pages from their submits' captures, 47040 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 814 dispatches, 37488 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6189 runs, 616.3 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17235 in captures, 0 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 4 retired still imported (164 MB), 47 retired given back (1132 MB)
[hos host] vk: replay reads of guest memory as submitted: 11400000 (6944183 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 600 ranges, 3436 MB), 236 pages written, 236 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 24 kept (106 MB), 12 spare buffers, 5479 texture pages copied (21 MB)
[hos host] vk: replay reads of guest memory as submitted: 11420000 (6956828 pages from the captures)
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 500000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 851, 3776.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2815 dirty pages from their submits' captures, 45553 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 789 dispatches, 36614 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6168 runs, 616.3 MB compared; 72 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16886 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 11440000 (6969720 pages from the captures)
[hos host] vk: texture slots registered again at another size: 800000 (of 1006629 registrations; slot 18985 640x360 -> 60x33)
[hos host] vk: 16 textures retired (1260 left, 2365 MB): budget 16, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 11460000 (6982606 pages from the captures)
[hos host] vk: program picks: 16404134 kept (6762156 by the last picks' textures, a uniform buffer bound since), 26401 new, 6169464 with other textures (of 22600000)
[hos host] vk: replay reads of guest memory as submitted: 11480000 (6995385 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 5280360 pages so far (0 held: may be written ahead in their submission)
[hos BaseProcExecutor5] vcores: 152927 takes, 24313 waits (mean 77 us, longest 37.6 ms, 0 over HOS_VCORE_LOG_MS), 11692 kicks, 11916 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: sparse mirrors: 114, 1111 MB with memory of 1552 MB (440 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1242 binds (297.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 350 kept as spares, 0 destroyed; 51 spares now, 1268 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3040 dirty pages from their submits' captures, 46821 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 810 dispatches, 37661 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6192 runs, 616.4 MB compared; 70 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17414 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 4 retired still imported (164 MB), 47 retired given back (1132 MB)
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.24, present 0.17, draw 7.34, descriptors 3.07, texture lookups 0.08, uploads 0.00, pipelines 0.30, programs 1.28, commands 2.10, dispatch 0.81, memory sync 0.40, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 464 (1111 MB): 5.54 MB synced in 1072.2 copies, 17.61 MB written back
[hos host] vk: 16440 presents, 59.3 fps; a frame: 2304 draws (0 skipped), 135 dispatches, 197 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1260 textures (2365 MB)
[hos host] vk: host time in GPU resource waits: 6.61 ms a frame (fences 2.34, timelines 4.27, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 699.3 full barriers recorded, 308.6 left out (nothing recorded since the one before), 23.3 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1371.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.6 ms a frame (the longest submission 3.3 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15374 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 11500000 (7008092 pages from the captures)
[hos sead::MainThread] vk: readbacks of GPU results: 1085 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1519 kept their virtual core (nothing there to wait for), 3089 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 352 mirrors given back (2397 MB in all); 1101 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: reaper: 124 images (157 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 33 (28 MB)
[hos host] vk: late pages found at the replay (600 takes): 20 pages in 20 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 11520000 (7020741 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 602 ranges, 3356 MB), 241 pages written, 241 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 24 kept (112 MB), 12 spare buffers, 5189 texture pages copied (20 MB)
[hos host] vk: write-backs left out (600 submits): 874, 3883.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1908 dirty pages from their submits' captures, 47678 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 890 dispatches, 37679 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6098 runs, 615.0 MB compared; 69 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17203 in captures, 0 only in the collection, 27 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 11540000 (7033469 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 11560000 (7046354 pages from the captures)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #295000: event 0x100d6c802c mode 0 value 213782 (now 213763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1243 binds (297.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 352 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2709 dirty pages from their submits' captures, 46749 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 808 dispatches, 37589 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5825 runs, 595.2 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17206 in captures, 0 only in the collection, 25 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 4 retired still imported (164 MB), 47 retired given back (1132 MB)
[hos host] vk: replay reads of guest memory as submitted: 11580000 (7058958 pages from the captures)
[hos host] vk: program picks: 16556232 kept (6831003 by the last picks' textures, a uniform buffer bound since), 26401 new, 6217366 with other textures (of 22800000)
[hos host] vk: replay reads of guest memory as submitted: 11600000 (7072446 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 294 sync waits (1222 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 11620000 (7084963 pages from the captures)
[hos host] vk: 13 textures retired (1247 left, 2359 MB): budget 13, drawn over 0, dead targets 0
[hos host] vk: late pages' fix (600 recordings with one): of 759 captures with late pages 759 fixed (75.1 pages, 75.1 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 503 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 5 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 15, of an earlier recording 483, not recorded yet 0, of this part 0; pages fixed from stamps 56039, from snapshots 927; the shadows: 288843 pages synced from captures, 287328 made known by it, 1515 left unknown because the replay or the collection wrote them since
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2749 dirty pages from their submits' captures, 46959 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 828 dispatches, 37620 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6101 runs, 629.6 MB compared; 71 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17177 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 11640000 (7097520 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 601 ranges, 3350 MB), 238 pages written, 238 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 24 kept (88 MB), 12 spare buffers, 5189 texture pages copied (20 MB)
[hos host] vk: 28 textures retired (1219 left, 2322 MB): budget 28, drawn over 0, dead targets 0
[hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.81, present 0.14, draw 7.00, descriptors 2.88, texture lookups 0.08, uploads 0.00, pipelines 0.30, programs 1.25, commands 2.01, dispatch 0.77, memory sync 0.38, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 464 (1100 MB): 5.52 MB synced in 1073.0 copies, 17.52 MB written back
[hos host] vk: 16560 presents, 60.1 fps; a frame: 2302 draws (0 skipped), 129 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1219 textures (2322 MB)
[hos host] vk: host time in GPU resource waits: 5.37 ms a frame (fences 2.23, timelines 3.14, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 692.7 full barriers recorded, 308.5 left out (nothing recorded since the one before), 18.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.4 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #215000: event 0x100d6c8030 mode 0 value 214999 (lr 0x806612a8)
[hos host] vk: replay reads of guest memory as submitted: 11660000 (7110282 pages from the captures)
[hos host] vk: 4 textures retired (1215 left, 2321 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 685 dirty pages from their submits' captures, 50801 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 856 dispatches, 39629 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6296 runs, 647.3 MB compared; 89 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17744 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 4 retired still imported (164 MB), 47 retired given back (1132 MB)
[hos host] vk: replay reads of guest memory as submitted: 11680000 (7122987 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11860 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 11700000 (7135709 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: program picks: 16708808 kept (6900163 by the last picks' textures, a uniform buffer bound since), 26401 new, 6264790 with other textures (of 23000000)
[hos host] vk: late pages found at the replay (600 takes): 11 pages in 11 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 11720000 (7148533 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1678 kept their virtual core (nothing there to wait for), 3186 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: readback pages refreshed from guest memory: 5347214 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 783 dirty pages from their submits' captures, 49416 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 38605 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6128 runs, 629.9 MB compared; 90 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17216 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 11740000 (7161267 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 11760000 (7174013 pages from the captures)
[hos Presentation Thread] vi: 16640 frames presented (60.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 15, 15.5-17.8 262, 17.8-22 22, 22-30 2, 30-40 0, over 40 0
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 600 ranges, 3346 MB), 256 pages written, 256 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 24 kept (84 MB), 10 spare buffers, 6288 texture pages copied (24 MB)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1512 dirty pages from their submits' captures, 49642 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 846 dispatches, 39066 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6129 runs, 629.7 MB compared; 88 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17899 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: replay reads of guest memory as submitted: 11780000 (7186732 pages from the captures)
[hos ModuleSystemWorker2] nvn: 2200000 recordings made, 71930 live copies (24.5 MB), 6246 texture copies; 6 command buffers' segments (4.4 MB)
[hos Presentation Thread] nvn: QueueSubmitCommands #100000: queue 000010100D6AA020 (its #16664), 1 handles (lr 0x805282f4)
[hos host] vk: replay reads of guest memory as submitted: 11800000 (7199432 pages from the captures)
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.62, present 0.14, draw 6.99, descriptors 2.88, texture lookups 0.07, uploads 0.00, pipelines 0.29, programs 1.25, commands 2.00, dispatch 0.76, memory sync 0.42, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 464 (1100 MB): 5.45 MB synced in 1069.6 copies, 17.50 MB written back
[hos host] vk: 16680 presents, 59.9 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1215 textures (2321 MB)
[hos host] vk: host time in GPU resource waits: 3.30 ms a frame (fences 2.46, timelines 0.84, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 689.5 full barriers recorded, 308.2 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1366.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 11820000 (7211918 pages from the captures)
[hos host] vk: program picks: 16861226 kept (6969208 by the last picks' textures, a uniform buffer bound since), 26401 new, 6312372 with other textures (of 23200000)
[hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1354 dirty pages from their submits' captures, 49392 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 838 dispatches, 39097 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6274 runs, 647.2 MB compared; 77 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17454 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #300000: event 0x100d6c802c mode 0 value 217582 (now 217563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 11840000 (7224452 pages from the captures)
[hos host] vk: recorder: 7345139 packets run in 10.0 s (busy 1.83 s, spun idle 0.14 s, slept 13218 times, 4912 of them under 200 us), 0 calls made directly (no room); 19838 batches over 598 frames (33.2 a frame, 370 packets a batch), 10308 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 26 times, 2.26 ms (0.0 times, 0.00 ms a frame); every thread 648 times, 61.96 ms in all, the longest 0.52 ms; vkWaitForFences 568/56.19 ms (max 0.52), the queue's lock 38/2.64 ms (max 0.22), vmaFreeMemory 9/1.69 ms (max 0.47), vkDestroyImageView 7/0.79 ms (max 0.21), vmaDestroyBuffer 6/0.30 ms (max 0.25), vkQueueBindSparse 10/0.16 ms (max 0.06)
[hos host] vk: recorder: of its busy 1.83 s, 0.45 s in parts' ends and submits (the longest 1.45 ms), 0.13 s of that blocked on the queue's lock (199 times, the longest 1.38 ms), so 0.32 s in the ends, submits and holds; 1.38 s running the other packets; 12283 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 11860000 (7236942 pages from the captures)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1040 dirty pages from their submits' captures, 49247 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 834 dispatches, 38614 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6133 runs, 630.0 MB compared; 89 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17444 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11140 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 11880000 (7249701 pages from the captures)
[hos ModuleSystemWorker2] vcores: 149193 takes, 21060 waits (mean 67 us, longest 15.7 ms, 0 over HOS_VCORE_LOG_MS), 10290 kicks, 10381 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 598 ranges, 3286 MB), 246 pages written, 246 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 6991 texture pages copied (27 MB)
[hos host] vk: replay reads of guest memory as submitted: 11900000 (7262443 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 11920000 (7275116 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1028 dirty pages from their submits' captures, 49156 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 833 dispatches, 38543 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6124 runs, 629.8 MB compared; 99 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17262 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: presenter thread: the copy's fence 2.44 ms, the present 0.84 ms a present
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1580 kept their virtual core (nothing there to wait for), 3028 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: program picks: 17013594 kept (7038169 by the last picks' textures, a uniform buffer bound since), 26401 new, 6360004 with other textures (of 23400000)
[hos host] vk: replay reads of guest memory as submitted: 11940000 (7287868 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 11960000 (7300572 pages from the captures)
[hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 520000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040)
[hos host] vk: readback pages refreshed from guest memory: 5414520 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1524 dirty pages from their submits' captures, 48773 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 832 dispatches, 38617 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6122 runs, 629.7 MB compared; 76 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17532 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: memory: working set 15317 MB (peak 16021), private 21494 MB, staging 48 MB in 3 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 848 MB; used 3304 of 15579 MB; heap 1 (device): 161 blocks 4855 MB, 19836 allocations 4095 MB; used 6978 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 18 blocks 576 MB, 133 images 306 MB; 70 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1076 images 1379 MB; 395 free ranges, the largest 32 MB
[hos host] vk: render target images: 5911 reused from the pool, 4243 made new, 4108 destroyed from it, 1160 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1845 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 581 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3916 idle
[hos host] vk: VRAM: 6978 MB used of 11474 MB allowed (759 MB of it free in the blocks); textures 2321 MB (budget 7158 MB), 1215 textures
[hos host] vk: budget: VRAM 4855 of 11474 MB (6978 MB by the driver, 2123 MB unaccounted): textures 2321, mirrors 1100, blocks free 759, shadows 284, other buffers 229, other images 111, staging 48; RAM 3239 MB (the process's private memory 21494 MB): imports 2236, capture spares 742, retired imports 120, captures 102, capture table 35, other buffers 4, others 0
[hos host] vk: textures: decoded ASTC 1476 MB, rendered 516 MB (84 drawn in one frame, kept: 47616 KB), other 298 MB; used in the last 8 presents 1970 MB
[hos host] vk: decoded ASTC used again: 429 textures (949 MB) copied from ones the game finalized, not decoded; 429 of those retired, 1 listed; 1941 MB of blocks hashed in 333 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 11653 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 68 (14164 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 479 waited for it, 3.27 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: replay reads of guest memory as submitted: 11980000 (7313211 pages from the captures)
[hos host] vk: status: 600 frames, the slowest 25.4 ms; VRAM 6.8 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2312 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.70, present 0.13, draw 7.01, descriptors 2.87, texture lookups 0.07, uploads 0.00, pipelines 0.30, programs 1.25, commands 2.01, dispatch 0.76, memory sync 0.42, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 464 (1100 MB): 5.45 MB synced in 1069.9 copies, 17.50 MB written back
[hos host] vk: 16800 presents, 60.0 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1215 textures (2321 MB)
[hos host] vk: host time in GPU resource waits: 3.27 ms a frame (fences 2.33, timelines 0.93, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 690.3 full barriers recorded, 308.7 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1366.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.4 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 16.74 ms a frame): its jobs 15.77 ms a frame (submits 2.85, presents 4.65, others 8.26), idle 0.97; the game waited for room at 73 presents, 0.09 ms each (0.01 a frame; 19.81 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 482 captures at a CPU signal; the submits' waits read in 0.042 ms a frame; readbacks: 2425 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.05 ms after its hand-over (20.03 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.45 + 0.13 (the window's copy is in the idle), after it 2.14 + 0.47; after a frame's last recording that reads or writes the game's memory come 1.23 recordings (0.00 ms of GPU time), after its last that reads it 1.23 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 9.91 parts a frame recorded in the signal's job, 70.5% of them took nothing that came with the signal; of the others, 2.92 a frame took pages of the signal's capture into a mirror (70.6 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 11 of 482 frames, its first three in 10
[hos host] vk: frame ahead, what came with the signal: 482 signals with a capture, each of 95.0 pages in 13.6 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.98+0.17, 0.09+0.03, 0.28+0.11, 0.46+0.07, 0.02+0.01, 0.07+0.01, 0.09+0.01, 0.06+0.02, 0.05+0.02, 0.03+0.01, 0.01+0.00, 0.00+0.00
[hos host] vk: queues ordered by their events: 108000 submissions stopped at a wait, 107936 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 302622, not met 0; captures at a CPU event's wait 10576, 0 of them at a wait met on arrival
[hos host] vk: render targets drawn over from another start, let go: 10689 so far (50405 MB), 9614 of them (45466 MB) between the passes
[hos host] vk: replay reads of guest memory as submitted: 12000000 (7326124 pages from the captures)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #220000: event 0x100d6c803c mode 0 value 219999 (lr 0x80661bfc)
[hos host] vk: (its written part: 1 of 1 KB readable, 759 bytes not zero; levels L0)
[hos host] vk: texture 0x112a256a00 2448x3056x1 x1 levels 1 NVN 0x4b (7344 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 1 KB of it
[hos host] vk: (its written part: 1 of 1 KB readable, 759 bytes not zero; levels L0)
[hos host] vk: texture 0x112a256a00 2448x3056x1 x1 levels 1 NVN 0x4b (7344 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 1 KB of it
[hos host] vk: replay reads of guest memory as submitted: 12020000 (7338934 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 594 ranges, 3216 MB), 247 pages written, 247 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (86 MB), 10 spare buffers, 7001 texture pages copied (27 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 150 sync waits (202 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1064 dirty pages from their submits' captures, 49440 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 840 dispatches, 38669 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6128 runs, 629.8 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17390 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos ModuleSystemWorker5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 14555558 lookups, 119244 missed (0.82 %)
[hos host] vk: replay reads of guest memory as submitted: 12040000 (7351517 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5986 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 17166298 kept (7107258 by the last picks' textures, a uniform buffer bound since), 26402 new, 6407299 with other textures (of 23600000)
[hos host] vk: replay reads of guest memory as submitted: 12060000 (7364301 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 635 dirty pages from their submits' captures, 50841 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 849 dispatches, 39658 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6325 runs, 647.9 MB compared; 83 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17791 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 192 live (2236 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: replay reads of guest memory as submitted: 12080000 (7377173 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 12100000 (7389960 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #305000: event 0x100d6c802c mode 0 value 221382 (now 221363) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 12120000 (7402824 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1690 dirty pages from their submits' captures, 47330 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 812 dispatches, 37530 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6041 runs, 613.8 MB compared; 75 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17548 in captures, 0 only in the collection, 27 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 12140000 (7415590 pages from the captures)
[hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.63, present 0.15, draw 6.87, descriptors 2.79, texture lookups 0.07, uploads 0.01, pipelines 0.29, programs 1.23, commands 1.99, dispatch 0.77, memory sync 0.41, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 465 (1101 MB): 5.48 MB synced in 1071.6 copies, 17.52 MB written back
[hos host] vk: 16920 presents, 60.0 fps; a frame: 2305 draws (0 skipped), 129 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1218 textures (2322 MB)
[hos host] vk: host time in GPU resource waits: 3.91 ms a frame (fences 2.28, timelines 1.63, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 690.5 full barriers recorded, 308.1 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.4 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1083 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1349 kept their virtual core (nothing there to wait for), 2747 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 581 ranges, 3224 MB), 243 pages written, 243 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 7001 texture pages copied (27 MB)
[hos host] vk: replay reads of guest memory as submitted: 12160000 (7428406 pages from the captures)
[hos Presentation Thread] vi: 16940 frames presented (59.9 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 9, 15.5-17.8 280, 17.8-22 11, 22-30 0, 30-40 0, over 40 0
[hos host] vk: program picks: 17318891 kept (7176301 by the last picks' textures, a uniform buffer bound since), 26406 new, 6454702 with other textures (of 23800000)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2635 dirty pages from their submits' captures, 48375 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 831 dispatches, 38730 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6191 runs, 631.3 MB compared; 71 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17341 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: replay reads of guest memory as submitted: 12180000 (7441276 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 12200000 (7453992 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 5481596 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 12220000 (7466827 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2550 dirty pages from their submits' captures, 46968 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 812 dispatches, 37584 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6038 runs, 613.8 MB compared; 72 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17232 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 12240000 (7479667 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13386 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 12260000 (7492344 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 576 ranges, 3186 MB), 246 pages written, 246 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (150 MB), 11 spare buffers, 6019 texture pages copied (23 MB)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1643 dirty pages from their submits' captures, 50427 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 854 dispatches, 39785 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6351 runs, 648.5 MB compared; 86 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18173 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: replay reads of guest memory as submitted: 12280000 (7504967 pages from the captures)
[hos host] vk: program picks: 17471479 kept (7245400 by the last picks' textures, a uniform buffer bound since), 26406 new, 6502114 with other textures (of 24000000)
[hos host] vk: late pages found at the replay (600 takes): 14 pages in 14 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 12300000 (7517712 pages from the captures)
[hos ModuleSystemWorker1] vcores: 148897 takes, 20983 waits (mean 64 us, longest 7.6 ms, 0 over HOS_VCORE_LOG_MS), 10181 kicks, 10283 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.98, present 0.14, draw 7.14, descriptors 2.94, texture lookups 0.07, uploads 0.00, pipelines 0.30, programs 1.26, commands 2.05, dispatch 0.80, memory sync 0.41, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 465 (1101 MB): 5.52 MB synced in 1072.6 copies, 17.54 MB written back
[hos host] vk: 17040 presents, 60.0 fps; a frame: 2308 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1220 textures (2322 MB)
[hos host] vk: host time in GPU resource waits: 4.59 ms a frame (fences 2.38, timelines 2.21, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 692.7 full barriers recorded, 308.2 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2479 dirty pages from their submits' captures, 46903 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 37557 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6024 runs, 613.6 MB compared; 80 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17136 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 12320000 (7530495 pages from the captures)
[hos ModuleSystemWorker4] nvn: 2250000 recordings made, 71929 live copies (24.6 MB), 6246 texture copies; 5 command buffers' segments (4.3 MB)
[hos host] vk: replay reads of guest memory as submitted: 12340000 (7543294 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1587 kept their virtual core (nothing there to wait for), 3021 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 12360000 (7556055 pages from the captures)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2235 dirty pages from their submits' captures, 48668 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 836 dispatches, 38703 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6195 runs, 631.2 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17697 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #225000: event 0x100d6c8048 mode 0 value 224999 (lr 0x80666598)
[hos host] vk: replay reads of guest memory as submitted: 12380000 (7568901 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 563 ranges, 3064 MB), 239 pages written, 239 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 24 kept (104 MB), 12 spare buffers, 5592 texture pages copied (21 MB)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #310000: event 0x100d6c802c mode 0 value 225182 (now 225163) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 12400000 (7581808 pages from the captures)
[hos host] vk: program picks: 17624108 kept (7314397 by the last picks' textures, a uniform buffer bound since), 26406 new, 6549485 with other textures (of 24200000)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1673 dirty pages from their submits' captures, 47815 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 814 dispatches, 37668 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6038 runs, 613.9 MB compared; 77 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17147 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 12420000 (7595023 pages from the captures)
[hos host] vk: reaper: 47 images (44 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 27 (37 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 222 sync waits (601 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 12440000 (7607695 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 12460000 (7620312 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 5549094 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2337 dirty pages from their submits' captures, 48525 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 38689 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6197 runs, 631.1 MB compared; 73 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17701 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.90, present 0.14, draw 7.03, descriptors 2.85, texture lookups 0.09, uploads 0.01, pipelines 0.31, programs 1.26, commands 2.02, dispatch 0.79, memory sync 0.43, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 465 (1101 MB): 5.52 MB synced in 1073.1 copies, 17.54 MB written back
[hos host] vk: 17160 presents, 60.1 fps; a frame: 2310 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1226 textures (2323 MB)
[hos host] vk: host time in GPU resource waits: 4.60 ms a frame (fences 2.41, timelines 2.19, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 693.3 full barriers recorded, 308.5 left out (nothing recorded since the one before), 18.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1366.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos ModuleSystemWorker3] barrier: 267027 looks: 103187 found the masks moved already, 161078 waited until they moved, 2762 for the whole 1 ms (a wait 337 us on average); 250213 wakes
[hos host] vk: replay reads of guest memory as submitted: 12480000 (7633063 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11957 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 12500000 (7645673 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 6 pages in 6 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 562 ranges, 3062 MB), 256 pages written, 256 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (166 MB), 12 spare buffers, 5992 texture pages copied (23 MB)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1538 dirty pages from their submits' captures, 49520 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 845 dispatches, 38942 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6197 runs, 631.2 MB compared; 83 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17973 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 540000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040)
[hos host] vk: program picks: 17776761 kept (7383482 by the last picks' textures, a uniform buffer bound since), 26406 new, 6596832 with other textures (of 24400000)
[hos host] vk: replay reads of guest memory as submitted: 12520000 (7658128 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 12540000 (7670809 pages from the captures)
[hos rmbl::VibrationThread] hid: 57344 vibration values, strongest 0.00 since the last count
[hos host] vk: replay reads of guest memory as submitted: 12560000 (7683576 pages from the captures)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1353 dirty pages from their submits' captures, 48716 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 820 dispatches, 38361 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6206 runs, 631.4 MB compared; 77 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17032 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1082 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1560 kept their virtual core (nothing there to wait for), 3048 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 12580000 (7696276 pages from the captures)
[hos Presentation Thread] vi: 17241 frames presented (60.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 4, 15.5-17.8 281, 17.8-22 16, 22-30 0, 30-40 0, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 12600000 (7709089 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1813 dirty pages from their submits' captures, 48907 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 857 dispatches, 38708 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6191 runs, 631.2 MB compared; 85 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17600 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 12620000 (7721778 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 553 ranges, 3018 MB), 246 pages written, 246 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 6039 texture pages copied (23 MB)
[hos host] vk: program picks: 17929255 kept (7452566 by the last picks' textures, a uniform buffer bound since), 26406 new, 6644338 with other textures (of 24600000)
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.84, present 0.13, draw 6.92, descriptors 2.80, texture lookups 0.07, uploads 0.01, pipelines 0.30, programs 1.24, commands 2.01, dispatch 0.78, memory sync 0.44, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 465 (1101 MB): 5.50 MB synced in 1072.5 copies, 17.54 MB written back
[hos host] vk: 17280 presents, 59.8 fps; a frame: 2308 draws (0 skipped), 130 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1228 textures (2326 MB)
[hos host] vk: host time in GPU resource waits: 3.55 ms a frame (fences 2.43, timelines 1.12, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 692.8 full barriers recorded, 308.7 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1366.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 12640000 (7734472 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 12660000 (7747210 pages from the captures)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #315000: event 0x100d6c802c mode 0 value 228982 (now 228963) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2561 dirty pages from their submits' captures, 48378 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 833 dispatches, 38717 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6190 runs, 631.1 MB compared; 73 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17755 in captures, 0 only in the collection, 26 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: recorder: 7311847 packets run in 10.0 s (busy 1.84 s, spun idle 0.14 s, slept 13083 times, 4757 of them under 200 us), 0 calls made directly (no room); 19469 batches over 599 frames (32.5 a frame, 376 packets a batch), 10019 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 2 times, 0.29 ms (0.0 times, 0.00 ms a frame); every thread 612 times, 57.44 ms in all, the longest 0.59 ms; vkWaitForFences 567/55.06 ms (max 0.59), the queue's lock 35/1.55 ms (max 0.18), a part's submit 2/0.33 ms (max 0.18), vmaDestroyBuffer 3/0.21 ms (max 0.18), vkDestroyImageView 2/0.15 ms (max 0.09), vkQueueBindSparse 1/0.11 ms (max 0.11)
[hos host] vk: recorder: of its busy 1.84 s, 0.49 s in parts' ends and submits (the longest 1.68 ms), 0.17 s of that blocked on the queue's lock (268 times, the longest 1.63 ms), so 0.32 s in the ends, submits and holds; 1.35 s running the other packets; 12207 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 12680000 (7759920 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 29341 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 12700000 (7772585 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 5616185 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2704 dirty pages from their submits' captures, 48222 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 832 dispatches, 38690 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6185 runs, 631.0 MB compared; 76 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17726 in captures, 0 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: late pages found at the replay (600 takes): 8 pages in 8 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos nn.npln.Worker] vcores: 148822 takes, 21324 waits (mean 64 us, longest 7.0 ms, 0 over HOS_VCORE_LOG_MS), 10481 kicks, 10603 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 12720000 (7785098 pages from the captures)
[hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #230000: event 0x100d6c800c mode 0 value 230000 (lr 0x802915c8)
[hos host] vk: late pages' fix (600 recordings with one): of 859 captures with late pages 859 fixed (75.4 pages, 75.4 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 533 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 70, of an earlier recording 463, not recorded yet 0, of this part 0; pages fixed from stamps 63620, from snapshots 1147; the shadows: 399955 pages synced from captures, 393578 made known by it, 6377 left unknown because the replay or the collection wrote them since
[hos host] vk: replay reads of guest memory as submitted: 12740000 (7797512 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 539 ranges, 2920 MB), 253 pages written, 253 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.036; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 24 kept (90 MB), 12 spare buffers, 8439 texture pages copied (32 MB)
[hos host] vk: presenter thread: the copy's fence 2.28 ms, the present 0.82 ms a present
[hos host] vk: program picks: 18081408 kept (7521617 by the last picks' textures, a uniform buffer bound since), 26406 new, 6692185 with other textures (of 24800000)
[hos host] vk: replay reads of guest memory as submitted: 12760000 (7810152 pages from the captures)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2431 dirty pages from their submits' captures, 46989 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 810 dispatches, 37613 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6027 runs, 613.8 MB compared; 76 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17220 in captures, 0 only in the collection, 25 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos host] vk: replay reads of guest memory as submitted: 12780000 (7822894 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1679 kept their virtual core (nothing there to wait for), 3185 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 12800000 (7835625 pages from the captures)
[hos host] vk: memory: working set 15323 MB (peak 16021), private 21499 MB, staging 32 MB in 2 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 848 MB; used 3308 of 15579 MB; heap 1 (device): 161 blocks 4855 MB, 19868 allocations 4116 MB; used 6978 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 18 blocks 576 MB, 133 images 306 MB; 70 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1097 images 1415 MB; 396 free ranges, the largest 32 MB
[hos host] vk: render target images: 5911 reused from the pool, 4243 made new, 4108 destroyed from it, 1160 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1845 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 581 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3916 idle
[hos host] vk: VRAM: 6978 MB used of 11474 MB allowed (738 MB of it free in the blocks); textures 2358 MB (budget 7158 MB), 1236 textures
[hos host] vk: budget: VRAM 4855 of 11474 MB (6978 MB by the driver, 2123 MB unaccounted): textures 2358, mirrors 1101, blocks free 738, shadows 284, other buffers 229, other images 111, staging 32; RAM 3243 MB (the process's private memory 21499 MB): imports 2240, capture spares 736, retired imports 120, captures 112, capture table 35, others 0
[hos host] vk: textures: decoded ASTC 1476 MB, rendered 516 MB (84 drawn in one frame, kept: 47616 KB), other 334 MB; used in the last 8 presents 1968 MB
[hos host] vk: decoded ASTC used again: 437 textures (955 MB) copied from ones the game finalized, not decoded; 437 of those retired, 2 listed; 1943 MB of blocks hashed in 333 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 12253 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 68 (14164 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 474 waited for it, 2.72 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 24.8 ms; VRAM 6.8 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2315 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.07, present 0.16, draw 7.14, descriptors 2.96, texture lookups 0.15, uploads 0.07, pipelines 0.31, programs 1.24, commands 2.05, dispatch 0.79, memory sync 0.48, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 465 (1101 MB): 5.53 MB synced in 1073.0 copies, 17.54 MB written back
[hos host] vk: 17400 presents, 60.0 fps; a frame: 2308 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.1 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1236 textures (2358 MB)
[hos host] vk: host time in GPU resource waits: 6.76 ms a frame (fences 2.30, timelines 4.45, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 693.2 full barriers recorded, 308.4 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.26 MB unswizzled on the CPU in 0.03 ms
[hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 15.72 ms a frame (submits 2.94, presents 4.60, others 8.18), idle 0.96; the game waited for room at 47 presents, 0.09 ms each (0.01 a frame; 14.49 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 477 captures at a CPU signal; the submits' waits read in 0.043 ms a frame; readbacks: 2352 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.95 ms after its hand-over (21.30 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.41 + 0.13 (the window's copy is in the idle), after it 2.16 + 0.47; after a frame's last recording that reads or writes the game's memory come 1.16 recordings (0.00 ms of GPU time), after its last that reads it 1.16 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 9.77 parts a frame recorded in the signal's job, 70.7% of them took nothing that came with the signal; of the others, 2.86 a frame took pages of the signal's capture into a mirror (69.9 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 19 of 477 frames, its first three in 19
[hos host] vk: frame ahead, what came with the signal: 477 signals with a capture, each of 88.0 pages in 13.4 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.98+0.17, 0.09+0.03, 0.30+0.11, 0.46+0.07, 0.02+0.01, 0.08+0.01, 0.10+0.01, 0.05+0.02, 0.05+0.02, 0.02+0.01, 0.01+0.00, 0.00+0.00
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2896 dirty pages from their submits' captures, 46860 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 809 dispatches, 37644 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6017 runs, 613.6 MB compared; 70 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17831 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: render targets drawn over from another start, let go: 10689 so far (50405 MB), 9614 of them (45466 MB) between the passes
[hos host] vk: replay reads of guest memory as submitted: 12820000 (7848301 pages from the captures)
[hos sead::MainThread] nvn: 2300000 recordings made, 71990 live copies (25.8 MB), 6246 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: replay reads of guest memory as submitted: 12840000 (7861840 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 240 sync waits (738 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 14777399 lookups, 120813 missed (0.82 %)
[hos host] vk: replay reads of guest memory as submitted: 12860000 (7874683 pages from the captures)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1927 dirty pages from their submits' captures, 49017 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 843 dispatches, 38762 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6190 runs, 631.3 MB compared; 77 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17074 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 539 ranges, 2920 MB), 250 pages written, 250 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (90 MB), 11 spare buffers, 3450 texture pages copied (13 MB)
[hos host] vk: program picks: 18233822 kept (7590642 by the last picks' textures, a uniform buffer bound since), 26406 new, 6739771 with other textures (of 25000000)
[hos host] vk: replay reads of guest memory as submitted: 12880000 (7887298 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 12900000 (7899919 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1641 dirty pages from their submits' captures, 48775 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 824 dispatches, 38619 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6180 runs, 631.1 MB compared; 79 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17625 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10861 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 12920000 (7912541 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 11 pages in 11 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #320000: event 0x100d6c802c mode 0 value 232782 (now 232763) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos sead::MainThread] nvn: EventSignal (CPU) #35000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: replay reads of guest memory as submitted: 12940000 (7925173 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 5683644 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1435 dirty pages from their submits' captures, 49058 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 836 dispatches, 38625 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6204 runs, 631.3 MB compared; 84 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17475 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 1 retired still imported (60 MB), 50 retired given back (1236 MB)
[hos host] vk: replay reads of guest memory as submitted: 12960000 (7937990 pages from the captures)
[hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.74, present 0.14, draw 6.95, descriptors 2.84, texture lookups 0.07, uploads 0.00, pipelines 0.29, programs 1.23, commands 2.02, dispatch 0.77, memory sync 0.41, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 465 (1101 MB): 5.50 MB synced in 1072.6 copies, 17.54 MB written back
[hos host] vk: 17520 presents, 60.2 fps; a frame: 2309 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1236 textures (2358 MB)
[hos host] vk: host time in GPU resource waits: 4.09 ms a frame (fences 2.35, timelines 1.74, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 692.3 full barriers recorded, 308.2 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 12980000 (7950885 pages from the captures)
[hos host] vk: program picks: 18386616 kept (7659725 by the last picks' textures, a uniform buffer bound since), 26406 new, 6786977 with other textures (of 25200000)
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 534 ranges, 2878 MB), 239 pages written, 239 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (108 MB), 11 spare buffers, 3450 texture pages copied (13 MB)
[hos Presentation Thread] vi: 17541 frames presented (59.9 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 9, 15.5-17.8 270, 17.8-22 19, 22-30 2, 30-40 0, over 40 0
[hos sead::MainThread] vk: readbacks of GPU results: 1085 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1380 kept their virtual core (nothing there to wait for), 2716 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 13000000 (7963705 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2647 dirty pages from their submits' captures, 47739 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 827 dispatches, 38128 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6037 runs, 613.8 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17746 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 13020000 (7976642 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 13040000 (7989432 pages from the captures)
[hos host] vk: queues ordered by their events: 114000 submissions stopped at a wait, 113938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 321855, not met 0; captures at a CPU event's wait 11192, 0 of them at a wait met on arrival
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2705 dirty pages from their submits' captures, 46145 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 797 dispatches, 37093 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6026 runs, 613.8 MB compared; 72 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16773 in captures, 0 only in the collection, 26 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 13060000 (8002398 pages from the captures)
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 560000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 13080000 (8015201 pages from the captures)
[hos ModuleSystemWorker2] nvn: CommandBufferSignalEvent #235000: event 0x100d6c8004 mode 0 value 234995 (lr 0x80291504)
[hos host] vk: replay reads of guest memory as submitted: 13100000 (8028049 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2899 dirty pages from their submits' captures, 46766 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 810 dispatches, 37600 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6019 runs, 613.7 MB compared; 73 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17242 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: program picks: 18539114 kept (7728820 by the last picks' textures, a uniform buffer bound since), 26406 new, 6834479 with other textures (of 25400000)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 530 ranges, 2866 MB), 244 pages written, 244 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 24 kept (106 MB), 12 spare buffers, 3824 texture pages copied (14 MB)
[hos host] vk: replay reads of guest memory as submitted: 13120000 (8040713 pages from the captures)
[hos BaseProcExecutor1] vcores: 149385 takes, 21190 waits (mean 65 us, longest 6.8 ms, 0 over HOS_VCORE_LOG_MS), 10377 kicks, 10549 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.11, present 0.14, draw 7.20, descriptors 2.97, texture lookups 0.08, uploads 0.00, pipelines 0.31, programs 1.26, commands 2.07, dispatch 0.81, memory sync 0.37, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 465 (1101 MB): 5.54 MB synced in 1072.3 copies, 17.54 MB written back
[hos host] vk: 17640 presents, 60.0 fps; a frame: 2306 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1236 textures (2358 MB)
[hos host] vk: host time in GPU resource waits: 6.57 ms a frame (fences 2.55, timelines 4.02, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 694.3 full barriers recorded, 308.7 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1366.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.6 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 13140000 (8053310 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14061 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2026 dirty pages from their submits' captures, 50063 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 855 dispatches, 39741 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6352 runs, 648.6 MB compared; 82 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18105 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: late pages found at the replay (600 takes): 10 pages in 10 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 13160000 (8065933 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 13180000 (8078758 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 5750709 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1659 dirty pages from their submits' captures, 48171 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 826 dispatches, 38078 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6033 runs, 613.7 MB compared; 86 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17512 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 13200000 (8091608 pages from the captures)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1079 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1519 kept their virtual core (nothing there to wait for), 2833 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #325000: event 0x100d6c802c mode 0 value 236582 (now 236563) (lr 0x8028c554)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.04
[hos host] vk: program picks: 18691331 kept (7797693 by the last picks' textures, a uniform buffer bound since), 26406 new, 6882262 with other textures (of 25600000)
[hos host] vk: replay reads of guest memory as submitted: 13220000 (8104499 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.03 over 524 ranges, 2822 MB), 256 pages written, 256 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.030; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (108 MB), 10 spare buffers, 4527 texture pages copied (17 MB)
[hos host] vk: replay reads of guest memory as submitted: 13240000 (8117929 pages from the captures)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2800 dirty pages from their submits' captures, 47604 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 816 dispatches, 38254 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6173 runs, 631.1 MB compared; 70 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17295 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 242 sync waits (822 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 13260000 (8130695 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 13280000 (8143166 pages from the captures)
[hos host] vk: a submission's parts ended: 82794 where the game had ended its render pass, 17206 at 512 draws
[hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.44, present 0.14, draw 7.34, descriptors 2.99, texture lookups 0.09, uploads 0.01, pipelines 0.32, programs 1.31, commands 2.12, dispatch 0.83, memory sync 0.48, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 465 (1101 MB): 5.50 MB synced in 1072.1 copies, 17.53 MB written back
[hos host] vk: 17760 presents, 58.2 fps; a frame: 2303 draws (0 skipped), 130 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1238 textures (2365 MB)
[hos host] vk: host time in GPU resource waits: 5.24 ms a frame (fences 2.35, timelines 2.89, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 692.9 full barriers recorded, 308.9 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1366.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.7 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.06 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2568 dirty pages from their submits' captures, 46814 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 813 dispatches, 37512 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5994 runs, 613.1 MB compared; 74 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17166 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 13300000 (8155661 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 13320000 (8168140 pages from the captures)
[hos host] vk: program picks: 18843726 kept (7866634 by the last picks' textures, a uniform buffer bound since), 26406 new, 6929867 with other textures (of 25800000)
[hos host] vk: replay reads of guest memory as submitted: 13340000 (8180644 pages from the captures)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2514 dirty pages from their submits' captures, 48182 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 830 dispatches, 38627 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6122 runs, 629.7 MB compared; 73 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17703 in captures, 0 only in the collection, 24 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 523 ranges, 2804 MB), 240 pages written, 240 copied (0.9 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 24 kept (100 MB), 12 spare buffers, 3920 texture pages copied (15 MB)
[hos host] vk: replay reads of guest memory as submitted: 13360000 (8193391 pages from the captures)
[hos sead::MainThread] nvn: 2350000 recordings made, 71998 live copies (25.9 MB), 6246 texture copies; 2 command buffers' segments (5.0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 23138 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 13380000 (8205889 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2993 dirty pages from their submits' captures, 46587 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 37459 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5963 runs, 612.4 MB compared; 71 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17165 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 13400000 (8218406 pages from the captures)
[hos Presentation Thread] vi: 17839 frames presented (59.6 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 17, 15.5-17.8 240, 17.8-22 37, 22-30 4, 30-40 0, over 40 0
[hos host] vk: late pages found at the replay (600 takes): 19 pages in 19 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1085 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1586 kept their virtual core (nothing there to wait for), 3022 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 13420000 (8231137 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 13440000 (8243874 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 5817798 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2260 dirty pages from their submits' captures, 48449 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 830 dispatches, 38626 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5978 runs, 612.3 MB compared; 88 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17582 in captures, 1 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: program picks: 18996431 kept (7935813 by the last picks' textures, a uniform buffer bound since), 26406 new, 6977162 with other textures (of 26000000)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #240000: event 0x100d6c8024 mode 0 value 240000 (lr 0x80665068)
[hos host] vk: reaper: 8 images (6 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 4 (3 MB)
[hos host] vk: a frame: 312 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.99, present 0.14, draw 7.29, descriptors 3.00, texture lookups 0.18, uploads 0.06, pipelines 0.30, programs 1.30, commands 2.09, dispatch 0.80, memory sync 0.39, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 465 (1102 MB): 5.49 MB synced in 1066.5 copies, 17.30 MB written back
[hos host] vk: 17880 presents, 60.3 fps; a frame: 2287 draws (0 skipped), 125 dispatches, 193 passes, 134 barriers, 6 submits, 0.8 uploads (0.2 MB), 85 ub updates (58 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1325 textures (2431 MB)
[hos host] vk: host time in GPU resource waits: 6.01 ms a frame (fences 2.37, timelines 3.64, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 687.2 full barriers recorded, 307.2 left out (nothing recorded since the one before), 19.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.1; 0x11: 26.7; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 15.9; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.6 from 29.8 batches at the game's barriers (48.6 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1355.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 25.8; 48: 7.9;
[hos host] vk: GPU time 12.3 ms a frame (the longest submission 3.2 ms); 1.1 textures made a frame: colour targets 0.8 (0.7 MB), depth 0.0 (0.4 MB), ASTC or BC 0.3 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.12 MB unswizzled on the CPU in 0.02 ms
[hos SceneLowPriority] vk: write-back of #16652 0x11636fd000: 652 KB the CPU wrote since the GPU kept
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 517 ranges, 3012 MB), 233 pages written, 233 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 4476 texture pages copied (17 MB)
[hos host] vk: 168 textures retired (1176 left, 2257 MB): budget 167, drawn over 1, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: recorder: 7041924 packets run in 10.0 s (busy 1.79 s, spun idle 0.14 s, slept 12963 times, 4922 of them under 200 us), 0 calls made directly (no room); 19264 batches over 599 frames (32.2 a frame, 366 packets a batch), 9899 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 7 times, 0.82 ms (0.0 times, 0.00 ms a frame); every thread 625 times, 68.40 ms in all, the longest 0.61 ms; vkWaitForFences 563/62.88 ms (max 0.61), the queue's lock 38/2.70 ms (max 0.27), vkDestroyImageView 9/1.25 ms (max 0.34), vmaDestroyImage 5/0.75 ms (max 0.57), vkDestroyBuffer 2/0.48 ms (max 0.47), vkQueueBindSparse 3/0.23 ms (max 0.20)
[hos host] vk: recorder: of its busy 1.79 s, 0.43 s in parts' ends and submits (the longest 1.93 ms), 0.11 s of that blocked on the queue's lock (165 times, the longest 1.89 ms), so 0.32 s in the ends, submits and holds; 1.37 s running the other packets; 11756 packets a frame
[hos host] vk: 8 textures retired (1168 left, 2225 MB): budget 8, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 13460000 (8257172 pages from the captures)
[hos host] vk: texture #21381 0x10002094a00 240x135 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (11 so far)
[hos host] vk: write-backs left out (600 submits): 463, 2147.9 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1093 dirty pages from their submits' captures, 48781 from earlier ones', 257 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 806 dispatches, 40984 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 2992 runs, 313.7 MB compared; 274 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13949 in captures, 250 only in the collection, 49 only by the replay itself
[hos host] vk: read live, by pool: 0x115c000000 (16 MB, captured, state 2): 133; 0x1163400000 (8 MB, captured, state 2): 107; 0x104e000000 (24 MB, captured, state 2): 14; 0x7ffe000000 (24 MB, captured, state 2): 1; 0x7f88000000 (12 MB, captured, state 2): 1; 0x1051800000 (4 MB, captured, state 2): 1;
[hos sead::MainThread] nvn: CommandBufferWaitEvent #330000: event 0x100d6c8000 mode 0 value 240329 (now 240310) (lr 0x82c0301c)
[hos host] vk: replay reads of guest memory as submitted: 13480000 (8269818 pages from the captures)
[hos NavMeshCAStepDepWorker1] vcores: 141827 takes, 21600 waits (mean 61 us, longest 9.4 ms, 0 over HOS_VCORE_LOG_MS), 10584 kicks, 10743 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 13 textures retired (1172 left, 2225 MB): budget 0, drawn over 13, dead targets 0
[hos host] vk: texture #21337 0x10000000000 761x177 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 17945 (drawn at 17881)
[hos host] vk: texture #21342 0x10000000000 725x220 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 17945 (drawn at 17881)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 18152 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: presenter thread: the copy's fence 2.60 ms, the present 0.82 ms a present
[hos host] vk: replay reads of guest memory as submitted: 13500000 (8282634 pages from the captures)
[hos host] vk: sparse mirrors: 113, 1104 MB with memory of 1548 MB (443 MB spared); blocks given memory later 13768, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1248 binds (298.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 52 spares now, 1272 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 670, 3031.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2237 dirty pages from their submits' captures, 49063 from earlier ones', 122 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 818 dispatches, 40480 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 4255 runs, 452.3 MB compared; 92 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16175 in captures, 61 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x1163400000 (8 MB, captured, state 2): 102; 0x104e000000 (24 MB, captured, state 2): 15; 0x7ffe000000 (24 MB, captured, state 2): 1; 0x7fb7c00000 (12 MB, captured, state 2): 1; 0x7f88000000 (12 MB, captured, state 2): 1; 0x1159c00000 (36 MB, captured, state 2): 1; 0x1051800000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: 1 textures retired (1203 left, 2235 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 8 pages in 8 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 513 ranges, 3000 MB), 226 pages written, 226 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 24 kept (136 MB), 12 spare buffers, 7041 texture pages copied (27 MB)
[hos BaseProcExecutor0] hid: game window not in front (foreground window: process 38376, class "TkTopLevel")
[hos host] vk: replay reads of guest memory as submitted: 13520000 (8295361 pages from the captures)
[hos host] vk: texture #21504 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (12 so far)
[hos host] vk: replay reads of guest memory as submitted: 13540000 (8308090 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: memory: working set 15446 MB (peak 16021), private 21691 MB, staging 32 MB in 2 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 848 MB; used 3188 of 15579 MB; heap 1 (device): 160 blocks 4825 MB, 19872 allocations 3996 MB; used 6949 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 16 blocks 512 MB, 144 images 287 MB; 75 free ranges, the largest 29 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1053 images 1278 MB; 447 free ranges, the largest 32 MB
[hos host] vk: render target images: 5916 reused from the pool, 4366 made new, 4220 destroyed from it, 1162 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1937 with the pool empty (0 of them with targets still in the dead queue), 340 with none in it a present yet, 46 with none of a kind they can use, 576 with all smaller (the largest 50% short on average), 588 with all bigger (over 1.5 times), 876 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4028 idle
[hos host] vk: VRAM: 6949 MB used of 11474 MB allowed (829 MB of it free in the blocks); textures 2235 MB (budget 7158 MB), 1204 textures
[hos host] vk: budget: VRAM 4825 of 11474 MB (6949 MB by the driver, 2123 MB unaccounted): textures 2235, mirrors 1103, blocks free 829, shadows 284, other buffers 229, other images 111, staging 32; RAM 3125 MB (the process's private memory 21691 MB): imports 2240, capture spares 744, captures 106, capture table 35, others 0
[hos host] vk: textures: decoded ASTC 1367 MB, rendered 537 MB (100 drawn in one frame, kept: 48482 KB), other 301 MB; used in the last 8 presents 1967 MB
[hos host] vk: decoded ASTC used again: 437 textures (955 MB) copied from ones the game finalized, not decoded; 437 of those retired, 0 listed; 1956 MB of blocks hashed in 336 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 37 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 12853 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 60 (11073 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 496 waited for it, 3.59 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 57.9 ms; VRAM 6.8 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2315 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 170 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 7.43, present 0.16, draw 4.34, descriptors 1.84, texture lookups 0.21, uploads 0.09, pipelines 0.18, programs 0.73, commands 1.23, dispatch 0.47, memory sync 0.57, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 466 (1102 MB): 4.30 MB synced in 867.7 copies, 9.50 MB written back
[hos host] vk: 18000 presents, 58.1 fps; a frame: 1347 draws (0 skipped), 67 dispatches, 166 passes, 111 barriers, 6 submits, 0.7 uploads (0.5 MB), 256 ub updates (122 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1204 textures (2235 MB)
[hos host] vk: host time in GPU resource waits: 7.94 ms a frame (fences 4.14, timelines 3.80, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 470.6 full barriers recorded, 213.6 left out (nothing recorded since the one before), 43.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.0; 0x8: 1.6; 0x9: 0.5; 0x10: 2.9; 0x11: 15.0; 0x12: 69.1; 0x13: 1.6; 0x21: 3.1; 0x28: 8.7; 0x92: 3.5; 0x97: 0.5;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 24.9 from 16.2 batches at the game's barriers (26.5 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 104.3, an op 759.7, the end 4.7, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.1; 5: 14.1; 48: 4.3;
[hos host] vk: GPU time 10.3 ms a frame (the longest submission 15.1 ms); 1.4 textures made a frame: colour targets 1.1 (1.8 MB), depth 0.2 (1.2 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.45 MB unswizzled on the CPU in 0.07 ms
[hos host] vk: queues ordered by their events: 117168 submissions stopped at a wait, 117108 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 331436, not met 0; captures at a CPU event's wait 11549, 0 of them at a wait met on arrival
[hos host] vk: the render thread (600 presents, 16.85 ms a frame): its jobs 15.10 ms a frame (submits 2.03, presents 4.51, others 8.55), idle 1.75; the game waited for room at 44 presents, 0.08 ms each (0.01 a frame; 13.61 jobs queued at a present); 13 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 3.75 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 499 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 2388 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.33 ms after its hand-over (25.35 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.08 + 0.15 (the window's copy is in the idle), after it 2.10 + 0.48; after a frame's last recording that reads or writes the game's memory come 1.14 recordings (0.00 ms of GPU time), after its last that reads it 1.14 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 9.62 parts a frame recorded in the signal's job, 70.4% of them took nothing that came with the signal; of the others, 2.85 a frame took pages of the signal's capture into a mirror (72.2 pages a frame) and 0.00 pages read as guest memory had them (0.3 pages a frame); a frame's first part took nothing in 22 of 499 frames, its first three in 22
[hos host] vk: frame ahead, what came with the signal: 499 signals with a capture, each of 94.6 pages in 13.5 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.94+0.20, 0.09+0.03, 0.29+0.10, 0.43+0.06, 0.02+0.01, 0.11+0.01, 0.11+0.01, 0.05+0.02, 0.05+0.02, 0.02+0.01, 0.01+0.00, 0.00+0.00
[hos host] vk: write-backs left out (600 submits): 851, 3776.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2727 dirty pages from their submits' captures, 46125 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 805 dispatches, 36925 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5814 runs, 595.1 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17159 in captures, 0 only in the collection, 26 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vi: real-time step on: ratio 1.002, game's frame rate 60
[hos host] vk: replay reads of guest memory as submitted: 13560000 (8320884 pages from the captures)
[hos host] vk: render targets drawn over from another start, let go: 10833 so far (50877 MB), 9758 of them (45937 MB) between the passes
[hos host] vk: program picks: 19144759 kept (8001344 by the last picks' textures, a uniform buffer bound since), 26406 new, 7028834 with other textures (of 26200000)
[hos host] vk: replay reads of guest memory as submitted: 13580000 (8334398 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 235 sync waits (998 ms), 0 polls timed out (0 ms), 12 queue finishes (80 ms)
[hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 13993789 lookups, 109195 missed (0.78 %)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 112, 1102 MB with memory of 1544 MB (441 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2861 dirty pages from their submits' captures, 46799 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 822 dispatches, 37725 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5951 runs, 612.4 MB compared; 70 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17222 in captures, 0 only in the collection, 24 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 13600000 (8347305 pages from the captures)
[hos host] vk: late pages' fix (600 recordings with one): of 770 captures with late pages 770 fixed (75.1 pages, 75.1 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 522 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 21, of an earlier recording 501, not recorded yet 0, of this part 0; pages fixed from stamps 56772, from snapshots 1028; the shadows: 325831 pages synced from captures, 322688 made known by it, 3143 left unknown because the replay or the collection wrote them since
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 580000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040)
[hos host] vk: replay reads of guest memory as submitted: 13620000 (8359943 pages from the captures)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1009 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1500 kept their virtual core (nothing there to wait for), 3108 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 513 ranges, 2796 MB), 237 pages written, 237 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (106 MB), 11 spare buffers, 7041 texture pages copied (27 MB)
[hos host] vk: replay reads of guest memory as submitted: 13640000 (8372687 pages from the captures)
[hos host] vk: 15 textures retired (1198 left, 2238 MB): budget 0, drawn over 15, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 5884792 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2919 dirty pages from their submits' captures, 47129 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 819 dispatches, 38045 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6115 runs, 629.8 MB compared; 71 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17153 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 13660000 (8385543 pages from the captures)
[hos host] vk: 12 textures retired (1188 left, 2237 MB): budget 0, drawn over 12, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 13680000 (8398467 pages from the captures)
[hos host] vk: program picks: 19296684 kept (8070234 by the last picks' textures, a uniform buffer bound since), 26406 new, 7076909 with other textures (of 26400000)
[hos host] vk: sparse mirrors: 112, 1102 MB with memory of 1544 MB (441 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1936 dirty pages from their submits' captures, 48196 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 38131 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5957 runs, 612.5 MB compared; 77 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17518 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 13700000 (8411246 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16263 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.93, present 0.17, draw 7.05, descriptors 2.89, texture lookups 0.09, uploads 0.01, pipelines 0.30, programs 1.26, commands 2.02, dispatch 0.79, memory sync 0.39, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 466 (1102 MB): 5.51 MB synced in 1071.6 copies, 17.50 MB written back
[hos host] vk: 18120 presents, 61.0 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 192 passes, 134 barriers, 6 submits, 0.2 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1190 textures (2237 MB)
[hos host] vk: host time in GPU resource waits: 5.80 ms a frame (fences 2.43, timelines 3.37, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 692.0 full barriers recorded, 308.5 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.3 ms a frame (the longest submission 4.1 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 13720000 (8424139 pages from the captures)
[hos host] vk: texture #21526 0x10002199c00 120x67 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (13 so far)
[hos host] vk: late pages found at the replay (600 takes): 21 pages in 21 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vi: 18137 frames presented (59.4 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 32, 15.5-17.8 227, 17.8-22 34, 22-30 4, 30-40 0, over 40 1
[hos host] vk: replay reads of guest memory as submitted: 13740000 (8437005 pages from the captures)
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #335000: event 0x100d6c8010 mode 0 value 244100 (now 244081) (lr 0x80660e38)
[hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1068 dirty pages from their submits' captures, 49713 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 838 dispatches, 39079 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6289 runs, 647.2 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17280 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 514 ranges, 2800 MB), 252 pages written, 252 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (134 MB), 10 spare buffers, 7041 texture pages copied (27 MB)
[hos host] vk: replay reads of guest memory as submitted: 13760000 (8449926 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 13780000 (8462782 pages from the captures)
[hos host] vk: sparse mirrors: 112, 1102 MB with memory of 1544 MB (441 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1515 dirty pages from their submits' captures, 48948 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 38622 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6130 runs, 629.8 MB compared; 79 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17446 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: program picks: 19449006 kept (8139295 by the last picks' textures, a uniform buffer bound since), 26406 new, 7124587 with other textures (of 26600000)
[hos host] vk: replay reads of guest memory as submitted: 13800000 (8475545 pages from the captures)
[hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #245000: event 0x100d6c8028 mode 0 value 244999 (lr 0x80667410)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 13820000 (8488163 pages from the captures)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1421 kept their virtual core (nothing there to wait for), 2931 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 13840000 (8500756 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1426 dirty pages from their submits' captures, 48852 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 835 dispatches, 38548 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6135 runs, 629.9 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17383 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos ModuleSystemWorker2] nvn: 2400000 recordings made, 72061 live copies (24.8 MB), 6301 texture copies; 3 command buffers' segments (4.2 MB)
[hos host] vk: replay reads of guest memory as submitted: 13860000 (8513504 pages from the captures)
[hos BaseProcExecutor1] vcores: 144698 takes, 22135 waits (mean 63 us, longest 11.9 ms, 0 over HOS_VCORE_LOG_MS), 10868 kicks, 10973 yields, 0 starved grants, 0 escapes in 5 s
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 515 ranges, 2806 MB), 242 pages written, 242 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 7041 texture pages copied (27 MB)
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.77, present 0.13, draw 7.07, descriptors 2.90, texture lookups 0.08, uploads 0.00, pipelines 0.30, programs 1.25, commands 2.05, dispatch 0.78, memory sync 0.42, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 466 (1102 MB): 5.47 MB synced in 1070.2 copies, 17.50 MB written back
[hos host] vk: 18240 presents, 59.9 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1193 textures (2238 MB)
[hos host] vk: host time in GPU resource waits: 4.06 ms a frame (fences 2.70, timelines 1.36, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 690.7 full barriers recorded, 308.9 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.4, an op 1367.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.3 ms a frame (the longest submission 3.7 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 13880000 (8526372 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 5951888 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 112, 1102 MB with memory of 1544 MB (441 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2388 dirty pages from their submits' captures, 47113 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 37563 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5953 runs, 612.3 MB compared; 72 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17160 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10240 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 13900000 (8539216 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 19601708 kept (8208437 by the last picks' textures, a uniform buffer bound since), 26406 new, 7171885 with other textures (of 26800000)
[hos host] vk: replay reads of guest memory as submitted: 13920000 (8552086 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 13940000 (8565010 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1111 dirty pages from their submits' captures, 50479 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 857 dispatches, 39652 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6291 runs, 647.2 MB compared; 91 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17761 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 13960000 (8577879 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 13980000 (8590511 pages from the captures)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2417 dirty pages from their submits' captures, 48518 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 833 dispatches, 38700 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6118 runs, 629.9 MB compared; 76 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17519 in captures, 0 only in the collection, 29 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 515 ranges, 2812 MB), 250 pages written, 250 copied (1.0 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 199 sync waits (590 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 24 kept (106 MB), 11 spare buffers, 7041 texture pages copied (27 MB)
[hos host] vk: replay reads of guest memory as submitted: 14000000 (8603763 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #340000: event 0x100d6c8010 mode 0 value 247900 (now 247881) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 14020000 (8616229 pages from the captures)
[hos host] vk: program picks: 19754177 kept (8277363 by the last picks' textures, a uniform buffer bound since), 26406 new, 7219416 with other textures (of 27000000)
[hos BaseProcExecutor4] hid: game window in front (foreground window: process 32080, class "Longshot")
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2895 dirty pages from their submits' captures, 46693 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 813 dispatches, 37522 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5974 runs, 612.5 MB compared; 71 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17221 in captures, 0 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 14040000 (8628769 pages from the captures)
[hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.08, present 0.14, draw 7.35, descriptors 3.06, texture lookups 0.10, uploads 0.01, pipelines 0.31, programs 1.30, commands 2.09, dispatch 0.78, memory sync 0.40, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 466 (1100 MB): 5.51 MB synced in 1070.9 copies, 17.50 MB written back
[hos host] vk: 18360 presents, 58.4 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 192 passes, 134 barriers, 6 submits, 0.2 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1193 textures (2238 MB)
[hos host] vk: host time in GPU resource waits: 6.05 ms a frame (fences 3.01, timelines 3.04, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 691.1 full barriers recorded, 308.2 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 13.0 ms a frame (the longest submission 14.8 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: queues ordered by their events: 120000 submissions stopped at a wait, 119938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 340545, not met 0; captures at a CPU event's wait 11828, 0 of them at a wait met on arrival
[hos ModuleSystemWorker5] vk: readbacks of GPU results: 1087 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1663 kept their virtual core (nothing there to wait for), 3201 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 14060000 (8641481 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 14080000 (8654310 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1252 binds (298.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2901 dirty pages from their submits' captures, 46680 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 818 dispatches, 37545 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5952 runs, 612.3 MB compared; 73 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17669 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: replay reads of guest memory as submitted: 14100000 (8666984 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.03 over 516 ranges, 2816 MB), 236 pages written, 236 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 24 kept (96 MB), 12 spare buffers, 7041 texture pages copied (27 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13134 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 14120000 (8679761 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 6019375 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3025 dirty pages from their submits' captures, 46751 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 816 dispatches, 37625 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5944 runs, 612.3 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17248 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 14140000 (8692492 pages from the captures)
[hos Presentation Thread] vi: 18436 frames presented (59.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 18, 15.5-17.8 248, 17.8-22 32, 22-30 0, 30-40 1, over 40 0
[hos host] vk: program picks: 19906557 kept (8346346 by the last picks' textures, a uniform buffer bound since), 26406 new, 7267036 with other textures (of 27200000)
[hos host] vk: late pages found at the replay (600 takes): 17 pages in 17 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos rmbl::VibrationThread] hid: 61440 vibration values, strongest 0.00 since the last count
[hos host] vk: replay reads of guest memory as submitted: 14160000 (8705180 pages from the captures)
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 600000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #250000: event 0x100d6c8034 mode 0 value 249999 (lr 0x806616b0)
[hos host] vk: replay reads of guest memory as submitted: 14180000 (8717932 pages from the captures)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1252 binds (298.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2175 dirty pages from their submits' captures, 49177 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 845 dispatches, 39023 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6118 runs, 629.6 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17470 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 14200000 (8730675 pages from the captures)
[hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.69, present 0.14, draw 6.95, descriptors 2.86, texture lookups 0.07, uploads 0.00, pipelines 0.29, programs 1.24, commands 1.99, dispatch 0.76, memory sync 0.39, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 466 (1100 MB): 5.52 MB synced in 1070.2 copies, 17.50 MB written back
[hos host] vk: 18480 presents, 60.9 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1196 textures (2240 MB)
[hos host] vk: host time in GPU resource waits: 5.53 ms a frame (fences 2.46, timelines 3.07, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 691.2 full barriers recorded, 308.3 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 124.9, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 14220000 (8743268 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: recorder: 6913407 packets run in 10.0 s (busy 1.75 s, spun idle 0.14 s, slept 12749 times, 4711 of them under 200 us), 0 calls made directly (no room); 18978 batches over 595 frames (31.9 a frame, 364 packets a batch), 9761 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 9 times, 1.16 ms (0.0 times, 0.00 ms a frame); every thread 683 times, 73.77 ms in all, the longest 0.49 ms; vkWaitForFences 587/65.92 ms (max 0.49), vkDestroyImageView 44/4.40 ms (max 0.27), the queue's lock 32/1.99 ms (max 0.18), vmaFreeMemory 4/1.06 ms (max 0.42), vmaDestroyImage 7/0.16 ms (max 0.04), a part's submit 2/0.14 ms (max 0.10)
[hos host] vk: recorder: of its busy 1.75 s, 0.43 s in parts' ends and submits (the longest 1.81 ms), 0.11 s of that blocked on the queue's lock (173 times, the longest 1.75 ms), so 0.32 s in the ends, submits and holds; 1.33 s running the other packets; 11619 packets a frame
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 952 dirty pages from their submits' captures, 48494 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 815 dispatches, 38063 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6124 runs, 630.0 MB compared; 83 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16911 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 517 ranges, 2818 MB), 249 pages written, 249 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 24 kept (92 MB), 12 spare buffers, 7212 texture pages copied (28 MB)
[hos host] vk: replay reads of guest memory as submitted: 14240000 (8755799 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 14260000 (8768319 pages from the captures)
[hos ModuleSystemWorker5] vk: readbacks of GPU results: 1083 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1473 kept their virtual core (nothing there to wait for), 2879 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: program picks: 20058700 kept (8415484 by the last picks' textures, a uniform buffer bound since), 26406 new, 7314893 with other textures (of 27400000)
[hos aal::LowPrioWorkerThread] vcores: 148794 takes, 21046 waits (mean 66 us, longest 7.8 ms, 0 over HOS_VCORE_LOG_MS), 10232 kicks, 10382 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 14280000 (8780915 pages from the captures)
[hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1252 binds (298.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1719 dirty pages from their submits' captures, 48871 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 844 dispatches, 38583 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6213 runs, 637.5 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17427 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #345000: event 0x100d6c8010 mode 0 value 251700 (now 251681) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 14300000 (8794070 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14392 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: presenter thread: the copy's fence 2.71 ms, the present 0.86 ms a present
[hos host] vk: replay reads of guest memory as submitted: 14320000 (8806517 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 367 dirty pages from their submits' captures, 51864 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 39997 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6351 runs, 650.1 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17954 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 14340000 (8819164 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 543 ranges, 2908 MB), 261 pages written, 261 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 7212 texture pages copied (28 MB)
[hos host] vk: slow upload 22.2 ms: #21660 0x110aed5200 1024x1024x1 x1 levels 11 NVN 0x8d ASTC (BC3), 475 of 475 KB
[hos host] vk: replay reads of guest memory as submitted: 14360000 (8832803 pages from the captures)
[hos host] vk: memory: working set 15304 MB (peak 16021), private 21450 MB, staging 32 MB in 2 arenas, 42 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 12 blocks 912 MB, 36 allocations 840 MB; used 3176 of 15579 MB; heap 1 (device): 160 blocks 4825 MB, 20657 allocations 4152 MB; used 6949 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 16 blocks 512 MB, 133 images 284 MB; 69 free ranges, the largest 29 MB
[hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1168 images 1394 MB; 436 free ranges, the largest 32 MB
[hos host] vk: render target images: 5916 reused from the pool, 4366 made new, 4231 destroyed from it, 1162 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 1937 with the pool empty (0 of them with targets still in the dead queue), 340 with none in it a present yet, 46 with none of a kind they can use, 576 with all smaller (the largest 50% short on average), 588 with all bigger (over 1.5 times), 876 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4039 idle
[hos host] vk: VRAM: 6949 MB used of 11474 MB allowed (672 MB of it free in the blocks); textures 2349 MB (budget 7158 MB), 1308 textures
[hos host] vk: budget: VRAM 4825 of 11474 MB (6949 MB by the driver, 2123 MB unaccounted): textures 2349, mirrors 1146, blocks free 672, shadows 284, other buffers 229, other images 111, staging 32; RAM 3119 MB (the process's private memory 21450 MB): imports 2244, capture spares 704, captures 86, other buffers 50, capture table 35, others 0
[hos host] vk: textures: decoded ASTC 1471 MB, rendered 533 MB (98 drawn in one frame, kept: 48302 KB), other 313 MB; used in the last 8 presents 2090 MB
[hos host] vk: decoded ASTC used again: 466 textures (1013 MB) copied from ones the game finalized, not decoded; 466 of those retired, 1 listed; 1989 MB of blocks hashed in 341 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 637 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 13453 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 60 (11073 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 437 waited for it, 3.11 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 47.6 ms; VRAM 6.8 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2320 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 306 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.88, present 0.14, draw 7.06, descriptors 2.94, texture lookups 0.32, uploads 0.23, pipelines 0.29, programs 1.19, commands 2.10, dispatch 0.79, memory sync 0.45, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 476 (1146 MB): 5.91 MB synced in 1076.6 copies, 17.62 MB written back
[hos host] vk: 18600 presents, 58.2 fps; a frame: 2262 draws (0 skipped), 133 dispatches, 188 passes, 134 barriers, 6 submits, 0.9 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1308 textures (2349 MB)
[hos host] vk: host time in GPU resource waits: 4.02 ms a frame (fences 2.75, timelines 1.27, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 684.9 full barriers recorded, 303.7 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1323.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.2 ms a frame (the longest submission 4.0 ms); 0.9 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.9 (0.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.07 MB unswizzled on the CPU in 0.01 ms
[hos PreActorModelCompose] nvn: 2450000 recordings made, 73192 live copies (27.2 MB), 6301 texture copies; 2 command buffers' segments (4.1 MB)
[hos host] vk: the render thread (600 presents, 16.76 ms a frame): its jobs 15.46 ms a frame (submits 3.43, presents 4.37, others 7.66), idle 1.30; the game waited for room at 18 presents, 0.15 ms each (0.00 a frame; 7.48 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 443 captures at a CPU signal; the submits' waits read in 0.044 ms a frame; readbacks: 2172 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.66 ms after its hand-over (41.13 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.42 + 0.13 (the window's copy is in the idle), after it 2.18 + 0.57; after a frame's last recording that reads or writes the game's memory come 1.18 recordings (0.00 ms of GPU time), after its last that reads it 1.18 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 9.85 parts a frame recorded in the signal's job, 71.0% of them took nothing that came with the signal; of the others, 2.86 a frame took pages of the signal's capture into a mirror (75.3 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 22 of 443 frames, its first three in 22
[hos host] vk: frame ahead, what came with the signal: 443 signals with a capture, each of 94.0 pages in 13.7 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.00+0.17, 0.11+0.03, 0.30+0.12, 0.41+0.06, 0.02+0.02, 0.09+0.03, 0.10+0.05, 0.06+0.06, 0.05+0.02, 0.02+0.01, 0.01+0.00, 0.00+0.00
[hos host] vk: readback pages refreshed from guest memory: 6086288 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 121, 1146 MB with memory of 1612 MB (465 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1262 binds (301.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 311 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 44 spares now, 1208 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 874, 3883.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2587 dirty pages from their submits' captures, 48438 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 38348 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6329 runs, 617.9 MB compared; 81 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17203 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: render targets drawn over from another start, let go: 10833 so far (50877 MB), 9758 of them (45937 MB) between the passes
[hos host] vk: replay reads of guest memory as submitted: 14380000 (8847635 pages from the captures)
[hos host] vk: program picks: 20210916 kept (8484766 by the last picks' textures, a uniform buffer bound since), 26416 new, 7362667 with other textures (of 27600000)
[hos sead::MainThread] nvn: GPU waits in 5 s: 217 sync waits (741 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 14400000 (8860960 pages from the captures)
[hos BaseProcExecutor1] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 14814359 lookups, 120760 missed (0.82 %)
[hos host] vk: 88 textures retired (1232 left, 2288 MB): budget 88, drawn over 0, dead targets 0
[hos host] vk: reaper: 383 images (754 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 182 (244 MB)
[hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1217 dirty pages from their submits' captures, 51041 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 891 dispatches, 39873 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6706 runs, 637.1 MB compared; 74 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17868 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 14420000 (8874338 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 14440000 (8887651 pages from the captures)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1114 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1635 kept their virtual core (nothing there to wait for), 2973 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: sparse mirrors: 121, 1145 MB with memory of 1612 MB (466 MB spared); blocks given memory later 13768, given back 16943; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1263 binds (301.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 311 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 44 spares now, 1208 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 500 dirty pages from their submits' captures, 57684 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 988 dispatches, 44350 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7196 runs, 705.2 MB compared; 76 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20105 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9583 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 556 ranges, 3139 MB), 262 pages written, 262 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (92 MB), 11 spare buffers, 6754 texture pages copied (26 MB)
[hos host] vk: replay reads of guest memory as submitted: 14460000 (8901294 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 20364946 kept (8557758 by the last picks' textures, a uniform buffer bound since), 26418 new, 7408635 with other textures (of 27800000)
[hos host] vk: replay reads of guest memory as submitted: 14480000 (8913741 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #255000: event 0x100d6c8040 mode 0 value 254999 (lr 0x806621e8)
[hos host] vk: a frame: 230 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.55, present 0.19, draw 6.52, descriptors 2.71, texture lookups 0.10, uploads 0.02, pipelines 0.28, programs 1.12, commands 1.88, dispatch 0.85, memory sync 0.55, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 476 (1146 MB): 5.63 MB synced in 1097.0 copies, 17.65 MB written back
[hos host] vk: 18720 presents, 60.6 fps; a frame: 2046 draws (0 skipped), 138 dispatches, 167 passes, 135 barriers, 6 submits, 0.3 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1253 textures (2297 MB)
[hos host] vk: host time in GPU resource waits: 2.91 ms a frame (fences 1.35, timelines 1.56, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 586.1 full barriers recorded, 280.7 left out (nothing recorded since the one before), 21.2 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.6, an op 1094.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.3 ms a frame (the longest submission 4.8 ms); 0.3 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1405 dirty pages from their submits' captures, 56384 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 987 dispatches, 44360 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7358 runs, 722.6 MB compared; 86 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19667 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 14500000 (8927103 pages from the captures)
[hos Presentation Thread] vi: 18734 frames presented (59.5 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 29, 15.5-17.8 240, 17.8-22 26, 22-30 2, 30-40 0, over 40 1
[hos sead::MainThread] nvn: CommandBufferWaitEvent #350000: event 0x100d6c800c mode 0 value 255499 (now 255480) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 14520000 (8940999 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 14540000 (8953888 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 123, 1151 MB with memory of 1624 MB (472 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1267 binds (301.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 313 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 42 spares now, 1196 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1590 dirty pages from their submits' captures, 55848 from earlier ones', 10 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 965 dispatches, 43802 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7224 runs, 705.9 MB compared; 64 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19706 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x1042000000 (4 MB, captured, state 2): 6; 0x1041c00000 (4 MB, captured, state 2): 3; 0x7fb1800000 (20 MB, captured, state 2): 1;
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 561 ranges, 3146 MB), 264 pages written, 264 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.035; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.019, bytes copied 0.000; 24 kept (98 MB), 11 spare buffers, 7453 texture pages copied (29 MB)
[hos host] vk: replay reads of guest memory as submitted: 14560000 (8967351 pages from the captures)
[hos host] vk: slow upload 16.2 ms: #21720 0x10ffefbe00 512x512x1 x1 levels 10 NVN 0x8e ASTC (BC3), 88 of 88 KB
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.07
[hos host] vk: texture #21786 0x1000a473400 1280x720 NVN 0x29 evicted, made again over storage render targets took: uploaded from the game's memory (14 so far)
[hos host] vk: texture #21787 0x10001fe0800 480x270 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (15 so far)
[hos host] vk: 1 textures retired (1310 left, 2423 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: readback pages refreshed from guest memory: 6153837 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: program picks: 20517721 kept (8627562 by the last picks' textures, a uniform buffer bound since), 26426 new, 7455852 with other textures (of 28000000)
[hos host] vk: replay reads of guest memory as submitted: 14580000 (8981492 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1861 dirty pages from their submits' captures, 53302 from earlier ones', 39 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 946 dispatches, 41958 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7198 runs, 674.6 MB compared; 100 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18810 in captures, 2 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb4400000 (24 MB, captured, state 2): 39;
[hos host] vk: 2 textures retired (1320 left, 2438 MB): budget 0, drawn over 0, dead targets 2
[hos aal::LowPrioWorkerThread] vcores: 152955 takes, 26089 waits (mean 76 us, longest 39.0 ms, 0 over HOS_VCORE_LOG_MS), 12520 kicks, 12679 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 14600000 (8993870 pages from the captures)
[hos host] vk: 6 textures retired (1401 left, 2573 MB): budget 0, drawn over 0, dead targets 6
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17044 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a frame: 220 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.03, present 0.21, draw 7.76, descriptors 3.62, texture lookups 0.87, uploads 0.54, pipelines 0.31, programs 1.15, commands 2.12, dispatch 0.96, memory sync 0.57, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 489 (1180 MB): 5.93 MB synced in 1097.4 copies, 17.71 MB written back
[hos host] vk: 18840 presents, 55.1 fps; a frame: 1958 draws (0 skipped), 141 dispatches, 171 passes, 134 barriers, 6 submits, 2.0 uploads (1.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1467 textures (2679 MB)
[hos host] vk: host time in GPU resource waits: 5.73 ms a frame (fences 1.75, timelines 3.99, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 638.9 full barriers recorded, 285.7 left out (nothing recorded since the one before), 18.2 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1063.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.5 ms a frame (the longest submission 4.5 ms); 5.2 textures made a frame: colour targets 2.7 (9.0 MB), depth 0.8 (4.6 MB), ASTC or BC 1.7 (2.8 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.44 MB unswizzled on the CPU in 0.07 ms
[hos host] vk: sparse mirrors: 136, 1182 MB with memory of 1692 MB (509 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1280 binds (304.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 326 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 29 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1964 dirty pages from their submits' captures, 51658 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 923 dispatches, 40701 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6996 runs, 656.9 MB compared; 72 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17951 in captures, 1 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x7fe1400000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1100 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 13 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1446 kept their virtual core (nothing there to wait for), 3162 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 14620000 (9009288 pages from the captures)
[hos host] vk: texture 0x110750be00 171x171x1 x1 levels 8 NVN 0x8d -> vk 37 (ASTC decoded) (1500 textures, 2714 MB)
[hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 620000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040)
[hos host] vk: 77 textures retired (1453 left, 2667 MB): budget 77, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 14640000 (9024977 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 125 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 590 ranges, 3384 MB), 256 pages written, 256 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (134 MB), 8 spare buffers, 7453 texture pages copied (29 MB)
[hos host] vk: late pages' fix (600 recordings with one): of 818 captures with late pages 818 fixed (78.6 pages, 78.9 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 505 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 1 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 95, of an earlier recording 409, not recorded yet 0, of this part 0; pages fixed from stamps 63202, from snapshots 1084; the shadows: 407301 pages synced from captures, 398333 made known by it, 8968 left unknown because the replay or the collection wrote them since
[hos host] vk: 8 textures retired (1449 left, 2665 MB): budget 8, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 14660000 (9040279 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2285 dirty pages from their submits' captures, 50221 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 879 dispatches, 39571 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6478 runs, 637.0 MB compared; 73 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17946 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool: 0x7fdfc00000 (12 MB, captured, state 2): 1;
[hos host] vk: 1 textures retired (1451 left, 2669 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 14680000 (9055384 pages from the captures)
[hos host] vk: texture 0x10ca831e00 1024x1024x1 x1 levels 11 NVN 0x8d -> vk 137 (ASTC decoded) (1500 textures, 2749 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: 5 textures retired (1590 left, 2911 MB): budget 4, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 14700000 (9069322 pages from the captures)
[hos host] vk: program picks: 20668408 kept (8690778 by the last picks' textures, a uniform buffer bound since), 26433 new, 7505158 with other textures (of 28200000)
[hos sead::MainThread] nvn: GPU waits in 5 s: 221 sync waits (895 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 21 textures retired (1569 left, 2873 MB): budget 21, drawn over 0, dead targets 0
[hos host] vk: sparse mirrors: 138, 1186 MB with memory of 1700 MB (514 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1282 binds (304.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 328 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 27 spares now, 1120 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 926, 4109.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2004 dirty pages from their submits' captures, 51629 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 905 dispatches, 40651 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6831 runs, 656.5 MB compared; 64 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18237 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 14720000 (9081733 pages from the captures)
[hos host] vk: 40 textures retired (1530 left, 2789 MB): budget 40, drawn over 0, dead targets 0
[hos host] vk: 4 textures retired (1530 left, 2792 MB): budget 4, drawn over 0, dead targets 0
[hos ModuleSystemWorker5] nvn: 2500000 recordings made, 75229 live copies (26.1 MB), 6322 texture copies; 6 command buffers' segments (4.4 MB)
[hos host] vk: replay reads of guest memory as submitted: 14740000 (9096421 pages from the captures)
[hos host] vk: 6 textures retired (1524 left, 2765 MB): budget 6, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: CommandBufferWaitEvent #355000: event 0x100d6c800c mode 0 value 259299 (now 259280) (lr 0x80660e38)
[hos host] vk: texture #21098 0x117103b000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 18945 (drawn at 16019)
[hos host] vk: replay reads of guest memory as submitted: 14760000 (9111586 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 596 ranges, 3510 MB), 262 pages written, 262 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.033; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (90 MB), 10 spare buffers, 7873 texture pages copied (30 MB)
[hos host] vk: 6 textures retired (1522 left, 2761 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2416 dirty pages from their submits' captures, 50022 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 876 dispatches, 39568 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6402 runs, 641.8 MB compared; 65 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17823 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 14780000 (9126981 pages from the captures)
[hos host] vk: a frame: 288 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.27, present 0.22, draw 8.04, descriptors 3.79, texture lookups 0.80, uploads 0.48, pipelines 0.32, programs 1.26, commands 2.09, dispatch 0.89, memory sync 0.50, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 493 (1186 MB): 5.74 MB synced in 1097.1 copies, 17.74 MB written back
[hos host] vk: 18960 presents, 55.4 fps; a frame: 2018 draws (0 skipped), 140 dispatches, 200 passes, 133 barriers, 6 submits, 2.0 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1533 textures (2769 MB)
[hos host] vk: host time in GPU resource waits: 6.04 ms a frame (fences 1.85, timelines 4.18, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 755.2 full barriers recorded, 315.2 left out (nothing recorded since the one before), 16.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.9; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 124.1, an op 1252.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.2 ms a frame (the longest submission 4.8 ms); 2.2 textures made a frame: colour targets 0.2 (0.9 MB), depth 0.1 (0.4 MB), ASTC or BC 1.9 (3.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.13 MB unswizzled on the CPU in 0.16 ms
[hos host] vk: 45 textures retired (1488 left, 2698 MB): budget 45, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 14800000 (9142325 pages from the captures)
[hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #260000: event 0x100d6c8000 mode 0 value 259999 (lr 0x80291440)
[hos host] vk: 40 textures retired (1449 left, 2656 MB): budget 40, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 6220788 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 135, 1173 MB with memory of 1680 MB (506 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1283 binds (304.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 328 taken from the spares, 165 made new; 358 kept as spares, 0 destroyed; 30 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2076 dirty pages from their submits' captures, 51286 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 40331 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6284 runs, 640.7 MB compared; 59 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18126 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: 6 textures retired (1443 left, 2655 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 14820000 (9157140 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.04
[hos host] vk: program picks: 20816396 kept (8749198 by the last picks' textures, a uniform buffer bound since), 26433 new, 7557170 with other textures (of 28400000)
[hos host] vk: texture #21785 0x10005280e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21748 0x10005191c00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21755 0x10004741200 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21756 0x10004bf1600 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21757 0x10004ce1800 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21758 0x10004b01400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 39 textures retired (1404 left, 2589 MB): budget 39, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 6 finalized, 34 MB
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1100 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1471 kept their virtual core (nothing there to wait for), 2881 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 14840000 (9172227 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21715 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 66 textures retired (1348 left, 2531 MB): budget 66, drawn over 0, dead targets 0
[hos Presentation Thread] vi: 19015 frames presented (56.2 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 51, 15.5-17.8 152, 17.8-22 52, 22-30 17, 30-40 8, over 40 1
[hos host] vk: replay reads of guest memory as submitted: 14860000 (9187697 pages from the captures)
[hos host] vk: 46 textures retired (1293 left, 2437 MB): budget 45, drawn over 0, dead targets 1
[hos host] vk: texture #22570 0x10008ca2e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19020 (drawn at 18959)
[hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2117 dirty pages from their submits' captures, 48220 from earlier ones', 1 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 851 dispatches, 38043 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6287 runs, 637.3 MB compared; 71 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17257 in captures, 0 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 1;
[hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 5 textures retired (1346 left, 2512 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 14880000 (9202761 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 600 ranges, 3538 MB), 253 pages written, 253 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (148 MB), 9 spare buffers, 7873 texture pages copied (30 MB)
[hos host] vk: 1 textures retired (1422 left, 2580 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: slow upload 15.8 ms: #22837 0x10bfa91e00 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB
[hos host] vk: replay reads of guest memory as submitted: 14900000 (9217828 pages from the captures)
[hos host] vk: 1 textures retired (1488 left, 2751 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: texture 0x10f24ecc00 128x128x1 x1 levels 8 NVN 0x49 -> vk 139 (1500 textures, 2763 MB)
[hos host] vk: sparse mirrors: 135, 1173 MB with memory of 1680 MB (506 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1283 binds (304.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 328 taken from the spares, 165 made new; 358 kept as spares, 0 destroyed; 30 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2948 dirty pages from their submits' captures, 48467 from earlier ones', 6 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 916 dispatches, 38505 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6433 runs, 621.3 MB compared; 67 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17729 in captures, 0 only in the collection, 25 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb4400000 (24 MB, captured, state 2): 2; 0x7fdfc00000 (12 MB, captured, state 2): 1; 0x7fb7c00000 (12 MB, captured, state 2): 1; 0x7f90800000 (4 MB, captured, state 2): 1; 0x1044000000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: 9 textures retired (1520 left, 2795 MB): budget 1, drawn over 0, dead targets 8
[hos host] vk: recorder: 6319602 packets run in 10.0 s (busy 1.84 s, spun idle 0.13 s, slept 11675 times, 4042 of them under 200 us), 0 calls made directly (no room); 18017 batches over 567 frames (31.8 a frame, 351 packets a batch), 9199 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 54 times, 6.73 ms (0.1 times, 0.01 ms a frame); every thread 672 times, 57.97 ms in all, the longest 0.58 ms; vkWaitForFences 497/42.15 ms (max 0.47), vkDestroyImageView 68/5.36 ms (max 0.29), vkQueueBindSparse 32/3.34 ms (max 0.29), the queue's lock 34/3.07 ms (max 0.28), vkEndCommandBuffer 12/1.88 ms (max 0.28), vmaFreeMemory 4/1.08 ms (max 0.58)
[hos host] vk: recorder: of its busy 1.84 s, 0.52 s in parts' ends and submits (the longest 1.80 ms), 0.20 s of that blocked on the queue's lock (350 times, the longest 1.75 ms), so 0.32 s in the ends, submits and holds; 1.33 s running the other packets; 11146 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 14920000 (9231034 pages from the captures)
[hos host] vk: 5 textures retired (1581 left, 2936 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: 3 textures retired (1606 left, 2964 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: a frame: 256 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.98, present 0.19, draw 8.24, descriptors 4.26, texture lookups 1.25, uploads 0.91, pipelines 0.31, programs 1.13, commands 1.99, dispatch 0.83, memory sync 0.44, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 493 (1172 MB): 5.64 MB synced in 1090.9 copies, 17.74 MB written back
[hos host] vk: 19080 presents, 54.7 fps; a frame: 2084 draws (0 skipped), 136 dispatches, 199 passes, 134 barriers, 6 submits, 3.0 uploads (1.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1606 textures (2964 MB)
[hos host] vk: host time in GPU resource waits: 6.99 ms a frame (fences 2.66, timelines 4.32, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 759.9 full barriers recorded, 315.9 left out (nothing recorded since the one before), 18.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1326.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.8 ms a frame (the longest submission 4.0 ms); 3.6 textures made a frame: colour targets 0.5 (2.8 MB), depth 0.3 (1.5 MB), ASTC or BC 2.9 (5.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.47 MB unswizzled on the CPU in 0.08 ms
[hos host] vk: replay reads of guest memory as submitted: 14940000 (9243939 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2348 dirty pages from their submits' captures, 48868 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 39231 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6508 runs, 634.6 MB compared; 64 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17218 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: program picks: 20967720 kept (8815453 by the last picks' textures, a uniform buffer bound since), 26433 new, 7605846 with other textures (of 28600000)
[hos tera::FileResourceMgr] vcores: 129856 takes, 20930 waits (mean 83 us, longest 10.4 ms, 0 over HOS_VCORE_LOG_MS), 10064 kicks, 10251 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 8 textures retired (1628 left, 3048 MB): budget 8, drawn over 0, dead targets 0
[hos host] vk: 79 textures retired (1557 left, 2933 MB): budget 79, drawn over 0, dead targets 0
[hos sead::MainThread] barrier: 295431 looks: 115207 found the masks moved already, 177257 waited until they moved, 2967 for the whole 1 ms (a wait 349 us on average); 271660 wakes
[hos host] vk: 64 textures retired (1494 left, 2834 MB): budget 64, drawn over 0, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.04 over 594 ranges, 3624 MB), 246 pages written, 246 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 4919 texture pages copied (19 MB)
[hos host] vk: 4 textures retired (1490 left, 2826 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 14960000 (9252834 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11062 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: queues ordered by their events: 126000 submissions stopped at a wait, 125938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 359830, not met 0; captures at a CPU event's wait 12429, 0 of them at a wait met on arrival
[hos host] vk: 18 textures retired (1472 left, 2763 MB): budget 18, drawn over 0, dead targets 0
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #360000: event 0x100d6c800c mode 0 value 263099 (now 263080) (lr 0x80660e38)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1287 binds (305.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 331 taken from the spares, 165 made new; 359 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1225, 5439.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 167 dirty pages from their submits' captures, 65927 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1175 dispatches, 51851 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8142 runs, 853.2 MB compared; 113 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23476 in captures, 3 only in the collection, 49 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: texture #22127 0x10004bf1c00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 19143 (drawn at 19082)
[hos host] vk: texture #21473 0x10004a11800 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19143 (drawn at 19082)
[hos host] vk: texture #21490 0x10002761000 960x540 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 19143 (drawn at 19082)
[hos host] vk: late pages found at the replay (600 takes): 6 pages in 6 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 2 textures retired (1474 left, 2756 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: texture #23122 0x100089d2a00 768x768 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (16 so far)
[hos host] vk: texture 0x1000ad13a00 1280x720x1 x1 levels 1 NVN 0x29 -> vk 97 (1500 textures, 2856 MB)
[hos host] vk: 41 textures retired (1474 left, 2734 MB): budget 1, drawn over 0, dead targets 40
[hos sead::MainThread] vk: readbacks of GPU results: 1107 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 8 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1477 kept their virtual core (nothing there to wait for), 3131 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #23165 0x10008852400 384x384 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (17 so far)
[hos host] vk: texture #23172 0x10002199c00 120x67 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (18 so far)
[hos host] vk: presenter thread: the copy's fence 1.80 ms, the present 0.76 ms a present
[hos host] vk: replay reads of guest memory as submitted: 14980000 (9263935 pages from the captures)
[hos host] vk: 3 textures retired (1479 left, 2737 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 6288325 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1533 dirty pages from their submits' captures, 49694 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 891 dispatches, 38984 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6609 runs, 637.1 MB compared; 67 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17565 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb7c00000 (12 MB, captured, state 2): 1;
[hos host] vk: 6 textures retired (1471 left, 2719 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15000000 (9277320 pages from the captures)
[hos host] vk: 12 textures retired (1459 left, 2711 MB): budget 12, drawn over 0, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 15020000 (9290119 pages from the captures)
[hos host] vk: memory: working set 14675 MB (peak 16021), private 21959 MB, staging 80 MB in 5 arenas, 45 made so far (205 ms)
[hos host] vk: memory heaps: heap 0 (system): 12 blocks 912 MB, 36 allocations 840 MB; used 3176 of 15579 MB; heap 1 (device): 176 blocks 5335 MB, 21432 allocations 4655 MB; used 7458 of 8472 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 18 blocks 576 MB, 144 images 368 MB; 79 free ranges, the largest 23 MB
[hos host] vk: image blocks (textures, compacted): 67 blocks 2144 MB, 1320 images 1760 MB; 533 free ranges, the largest 32 MB
[hos host] vk: render target images: 6347 reused from the pool, 4556 made new, 4410 destroyed from it, 1171 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 155 missed after their image was taken by size; it holds 6 (45 MB)
[hos host] vk: render targets not reused by size: 2086 with the pool empty (0 of them with targets still in the dead queue), 342 with none in it a present yet, 48 with none of a kind they can use, 582 with all smaller (the largest 50% short on average), 606 with all bigger (over 1.5 times), 889 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4218 idle
[hos host] vk: VRAM: 7458 MB used of 8472 MB allowed (680 MB of it free in the blocks); textures 2712 MB (budget 2106 MB), 1462 textures
[hos host] vk: budget: VRAM 5327 of 8472 MB (7458 MB by the driver, 2131 MB unaccounted): textures 2712, mirrors 1184, blocks free 680, shadows 284, other buffers 229, other images 111, staging 80, target pool 45; RAM 3119 MB (the process's private memory 21959 MB): imports 2244, capture spares 750, captures 90, capture table 35, others 0
[hos host] vk: textures: decoded ASTC 1778 MB, rendered 584 MB (100 drawn in one frame, kept: 49346 KB), other 314 MB; used in the last 8 presents 1996 MB
[hos host] vk: decoded ASTC used again: 491 textures (1093 MB) copied from ones the game finalized, not decoded; 491 of those retired, 0 listed; 2225 MB of blocks hashed in 370 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 14053 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 64 (12501 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB);
[hos host] vk: 8 textures retired (1454 left, 2706 MB): budget 8, drawn over 0, dead targets 0
[hos Presentation Thread] vk: presents after the CPU's signal (600): 463 waited for it, 3.73 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 44.9 ms; VRAM 7.3 of 8.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6476 queued, 6476 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16362 ms and 348.1 MB read in all); 8781 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12295 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2336 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9632 (6476 ready, 0 loading), 17767 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13980 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 182 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.37, present 0.20, draw 5.02, descriptors 2.18, texture lookups 0.28, uploads 0.06, pipelines 0.23, programs 0.79, commands 1.44, dispatch 0.82, memory sync 0.42, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 496 (1184 MB): 5.53 MB synced in 1073.0 copies, 17.57 MB written back
[hos host] vk: 19200 presents, 60.7 fps; a frame: 1525 draws (0 skipped), 134 dispatches, 167 passes, 133 barriers, 6 submits, 0.7 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1454 textures (2706 MB)
[hos host] vk: host time in GPU resource waits: 3.56 ms a frame (fences 2.78, timelines 0.78, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 604.4 full barriers recorded, 281.5 left out (nothing recorded since the one before), 16.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 2.8; 0x9: 1.0; 0x10: 3.0; 0x11: 25.8; 0x12: 69.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.7, an op 745.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.5 ms a frame (the longest submission 4.7 ms); 1.6 textures made a frame: colour targets 0.9 (3.8 MB), depth 0.3 (2.0 MB), ASTC or BC 0.4 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.27 MB unswizzled on the CPU in 0.04 ms
[hos host] vk: the render thread (600 presents, 17.49 ms a frame): its jobs 16.40 ms a frame (submits 2.90, presents 5.12, others 8.38), idle 1.09; the game waited for room at 39 presents, 0.66 ms each (0.04 a frame; 14.15 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 466 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 2393 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.76 ms after its hand-over (42.72 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.45 + 0.22 (the window's copy is in the idle), after it 2.10 + 1.27; after a frame's last recording that reads or writes the game's memory come 1.02 recordings (0.00 ms of GPU time), after its last that reads it 1.02 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 9.34 parts a frame recorded in the signal's job, 71.0% of them took nothing that came with the signal; of the others, 2.69 a frame took pages of the signal's capture into a mirror (58.2 pages a frame) and 0.02 pages read as guest memory had them (0.1 pages a frame); a frame's first part took nothing in 12 of 466 frames, its first three in 12
[hos host] vk: frame ahead, what came with the signal: 466 signals with a capture, each of 90.5 pages in 12.7 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.93+0.17, 0.32+0.08, 0.28+0.09, 0.08+0.03, 0.09+0.05, 0.13+0.34, 0.17+0.38, 0.05+0.09, 0.03+0.02, 0.01+0.00, 0.01+0.00, 0.00+0.00
[hos host] vk: texture #20951 0x117047f000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 19201 (drawn at 15578)
[hos ModuleSystemWorker1] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 41638293 lookups, 119840 missed (0.29 %)
[hos sead::MainThread] nvn: GPU waits in 5 s: 192 sync waits (699 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: texture #23050 0x10008942600 640x360 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 19208 (drawn at 19147)
[hos host] vk: texture #23060 0x10009932a00 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 19208 (drawn at 19147)
[hos host] vk: render targets drawn over from another start, let go: 11509 so far (53914 MB), 10376 of them (48723 MB) between the passes
[hos host] vk: 33 textures retired (1430 left, 2675 MB): budget 33, drawn over 0, dead targets 0
[hos host] vk: reaper: 943 images (2011 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 87 (130 MB)
[hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1287 binds (305.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 331 taken from the spares, 165 made new; 359 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 575 dirty pages from their submits' captures, 54509 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 944 dispatches, 42365 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7008 runs, 688.8 MB compared; 69 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18949 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 595 ranges, 3628 MB), 270 pages written, 270 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 6820 texture pages copied (26 MB)
[hos host] vk: program picks: 21118084 kept (8879861 by the last picks' textures, a uniform buffer bound since), 26433 new, 7655482 with other textures (of 28800000)
[hos host] vk: texture #23034 0x10004380c00 1280x720 x1 NVN 0x5 the GPU drew, finalized by the game, sampled again at frame 19219 (drawn at 19158)
[hos host] vk: texture #23059 0x10004831400 1280x720 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19219 (drawn at 19158)
[hos host] vk: 54 textures retired (1391 left, 2644 MB): budget 51, drawn over 1, dead targets 2
[hos host] vk: replay reads of guest memory as submitted: 15040000 (9303550 pages from the captures)
[hos host] vk: texture #23053 0x10004ce1800 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 19220 (drawn at 19158)
[hos host] vk: texture #23044 0x10005191c00 1600x1600 x3 NVN 0x32 the GPU drew, finalized by the game, sampled again at frame 19220 (drawn at 19158)
[hos host] vk: texture #23141 0x10008ca2e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19227 (drawn at 19166)
[hos host] vk: 6 textures retired (1394 left, 2703 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15060000 (9318008 pages from the captures)
[hos host] vk: commit: 1824 MB left under Windows' limit as it is now, and the page files can grow by 79851 MB more: counted as room
[hos host] vk: 6 textures retired (1394 left, 2745 MB): budget 6, drawn over 0, dead targets 0
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #265000: event 0x100d6c8008 mode 0 value 264998 (lr 0x80660f34)
[hos host] vk: texture #23051 0x10004741200 2560x1440 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 19246 (drawn at 19158)
[hos host] vk: texture #23052 0x10004bf1600 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 19246 (drawn at 19158)
[hos host] vk: texture #23054 0x10004b01400 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 19246 (drawn at 19158)
[hos host] vk: texture #23078 0x10005280e00 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 19246 (drawn at 19158)
[hos host] vk: texture #23340 0x1000738e800 2400x2400 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (19 so far)
[hos host] [debug] vk: a staging arena of 22 MB took 6.5 ms to make
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 640000 so far (x1 0x11791f0000 x2 0x2 x3 0x11791f0000 x4 0x654040)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14041 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 1 textures retired (1378 left, 2689 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: texture #23140 0x10008c12c00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19260 (drawn at 19199)
[hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 1125, 4995.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 206 dirty pages from their submits' captures, 61885 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1052 dispatches, 48030 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8215 runs, 793.4 MB compared; 96 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21914 in captures, 0 only in the collection, 46 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; 0x7f96800000 (4 MB, captured, state 2): 1;
[hos host] vk: texture #23122 0x100089d2a00 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19265 (drawn at 19204)
[hos host] vk: 4 textures retired (1374 left, 2689 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15080000 (9326305 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: texture #21307 0x1001e23cc00 1120x1333x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1421 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21306 0x1001e840e00 1120x1333x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1387 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 3 textures retired (1371 left, 2677 MB): budget 2, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 2 finalized, 11 MB
[hos Presentation Thread] vi: 19304 frames presented (57.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 24, 15.5-17.8 227, 17.8-22 19, 22-30 14, 30-40 4, over 40 1
[hos host] vk: texture #23127 0x100020d0a00 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19306 (drawn at 19245)
[hos host] vk: texture #23306 0x10008612200 1722x1722 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19308 (drawn at 19246)
[hos host] vk: texture #23332 0x10008612200 1764x1764 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19309 (drawn at 19247)
[hos host] vk: texture #21205 0x1001ef35200 640x360x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1433 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 11 textures retired (1360 left, 2604 MB): budget 7, drawn over 4, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.05 over 595 ranges, 3630 MB), 254 pages written, 254 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 7009 texture pages copied (27 MB)
[hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1287 binds (305.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 331 taken from the spares, 165 made new; 359 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1300, 5772.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 215 dirty pages from their submits' captures, 69178 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1152 dispatches, 54443 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9332 runs, 915.8 MB compared; 91 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 25178 in captures, 0 only in the collection, 52 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1082 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1316 kept their virtual core (nothing there to wait for), 2780 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 1 textures retired (1359 left, 2590 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: a frame: 134 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 7.24, present 0.21, draw 4.00, descriptors 1.85, texture lookups 0.48, uploads 0.12, pipelines 0.17, programs 0.67, commands 1.02, dispatch 0.79, memory sync 0.43, queue submit 0.00, slot wait 0.21
[hos host] vk: a frame: mirrors 496 (1184 MB): 5.32 MB synced in 1058.8 copies, 17.63 MB written back
[hos host] vk: 19320 presents, 60.2 fps; a frame: 967 draws (0 skipped), 141 dispatches, 171 passes, 133 barriers, 6 submits, 0.2 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1359 textures (2590 MB)
[hos host] vk: host time in GPU resource waits: 2.65 ms a frame (fences 2.42, timelines 0.22, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 625.3 full barriers recorded, 286.5 left out (nothing recorded since the one before), 15.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 2.5; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 68.9; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 48.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.9, an op 592.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 9.7 ms a frame (the longest submission 5.6 ms); 1.2 textures made a frame: colour targets 0.5 (2.4 MB), depth 0.6 (3.9 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.30 MB unswizzled on the CPU in 0.06 ms
[hos ModuleSystemWorker3] nvn: 2550000 recordings made, 74830 live copies (22.5 MB), 6411 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: 2 textures retired (1359 left, 2590 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: 17 textures retired (1343 left, 2571 MB): budget 17, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: CommandBufferWaitEvent #365000: event 0x100d6c800c mode 0 value 266899 (now 266880) (lr 0x80660e38)
[hos host] vk: 37 textures retired (1306 left, 2544 MB): budget 37, drawn over 0, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11082 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 15100000 (9333992 pages from the captures)
[hos host] vk: 64 textures retired (1242 left, 2466 MB): budget 64, drawn over 0, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1250, 5550.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 131 dirty pages from their submits' captures, 66764 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1106 dispatches, 52512 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8765 runs, 876.3 MB compared; 110 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 24399 in captures, 0 only in the collection, 50 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: texture #21220 0x1001f0ca400 105x33x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21224 0x1001f0bb000 133x42x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21219 0x1001f0d1600 272x106x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21223 0x1001f0c4200 169x97x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21227 0x1001f0b2e00 197x106x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 121 textures retired (1121 left, 2251 MB): budget 121, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 5 finalized, 0 MB
[hos host] vk: texture #21243 0x1001f08c000 62x32x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21254 0x1001f07a000 54x46x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21246 0x1001f087e00 126x96x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21239 0x1001f092400 18x33x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21238 0x1001f094600 96x112x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21234 0x1001f098800 32x48x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21250 0x1001f084400 102x30x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21242 0x1001f08e200 82x97x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21233 0x1001f09aa00 160x160x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21253 0x1001f07e200 166x94x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21230 0x1001f0a6c00 96x96x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 53 textures retired (1124 left, 2234 MB): budget 11, drawn over 0, dead targets 42; GPU-drawn with no copy kept: 0 live, 11 finalized, 0 MB
[hos host] vk: slow upload 16.8 ms: #23505 0x10f4eb1600 768x768x1 x1 levels 10 NVN 0x7b ASTC, 528 of 528 KB
[hos host] vk: texture #21400 0x1001ef51600 640x360x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1449 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21456 0x1001e24ba00 1120x1333x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1447 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21258 0x1001f072c00 96x32x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1427 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: texture #21262 0x1001f069800 42x41x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1427 frames, over the budget): made again over this storage, it starts empty
[hos host] vk: 112 textures retired (1106 left, 2289 MB): budget 112, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 9 finalized, 12 MB
[hos host] vk: replay reads of guest memory as submitted: 15120000 (9345507 pages from the captures)
[hos tera::FileResourceMgr] vcores: 145649 takes, 20028 waits (mean 61 us, longest 9.1 ms, 0 over HOS_VCORE_LOG_MS), 9781 kicks, 9753 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: 43 textures retired (1191 left, 2391 MB): budget 43, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 6355714 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1291 binds (305.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 333 taken from the spares, 165 made new; 361 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1535 dirty pages from their submits' captures, 56253 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 971 dispatches, 44320 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6919 runs, 716.3 MB compared; 79 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19928 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa8400000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 15140000 (9360771 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 593 ranges, 3644 MB), 259 pages written, 259 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 7355 texture pages copied (28 MB)
[hos host] vk: 2 textures retired (1280 left, 2495 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: program picks: 21250491 kept (8915218 by the last picks' textures, a uniform buffer bound since), 26457 new, 7723051 with other textures (of 29000000)
[hos host] vk: 40 textures retired (1245 left, 2458 MB): budget 40, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15160000 (9375341 pages from the captures)
[hos host] vk: 15 textures retired (1230 left, 2429 MB): budget 15, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15180000 (9389378 pages from the captures)
[hos host] vk: ASTC texture #23788 0x1132ba6a00 384x384 x1 levels 9: the image of #23521 0x10cddbe000 (the same blocks, finalized by the game) copied, not decoded again (500 so far, 1101 MB)
[hos host] vk: 1 textures retired (1238 left, 2429 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: a frame: 188 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.19, present 0.18, draw 5.76, descriptors 3.13, texture lookups 1.22, uploads 0.99, pipelines 0.21, programs 0.76, commands 1.30, dispatch 0.81, memory sync 0.46, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 498 (1184 MB): 5.45 MB synced in 1070.5 copies, 17.56 MB written back
[hos host] vk: 19440 presents, 55.5 fps; a frame: 1237 draws (0 skipped), 133 dispatches, 181 passes, 132 barriers, 6 submits, 3.0 uploads (2.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1238 textures (2429 MB)
[hos host] vk: host time in GPU resource waits: 6.19 ms a frame (fences 3.40, timelines 2.79, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 687.5 full barriers recorded, 296.9 left out (nothing recorded since the one before), 15.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 2.6; 0x9: 1.0; 0x10: 2.7; 0x11: 26.3; 0x12: 69.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.3, an op 761.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.3 ms a frame (the longest submission 4.6 ms); 3.7 textures made a frame: colour targets 0.4 (1.9 MB), depth 0.3 (2.4 MB), ASTC or BC 2.9 (4.8 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.17 MB unswizzled on the CPU in 0.17 ms
[hos host] vk: write-backs left out (600 submits): 926, 4109.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1588 dirty pages from their submits' captures, 51085 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 865 dispatches, 40176 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6570 runs, 654.7 MB compared; 76 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17624 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 15200000 (9403810 pages from the captures)
[hos host] vk: 1 textures retired (1242 left, 2425 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15220000 (9417858 pages from the captures)
[hos sead::MainThread] vk: readbacks of GPU results: 1092 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1368 kept their virtual core (nothing there to wait for), 2984 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12344 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1291 binds (305.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 333 taken from the spares, 165 made new; 361 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 603 dirty pages from their submits' captures, 56178 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 929 dispatches, 43445 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6970 runs, 701.9 MB compared; 57 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19313 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 15240000 (9431880 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 15260000 (9445808 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 592 ranges, 3638 MB), 259 pages written, 259 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (136 MB), 10 spare buffers, 3629 texture pages copied (14 MB)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #270000: event 0x100d6c801c mode 0 value 270000 (lr 0x80665068)
[hos sead::MainThread] nvn: GPU waits in 5 s: 76 sync waits (364 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 4 textures retired (1252 left, 2419 MB): budget 3, drawn over 0, dead targets 1
[hos host] vk: replay reads of guest memory as submitted: 15280000 (9459725 pages from the captures)
[hos host] vk: texture slots registered again at another size: 1000000 (of 1248706 registrations; slot 28277 512x512 -> 120x67)
[hos host] vk: program picks: 21395627 kept (8968541 by the last picks' textures, a uniform buffer bound since), 26457 new, 7777915 with other textures (of 29200000)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 199 dirty pages from their submits' captures, 56732 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 946 dispatches, 43576 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7063 runs, 703.4 MB compared; 64 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19458 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 15300000 (9474082 pages from the captures)
[hos host] vk: texture #23826 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (20 so far)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #370000: event 0x100d6c8010 mode 0 value 270700 (now 270681) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 15320000 (9488525 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 15340000 (9503437 pages from the captures)
[hos host] vk: 6 textures retired (1271 left, 2430 MB): budget 6, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 19 MB
[hos host] vk: a frame: 290 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.20, present 0.17, draw 6.38, descriptors 2.75, texture lookups 0.15, uploads 0.03, pipelines 0.28, programs 1.06, commands 1.79, dispatch 0.81, memory sync 0.45, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 498 (1166 MB): 5.58 MB synced in 1090.1 copies, 17.59 MB written back
[hos host] vk: 19560 presents, 60.0 fps; a frame: 1851 draws (0 skipped), 136 dispatches, 196 passes, 132 barriers, 6 submits, 0.5 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1273 textures (2433 MB)
[hos host] vk: host time in GPU resource waits: 1.68 ms a frame (fences 1.32, timelines 0.36, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 721.5 full barriers recorded, 309.4 left out (nothing recorded since the one before), 16.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.9, an op 1150.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.6 ms a frame (the longest submission 4.5 ms); 0.6 textures made a frame: colour targets 0.1 (0.4 MB), depth 0.1 (0.2 MB), ASTC or BC 0.4 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 133, 1166 MB with memory of 1668 MB (501 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1294 binds (306.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 333 taken from the spares, 165 made new; 365 kept as spares, 0 destroyed; 32 spares now, 1152 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 382 dirty pages from their submits' captures, 57495 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 964 dispatches, 43959 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7144 runs, 705.4 MB compared; 54 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19599 in captures, 1 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB)
[hos host] vk: 1 textures retired (1281 left, 2455 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15360000 (9517661 pages from the captures)
[hos host] vk: 24 textures retired (1264 left, 2446 MB): budget 24, drawn over 0, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9616 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 15380000 (9531915 pages from the captures)
[hos Presentation Thread] vi: 19595 frames presented (58.2 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 22, 15.5-17.8 242, 17.8-22 16, 22-30 6, 30-40 5, over 40 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 592 ranges, 3642 MB), 268 pages written, 268 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 4661 texture pages copied (18 MB)
[hos host] vk: 28 textures retired (1253 left, 2482 MB): budget 28, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 6423220 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 400 dirty pages from their submits' captures, 56692 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 956 dispatches, 43621 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7292 runs, 707.7 MB compared; 47 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19674 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fdfc00000 (12 MB, captured, state 2): 1;
[hos host] vk: replay reads of guest memory as submitted: 15400000 (9545975 pages from the captures)
[hos host] vk: 368 mirrors given back (2441 MB in all); 1177 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: 25 textures retired (1292 left, 2605 MB): budget 21, drawn over 4, dead targets 0
[hos host] vk: texture #22534 0x1171b2b000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 19622 (drawn at 18893)
[hos host] vk: replay reads of guest memory as submitted: 15420000 (9558583 pages from the captures)
[hos host] vk: program picks: 21540426 kept (9021988 by the last picks' textures, a uniform buffer bound since), 26458 new, 7833115 with other textures (of 29400000)
[hos rmbl::VibrationThread] hid: 65536 vibration values, strongest 0.00 since the last count
[hos host] vk: 14 textures retired (1292 left, 2615 MB): budget 6, drawn over 8, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1117 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1535 kept their virtual core (nothing there to wait for), 3073 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 4 textures retired (1297 left, 2649 MB): budget 0, drawn over 4, dead targets 0
[hos host] vk: sparse mirrors: 127, 1134 MB with memory of 1620 MB (485 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1303 binds (306.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 336 taken from the spares, 165 made new; 374 kept as spares, 0 destroyed; 38 spares now, 1200 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1071 dirty pages from their submits' captures, 53780 from earlier ones', 1 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 932 dispatches, 41589 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7036 runs, 673.6 MB compared; 56 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18500 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 193 live (2244 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 15440000 (9569598 pages from the captures)
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 660000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040)
[hos host] vk: 9 textures retired (1294 left, 2648 MB): budget 1, drawn over 8, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB
[hos host] vk: recorder: 5246760 packets run in 10.0 s (busy 1.59 s, spun idle 0.12 s, slept 10632 times, 3163 of them under 200 us), 0 calls made directly (no room); 17177 batches over 591 frames (29.1 a frame, 305 packets a batch), 8485 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 67 times, 10.21 ms (0.1 times, 0.02 ms a frame); every thread 742 times, 67.12 ms in all, the longest 0.59 ms; vkWaitForFences 521/43.27 ms (max 0.55), vkDestroyImageView 143/17.45 ms (max 0.57), vmaFreeMemory 13/2.59 ms (max 0.59), vkEndCommandBuffer 6/1.37 ms (max 0.45), vkQueueBindSparse 21/1.03 ms (max 0.30), the queue's lock 8/0.69 ms (max 0.18)
[hos host] vk: recorder: of its busy 1.59 s, 0.41 s in parts' ends and submits (the longest 1.37 ms), 0.12 s of that blocked on the queue's lock (250 times, the longest 1.33 ms), so 0.30 s in the ends, submits and holds; 1.18 s running the other packets; 8878 packets a frame
[hos host] vk: 11 textures retired (1290 left, 2643 MB): budget 0, drawn over 11, dead targets 0
[hos host] vk: texture #23902 0x100089d2a00 953x953 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19668 (drawn at 19606)
[hos host] vk: 9 textures retired (1283 left, 2618 MB): budget 0, drawn over 9, dead targets 0
[hos host] vk: texture #23906 0x100089d2a00 1018x1018 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19672 (drawn at 19610)
[hos host] vk: replay reads of guest memory as submitted: 15460000 (9580203 pages from the captures)
[hos host] vk: texture #23909 0x100089d2a00 1046x1046 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19674 (drawn at 19612)
[hos host] vk: 9 textures retired (1280 left, 2614 MB): budget 2, drawn over 7, dead targets 0
[hos host] vk: a frame: 274 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.40, present 0.20, draw 6.50, descriptors 2.90, texture lookups 0.42, uploads 0.09, pipelines 0.27, programs 1.07, commands 1.76, dispatch 0.82, memory sync 0.44, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 501 (1116 MB): 5.77 MB synced in 1088.3 copies, 17.72 MB written back
[hos host] vk: 19680 presents, 60.2 fps; a frame: 1817 draws (0 skipped), 145 dispatches, 192 passes, 132 barriers, 6 submits, 0.5 uploads (0.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1281 textures (2619 MB)
[hos host] vk: host time in GPU resource waits: 2.27 ms a frame (fences 1.30, timelines 0.97, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 719.8 full barriers recorded, 306.3 left out (nothing recorded since the one before), 15.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.0, an op 1137.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.1 ms a frame (the longest submission 4.5 ms); 1.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.8 (2.9 MB), ASTC or BC 0.4 (0.8 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.37 MB unswizzled on the CPU in 0.06 ms
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 281 dirty pages from their submits' captures, 57726 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 952 dispatches, 44454 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7514 runs, 710.0 MB compared; 62 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20163 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 6 textures retired (1279 left, 2619 MB): budget 2, drawn over 4, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 589 ranges, 3446 MB), 267 pages written, 267 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (102 MB), 10 spare buffers, 6194 texture pages copied (24 MB)
[hos rmbl::VibrationThread] vcores: 140103 takes, 21119 waits (mean 72 us, longest 8.1 ms, 0 over HOS_VCORE_LOG_MS), 10093 kicks, 10095 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos sead::MainThread] nvn: 2600000 recordings made, 74825 live copies (24.1 MB), 6530 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: replay reads of guest memory as submitted: 15480000 (9590690 pages from the captures)
[hos host] vk: 8 textures retired (1276 left, 2608 MB): budget 1, drawn over 7, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11077 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 6 textures retired (1278 left, 2619 MB): budget 1, drawn over 5, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 15500000 (9602292 pages from the captures)
[hos host] vk: 6 textures retired (1306 left, 2571 MB): budget 0, drawn over 0, dead targets 6
[hos host] vk: pool 0x10140a5600+0xf00000 imported in 4.0 ms as 0x1014000000..0x1015000000 (240 imports, 724 ms in all)
[hos host] vk: sparse mirrors: 122, 1119 MB with memory of 1596 MB (476 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1309 binds (307.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 338 taken from the spares, 165 made new; 381 kept as spares, 0 destroyed; 43 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 729 dirty pages from their submits' captures, 54299 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 920 dispatches, 42340 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7336 runs, 709.0 MB compared; 63 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19494 in captures, 2 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: program picks: 21684454 kept (9077147 by the last picks' textures, a uniform buffer bound since), 26458 new, 7889087 with other textures (of 29600000)
[hos host] vk: 19 textures retired (1364 left, 2828 MB): budget 19, drawn over 0, dead targets 0
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #375000: event 0x100d6c800c mode 0 value 274499 (now 274480) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 15520000 (9614241 pages from the captures)
[hos host] vk: 1 textures retired (1392 left, 2926 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: 1 textures retired (1407 left, 2967 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1398 dirty pages from their submits' captures, 54501 from earlier ones', 1 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1009 dispatches, 42993 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8031 runs, 715.9 MB compared; 57 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18605 in captures, 0 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa8400000 (4 MB, captured, state 2): 1;
[hos host] vk: presenter thread: the copy's fence 1.87 ms, the present 0.72 ms a present
[hos host] vk: replay reads of guest memory as submitted: 15540000 (9624787 pages from the captures)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #275000: event 0x100d6c801c mode 0 value 274997 (lr 0x802911b8)
[hos host] vk: texture #23410 0x100020e5000 30x16 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 19778 (drawn at 19717)
[hos host] vk: texture #23382 0x1000589ac00 640x360 x1 NVN 0x11 the GPU drew, finalized by the game, sampled again at frame 19778 (drawn at 19717)
[hos host] vk: texture #23412 0x100020e5a00 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 19778 (drawn at 19717)
[hos host] vk: texture #23394 0x10004bf1c00 1280x720 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19778 (drawn at 19717)
[hos host] vk: 10 textures retired (1417 left, 3034 MB): budget 7, drawn over 3, dead targets 0
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 584 ranges, 3522 MB), 256 pages written, 256 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (116 MB), 10 spare buffers, 8192 texture pages copied (32 MB)
[hos host] vk: 52 textures retired (1371 left, 2913 MB): budget 44, drawn over 8, dead targets 0
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1096 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 24 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1451 kept their virtual core (nothing there to wait for), 3157 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 15560000 (9633931 pages from the captures)
[hos ModuleSystemWorker4] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 15444036 lookups, 120643 missed (0.78 %)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: memory: working set 14749 MB (peak 16021), private 22097 MB, staging 80 MB in 5 arenas, 47 made so far (212 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 944 MB, 35 allocations 864 MB; used 3232 of 15579 MB; heap 1 (device): 177 blocks 5399 MB, 20478 allocations 4790 MB; used 7522 of 8426 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 25 blocks 800 MB, 197 images 678 MB; 129 free ranges, the largest 15 MB
[hos host] vk: image blocks (textures, compacted): 60 blocks 1920 MB, 1176 images 1639 MB; 510 free ranges, the largest 24 MB
[hos host] vk: render target images: 6408 reused from the pool, 4918 made new, 4719 destroyed from it, 1224 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 164 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 2276 with the pool empty (0 of them with targets still in the dead queue), 363 with none in it a present yet, 59 with none of a kind they can use, 637 with all smaller (the largest 50% short on average), 629 with all bigger (over 1.5 times), 951 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4527 idle
[hos host] vk: VRAM: 7522 MB used of 8426 MB allowed (605 MB of it free in the blocks); textures 2954 MB (budget 2106 MB), 1379 textures
[hos host] vk: budget: VRAM 5395 of 8426 MB (7522 MB by the driver, 2127 MB unaccounted): textures 2954, mirrors 1130, blocks free 605, shadows 284, other buffers 229, other images 111, staging 80; RAM 3167 MB (the process's private memory 22097 MB): imports 2260, capture spares 728, captures 110, capture table 35, other buffers 26, retired imports 8, others 0
[hos host] vk: textures: decoded ASTC 1490 MB, rendered 840 MB (73 drawn in one frame, kept: 48107 KB), other 531 MB; used in the last 8 presents 1821 MB
[hos host] vk: decoded ASTC used again: 528 textures (1160 MB) copied from ones the game finalized, not decoded; 528 of those retired, 4 listed; 2374 MB of blocks hashed in 390 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 14653 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34, last used 2 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 78 (13794 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB);
[hos host] vk: 52 textures retired (1327 left, 2801 MB): budget 44, drawn over 8, dead targets 0
[hos Presentation Thread] vk: presents after the CPU's signal (600): 214 waited for it, 3.22 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 41.2 ms; VRAM 7.3 of 8.3 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: reaper: 1102 images (2840 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 120 (214 MB)
[hos host] vk: program loads at create (play): 6476 queued, 6476 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16362 ms and 348.1 MB read in all); 8781 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12295 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2361 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9632 (6476 ready, 0 loading), 17767 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13980 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 238 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.51, present 0.20, draw 6.70, descriptors 3.24, texture lookups 0.92, uploads 0.57, pipelines 0.25, programs 1.08, commands 1.66, dispatch 0.84, memory sync 0.47, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 509 (1130 MB): 5.58 MB synced in 1074.0 copies, 17.82 MB written back
[hos host] vk: 19800 presents, 59.2 fps; a frame: 1755 draws (0 skipped), 149 dispatches, 174 passes, 132 barriers, 6 submits, 2.1 uploads (2.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1328 textures (2806 MB)
[hos host] vk: host time in GPU resource waits: 3.27 ms a frame (fences 1.50, timelines 1.77, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 671.6 full barriers recorded, 291.5 left out (nothing recorded since the one before), 13.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.8, an op 1010.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.4 ms a frame (the longest submission 4.0 ms); 2.8 textures made a frame: colour targets 0.5 (1.8 MB), depth 1.0 (5.4 MB), ASTC or BC 1.3 (2.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.66 MB unswizzled on the CPU in 0.23 ms
[hos host] vk: the render thread (600 presents, 16.96 ms a frame): its jobs 14.63 ms a frame (submits 6.01, presents 4.91, others 3.70), idle 2.33; the game waited for room at 29 presents, 0.57 ms each (0.03 a frame; 9.69 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 222 captures at a CPU signal; the submits' waits read in 0.040 ms a frame; readbacks: 1267 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 13.84 ms after its hand-over (36.13 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.66 + 0.21 (the window's copy is in the idle), after it 1.56 + 0.90; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.33 parts a frame recorded in the signal's job, 71.3% of them took nothing that came with the signal; of the others, 2.38 a frame took pages of the signal's capture into a mirror (36.9 pages a frame) and 0.01 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 25 of 222 frames, its first three in 25
[hos host] vk: frame ahead, what came with the signal: 222 signals with a capture, each of 51.8 pages in 10.8 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.73+0.16, 0.20+0.10, 0.21+0.07, 0.08+0.03, 0.09+0.12, 0.11+0.21, 0.09+0.18, 0.04+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: readback pages refreshed from guest memory: 6490330 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 128, 1130 MB with memory of 1624 MB (493 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1315 binds (308.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 344 taken from the spares, 165 made new; 381 kept as spares, 0 destroyed; 37 spares now, 1196 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 155 dirty pages from their submits' captures, 59343 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1036 dispatches, 45790 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8356 runs, 766.0 MB compared; 57 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20654 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: render targets drawn over from another start, let go: 11824 so far (55528 MB), 10639 of them (50075 MB) between the passes
[hos host] vk: 11 textures retired (1325 left, 2760 MB): budget 1, drawn over 10, dead targets 0
[hos host] vk: 10 textures retired (1323 left, 2741 MB): budget 2, drawn over 7, dead targets 1
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12610 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 15580000 (9642328 pages from the captures)
[hos host] vk: 10 textures retired (1323 left, 2719 MB): budget 0, drawn over 10, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 6 textures retired (1322 left, 2716 MB): budget 0, drawn over 6, dead targets 0
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 163 dirty pages from their submits' captures, 59838 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1028 dispatches, 45744 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8354 runs, 752.2 MB compared; 49 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21198 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa6800000 (4 MB, captured, state 2): 2;
[hos host] vk: program picks: 21829563 kept (9142072 by the last picks' textures, a uniform buffer bound since), 26458 new, 7943978 with other textures (of 29800000)
[hos host] vk: 4 textures retired (1330 left, 2741 MB): budget 0, drawn over 4, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15600000 (9652267 pages from the captures)
[hos host] vk: 9 textures retired (1326 left, 2715 MB): budget 0, drawn over 9, dead targets 0
[hos sead::MainThread] nvn: GPU waits in 5 s: 94 sync waits (298 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 8 textures retired (1328 left, 2718 MB): budget 0, drawn over 8, dead targets 0
[hos host] vk: 2 textures retired (1299 left, 2596 MB): budget 0, drawn over 2, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 582 ranges, 3521 MB), 271 pages written, 271 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (110 MB), 9 spare buffers, 2477 texture pages copied (9 MB)
[hos host] vk: replay reads of guest memory as submitted: 15620000 (9661932 pages from the captures)
[hos host] vk: sparse mirrors: 127, 1128 MB with memory of 1620 MB (491 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1316 binds (308.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 344 taken from the spares, 165 made new; 382 kept as spares, 0 destroyed; 38 spares now, 1200 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 207 dirty pages from their submits' captures, 59867 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1070 dispatches, 45981 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8721 runs, 773.7 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20865 in captures, 2 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa6800000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: 14 textures retired (1286 left, 2586 MB): budget 12, drawn over 2, dead targets 0
[hos Presentation Thread] vi: 19894 frames presented (59.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 41, 15.5-17.8 236, 17.8-22 14, 22-30 6, 30-40 2, over 40 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: 98 textures retired (1193 left, 2410 MB): budget 98, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB
[hos host] vk: 69 textures retired (1130 left, 2295 MB): budget 61, drawn over 8, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15640000 (9671298 pages from the captures)
[hos host] vk: 70 textures retired (1063 left, 2182 MB): budget 64, drawn over 6, dead targets 0
[hos host] vk: a frame: 214 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.48, present 0.22, draw 5.99, descriptors 2.70, texture lookups 0.27, uploads 0.26, pipelines 0.25, programs 1.01, commands 1.58, dispatch 0.85, memory sync 0.47, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 509 (1124 MB): 5.53 MB synced in 1079.0 copies, 18.00 MB written back
[hos host] vk: 19920 presents, 60.0 fps; a frame: 1744 draws (0 skipped), 159 dispatches, 168 passes, 135 barriers, 6 submits, 1.1 uploads (1.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1064 textures (2185 MB)
[hos host] vk: host time in GPU resource waits: 1.44 ms a frame (fences 1.42, timelines 0.02, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 651.0 full barriers recorded, 289.2 left out (nothing recorded since the one before), 13.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 28.6; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.5, an op 886.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.4 ms a frame (the longest submission 3.3 ms); 1.0 textures made a frame: colour targets 0.2 (1.1 MB), depth 0.7 (2.5 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.89 MB unswizzled on the CPU in 0.25 ms
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11067 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 311 dirty pages from their submits' captures, 56662 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1031 dispatches, 43351 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8372 runs, 724.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19984 in captures, 2 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 20 textures retired (1060 left, 2194 MB): budget 20, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15660000 (9679863 pages from the captures)
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #380000: event 0x100d6c800c mode 0 value 278299 (now 278280) (lr 0x80660e38)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: VRAM compaction 7 begins: the textures' blocks hold 1920 MB, 746 MB of it free space
[hos host] vk: 2 textures retired (1040 left, 2125 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: program picks: 21975766 kept (9208829 by the last picks' textures, a uniform buffer bound since), 26458 new, 7997775 with other textures (of 30000000)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1150 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1412 kept their virtual core (nothing there to wait for), 2940 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: queues ordered by their events: 132000 submissions stopped at a wait, 131938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 380286, not met 0; captures at a CPU event's wait 12701, 0 of them at a wait met on arrival
[hos host] vk: replay reads of guest memory as submitted: 15680000 (9689300 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 128, 1130 MB with memory of 1628 MB (497 MB spared); blocks given memory later 13773, given back 17008; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1319 binds (309.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 345 taken from the spares, 165 made new; 382 kept as spares, 0 destroyed; 37 spares now, 1192 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 129 dirty pages from their submits' captures, 58229 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1041 dispatches, 44371 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8451 runs, 740.7 MB compared; 53 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20380 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 581 ranges, 3531 MB), 284 pages written, 284 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (132 MB), 9 spare buffers, 3650 texture pages copied (14 MB)
[hos BaseProcExecutor1] vcores: 152285 takes, 23128 waits (mean 72 us, longest 7.3 ms, 0 over HOS_VCORE_LOG_MS), 10974 kicks, 10935 yields, 0 starved grants, 0 escapes in 5 s
[hos ModuleSystemWorker2] ocb: 28 choices of another neighbour held back, 13 taken, 0 updates refused, in 18000 updates
[hos sead::MainThread] nvn: QueueSubmitCommands #120000: queue 000010100D695020 (its #19996), 1 handles (lr 0x80666be0)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: replay reads of guest memory as submitted: 15700000 (9699649 pages from the captures)
[hos sead::MainThread] nvn: EventSignal (CPU) #40000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: readback pages refreshed from guest memory: 6557874 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 241 dirty pages from their submits' captures, 64070 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1061 dispatches, 45069 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8765 runs, 745.9 MB compared; 63 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20878 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2;
[hos host] vk: VRAM compaction 7 over: 389 textures moved (280 MB) in 34 passes, 0 left where they were; 21 blocks (672 MB) given back to the driver, 4480 MB in all so far
[hos host] vk: texture #20735 0x116ff5c000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 20017 (drawn at 15243)
[hos host] vk: 384 mirrors given back (2517 MB in all); 1123 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: replay reads of guest memory as submitted: 15720000 (9709917 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #280000: event 0x100d6c8030 mode 0 value 279999 (lr 0x806612a8)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5217 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a frame: 190 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.01, present 0.44, draw 6.20, descriptors 2.90, texture lookups 0.44, uploads 0.28, pipelines 0.25, programs 1.00, commands 1.60, dispatch 0.91, memory sync 0.52, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 510 (1116 MB): 5.81 MB synced in 1084.4 copies, 18.16 MB written back
[hos host] vk: 20040 presents, 60.0 fps; a frame: 1723 draws (0 skipped), 170 dispatches, 171 passes, 144 barriers, 6 submits, 0.9 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1072 textures (2159 MB)
[hos host] vk: host time in GPU resource waits: 1.78 ms a frame (fences 1.44, timelines 0.34, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 665.4 full barriers recorded, 300.4 left out (nothing recorded since the one before), 17.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.1, an op 895.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.7 ms a frame (the longest submission 4.4 ms); 1.5 textures made a frame: colour targets 0.4 (2.0 MB), depth 0.7 (2.4 MB), ASTC or BC 0.5 (0.8 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.55 MB unswizzled on the CPU in 0.23 ms
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 680000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: sparse mirrors: 124, 1116 MB with memory of 1604 MB (487 MB spared); blocks given memory later 13773, given back 17009; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1324 binds (309.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 345 taken from the spares, 165 made new; 386 kept as spares, 0 destroyed; 41 spares now, 1216 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1001, 4442.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 418 dirty pages from their submits' captures, 56919 from earlier ones', 28 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1043 dispatches, 43485 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8846 runs, 731.3 MB compared; 53 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20674 in captures, 0 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 28;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 15740000 (9719726 pages from the captures)
[hos host] vk: program picks: 22124245 kept (9277010 by the last picks' textures, a uniform buffer bound since), 26461 new, 8049293 with other textures (of 30200000)
[hos ModuleSystemWorker1] tex: the game's texture heap: 331 compaction rounds asked for in 18000 frames without one running
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 579 ranges, 3531 MB), 292 pages written, 292 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (122 MB), 9 spare buffers, 6319 texture pages copied (24 MB)
[hos host] vk: replay reads of guest memory as submitted: 15760000 (9726878 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 277 dirty pages from their submits' captures, 59599 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1035 dispatches, 45638 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9360 runs, 770.6 MB compared; 72 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21839 in captures, 2 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7f86400000 (8 MB, captured, state 2): 1;
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: replay reads of guest memory as submitted: 15780000 (9736021 pages from the captures)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1140 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1303 kept their virtual core (nothing there to wait for), 3305 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture #24295 0x10009c02a00 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 20126 (drawn at 20065)
[hos host] vk: texture #24291 0x10008c12600 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20126 (drawn at 20065)
[hos host] vk: sparse mirrors: 120, 1062 MB with memory of 1528 MB (465 MB spared); blocks given memory later 13773, given back 17017; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1329 binds (310.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 345 taken from the spares, 165 made new; 390 kept as spares, 0 destroyed; 45 spares now, 1292 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 283 dirty pages from their submits' captures, 57059 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 971 dispatches, 43677 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8960 runs, 732.9 MB compared; 47 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20920 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos ModuleSystemWorker5] nvn: 2650000 recordings made, 74449 live copies (24.5 MB), 6593 texture copies; 6 command buffers' segments (4.3 MB)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #385000: event 0x100d6c800c mode 0 value 282099 (now 282080) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 15800000 (9744428 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14238 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a frame: 172 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.27, present 0.18, draw 6.38, descriptors 2.86, texture lookups 0.26, uploads 0.28, pipelines 0.27, programs 1.05, commands 1.73, dispatch 1.01, memory sync 0.48, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 511 (1054 MB): 5.60 MB synced in 1080.6 copies, 18.34 MB written back
[hos host] vk: 20160 presents, 60.0 fps; a frame: 1682 draws (0 skipped), 180 dispatches, 168 passes, 144 barriers, 6 submits, 0.8 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1130 textures (2265 MB)
[hos host] vk: host time in GPU resource waits: 1.90 ms a frame (fences 1.42, timelines 0.49, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 662.7 full barriers recorded, 297.6 left out (nothing recorded since the one before), 17.2 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.9, an op 897.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.2 ms); 1.1 textures made a frame: colour targets 0.3 (1.9 MB), depth 0.4 (1.7 MB), ASTC or BC 0.4 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.53 MB unswizzled on the CPU in 0.24 ms
[hos sead::MainThread] nvn: GPU waits in 5 s: 37 sync waits (68 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 578 ranges, 3535 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (126 MB), 10 spare buffers, 7624 texture pages copied (29 MB)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 764 dirty pages from their submits' captures, 54800 from earlier ones', 3 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 938 dispatches, 41514 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8718 runs, 714.3 MB compared; 59 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20212 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 3;
[hos host] vk: replay reads of guest memory as submitted: 15820000 (9754637 pages from the captures)
[hos host] vk: program picks: 22272273 kept (9344177 by the last picks' textures, a uniform buffer bound since), 26490 new, 8101236 with other textures (of 30400000)
[hos Presentation Thread] vi: 20194 frames presented (59.9 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 26, 15.5-17.8 251, 17.8-22 20, 22-30 3, 30-40 0, over 40 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06
[hos host] vk: texture #24398 0x10008942600 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20196 (drawn at 20135)
[hos host] vk: replay reads of guest memory as submitted: 15840000 (9764386 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 6625116 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 115, 1045 MB with memory of 1496 MB (450 MB spared); blocks given memory later 13773, given back 17082; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1338 binds (312.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 347 taken from the spares, 165 made new; 397 kept as spares, 0 destroyed; 50 spares now, 1324 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 423 dirty pages from their submits' captures, 55479 from earlier ones', 82 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 933 dispatches, 42292 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9164 runs, 727.6 MB compared; 51 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21058 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 80; 0x7f8a000000 (4 MB, captured, state 2): 2;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 15860000 (9773230 pages from the captures)
[hos host] vk: texture #24751 0x10008612200 860x860 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20248 (drawn at 20187)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 285 dirty pages from their submits' captures, 55374 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 933 dispatches, 42317 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9011 runs, 724.1 MB compared; 50 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20907 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15397 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: recorder: 5663800 packets run in 10.0 s (busy 1.70 s, spun idle 0.12 s, slept 10672 times, 3229 of them under 200 us), 0 calls made directly (no room); 17934 batches over 599 frames (29.9 a frame, 316 packets a batch), 8843 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 182 times, 20.98 ms (0.3 times, 0.04 ms a frame); every thread 790 times, 57.91 ms in all, the longest 0.85 ms; vkWaitForFences 510/29.61 ms (max 0.49), vkDestroyImageView 189/20.30 ms (max 0.53), vmaFreeMemory 23/2.64 ms (max 0.32), vkQueueBindSparse 33/2.25 ms (max 0.41), vmaDestroyBuffer 10/1.24 ms (max 0.85), a part's submit 5/0.73 ms (max 0.28)
[hos host] vk: recorder: of its busy 1.70 s, 0.41 s in parts' ends and submits (the longest 1.59 ms), 0.08 s of that blocked on the queue's lock (164 times, the longest 1.53 ms), so 0.33 s in the ends, submits and holds; 1.29 s running the other packets; 9455 packets a frame
[hos host] vk: 4 textures retired (1216 left, 2428 MB): budget 0, drawn over 4, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15880000 (9781472 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 576 ranges, 3538 MB), 278 pages written, 278 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.015, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 5954 texture pages copied (23 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 7 textures retired (1211 left, 2413 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: texture #24832 0x10008612200 885x885 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20275 (drawn at 20213)
[hos host] vk: program picks: 22420382 kept (9407784 by the last picks' textures, a uniform buffer bound since), 26538 new, 8153079 with other textures (of 30600000)
[hos host] vk: 3 textures retired (1212 left, 2416 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: a frame: 253 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.91, present 0.20, draw 6.83, descriptors 3.03, texture lookups 0.23, uploads 0.30, pipelines 0.28, programs 1.19, commands 1.80, dispatch 1.03, memory sync 0.54, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 512 (1045 MB): 5.57 MB synced in 1077.6 copies, 18.47 MB written back
[hos host] vk: 20280 presents, 59.9 fps; a frame: 1815 draws (0 skipped), 183 dispatches, 165 passes, 145 barriers, 6 submits, 0.8 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1212 textures (2416 MB)
[hos host] vk: host time in GPU resource waits: 1.46 ms a frame (fences 1.18, timelines 0.28, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 667.0 full barriers recorded, 299.6 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 38.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.7, an op 932.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.0 ms); 0.9 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.5 (1.3 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.56 MB unswizzled on the CPU in 0.26 ms
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1150 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 11 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1056 kept their virtual core (nothing there to wait for), 2784 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos BaseProcExecutor1] vcores: 154079 takes, 22438 waits (mean 71 us, longest 7.6 ms, 0 over HOS_VCORE_LOG_MS), 10871 kicks, 10833 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 2 textures retired (1218 left, 2412 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15900000 (9791960 pages from the captures)
[hos host] vk: sparse mirrors: 115, 1045 MB with memory of 1496 MB (450 MB spared); blocks given memory later 13774, given back 17083; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1340 binds (312.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 347 taken from the spares, 165 made new; 397 kept as spares, 0 destroyed; 50 spares now, 1324 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 473 dirty pages from their submits' captures, 57027 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 43820 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8754 runs, 743.6 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20583 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #285000: event 0x100d6c803c mode 0 value 284999 (lr 0x80661bfc)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: a submission's parts ended: 98860 where the game had ended its render pass, 21140 at 512 draws
[hos host] vk: 8 textures retired (1219 left, 2402 MB): budget 0, drawn over 8, dead targets 0
[hos host] vk: 5 textures retired (1221 left, 2406 MB): budget 0, drawn over 5, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 15920000 (9801987 pages from the captures)
[hos host] vk: 7 textures retired (1219 left, 2400 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: texture #20276 0x116fb36000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 20321 (drawn at 14859)
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 665 dirty pages from their submits' captures, 57417 from earlier ones', 12 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1000 dispatches, 44083 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8683 runs, 729.4 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21155 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 12;
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #390000: event 0x100d6c800c mode 0 value 285899 (now 285880) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 15940000 (9811727 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.05 over 576 ranges, 3540 MB), 277 pages written, 277 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.033; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.022, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 4162 texture pages copied (16 MB)
[hos host] vk: presenter thread: the copy's fence 1.33 ms, the present 0.73 ms a present
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10276 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 117, 1050 MB with memory of 1504 MB (453 MB spared); blocks given memory later 13774, given back 17084; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1343 binds (313.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 349 taken from the spares, 165 made new; 397 kept as spares, 0 destroyed; 48 spares now, 1316 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 110 dirty pages from their submits' captures, 60353 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1017 dispatches, 45740 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9185 runs, 770.3 MB compared; 56 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22467 in captures, 3 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 3;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 15960000 (9820548 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 22568243 kept (9471445 by the last picks' textures, a uniform buffer bound since), 26538 new, 8205218 with other textures (of 30800000)
[hos host] vk: 2 textures retired (1168 left, 2224 MB): budget 0, drawn over 0, dead targets 2
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos BaseProcExecutor1] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 15888477 lookups, 129796 missed (0.82 %)
[hos host] vk: memory: working set 14733 MB (peak 16021), private 21447 MB, staging 80 MB in 5 arenas, 47 made so far (212 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 858 MB; used 3216 of 15579 MB; heap 1 (device): 157 blocks 4759 MB, 18934 allocations 4056 MB; used 6882 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 22 blocks 704 MB, 159 images 390 MB; 95 free ranges, the largest 31 MB
[hos host] vk: image blocks (textures, compacted): 43 blocks 1376 MB, 1019 images 1278 MB; 321 free ranges, the largest 24 MB
[hos host] vk: render target images: 6493 reused from the pool, 5216 made new, 5055 destroyed from it, 1309 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 166 missed after their image was taken by size; it holds 6 (51 MB)
[hos host] vk: render targets not reused by size: 2467 with the pool empty (0 of them with targets still in the dead queue), 380 with none in it a present yet, 59 with none of a kind they can use, 651 with all smaller (the largest 50% short on average), 705 with all bigger (over 1.5 times), 951 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4863 idle
[hos host] vk: VRAM: 6882 MB used of 11474 MB allowed (703 MB of it free in the blocks); textures 2253 MB (budget 7158 MB), 1178 textures
[hos host] vk: budget: VRAM 4759 of 11474 MB (6882 MB by the driver, 2123 MB unaccounted): textures 2253, mirrors 1046, blocks free 703, shadows 284, other buffers 229, other images 111, staging 80, target pool 51; RAM 3161 MB (the process's private memory 21447 MB): imports 2260, capture spares 750, captures 96, capture table 35, other buffers 12, retired imports 8, others 0
[hos host] vk: textures: decoded ASTC 1179 MB, rendered 552 MB (77 drawn in one frame, kept: 49159 KB), other 471 MB; used in the last 8 presents 1644 MB
[hos host] vk: decoded ASTC used again: 577 textures (1277 MB) copied from ones the game finalized, not decoded; 577 of those retired, 4 listed; 2411 MB of blocks hashed in 397 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 15253 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34, last used 230 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 82 (15946 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 136 waited for it, 1.21 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 24.7 ms; VRAM 6.7 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6476 queued, 6476 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16362 ms and 348.1 MB read in all); 8781 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12295 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2395 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9632 (6476 ready, 0 loading), 17569 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13980 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 291 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.88, present 0.22, draw 6.81, descriptors 3.10, texture lookups 0.34, uploads 0.30, pipelines 0.28, programs 1.11, commands 1.80, dispatch 0.99, memory sync 0.55, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 514 (1046 MB): 5.62 MB synced in 1082.2 copies, 18.29 MB written back
[hos host] vk: 20400 presents, 60.1 fps; a frame: 1854 draws (0 skipped), 176 dispatches, 166 passes, 144 barriers, 6 submits, 0.7 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1179 textures (2256 MB)
[hos host] vk: host time in GPU resource waits: 1.68 ms a frame (fences 1.32, timelines 0.36, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 660.9 full barriers recorded, 298.8 left out (nothing recorded since the one before), 13.2 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.1, an op 930.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.7 ms a frame (the longest submission 3.0 ms); 1.2 textures made a frame: colour targets 0.4 (1.9 MB), depth 0.7 (2.3 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.62 MB unswizzled on the CPU in 0.26 ms
[hos host] vk: the render thread (600 presents, 16.66 ms a frame): its jobs 14.50 ms a frame (submits 8.07, presents 3.98, others 2.46), idle 2.16; the game waited for room at 17 presents, 0.18 ms each (0.01 a frame; 7.25 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 142 captures at a CPU signal; the submits' waits read in 0.041 ms a frame; readbacks: 1006 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 13.62 ms after its hand-over (18.67 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.31 + 0.31 (the window's copy is in the idle), after it 1.08 + 0.35; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.57 parts a frame recorded in the signal's job, 79.9% of them took nothing that came with the signal; of the others, 1.72 a frame took pages of the signal's capture into a mirror (25.8 pages a frame) and 0.05 pages read as guest memory had them (0.7 pages a frame); a frame's first part took nothing in 18 of 142 frames, its first three in 18
[hos host] vk: frame ahead, what came with the signal: 142 signals with a capture, each of 65.7 pages in 8.2 pools; between a frame's last submit and that signal 0.01 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.13+0.06, 0.44+0.14, 0.04+0.02, 0.06+0.02, 0.12+0.03, 0.16+0.03, 0.08+0.02, 0.05+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: replay reads of guest memory as submitted: 15980000 (9830084 pages from the captures)
[hos host] vk: render targets drawn over from another start, let go: 12228 so far (57365 MB), 11040 of them (51903 MB) between the passes
[hos host] vk: 2 textures retired (1184 left, 2270 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 6692530 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1001, 4442.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 217 dirty pages from their submits' captures, 57275 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1021 dispatches, 43817 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9070 runs, 752.0 MB compared; 53 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21633 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2;
[hos host] vk: 2 textures retired (1189 left, 2274 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 16000000 (9840143 pages from the captures)
[hos host] vk: reaper: 703 images (2217 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 86 (142 MB)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1160 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1164 kept their virtual core (nothing there to wait for), 2932 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos ModuleSystemWorker1] nvn: DrawMeshTasksIndirect: 700000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040)
[hos host] vk: sparse mirrors: 116, 1046 MB with memory of 1500 MB (453 MB spared); blocks given memory later 13774, given back 17123; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1346 binds (313.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 349 taken from the spares, 165 made new; 398 kept as spares, 0 destroyed; 49 spares now, 1320 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 59660 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1035 dispatches, 45414 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8818 runs, 767.6 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21630 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 3;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 570 ranges, 3528 MB), 281 pages written, 281 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.015; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 6665 texture pages copied (26 MB)
[hos host] vk: replay reads of guest memory as submitted: 16020000 (9849568 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13170 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] nvn: GPU waits in 5 s: 48 sync waits (70 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 22716214 kept (9536998 by the last picks' textures, a uniform buffer bound since), 26545 new, 8257240 with other textures (of 31000000)
[hos host] vk: replay reads of guest memory as submitted: 16040000 (9858899 pages from the captures)
[hos Presentation Thread] vi: 20495 frames presented (60.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 17, 15.5-17.8 269, 17.8-22 15, 22-30 0, 30-40 0, over 40 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 163 dirty pages from their submits' captures, 58994 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1010 dispatches, 44988 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8769 runs, 770.5 MB compared; 70 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21866 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; 0x7fa9000000 (4 MB, captured, state 2): 1;
[hos host] vk: replay reads of guest memory as submitted: 16060000 (9868323 pages from the captures)
[hos host] vk: a frame: 253 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.66, present 0.22, draw 5.94, descriptors 2.61, texture lookups 0.32, uploads 0.12, pipelines 0.25, programs 1.01, commands 1.60, dispatch 0.90, memory sync 0.51, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 514 (1046 MB): 5.55 MB synced in 1077.0 copies, 18.33 MB written back
[hos host] vk: 20520 presents, 59.5 fps; a frame: 1799 draws (0 skipped), 168 dispatches, 167 passes, 144 barriers, 6 submits, 0.3 uploads (0.5 MB), 81 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1208 textures (2296 MB)
[hos host] vk: host time in GPU resource waits: 1.92 ms a frame (fences 1.44, timelines 0.48, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 656.1 full barriers recorded, 299.2 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.3, an op 923.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.8 ms a frame (the longest submission 5.1 ms); 1.1 textures made a frame: colour targets 0.3 (1.2 MB), depth 0.7 (2.1 MB), ASTC or BC 0.1 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.54 MB unswizzled on the CPU in 0.10 ms
[hos host] vk: sparse mirrors: 116, 1046 MB with memory of 1500 MB (453 MB spared); blocks given memory later 13774, given back 17127; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1348 binds (314.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 349 taken from the spares, 165 made new; 398 kept as spares, 0 destroyed; 49 spares now, 1320 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1169 dirty pages from their submits' captures, 53943 from earlier ones', 4 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 956 dispatches, 41809 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8223 runs, 715.7 MB compared; 88 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19892 in captures, 2 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa9000000 (4 MB, captured, state 2): 4;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #395000: event 0x100d6c800c mode 0 value 289699 (now 289680) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 16080000 (9878570 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.18 ms (lock 0.00, scan 0.05 over 568 ranges, 3566 MB), 309 pages written, 309 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.030; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 25 kept (114 MB), 10 spare buffers, 6452 texture pages copied (25 MB)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #290000: event 0x100d6c8048 mode 0 value 289999 (lr 0x80666598)
[hos host] vk: replay reads of guest memory as submitted: 16100000 (9889211 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 231 dirty pages from their submits' captures, 58426 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1015 dispatches, 44291 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8518 runs, 748.9 MB compared; 58 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21566 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1;
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11012 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos rmbl::VibrationThread] vcores: 154113 takes, 24103 waits (mean 77 us, longest 9.4 ms, 0 over HOS_VCORE_LOG_MS), 11632 kicks, 11559 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: late pages found at the replay (600 takes): 155 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 22862211 kept (9603557 by the last picks' textures, a uniform buffer bound since), 26549 new, 8311239 with other textures (of 31200000)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: replay reads of guest memory as submitted: 16120000 (9898877 pages from the captures)
[hos sead::MainThread] nvn: 2700000 recordings made, 74154 live copies (24.8 MB), 6604 texture copies; 3 command buffers' segments (5.1 MB)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1116 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1123 kept their virtual core (nothing there to wait for), 2973 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: readback pages refreshed from guest memory: 6760273 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 118, 1045 MB with memory of 1508 MB (462 MB spared); blocks given memory later 13774, given back 17142; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1353 binds (315.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 351 taken from the spares, 165 made new; 398 kept as spares, 0 destroyed; 47 spares now, 1312 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 253 dirty pages from their submits' captures, 57246 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 996 dispatches, 44203 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8654 runs, 761.3 MB compared; 67 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21028 in captures, 2 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 16140000 (9907662 pages from the captures)
[hos host] vk: a frame: 229 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.81, present 0.19, draw 6.16, descriptors 2.77, texture lookups 0.23, uploads 0.18, pipelines 0.25, programs 1.03, commands 1.63, dispatch 0.90, memory sync 0.50, queue submit 0.00, slot wait 0.22
[hos host] vk: a frame: mirrors 516 (1045 MB): 5.47 MB synced in 1063.9 copies, 18.19 MB written back
[hos host] vk: 20640 presents, 60.5 fps; a frame: 1759 draws (0 skipped), 165 dispatches, 167 passes, 144 barriers, 6 submits, 0.5 uploads (1.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1209 textures (2270 MB)
[hos host] vk: host time in GPU resource waits: 1.48 ms a frame (fences 1.20, timelines 0.29, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 655.9 full barriers recorded, 299.9 left out (nothing recorded since the one before), 13.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.0, an op 850.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.1 ms); 1.0 textures made a frame: colour targets 0.2 (1.5 MB), depth 0.6 (2.1 MB), ASTC or BC 0.1 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.03 MB unswizzled on the CPU in 0.17 ms
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 563 ranges, 3530 MB), 270 pages written, 270 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (162 MB), 9 spare buffers, 8065 texture pages copied (31 MB)
[hos host] vk: replay reads of guest memory as submitted: 16160000 (9917325 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 195 dirty pages from their submits' captures, 58028 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1018 dispatches, 44090 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8656 runs, 728.1 MB compared; 58 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21721 in captures, 4 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1;
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14894 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 16180000 (9927664 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: program picks: 23008893 kept (9666025 by the last picks' textures, a uniform buffer bound since), 26567 new, 8364539 with other textures (of 31400000)
[hos host] vk: sparse mirrors: 118, 1045 MB with memory of 1508 MB (463 MB spared); blocks given memory later 13774, given back 17157; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1355 binds (315.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 351 taken from the spares, 165 made new; 398 kept as spares, 0 destroyed; 47 spares now, 1312 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 137 dirty pages from their submits' captures, 58027 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 976 dispatches, 43588 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9748 runs, 759.8 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22669 in captures, 2 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 3;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 16200000 (9937180 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 227 dirty pages from their submits' captures, 57256 from earlier ones', 10 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 986 dispatches, 43744 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9052 runs, 746.9 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21764 in captures, 1 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 10;
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #400000: event 0x100d6c800c mode 0 value 293499 (now 293480) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 16220000 (9946072 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 564 ranges, 3530 MB), 280 pages written, 280 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.015, bytes copied 0.000; 25 kept (108 MB), 10 spare buffers, 5013 texture pages copied (19 MB)
[hos host] vk: a frame: 280 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.74, present 0.17, draw 6.03, descriptors 2.66, texture lookups 0.15, uploads 0.22, pipelines 0.26, programs 1.00, commands 1.62, dispatch 0.95, memory sync 0.47, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 517 (1042 MB): 5.66 MB synced in 1092.5 copies, 18.63 MB written back
[hos host] vk: 20760 presents, 60.0 fps; a frame: 1794 draws (0 skipped), 185 dispatches, 166 passes, 144 barriers, 6 submits, 0.5 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1269 textures (2428 MB)
[hos host] vk: host time in GPU resource waits: 1.38 ms a frame (fences 1.35, timelines 0.03, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 672.6 full barriers recorded, 300.3 left out (nothing recorded since the one before), 13.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.5, an op 812.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.5 ms a frame (the longest submission 3.6 ms); 0.9 textures made a frame: colour targets 0.1 (0.9 MB), depth 0.7 (2.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.44 MB unswizzled on the CPU in 0.22 ms
[hos sead::MainThread] nvn: GPU waits in 5 s: 29 sync waits (84 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 400 mirrors given back (2601 MB in all); 1040 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1140 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1195 kept their virtual core (nothing there to wait for), 3157 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 16240000 (9955329 pages from the captures)
[hos host] vk: texture #25259 0x10008612200 843x843 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20777 (drawn at 20716)
[hos host] vk: 1 textures retired (1289 left, 2528 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 117, 1040 MB with memory of 1500 MB (459 MB spared); blocks given memory later 13774, given back 17159; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1360 binds (316.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 352 taken from the spares, 165 made new; 400 kept as spares, 0 destroyed; 48 spares now, 1320 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 190 dirty pages from their submits' captures, 58337 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1023 dispatches, 44618 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8525 runs, 756.3 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21508 in captures, 4 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; 0x7fa9000000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: queues ordered by their events: 138000 submissions stopped at a wait, 137936 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 401012, not met 0; captures at a CPU event's wait 12896, 0 of them at a wait met on arrival
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15032 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: texture #25323 0x10008612200 822x822 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20790 (drawn at 20728)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05
[hos Presentation Thread] vi: 20796 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 21, 15.5-17.8 270, 17.8-22 8, 22-30 2, 30-40 0, over 40 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 199 textures retired (1111 left, 2383 MB): budget 172, drawn over 27, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 6 MB
[hos host] vk: replay reads of guest memory as submitted: 16260000 (9964133 pages from the captures)
[hos host] vk: program picks: 23156325 kept (9731044 by the last picks' textures, a uniform buffer bound since), 26567 new, 8417107 with other textures (of 31600000)
[hos host] vk: 46 textures retired (1075 left, 2364 MB): budget 37, drawn over 9, dead targets 0
[hos host] vk: texture #25291 0x10008612200 811x811 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20815 (drawn at 20753)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #295000: event 0x100d6c8010 mode 0 value 295001 (lr 0x80660bcc)
[hos host] vk: 9 textures retired (1057 left, 2226 MB): budget 0, drawn over 9, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 6827562 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 175 dirty pages from their submits' captures, 58259 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1029 dispatches, 44496 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8569 runs, 756.1 MB compared; 59 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21281 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1;
[hos host] vk: 10 textures retired (1056 left, 2232 MB): budget 0, drawn over 10, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 16280000 (9972653 pages from the captures)
[hos host] vk: 8 textures retired (1055 left, 2216 MB): budget 0, drawn over 8, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 564 ranges, 3480 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.019, bytes copied 0.000; 24 kept (100 MB), 11 spare buffers, 8046 texture pages copied (31 MB)
[hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 720000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040)
[hos host] vk: 2 textures retired (1062 left, 2239 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: recorder: 5650235 packets run in 10.0 s (busy 1.68 s, spun idle 0.12 s, slept 10557 times, 3178 of them under 200 us), 0 calls made directly (no room); 17963 batches over 600 frames (29.9 a frame, 315 packets a batch), 8767 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 133 times, 17.73 ms (0.2 times, 0.03 ms a frame); every thread 719 times, 52.69 ms in all, the longest 0.56 ms; vkWaitForFences 505/29.97 ms (max 0.56), vkDestroyImageView 154/19.26 ms (max 0.44), vmaFreeMemory 18/1.57 ms (max 0.53), vkQueueBindSparse 24/1.40 ms (max 0.38), vmaDestroyBuffer 12/0.22 ms (max 0.05), the queue's lock 2/0.16 ms (max 0.13)
[hos host] vk: recorder: of its busy 1.68 s, 0.37 s in parts' ends and submits (the longest 1.32 ms), 0.05 s of that blocked on the queue's lock (107 times, the longest 1.28 ms), so 0.32 s in the ends, submits and holds; 1.31 s running the other packets; 9417 packets a frame
[hos rmbl::VibrationThread] hid: 69632 vibration values, strongest 0.02 since the last count
[hos host] vk: 17 textures retired (1053 left, 2239 MB): budget 13, drawn over 4, dead targets 0
[hos host] vk: sparse mirrors: 118, 1043 MB with memory of 1508 MB (464 MB spared); blocks given memory later 13778, given back 17167; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1366 binds (317.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 353 taken from the spares, 165 made new; 400 kept as spares, 0 destroyed; 47 spares now, 1312 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: texture #21167 0x117144e000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 20866 (drawn at 16210)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 178 dirty pages from their submits' captures, 58894 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1031 dispatches, 44856 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8688 runs, 757.0 MB compared; 54 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21553 in captures, 3 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 16300000 (9981543 pages from the captures)
[hos host] vk: 9 textures retired (1052 left, 2207 MB): budget 0, drawn over 9, dead targets 0
[hos host] vk: a frame: 240 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.54, present 0.20, draw 5.73, descriptors 2.52, texture lookups 0.40, uploads 0.18, pipelines 0.24, programs 0.96, commands 1.56, dispatch 0.89, memory sync 0.49, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 518 (1043 MB): 5.47 MB synced in 1057.9 copies, 18.01 MB written back
[hos host] vk: 20880 presents, 60.0 fps; a frame: 1720 draws (0 skipped), 164 dispatches, 164 passes, 144 barriers, 6 submits, 0.4 uploads (1.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1071 textures (2251 MB)
[hos host] vk: host time in GPU resource waits: 1.68 ms a frame (fences 1.47, timelines 0.21, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 640.8 full barriers recorded, 296.3 left out (nothing recorded since the one before), 13.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.8 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.9, an op 816.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.1 ms a frame (the longest submission 4.1 ms); 1.2 textures made a frame: colour targets 0.2 (1.2 MB), depth 0.9 (4.9 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.16 MB unswizzled on the CPU in 0.18 ms
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6656 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos NavMeshCAStepDepWorker1] vcores: 152759 takes, 23202 waits (mean 69 us, longest 7.1 ms, 0 over HOS_VCORE_LOG_MS), 11116 kicks, 10990 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 15 textures retired (1061 left, 2189 MB): budget 0, drawn over 15, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 16320000 (9991066 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 283 dirty pages from their submits' captures, 55335 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 964 dispatches, 42312 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8372 runs, 721.5 MB compared; 50 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20489 in captures, 1 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x1043000000 (4 MB, captured, state 2): 3;
[hos host] vk: 14 textures retired (1069 left, 2195 MB): budget 0, drawn over 14, dead targets 0
[hos host] vk: program picks: 23302135 kept (9796634 by the last picks' textures, a uniform buffer bound since), 26569 new, 8471295 with other textures (of 31800000)
[hos host] vk: replay reads of guest memory as submitted: 16340000 (9999977 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1140 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1148 kept their virtual core (nothing there to wait for), 2948 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #405000: event 0x100d6c800c mode 0 value 297299 (now 297280) (lr 0x80660e38)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 562 ranges, 3395 MB), 276 pages written, 276 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (150 MB), 10 spare buffers, 4312 texture pages copied (16 MB)
[hos host] vk: sparse mirrors: 117, 1041 MB with memory of 1496 MB (454 MB spared); blocks given memory later 13778, given back 17167; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1367 binds (318.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 353 taken from the spares, 165 made new; 401 kept as spares, 0 destroyed; 48 spares now, 1324 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 187 dirty pages from their submits' captures, 56443 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 970 dispatches, 42926 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8351 runs, 721.1 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20310 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 16360000 (10009149 pages from the captures)
[hos host] vk: 22 textures retired (1087 left, 2201 MB): budget 0, drawn over 22, dead targets 0
[hos host] vk: 8 textures retired (1087 left, 2193 MB): budget 0, drawn over 8, dead targets 0
[hos host] vk: presenter thread: the copy's fence 1.35 ms, the present 0.65 ms a present
[hos host] vk: replay reads of guest memory as submitted: 16380000 (10019497 pages from the captures)
[hos host] vk: 13 textures retired (1086 left, 2192 MB): budget 0, drawn over 13, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1001, 4442.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 302 dirty pages from their submits' captures, 56648 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1006 dispatches, 43240 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8568 runs, 725.2 MB compared; 46 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21443 in captures, 3 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 7 textures retired (1083 left, 2180 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 18676 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 15527561 lookups, 126666 missed (0.82 %)
[hos host] vk: memory: working set 14750 MB (peak 16021), private 21721 MB, staging 80 MB in 5 arenas, 47 made so far (212 ms)
[hos host] vk: memory heaps: heap 0 (system): 14 blocks 944 MB, 34 allocations 856 MB; used 3232 of 15579 MB; heap 1 (device): 160 blocks 4855 MB, 19536 allocations 4016 MB; used 6978 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 24 blocks 768 MB, 193 images 504 MB; 124 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 44 blocks 1408 MB, 902 images 1082 MB; 370 free ranges, the largest 23 MB
[hos host] vk: render target images: 6549 reused from the pool, 5581 made new, 5386 destroyed from it, 1386 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 172 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 2717 with the pool empty (0 of them with targets still in the dead queue), 394 with none in it a present yet, 59 with none of a kind they can use, 687 with all smaller (the largest 50% short on average), 753 with all bigger (over 1.5 times), 968 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 5194 idle
[hos host] vk: VRAM: 6978 MB used of 11474 MB allowed (839 MB of it free in the blocks); textures 2221 MB (budget 7158 MB), 1101 textures
[hos host] vk: budget: VRAM 4855 of 11474 MB (6978 MB by the driver, 2123 MB unaccounted): textures 2221, mirrors 1089, blocks free 839, shadows 284, other buffers 229, other images 111, staging 80; RAM 3167 MB (the process's private memory 21721 MB): imports 2260, capture spares 746, captures 118, capture table 35, retired imports 8, others 0
[hos host] vk: textures: decoded ASTC 1081 MB, rendered 687 MB (77 drawn in one frame, kept: 54095 KB), other 385 MB; used in the last 8 presents 1716 MB
[hos host] vk: decoded ASTC used again: 581 textures (1288 MB) copied from ones the game finalized, not decoded; 581 of those retired, 1 listed; 2424 MB of blocks hashed in 399 ms
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 15853 presents ago
[hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 446 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 87 (15377 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 111 waited for it, 1.30 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 26.7 ms; VRAM 6.8 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6476 queued, 6476 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16362 ms and 348.1 MB read in all); 8781 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12295 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2407 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9632 (6476 ready, 0 loading), 17567 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 13980 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 216 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.77, present 0.19, draw 6.12, descriptors 2.59, texture lookups 0.19, uploads 0.02, pipelines 0.27, programs 1.06, commands 1.72, dispatch 0.90, memory sync 0.48, queue submit 0.00, slot wait 0.24
[hos host] vk: a frame: mirrors 519 (1089 MB): 5.90 MB synced in 1068.5 copies, 18.10 MB written back
[hos host] vk: 21000 presents, 60.0 fps; a frame: 1828 draws (0 skipped), 169 dispatches, 166 passes, 144 barriers, 6 submits, 0.3 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1102 textures (2226 MB)
[hos host] vk: host time in GPU resource waits: 1.70 ms a frame (fences 1.49, timelines 0.21, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 651.2 full barriers recorded, 297.9 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.0, an op 895.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.4 ms a frame (the longest submission 3.7 ms); 0.9 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.6 (2.1 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.07 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 13.90 ms a frame (submits 8.10, presents 3.99, others 1.82), idle 2.76; the game waited for room at 7 presents, 0.57 ms each (0.01 a frame; 4.32 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 114 captures at a CPU signal; the submits' waits read in 0.042 ms a frame; readbacks: 704 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 13.02 ms after its hand-over (25.09 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.71 + 0.30 (the window's copy is in the idle), after it 0.81 + 0.35; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.54 parts a frame recorded in the signal's job, 80.1% of them took nothing that came with the signal; of the others, 1.70 a frame took pages of the signal's capture into a mirror (22.8 pages a frame) and 0.02 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 18 of 114 frames, its first three in 18
[hos host] vk: frame ahead, what came with the signal: 114 signals with a capture, each of 37.5 pages in 7.9 pools; between a frame's last submit and that signal 0.03 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (2 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.14+0.06, 0.35+0.13, 0.03+0.02, 0.07+0.03, 0.05+0.03, 0.07+0.03, 0.06+0.02, 0.04+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: replay reads of guest memory as submitted: 16400000 (10028853 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vi: real-time step on: ratio 1.122, game's frame rate 60
[hos host] vk: render targets drawn over from another start, let go: 12537 so far (58776 MB), 11348 of them (53314 MB) between the passes
[hos host] vk: program picks: 23448414 kept (9864410 by the last picks' textures, a uniform buffer bound since), 26573 new, 8525012 with other textures (of 32000000)
[hos host] vk: readback pages refreshed from guest memory: 6895265 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 16420000 (10040469 pages from the captures)
[hos host] vk: sparse mirrors: 119, 1090 MB with memory of 1560 MB (469 MB spared); blocks given memory later 13784, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1371 binds (318.6 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 355 taken from the spares, 165 made new; 401 kept as spares, 0 destroyed; 46 spares now, 1260 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 595 dirty pages from their submits' captures, 57557 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1047 dispatches, 44163 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9056 runs, 749.7 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21639 in captures, 3 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa6800000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos sead::MainThread] nvn: 2750000 recordings made, 74295 live copies (24.7 MB), 6657 texture copies; 6 command buffers' segments (4.3 MB)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 560 ranges, 3395 MB), 290 pages written, 290 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.020, bytes copied 0.000; 25 kept (104 MB), 10 spare buffers, 7312 texture pages copied (28 MB)
[hos host] vk: 29 textures retired (1117 left, 2231 MB): budget 0, drawn over 29, dead targets 0
[hos host] vk: reaper: 565 images (1933 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 152 (281 MB)
[hos host] vk: 10 textures retired (1114 left, 2215 MB): budget 0, drawn over 10, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 16440000 (10050102 pages from the captures)
[hos host] vk: 7 textures retired (1113 left, 2210 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 101 dirty pages from their submits' captures, 56860 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1026 dispatches, 43334 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8659 runs, 727.1 MB compared; 48 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21362 in captures, 2 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 8 textures retired (1112 left, 2206 MB): budget 0, drawn over 8, dead targets 0
[hos sead::MainThread] nvn: GPU waits in 5 s: 26 sync waits (40 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: replay reads of guest memory as submitted: 16460000 (10061711 pages from the captures)
[hos host] vk: 9 textures retired (1121 left, 2211 MB): budget 2, drawn over 7, dead targets 0
[hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #300000: event 0x100d6c8018 mode 0 value 300000 (lr 0x802915c8)
[hos host] vk: 3 textures retired (1130 left, 2273 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos Presentation Thread] vi: 21096 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 25, 15.5-17.8 259, 17.8-22 13, 22-30 3, 30-40 0, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 16480000 (10072325 pages from the captures)
[hos host] vk: 10 textures retired (1136 left, 2338 MB): budget 3, drawn over 7, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12055 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1128 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 9 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1104 kept their virtual core (nothing there to wait for), 2992 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: sparse mirrors: 120, 1091 MB with memory of 1564 MB (472 MB spared); blocks given memory later 13784, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1372 binds (318.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 356 taken from the spares, 165 made new; 401 kept as spares, 0 destroyed; 45 spares now, 1256 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 215 dirty pages from their submits' captures, 56560 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 959 dispatches, 43155 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8438 runs, 708.6 MB compared; 52 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21151 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: 15 textures retired (1130 left, 2379 MB): budget 7, drawn over 8, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture #25562 0x10008612200 933x933 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 21113 (drawn at 21051)
[hos host] vk: program picks: 23594927 kept (9933142 by the last picks' textures, a uniform buffer bound since), 26582 new, 8578490 with other textures (of 32200000)
[hos host] vk: 11 textures retired (1110 left, 2238 MB): budget 3, drawn over 8, dead targets 0
[hos ModuleSystemWorker3] barrier: 324307 looks: 127699 found the masks moved already, 193538 waited until they moved, 3070 for the whole 1 ms (a wait 360 us on average); 292555 wakes
[hos ModuleSystemWorker5] barrier: 324307 looks: 127699 found the masks moved already, 193538 waited until they moved, 3070 for the whole 1 ms (a wait 360 us on average); 292555 wakes
[hos host] vk: a frame: 226 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.31, present 0.20, draw 6.36, descriptors 2.75, texture lookups 0.33, uploads 0.08, pipelines 0.26, programs 1.11, commands 1.74, dispatch 0.93, memory sync 0.48, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 521 (1091 MB): 5.60 MB synced in 1084.3 copies, 18.21 MB written back
[hos host] vk: 21120 presents, 60.0 fps; a frame: 1887 draws (0 skipped), 176 dispatches, 167 passes, 144 barriers, 6 submits, 0.4 uploads (0.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14189 pipelines (3740 ms), 0 programs (10 failed), 1110 textures (2238 MB)
[hos host] vk: host time in GPU resource waits: 1.79 ms a frame (fences 1.39, timelines 0.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 662.5 full barriers recorded, 298.7 left out (nothing recorded since the one before), 13.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.9 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.9, an op 990.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.9 ms a frame (the longest submission 3.7 ms); 1.2 textures made a frame: colour targets 0.1 (0.4 MB), depth 0.8 (3.7 MB), ASTC or BC 0.3 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.39 MB unswizzled on the CPU in 0.05 ms
[hos host] vk: replay reads of guest memory as submitted: 16500000 (10082851 pages from the captures)
[hos host] vk: texture #25681 0x10008612200 937x937 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 21126 (drawn at 21065)
[hos host] vk: 15 textures retired (1111 left, 2267 MB): budget 4, drawn over 6, dead targets 5
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 559 ranges, 3393 MB), 279 pages written, 279 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (156 MB), 7 spare buffers, 8939 texture pages copied (34 MB)
[hos host] vk: 13 textures retired (1110 left, 2239 MB): budget 2, drawn over 11, dead targets 0
[hos sead::MainThread] nvn: CommandBufferWaitEvent #410000: event 0x100d6c800c mode 0 value 301099 (now 301080) (lr 0x80660e38)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 193 dirty pages from their submits' captures, 56516 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 937 dispatches, 42830 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8792 runs, 730.4 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20813 in captures, 1 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 16520000 (10094154 pages from the captures)
[hos host] vk: 7 textures retired (1111 left, 2225 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: 6 textures retired (1113 left, 2215 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: 8 textures retired (1114 left, 2215 MB): budget 4, drawn over 4, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 16540000 (10105063 pages from the captures)
[hos host] vk: 7 textures retired (1116 left, 2208 MB): budget 0, drawn over 7, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 120, 1091 MB with memory of 1564 MB (472 MB spared); blocks given memory later 13788, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1375 binds (318.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 356 taken from the spares, 165 made new; 401 kept as spares, 0 destroyed; 45 spares now, 1256 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 111 dirty pages from their submits' captures, 57296 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 914 dispatches, 43293 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8949 runs, 720.4 MB compared; 51 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21168 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool: 0x7fff800000 (8 MB, captured, state 2): 2;
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos ModuleSystemWorker1] vcores: 157468 takes, 27074 waits (mean 74 us, longest 28.1 ms, 0 over HOS_VCORE_LOG_MS), 12886 kicks, 12745 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: 11 textures retired (1109 left, 2179 MB): budget 3, drawn over 8, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: texture #21020 0x11708f8000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 21197 (drawn at 15802)
[hos host] vk: 7 textures retired (1114 left, 2180 MB): budget 1, drawn over 6, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 16560000 (10115140 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13225 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 10 textures retired (1116 left, 2186 MB): budget 0, drawn over 10, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 6962371 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 418 dirty pages from their submits' captures, 55819 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 42368 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8621 runs, 700.5 MB compared; 51 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20654 in captures, 4 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: program picks: 23737222 kept (9997105 by the last picks' textures, a uniform buffer bound since), 26584 new, 8636193 with other textures (of 32400000)
[hos host] vk: replay reads of guest memory as submitted: 16580000 (10126875 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.05 over 556 ranges, 3222 MB), 293 pages written, 293 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (104 MB), 10 spare buffers, 4675 texture pages copied (18 MB)
[hos host] vk: a frame: 199 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.12, present 0.19, draw 6.11, descriptors 2.71, texture lookups 0.30, uploads 0.16, pipelines 0.26, programs 1.03, commands 1.64, dispatch 0.94, memory sync 0.51, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 521 (1091 MB): 5.75 MB synced in 1107.5 copies, 18.41 MB written back
[hos host] vk: 21240 presents, 60.0 fps; a frame: 1729 draws (0 skipped), 182 dispatches, 164 passes, 144 barriers, 6 submits, 0.5 uploads (0.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14189 pipelines (3740 ms), 0 programs (10 failed), 1136 textures (2256 MB)
[hos host] vk: host time in GPU resource waits: 1.75 ms a frame (fences 1.62, timelines 0.12, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 660.8 full barriers recorded, 297.1 left out (nothing recorded since the one before), 13.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.5, an op 908.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.3 ms a frame (the longest submission 4.7 ms); 1.1 textures made a frame: colour targets 0.1 (0.5 MB), depth 0.7 (2.8 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.87 MB unswizzled on the CPU in 0.12 ms
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 740000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040)
[hos host] vk: replay reads of guest memory as submitted: 16600000 (10138535 pages from the captures)
[hos host] vk: sparse mirrors: 119, 1091 MB with memory of 1560 MB (469 MB spared); blocks given memory later 13790, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1377 binds (318.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 356 taken from the spares, 165 made new; 402 kept as spares, 0 destroyed; 46 spares now, 1260 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 370 dirty pages from their submits' captures, 55726 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 907 dispatches, 42373 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8644 runs, 699.3 MB compared; 57 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20772 in captures, 2 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1134 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1072 kept their virtual core (nothing there to wait for), 2768 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 16620000 (10150104 pages from the captures)
[hos host] vk: 5 textures retired (1102 left, 2096 MB): budget 0, drawn over 5, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: 9 textures retired (1104 left, 2126 MB): budget 7, drawn over 2, dead targets 0
[hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 805 dirty pages from their submits' captures, 56711 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 919 dispatches, 43580 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8543 runs, 712.9 MB compared; 52 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21179 in captures, 0 only in the collection, 39 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: pipeline cache: play file 10 saved (3 KB): 1 pipelines, 1 from the driver's cache, 0 compiled; 58 more from the start's cache
[hos host] vk: replay reads of guest memory as submitted: 16640000 (10161389 pages from the captures)
[hos host] vk: 2 textures retired (1117 left, 2177 MB): budget 0, drawn over 0, dead targets 2
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21767 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.04 over 555 ranges, 3232 MB), 295 pages written, 295 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.034; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.026, bytes copied 0.000; 25 kept (124 MB), 10 spare buffers, 9660 texture pages copied (37 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 2 textures retired (1137 left, 2368 MB): budget 0, drawn over 0, dead targets 2
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #415000: event 0x100d6c800c mode 0 value 304899 (now 304880) (lr 0x80660e38)
[hos host] vk: sparse mirrors: 118, 1090 MB with memory of 1556 MB (465 MB spared); blocks given memory later 13790, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1378 binds (318.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 356 taken from the spares, 165 made new; 403 kept as spares, 0 destroyed; 47 spares now, 1264 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1100, 4884.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 174 dirty pages from their submits' captures, 61202 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 990 dispatches, 46942 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8462 runs, 784.0 MB compared; 98 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22947 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB)
[hos host] vk: replay reads of guest memory as submitted: 16660000 (10172126 pages from the captures)
[hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #305000: event 0x100d6c8024 mode 0 value 305000 (lr 0x80665068)
[hos host] vk: program picks: 23878660 kept (10055818 by the last picks' textures, a uniform buffer bound since), 26584 new, 8694755 with other textures (of 32600000)
[hos host] vk: a frame: 189 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.89, present 0.18, draw 5.21, descriptors 2.38, texture lookups 0.49, uploads 0.22, pipelines 0.22, programs 0.85, commands 1.37, dispatch 0.89, memory sync 0.49, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 521 (1090 MB): 5.63 MB synced in 1093.6 copies, 18.08 MB written back
[hos host] vk: 21360 presents, 60.1 fps; a frame: 1405 draws (0 skipped), 165 dispatches, 156 passes, 144 barriers, 6 submits, 0.4 uploads (1.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14189 pipelines (3740 ms), 0 programs (10 failed), 1152 textures (2542 MB)
[hos host] vk: host time in GPU resource waits: 2.28 ms a frame (fences 1.76, timelines 0.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 605.9 full barriers recorded, 286.1 left out (nothing recorded since the one before), 13.4 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.9, an op 700.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.7 ms a frame (the longest submission 4.3 ms); 1.3 textures made a frame: colour targets 0.3 (1.8 MB), depth 0.8 (4.8 MB), ASTC or BC 0.2 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 1.00 MB unswizzled on the CPU in 0.15 ms
[hos host] vk: 120 textures retired (1040 left, 2427 MB): budget 89, drawn over 31, dead targets 0
[hos sead::MainThread] nvn: GPU waits in 5 s: 39 sync waits (82 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 60 textures retired (988 left, 2395 MB): budget 50, drawn over 10, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 16680000 (10182298 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 230 dirty pages from their submits' captures, 59415 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 951 dispatches, 45430 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8304 runs, 752.1 MB compared; 80 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22138 in captures, 1 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 29 textures retired (961 left, 2292 MB): budget 19, drawn over 10, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos Presentation Thread] vi: 21397 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 20, 15.5-17.8 269, 17.8-22 9, 22-30 3, 30-40 0, over 40 0
[hos host] vk: 9 textures retired (959 left, 2271 MB): budget 0, drawn over 9, dead targets 0
[hos host] vk: 11 textures retired (957 left, 2308 MB): budget 6, drawn over 5, dead targets 0
[hos host] vk: 17 textures retired (970 left, 2331 MB): budget 2, drawn over 7, dead targets 8
[hos host] vk: replay reads of guest memory as submitted: 16700000 (10193244 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 551 ranges, 3142 MB), 269 pages written, 269 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (136 MB), 8 spare buffers, 5636 texture pages copied (22 MB)
[hos host] vk: readback pages refreshed from guest memory: 7029903 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5597 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: sparse mirrors: 119, 1107 MB with memory of 1580 MB (472 MB spared); blocks given memory later 13790, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1379 binds (319.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 357 taken from the spares, 165 made new; 403 kept as spares, 0 destroyed; 46 spares now, 1240 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 446 dirty pages from their submits' captures, 59879 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 980 dispatches, 45584 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8300 runs, 752.9 MB compared; 72 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22356 in captures, 0 only in the collection, 43 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB)
[hos host] vk: 16 textures retired (975 left, 2367 MB): budget 0, drawn over 6, dead targets 10
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1168 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1197 kept their virtual core (nothing there to wait for), 3155 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: texture 0x10009516800 640x360x1 x1 levels 1 NVN 0x25 -> vk 37 (1000 textures, 2469 MB)
[hos host] vk: texture 0x10009422400 62x62x1 x1 levels 1 NVN 0x34 -> vk 126 (1000 textures, 2494 MB)
[hos host] vk: 17 textures retired (993 left, 2404 MB): budget 11, drawn over 6, dead targets 0
[hos host] vk: texture 0x1000a50cc00 2560x1440x1 x1 levels 1 NVN 0x3f -> vk 122 (1000 textures, 2450 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: texture 0x10008612200 1799x1799x1 x1 levels 1 NVN 0x34 -> vk 126 (1000 textures, 2462 MB)
[hos host] vk: texture 0x10008612200 1832x1832x1 x1 levels 1 NVN 0x34 -> vk 126 (1000 textures, 2492 MB)
[hos host] vk: 38 textures retired (970 left, 2359 MB): budget 32, drawn over 6, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 16720000 (10205981 pages from the captures)
[hos ModuleSystemWorker1] nvn: 2800000 recordings made, 72259 live copies (23.4 MB), 6822 texture copies; 3 command buffers' segments (4.2 MB)
[hos host] vk: VRAM compaction 8 begins: the textures' blocks hold 1376 MB, 638 MB of it free space
[hos host] vk: recorder: 5384180 packets run in 10.0 s (busy 1.62 s, spun idle 0.12 s, slept 10653 times, 3265 of them under 200 us), 0 calls made directly (no room); 17927 batches over 600 frames (29.9 a frame, 300 packets a batch), 8944 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 152 times, 27.20 ms (0.3 times, 0.05 ms a frame); every thread 768 times, 60.53 ms in all, the longest 1.24 ms; vkDestroyImageView 216/32.71 ms (max 0.52), vkWaitForFences 508/24.91 ms (max 0.45), a part's submit 3/1.51 ms (max 1.24), vkQueueBindSparse 15/0.54 ms (max 0.22), vmaDestroyImage 16/0.39 ms (max 0.04), vmaFreeMemory 4/0.34 ms (max 0.23)
[hos host] vk: recorder: of its busy 1.62 s, 0.39 s in parts' ends and submits (the longest 1.19 ms), 0.07 s of that blocked on the queue's lock (153 times, the longest 1.16 ms), so 0.32 s in the ends, submits and holds; 1.24 s running the other packets; 8974 packets a frame
[hos host] vk: 59 textures retired (924 left, 2337 MB): budget 50, drawn over 9, dead targets 0
[hos host] vk: 20 textures retired (913 left, 2336 MB): budget 11, drawn over 9, dead targets 0
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 278 dirty pages from their submits' captures, 61360 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 983 dispatches, 46298 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8758 runs, 768.8 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23003 in captures, 0 only in the collection, 42 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 16740000 (10218873 pages from the captures)
[hos host] vk: 12 textures retired (915 left, 2324 MB): budget 0, drawn over 12, dead targets 0
[hos host] vk: a frame: 226 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.78, present 0.34, draw 4.75, descriptors 2.13, texture lookups 0.84, uploads 0.05, pipelines 0.20, programs 0.81, commands 1.25, dispatch 0.88, memory sync 0.48, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 523 (1054 MB): 5.85 MB synced in 1093.0 copies, 18.09 MB written back
[hos host] vk: 21480 presents, 59.8 fps; a frame: 1316 draws (0 skipped), 159 dispatches, 156 passes, 144 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14195 pipelines (3742 ms), 0 programs (10 failed), 916 textures (2335 MB)
[hos host] vk: host time in GPU resource waits: 1.93 ms a frame (fences 1.68, timelines 0.25, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 595.5 full barriers recorded, 284.3 left out (nothing recorded since the one before), 13.3 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 69.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.5, an op 613.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.5 ms a frame (the longest submission 4.7 ms); 3.1 textures made a frame: colour targets 1.2 (7.4 MB), depth 1.6 (12.1 MB), ASTC or BC 0.4 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.11 MB unswizzled on the CPU in 0.01 ms
[hos sead::MainThread] vcores: 154245 takes, 22865 waits (mean 65 us, longest 6.8 ms, 0 over HOS_VCORE_LOG_MS), 10934 kicks, 10907 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: program picks: 24019710 kept (10105860 by the last picks' textures, a uniform buffer bound since), 26627 new, 8753662 with other textures (of 32800000)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: 23 textures retired (893 left, 1905 MB): budget 0, drawn over 15, dead targets 8
[hos host] vk: replay reads of guest memory as submitted: 16760000 (10232420 pages from the captures)
[hos host] vk: sparse mirrors: 112, 1051 MB with memory of 1492 MB (440 MB spared); blocks given memory later 13804, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1390 binds (327.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 359 taken from the spares, 165 made new; 412 kept as spares, 0 destroyed; 53 spares now, 1328 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 791 dirty pages from their submits' captures, 59446 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 997 dispatches, 44410 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8103 runs, 729.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21641 in captures, 0 only in the collection, 40 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 194 live (2260 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB)
[hos host] vk: VRAM compaction 8 over: 333 textures moved (239 MB) in 30 passes, 0 left where they were; 19 blocks (608 MB) given back to the driver, 5088 MB in all so far
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.05 over 558 ranges, 3138 MB), 290 pages written, 290 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (96 MB), 11 spare buffers, 8106 texture pages copied (31 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13505 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #420000: event 0x100d6c800c mode 0 value 308699 (now 308680) (lr 0x80660e38)
[hos host] vk: 1 textures retired (974 left, 1947 MB): budget 0, drawn over 0, dead targets 1
[hos host] vk: replay reads of guest memory as submitted: 16780000 (10243369 pages from the captures)
[hos host] vk: slow upload 18.3 ms: #26760 0x10df1aba00 1024x1024x1 x1 levels 11 NVN 0x8d ASTC (BC3), 475 of 475 KB
[hos host] vk: texture 0x10d7bfc600 768x768x1 x1 levels 10 NVN 0x7b -> vk 37 (ASTC decoded) (1000 textures, 1992 MB)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 652 dirty pages from their submits' captures, 59065 from earlier ones', 9 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 45350 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8221 runs, 768.0 MB compared; 58 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23846 in captures, 2 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x1043400000 (4 MB, captured, state 2): 7; 0x7fb8800000 (52 MB, captured, state 2): 1; 0x7f9f800000 (4 MB, captured, state 2): 1;
[hos host] vk: slow upload 15.1 ms: #26717 0x10d2577200 1024x1024x1 x1 levels 11 NVN 0x8d ASTC (BC3), 475 of 475 KB
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: presenter thread: the copy's fence 1.67 ms, the present 0.71 ms a present
[hos host] vk: replay reads of guest memory as submitted: 16800000 (10254228 pages from the captures)
[hos host] vk: texture storage read as submitted (4001 times; 166149 pages from the copies, 0 of them rewritten in guest memory since their submit; 8260 pages kept, 208 spans, 734 MB)
[hos host] vk: slow upload 17.2 ms: #26878 0x11038fba00 768x768x1 x1 levels 10 NVN 0x7e ASTC (BC3), 324 of 324 KB
[hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 41 made): 39253942 lookups, 125690 missed (0.32 %)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 120, 1118 MB with memory of 1584 MB (465 MB spared); blocks given memory later 13819, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1400 binds (335.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 367 taken from the spares, 165 made new; 412 kept as spares, 0 destroyed; 45 spares now, 1236 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1262 dirty pages from their submits' captures, 59881 from earlier ones', 362 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1072 dispatches, 45554 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8381 runs, 749.7 MB compared; 91 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22628 in captures, 2 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb4400000 (24 MB, captured, state 2): 355; 0x7fb8800000 (52 MB, captured, state 2): 3; 0x7fa0800000 (8 MB, captured, state 2): 2; 0x1043400000 (4 MB, captured, state 2): 2;
[hos host] vk: imports: 194 live (2260 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06
[hos host] vk: memory: working set 14753 MB (peak 16021), private 21338 MB, staging 64 MB in 4 arenas, 47 made so far (212 ms)
[hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 846 MB; used 3208 of 15579 MB; heap 1 (device): 154 blocks 4631 MB, 20005 allocations 3951 MB; used 6754 of 11474 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 26 blocks 832 MB, 155 images 344 MB; 96 free ranges, the largest 31 MB
[hos host] vk: image blocks (textures, compacted): 37 blocks 1184 MB, 942 images 1163 MB; 242 free ranges, the largest 13 MB
[hos host] vk: render target images: 6809 reused from the pool, 6046 made new, 5889 destroyed from it, 1530 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 201 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 2884 with the pool empty (0 of them with targets still in the dead queue), 421 with none in it a present yet, 61 with none of a kind they can use, 817 with all smaller (the largest 49% short on average), 827 with all bigger (over 1.5 times), 1033 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 5697 idle
[hos host] vk: VRAM: 6754 MB used of 11474 MB allowed (680 MB of it free in the blocks); textures 2143 MB (budget 7158 MB), 1103 textures
[hos host] vk: budget: VRAM 4631 of 11474 MB (6754 MB by the driver, 2123 MB unaccounted): textures 2143, mirrors 1118, blocks free 680, shadows 284, other buffers 229, other images 111, staging 64; RAM 3141 MB (the process's private memory 21338 MB): imports 2260, capture spares 746, captures 100, capture table 35, others 0
[hos host] vk: textures: decoded ASTC 1234 MB, rendered 567 MB (84 drawn in one frame, kept: 44665 KB), other 303 MB; used in the last 8 presents 1690 MB
[hos host] vk: decoded ASTC used again: 622 textures (1399 MB) copied from ones the game finalized, not decoded; 622 of those retired, 0 listed; 2531 MB of blocks hashed in 413 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 16453 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d503ee00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 89 (15729 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 180 waited for it, 2.08 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 53.4 ms; VRAM 6.6 of 11.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 13, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6517 queued, 6517 loaded (median 1.46 ms, 99th percentile 16.63 ms, worst 38.13 ms, 16444 ms and 350.3 MB read in all); 8822 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12331 graphics pipelines (1 not in a cache), 736 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2449 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9673 (6517 ready, 0 loading), 17647 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 14030 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 781 known, 24 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 344 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.52, present 0.26, draw 5.96, descriptors 3.25, texture lookups 1.54, uploads 1.01, pipelines 0.21, programs 0.78, commands 1.34, dispatch 0.86, memory sync 0.53, queue submit 0.00, slot wait 0.23
[hos host] vk: a frame: mirrors 532 (1118 MB): 6.30 MB synced in 1097.2 copies, 18.29 MB written back
[hos host] vk: 21600 presents, 56.9 fps; a frame: 1412 draws (0 skipped), 162 dispatches, 152 passes, 146 barriers, 6 submits, 2.1 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14199 pipelines (3744 ms), 0 programs (10 failed), 1103 textures (2143 MB)
[hos host] vk: host time in GPU resource waits: 5.60 ms a frame (fences 3.36, timelines 2.24, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 597.5 full barriers recorded, 282.1 left out (nothing recorded since the one before), 15.3 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 39.8; 0x12: 69.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 137.5, an op 563.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 10.8 ms a frame (the longest submission 4.3 ms); 3.8 textures made a frame: colour targets 0.8 (3.9 MB), depth 1.0 (4.4 MB), ASTC or BC 2.0 (4.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.65 MB unswizzled on the CPU in 0.11 ms
[hos host] vk: queues ordered by their events: 143941 submissions stopped at a wait, 143881 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 421436, not met 0; captures at a CPU event's wait 13122, 0 of them at a wait met on arrival
[hos host] vk: the render thread (600 presents, 16.86 ms a frame): its jobs 14.36 ms a frame (submits 6.71, presents 4.61, others 3.04), idle 2.50; the game waited for room at 10 presents, 1.42 ms each (0.02 a frame; 6.34 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 186 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 1213 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 13.50 ms after its hand-over (57.40 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.91 + 0.27 (the window's copy is in the idle), after it 1.13 + 0.71; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 8.50 parts a frame recorded in the signal's job, 78.6% of them took nothing that came with the signal; of the others, 1.81 a frame took pages of the signal's capture into a mirror (32.7 pages a frame) and 0.03 pages read as guest memory had them (0.1 pages a frame); a frame's first part took nothing in 20 of 186 frames, its first three in 20
[hos host] vk: frame ahead, what came with the signal: 186 signals with a capture, each of 59.0 pages in 8.9 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.41+0.11, 0.23+0.10, 0.18+0.05, 0.04+0.03, 0.06+0.30, 0.07+0.07, 0.09+0.03, 0.04+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1127 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1263 kept their virtual core (nothing there to wait for), 3089 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.05 over 587 ranges, 3198 MB), 270 pages written, 270 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 9914 texture pages copied (38 MB)
[hos host] vk: queues ordered by their events: 144000 submissions stopped at a wait, 143938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 421630, not met 0; captures at a CPU event's wait 13125, 0 of them at a wait met on arrival
[hos sead::MainThread] nvn: CommandBufferSignalEvent #310000: event 0x100d6c802c mode 0 value 309999 (lr 0x80667688)
[hos host] vk: render targets drawn over from another start, let go: 13160 so far (62285 MB), 11935 of them (56613 MB) between the passes
[hos host] vk: replay reads of guest memory as submitted: 16820000 (10265821 pages from the captures)
[hos host] vk: program picks: 24164830 kept (10146003 by the last picks' textures, a uniform buffer bound since), 26652 new, 8808517 with other textures (of 33000000)
[hos host] vk: readback pages refreshed from guest memory: 7097017 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 991 dirty pages from their submits' captures, 59981 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1030 dispatches, 45649 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8709 runs, 766.1 MB compared; 58 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23960 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1;
[hos host] vk: replay reads of guest memory as submitted: 16840000 (10277016 pages from the captures)
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 760000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13058 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 1 textures retired (1130 left, 2162 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: reaper: 864 images (3779 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 88 (275 MB)
[hos host] vk: replay reads of guest memory as submitted: 16860000 (10288143 pages from the captures)
[hos host] vk: sparse mirrors: 123, 1123 MB with memory of 1596 MB (472 MB spared); blocks given memory later 13822, given back 17188; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1406 binds (336.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 370 taken from the spares, 165 made new; 412 kept as spares, 0 destroyed; 42 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 422 dirty pages from their submits' captures, 65136 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 1038 dispatches, 47309 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8753 runs, 781.6 MB compared; 82 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 24230 in captures, 1 only in the collection, 41 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2;
[hos host] vk: imports: 195 live (2264 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 67 sync waits (239 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos Presentation Thread] vi: 21686 frames presented (57.6 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 66, 15.5-17.8 182, 17.8-22 25, 22-30 6, 30-40 7, over 40 3
[hos host] vk: replay reads of guest memory as submitted: 16880000 (10300666 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06
[hos host] vk: 417 mirrors given back (2682 MB in all); 1132 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone
[hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 600 ranges, 3294 MB), 291 pages written, 291 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (96 MB), 11 spare buffers, 5135 texture pages copied (20 MB)
[hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2650 dirty pages from their submits' captures, 53666 from earlier ones', 69 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 946 dispatches, 41642 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8383 runs, 676.9 MB compared; 70 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21175 in captures, 13 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9c400000 (32 MB, captured, state 2): 62; 0x7f98800000 (8 MB, captured, state 2): 2; 0x7f95000000 (4 MB, captured, state 2): 2; 0x7fa9000000 (4 MB, captured, state 2): 1; 0x7f88000000 (12 MB, captured, state 2): 1; 0x1041c00000 (4 MB, captured, state 2): 1;
[hos host] vk: replay reads of guest memory as submitted: 16900000 (10313419 pages from the captures)
[hos host] vk: a frame: 336 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.04, present 0.18, draw 6.04, descriptors 2.81, texture lookups 0.64, uploads 0.40, pipelines 0.26, programs 0.94, commands 1.57, dispatch 1.02, memory sync 0.60, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 540 (1132 MB): 6.28 MB synced in 1130.0 copies, 18.90 MB written back
[hos host] vk: 21720 presents, 56.5 fps; a frame: 1671 draws (0 skipped), 181 dispatches, 166 passes, 148 barriers, 6 submits, 1.6 uploads (0.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14230 pipelines (3751 ms), 0 programs (10 failed), 1290 textures (2332 MB)
[hos host] vk: host time in GPU resource waits: 5.16 ms a frame (fences 3.00, timelines 2.16, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 647.0 full barriers recorded, 296.1 left out (nothing recorded since the one before), 19.8 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 40.5; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 139.1, an op 708.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.7 ms a frame (the longest submission 5.0 ms); 1.6 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 1.6 (1.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.22 MB unswizzled on the CPU in 0.02 ms
[hos host] vk: replay reads of guest memory as submitted: 16920000 (10325944 pages from the captures)
[hos host] vk: program picks: 24314157 kept (10191167 by the last picks' textures, a uniform buffer bound since), 26671 new, 8859171 with other textures (of 33200000)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #425000: event 0x100d6c8010 mode 0 value 312500 (now 312481) (lr 0x80660e38)
[hos host] vk: sparse mirrors: 124, 1133 MB with memory of 1608 MB (474 MB spared); blocks given memory later 13830, given back 17188; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1419 binds (337.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 376 taken from the spares, 165 made new; 417 kept as spares, 0 destroyed; 41 spares now, 1212 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2176 dirty pages from their submits' captures, 81481 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 960 dispatches, 43291 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8334 runs, 696.4 MB compared; 62 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20439 in captures, 0 only in the collection, 35 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 196 live (2268 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB)
[hos host] vk: replay reads of guest memory as submitted: 16940000 (10338172 pages from the captures)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1127 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 15 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 995 kept their virtual core (nothing there to wait for), 2589 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 16960000 (10350547 pages from the captures)
[hos host] vk: late pages' fix (600 recordings with one): of 1076 captures with late pages 1076 fixed (75.4 pages, 76.1 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 439 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 153, of an earlier recording 286, not recorded yet 0, of this part 0; pages fixed from stamps 80084, from snapshots 1089; the shadows: 1527891 pages synced from captures, 1497967 made known by it, 29924 left unknown because the replay or the collection wrote them since
[hos tera::FileResourceMgr] vcores: 152738 takes, 28588 waits (mean 85 us, longest 28.8 ms, 0 over HOS_VCORE_LOG_MS), 13177 kicks, 13160 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1828 dirty pages from their submits' captures, 51509 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 881 dispatches, 40342 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7797 runs, 661.9 MB compared; 58 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19334 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: capture: a submit 0.25 ms (lock 0.01, scan 0.04 over 583 ranges, 3283 MB), 371 pages written, 371 copied (1.4 MB)
[hos host] vk: a capture taken (ms): all 0.042; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.021, bytes copied 0.000; 25 kept (110 MB), 10 spare buffers, 7738 texture pages copied (30 MB)
[hos host] vk: replay reads of guest memory as submitted: 16980000 (10363020 pages from the captures)
[hos host] vk: 138 textures retired (1163 left, 2128 MB): budget 138, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 5 finalized, 31 MB
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.07
[hos host] vk: 6 textures retired (1157 left, 2111 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 20817 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 17000000 (10375305 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 7164543 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: 7 textures retired (1156 left, 2108 MB): budget 7, drawn over 0, dead targets 0
[hos host] vk: sparse mirrors: 125, 1026 MB with memory of 1612 MB (585 MB spared); blocks given memory later 13831, given back 18910; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1422 binds (346.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 377 taken from the spares, 165 made new; 417 kept as spares, 0 destroyed; 40 spares now, 1208 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2223 dirty pages from their submits' captures, 55052 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 902 dispatches, 41567 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8035 runs, 680.9 MB compared; 61 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19631 in captures, 132 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 196 live (2268 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 17020000 (10387682 pages from the captures)
[hos host] vk: program picks: 24466918 kept (10240470 by the last picks' textures, a uniform buffer bound since), 26677 new, 8906404 with other textures (of 33400000)
[hos ModuleSystemWorker3] nvn: 2850000 recordings made, 68209 live copies (23.3 MB), 6920 texture copies; 3 command buffers' segments (4.1 MB)
[hos host] vk: a frame: 348 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.94, present 0.24, draw 6.63, descriptors 2.73, texture lookups 0.21, uploads 0.06, pipelines 0.31, programs 1.12, commands 1.96, dispatch 1.00, memory sync 0.72, queue submit 0.00, slot wait 0.27
[hos host] vk: a frame: mirrors 543 (1028 MB): 6.92 MB synced in 1126.2 copies, 18.46 MB written back
[hos host] vk: 21840 presents, 56.4 fps; a frame: 1917 draws (0 skipped), 177 dispatches, 177 passes, 148 barriers, 6 submits, 0.2 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14230 pipelines (3751 ms), 0 programs (10 failed), 1162 textures (2135 MB)
[hos host] vk: host time in GPU resource waits: 5.31 ms a frame (fences 2.00, timelines 3.32, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 676.7 full barriers recorded, 306.0 left out (nothing recorded since the one before), 20.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 40.6; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 138.8, an op 876.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.4 ms a frame (the longest submission 12.3 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.5 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.23 MB unswizzled on the CPU in 0.02 ms
[hos host] vk: replay reads of guest memory as submitted: 17040000 (10400194 pages from the captures)
[hos host] vk: 13 textures retired (1154 left, 2110 MB): budget 13, drawn over 0, dead targets 0
[hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2799 dirty pages from their submits' captures, 52300 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 885 dispatches, 40277 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7915 runs, 664.2 MB compared; 51 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19497 in captures, 0 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 17060000 (10412701 pages from the captures)
[hos host] vk: 7 textures retired (1152 left, 2104 MB): budget 7, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: CommandBufferSignalEvent #315000: event 0x100d6c8038 mode 0 value 314999 (lr 0x8066189c)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 576 ranges, 3290 MB), 284 pages written, 284 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.041; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.022, bytes copied 0.000; 25 kept (110 MB), 10 spare buffers, 12554 texture pages copied (49 MB)
[hos host] vk: replay reads of guest memory as submitted: 17080000 (10425233 pages from the captures)
[hos host] vk: sparse mirrors: 123, 1026 MB with memory of 1592 MB (565 MB spared); blocks given memory later 13832, given back 18911; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1430 binds (347.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 380 taken from the spares, 165 made new; 422 kept as spares, 0 destroyed; 42 spares now, 1228 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1565 dirty pages from their submits' captures, 53179 from earlier ones', 81 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 896 dispatches, 41047 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8423 runs, 681.7 MB compared; 52 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20227 in captures, 1 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 79; 0x7f90400000 (4 MB, captured, state 2): 2;
[hos host] vk: imports: 198 live (2276 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 17100000 (10437594 pages from the captures)
[hos host] vk: 4 textures retired (1152 left, 2108 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1447 dirty pages from their submits' captures, 52658 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 41013 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7908 runs, 672.0 MB compared; 57 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20230 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos sead::MainThread] vk: readbacks of GPU results: 1123 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 16 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1205 kept their virtual core (nothing there to wait for), 3147 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 17120000 (10449811 pages from the captures)
[hos host] vk: program picks: 24620868 kept (10289533 by the last picks' textures, a uniform buffer bound since), 26686 new, 8952445 with other textures (of 33600000)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #430000: event 0x100d6c8010 mode 0 value 316300 (now 316281) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 17140000 (10462503 pages from the captures)
[hos host] vk: a frame: 378 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.31, present 0.17, draw 6.11, descriptors 2.59, texture lookups 0.22, uploads 0.11, pipelines 0.29, programs 0.96, commands 1.77, dispatch 0.97, memory sync 0.66, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 545 (1014 MB): 5.85 MB synced in 1125.2 copies, 18.70 MB written back
[hos host] vk: 21960 presents, 60.5 fps; a frame: 1895 draws (0 skipped), 183 dispatches, 176 passes, 147 barriers, 6 submits, 0.3 uploads (0.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14230 pipelines (3751 ms), 0 programs (10 failed), 1157 textures (2129 MB)
[hos host] vk: host time in GPU resource waits: 4.01 ms a frame (fences 1.56, timelines 2.45, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 684.3 full barriers recorded, 304.9 left out (nothing recorded since the one before), 19.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 40.1; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 138.0, an op 820.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.8 ms a frame (the longest submission 4.3 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.9 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.33 MB unswizzled on the CPU in 0.04 ms
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 574 ranges, 3304 MB), 312 pages written, 312 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.056; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.039, bytes copied 0.000; 25 kept (234 MB), 11 spare buffers, 11683 texture pages copied (45 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 252 sync waits (841 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 120, 1014 MB with memory of 1568 MB (553 MB spared); blocks given memory later 13832, given back 18911; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1432 binds (347.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 380 taken from the spares, 165 made new; 425 kept as spares, 0 destroyed; 45 spares now, 1252 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 949, 4216.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1116 dirty pages from their submits' captures, 54234 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 921 dispatches, 41997 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8003 runs, 686.5 MB compared; 69 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19712 in captures, 1 only in the collection, 36 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 198 live (2276 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 62681 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: 8 textures retired (1151 left, 2111 MB): budget 8, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17160000 (10474471 pages from the captures)
[hos Presentation Thread] vi: 21977 frames presented (58.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 72, 15.5-17.8 166, 17.8-22 37, 22-30 11, 30-40 1, over 40 4
[hos host] vk: 1 textures retired (1154 left, 2110 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 17180000 (10487405 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.08, writes collected 0.05
[hos host] vk: readback pages refreshed from guest memory: 7231646 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 926, 4109.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1329 dirty pages from their submits' captures, 71189 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 904 dispatches, 42600 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8062 runs, 703.2 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19622 in captures, 0 only in the collection, 37 only by the replay itself
[hos host] vk: read live, by pool: 0x7f90400000 (4 MB, captured, state 2): 3;
[hos host] vk: replay reads of guest memory as submitted: 17200000 (10499496 pages from the captures)
[hos host] vk: 95 textures retired (1087 left, 2109 MB): budget 95, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 3 MB
[hos host] vk: program picks: 24774240 kept (10340971 by the last picks' textures, a uniform buffer bound since), 26686 new, 8999073 with other textures (of 33800000)
[hos host] vk: recorder: 5395212 packets run in 10.0 s (busy 1.68 s, spun idle 0.12 s, slept 10782 times, 3681 of them under 200 us), 0 calls made directly (no room); 17463 batches over 577 frames (30.3 a frame, 309 packets a batch), 9058 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 161 times, 21.55 ms (0.3 times, 0.04 ms a frame); every thread 767 times, 65.99 ms in all, the longest 0.82 ms; vkWaitForFences 507/37.22 ms (max 0.57), vkDestroyImageView 107/11.75 ms (max 0.52), vkQueueBindSparse 56/4.88 ms (max 0.48), vmaFreeMemory 20/4.64 ms (max 0.82), vkEndCommandBuffer 13/2.90 ms (max 0.44), vmaDestroyBuffer 21/1.94 ms (max 0.63)
[hos host] vk: recorder: of its busy 1.68 s, 0.49 s in parts' ends and submits (the longest 1.40 ms), 0.17 s of that blocked on the queue's lock (310 times, the longest 1.36 ms), so 0.31 s in the ends, submits and holds; 1.20 s running the other packets; 9350 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 17220000 (10511590 pages from the captures)
[hos host] vk: sparse mirrors: 121, 1024 MB with memory of 1596 MB (571 MB spared); blocks given memory later 13832, given back 18919; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1439 binds (349.8 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 383 taken from the spares, 165 made new; 427 kept as spares, 0 destroyed; 44 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1858 dirty pages from their submits' captures, 55992 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 924 dispatches, 42572 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8040 runs, 700.9 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20343 in captures, 0 only in the collection, 38 only by the replay itself
[hos host] vk: read live, by pool: 0x7f9c400000 (32 MB, captured, state 2): 3;
[hos host] vk: imports: 197 live (2276 MB), 2 retired still imported (8 MB), 53 retired given back (1304 MB)
[hos sead::MainThread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.04 over 572 ranges, 3314 MB), 339 pages written, 339 copied (1.3 MB)
[hos host] vk: a capture taken (ms): all 0.050; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.032, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 12270 texture pages copied (47 MB)
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 780000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040)
[hos host] vk: replay reads of guest memory as submitted: 17240000 (10524052 pages from the captures)
[hos host] vk: 5 textures retired (1098 left, 2110 MB): budget 5, drawn over 0, dead targets 0
[hos rmbl::VibrationThread] hid: 73728 vibration values, strongest 0.02 since the last count
[hos BaseProcExecutor3] vcores: 151872 takes, 23883 waits (mean 82 us, longest 12.4 ms, 0 over HOS_VCORE_LOG_MS), 11102 kicks, 11349 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 17260000 (10536900 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1826 dirty pages from their submits' captures, 52584 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 928 dispatches, 40871 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7729 runs, 667.5 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19914 in captures, 1 only in the collection, 34 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 9 textures retired (1139 left, 2152 MB): budget 9, drawn over 0, dead targets 0
[hos host] vk: a frame: 332 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 10.83, present 0.18, draw 6.78, descriptors 2.90, texture lookups 0.37, uploads 0.22, pipelines 0.29, programs 1.12, commands 1.97, dispatch 0.95, memory sync 0.66, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 548 (1024 MB): 6.58 MB synced in 1122.2 copies, 18.50 MB written back
[hos host] vk: 22080 presents, 57.7 fps; a frame: 2003 draws (0 skipped), 174 dispatches, 191 passes, 147 barriers, 6 submits, 0.9 uploads (0.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14231 pipelines (3752 ms), 0 programs (10 failed), 1139 textures (2152 MB)
[hos host] vk: host time in GPU resource waits: 4.55 ms a frame (fences 1.72, timelines 2.83, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 720.3 full barriers recorded, 314.1 left out (nothing recorded since the one before), 24.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 40.2; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 138.1, an op 965.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 11.9 ms a frame (the longest submission 5.2 ms); 0.9 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.8 (1.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.10 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: replay reads of guest memory as submitted: 17280000 (10549529 pages from the captures)
[hos host] vk: 8 textures retired (1158 left, 2177 MB): budget 8, drawn over 0, dead targets 0
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1139 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1264 kept their virtual core (nothing there to wait for), 3088 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 17300000 (10562696 pages from the captures)
[hos host] vk: 3 textures retired (1170 left, 2215 MB): budget 3, drawn over 0, dead targets 0
[hos host] vk: 6 textures retired (1164 left, 2214 MB): budget 6, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17320000 (10575029 pages from the captures)
[hos host] vk: sparse mirrors: 120, 1021 MB with memory of 1592 MB (570 MB spared); blocks given memory later 13832, given back 18930; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1445 binds (351.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 385 taken from the spares, 165 made new; 429 kept as spares, 0 destroyed; 44 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1966 dirty pages from their submits' captures, 51582 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 924 dispatches, 39986 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7580 runs, 650.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19561 in captures, 0 only in the collection, 33 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 197 live (2276 MB), 2 retired still imported (8 MB), 53 retired given back (1304 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 22394 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 24926062 kept (10394863 by the last picks' textures, a uniform buffer bound since), 26713 new, 9047224 with other textures (of 34000000)
[hos host] vk: late pages' fix: stamps for 0x1038bf0000 (pool 0x1038400000; 896 blocks, 57344 KB)
[hos host] vk: 433 mirrors given back (2737 MB in all); 1006 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone
[hos host] vk: replay reads of guest memory as submitted: 17340000 (10587185 pages from the captures)
[hos host] vk: 3 textures retired (1161 left, 2213 MB): budget 3, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: CommandBufferSignalEvent #320000: event 0x100d6c8044 mode 0 value 319999 (lr 0x802742f0)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 579 ranges, 3363 MB), 278 pages written, 278 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.037; GPU wait for late pages 0.000 (1 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (156 MB), 10 spare buffers, 7267 texture pages copied (28 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: pool 0x7fb8800000..0x7fbbc00000 (52 MB) trimmed to its 1 live part(s), 8 MB: the rest outside every game pool for 1856 submissions (5 trims, 204 MB in all)
[hos host] vk: 554 pools mirrored in VRAM (1008 MB)
[hos host] vk: pool 0x7fbd800000..0x7fc1400000 (60 MB) trimmed to its 1 live part(s), 20 MB: the rest outside every game pool for 1920 submissions (6 trims, 244 MB in all)
[hos host] vk: pool 0x7fbf400000..0x7fc0800000: 16 KB the CPU wrote since retired pool 0x7fbd800000's last sync kept from guest memory, not its mirror
[hos host] vk: 5 textures retired (1180 left, 2226 MB): budget 5, drawn over 0, dead targets 0
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #435000: event 0x100d6c8010 mode 0 value 320100 (now 320081) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 17360000 (10599530 pages from the captures)
[hos host] vk: pool 0x7fd6400000..0x7fd9400000 (48 MB) trimmed to its 1 live part(s), 20 MB: the rest outside every game pool for 1984 submissions (7 trims, 272 MB in all)
[hos host] vk: pool 0x7fd7000000..0x7fd8400000: 4 KB the CPU wrote since retired pool 0x7fd6400000's last sync kept from guest memory, not its mirror
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2114 dirty pages from their submits' captures, 53904 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 929 dispatches, 41812 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8664 runs, 669.8 MB compared; 392 pages read into the shadows; 288 MB of shadows (54 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21303 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 4 textures retired (1176 left, 2224 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17380000 (10611243 pages from the captures)
[hos host] vk: texture slots registered again at another size: 1200000 (of 1494067 registrations; slot 9736 480x270 -> 640x360)
[hos host] vk: 2 textures retired (1174 left, 2224 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: presenter thread: the copy's fence 1.94 ms, the present 0.80 ms a present
[hos host] vk: 1 textures retired (1173 left, 2223 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17400000 (10623476 pages from the captures)
[hos ModuleSystemWorker3] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 41 made): 20617862 lookups, 135034 missed (0.65 %)
[hos host] vk: 4 textures retired (1181 left, 2211 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 7299160 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: sparse mirrors: 120, 1007 MB with memory of 1464 MB (456 MB spared); blocks given memory later 13840, given back 18933; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1459 binds (352.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 390 taken from the spares, 166 made new; 436 kept as spares, 0 destroyed; 46 spares now, 1376 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1688 dirty pages from their submits' captures, 54613 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 918 dispatches, 41620 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9865 runs, 676.6 MB compared; 47 pages read into the shadows; 288 MB of shadows (52 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23064 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 198 live (2172 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos host] vk: replay reads of guest memory as submitted: 17420000 (10635479 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 17440000 (10647732 pages from the captures)
[hos host] vk: program picks: 25079019 kept (10447532 by the last picks' textures, a uniform buffer bound since), 26735 new, 9094245 with other textures (of 34200000)
[hos host] vk: memory: working set 14962 MB (peak 16021), private 21905 MB, staging 64 MB in 4 arenas, 47 made so far (212 ms)
[hos host] vk: memory heaps: heap 0 (system): 17 blocks 960 MB, 36 allocations 858 MB; used 3332 of 15579 MB; heap 1 (device): 165 blocks 4983 MB, 18449 allocations 3951 MB; used 7106 of 8475 MB;
[hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB
[hos host] vk: image blocks (render targets): 26 blocks 832 MB, 150 images 303 MB; 92 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 48 blocks 1536 MB, 1069 images 1305 MB; 343 free ranges, the largest 12 MB
[hos host] vk: render target images: 6809 reused from the pool, 6048 made new, 5896 destroyed from it, 1530 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 201 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 2886 with the pool empty (0 of them with targets still in the dead queue), 421 with none in it a present yet, 61 with none of a kind they can use, 817 with all smaller (the largest 49% short on average), 827 with all bigger (over 1.5 times), 1033 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 5704 idle
[hos host] vk: VRAM: 7106 MB used of 8475 MB allowed (1032 MB of it free in the blocks); textures 2244 MB (budget 2107 MB), 1225 textures
[hos host] vk: budget: VRAM 4983 of 8475 MB (7106 MB by the driver, 2123 MB unaccounted): textures 2244, blocks free 1032, mirrors 1009, shadows 288, other buffers 233, other images 111, staging 64; RAM 3253 MB (the process's private memory 21905 MB): imports 2176, capture spares 732, retired imports 176, captures 134, capture table 35, others 0
[hos host] vk: textures: decoded ASTC 1398 MB, rendered 508 MB (82 drawn in one frame, kept: 44485 KB), other 303 MB; used in the last 8 presents 2082 MB
[hos host] vk: decoded ASTC used again: 686 textures (1539 MB) copied from ones the game finalized, not decoded; 686 of those retired, 2 listed; 2655 MB of blocks hashed in 430 ms
[hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago
[hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 17053 presents ago
[hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk: 16 MB: 0x10d503ee00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 91 (16081 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB);
[hos host] vk: 6 textures retired (1219 left, 2243 MB): budget 6, drawn over 0, dead targets 0
[hos Presentation Thread] vk: presents after the CPU's signal (600): 502 waited for it, 3.26 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 48.3 ms; VRAM 6.9 of 8.2 GB
[hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 13, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (play): 6554 queued, 6554 loaded (median 1.46 ms, 99th percentile 16.63 ms, worst 38.13 ms, 16489 ms and 351.4 MB read in all); 8864 programs (1634 of them compute variants without their code, 0 graphics programs no note names), 12363 graphics pipelines (1 not in a cache), 744 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2469 pipelines made at create taken by draws; 3157 create calls with nothing kept for their stages, 0 before the index; units 9711 (6554 ready, 0 loading), 17807 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (play): 14090 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (play): 789 known, 32 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: a frame: 340 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.36, present 0.18, draw 7.74, descriptors 3.31, texture lookups 0.31, uploads 0.12, pipelines 0.33, programs 1.36, commands 2.15, dispatch 1.13, memory sync 0.57, queue submit 0.00, slot wait 0.26
[hos host] vk: a frame: mirrors 557 (1009 MB): 6.53 MB synced in 1194.8 copies, 19.40 MB written back
[hos host] vk: 22200 presents, 56.3 fps; a frame: 2316 draws (0 skipped), 230 dispatches, 213 passes, 148 barriers, 6 submits, 1.3 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14231 pipelines (3752 ms), 0 programs (10 failed), 1221 textures (2243 MB)
[hos host] vk: host time in GPU resource waits: 5.08 ms a frame (fences 1.48, timelines 3.60, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 815.1 full barriers recorded, 328.4 left out (nothing recorded since the one before), 35.1 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 40.9; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 139.5, an op 1278.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.4 ms a frame (the longest submission 4.2 ms); 1.4 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 1.3 (1.8 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.15 MB unswizzled on the CPU in 0.02 ms
[hos host] vk: the render thread (600 presents, 17.41 ms a frame): its jobs 16.39 ms a frame (submits 2.61, presents 5.23, others 8.55), idle 1.02; the game waited for room at 30 presents, 0.52 ms each (0.03 a frame; 15.61 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 503 captures at a CPU signal; the submits' waits read in 0.038 ms a frame; readbacks: 2590 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.68 ms after its hand-over (40.32 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.79 + 0.26 (the window's copy is in the idle), after it 1.55 + 0.77; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms)
[hos host] vk: frame ahead (600 presents): 9.67 parts a frame recorded in the signal's job, 77.2% of them took nothing that came with the signal; of the others, 2.19 a frame took pages of the signal's capture into a mirror (55.2 pages a frame) and 0.02 pages read as guest memory had them (0.2 pages a frame); a frame's first part took nothing in 15 of 503 frames, its first three in 15
[hos host] vk: frame ahead, what came with the signal: 503 signals with a capture, each of 85.9 pages in 10.7 pools; between a frame's last submit and that signal 0.03 texture and sampler registrations, 0.21 releases and 0.00 other CPU signals a signal (9 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.62+0.14, 0.24+0.11, 0.16+0.05, 0.03+0.02, 0.03+0.02, 0.12+0.16, 0.19+0.19, 0.09+0.05, 0.05+0.02, 0.02+0.01, 0.00+0.00, 0.00+0.00
[hos host] vk: render targets drawn over from another start, let go: 13163 so far (62304 MB), 11938 of them (56633 MB) between the passes
[hos host] vk: 4 textures retired (1222 left, 2242 MB): budget 4, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: 2900000 recordings made, 65646 live copies (27.5 MB), 6922 texture copies; 1 command buffers' segments (4.0 MB)
[hos host] vk: replay reads of guest memory as submitted: 17460000 (10659529 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1744 dirty pages from their submits' captures, 54337 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 903 dispatches, 41770 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 10295 runs, 672.0 MB compared; 73 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23053 in captures, 0 only in the collection, 25 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 4 textures retired (1229 left, 2255 MB): budget 4, drawn over 0, dead targets 0
[hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.03 over 578 ranges, 3288 MB), 296 pages written, 296 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.040; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (138 MB), 12 spare buffers, 3802 texture pages copied (14 MB)
[hos host] vk: replay reads of guest memory as submitted: 17480000 (10671203 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 17500000 (10682685 pages from the captures)
[hos host] vk: 7 textures retired (1226 left, 2252 MB): budget 7, drawn over 0, dead targets 0
[hos host] vk: sparse mirrors: 122, 1009 MB with memory of 1472 MB (462 MB spared); blocks given memory later 13843, given back 18933; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1462 binds (353.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 392 taken from the spares, 166 made new; 436 kept as spares, 0 destroyed; 44 spares now, 1368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2168 dirty pages from their submits' captures, 50590 from earlier ones', 916 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 873 dispatches, 39617 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 10278 runs, 641.1 MB compared; 69 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22800 in captures, 0 only in the collection, 23 only by the replay itself
[hos host] vk: read live, by pool: 0x1055800000 (4 MB, captured, state 2): 916;
[hos host] vk: imports: 199 live (2176 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 251 sync waits (842 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: reaper: 422 images (745 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 129 (175 MB)
[hos ModuleSystemWorker5] vk: readbacks of GPU results: 1085 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 21 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1100 kept their virtual core (nothing there to wait for), 3252 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 4 textures retired (1224 left, 2245 MB): budget 4, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17520000 (10694502 pages from the captures)
[hos host] vk: 10 textures retired (1218 left, 2247 MB): budget 10, drawn over 0, dead targets 0
[hos Presentation Thread] vi: 22263 frames presented (57.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 21, 15.5-17.8 200, 17.8-22 49, 22-30 12, 30-40 3, over 40 1
[hos host] vk: replay reads of guest memory as submitted: 17540000 (10705862 pages from the captures)
[hos host] vk: 11 textures retired (1208 left, 2234 MB): budget 11, drawn over 0, dead targets 0
[hos host] vk: program picks: 25233330 kept (10500052 by the last picks' textures, a uniform buffer bound since), 26735 new, 9139934 with other textures (of 34400000)
[hos host] vk: write-backs left out (600 submits): 799, 3550.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2325 dirty pages from their submits' captures, 50932 from earlier ones', 208 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 39594 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9370 runs, 623.7 MB compared; 63 pages read into the shadows; 288 MB of shadows (55 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22282 in captures, 0 only in the collection, 23 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb7c00000 (12 MB, captured, state 2): 118; 0x7fe2800000 (4 MB, captured, state 2): 90;
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 19477 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 17560000 (10718470 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 17580000 (10730361 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 15 pages in 15 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 575 ranges, 3289 MB), 290 pages written, 290 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.042; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.015, bytes copied 0.000; 25 kept (142 MB), 11 spare buffers, 5629 texture pages copied (21 MB)
[hos host] vk: replay reads of guest memory as submitted: 17600000 (10741613 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 123, 1005 MB with memory of 1472 MB (466 MB spared); blocks given memory later 13843, given back 18944; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1468 binds (354.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 394 taken from the spares, 166 made new; 437 kept as spares, 0 destroyed; 43 spares now, 1368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1771 dirty pages from their submits' captures, 49583 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 851 dispatches, 38969 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 9130 runs, 628.1 MB compared; 61 pages read into the shadows; 288 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21597 in captures, 0 only in the collection, 23 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 200 live (2180 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos host] vk: 2 textures retired (1223 left, 2244 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: 17 textures retired (1214 left, 2251 MB): budget 17, drawn over 0, dead targets 0
[hos host] vk: a frame: 358 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.38, present 0.19, draw 7.75, descriptors 3.16, texture lookups 0.19, uploads 0.05, pipelines 0.33, programs 1.41, commands 2.25, dispatch 1.21, memory sync 0.57, queue submit 0.00, slot wait 0.25
[hos host] vk: a frame: mirrors 560 (1005 MB): 6.59 MB synced in 1253.7 copies, 20.08 MB written back
[hos host] vk: 22320 presents, 57.9 fps; a frame: 2424 draws (0 skipped), 271 dispatches, 221 passes, 148 barriers, 6 submits, 0.5 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14232 pipelines (3752 ms), 0 programs (10 failed), 1215 textures (2251 MB)
[hos host] vk: host time in GPU resource waits: 5.23 ms a frame (fences 1.84, timelines 3.39, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 858.8 full barriers recorded, 330.2 left out (nothing recorded since the one before), 43.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 40.5; 0x12: 70.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 139.4, an op 1346.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.6 ms a frame (the longest submission 3.8 ms); 0.6 textures made a frame: colour targets 0.1 (0.3 MB), depth 0.0 (0.1 MB), ASTC or BC 0.5 (0.5 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.12 MB unswizzled on the CPU in 0.02 ms
[hos host] vk: replay reads of guest memory as submitted: 17620000 (10752559 pages from the captures)
[hos host] vk: 1 textures retired (1214 left, 2251 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: pool 0x7fb3db5d00+0x74b8 imported in 2.4 ms as 0x7fb3c00000..0x7fb4000000 (251 imports, 737 ms in all)
[hos host] vk: replay reads of guest memory as submitted: 17640000 (10764321 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #440000: event 0x100d6c8010 mode 0 value 323900 (now 323881) (lr 0x80660e38)
[hos host] vk: readback pages refreshed from guest memory: 7368204 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1881 dirty pages from their submits' captures, 52335 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 913 dispatches, 40935 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8757 runs, 651.6 MB compared; 69 pages read into the shadows; 288 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21083 in captures, 0 only in the collection, 28 only by the replay itself
[hos host] vk: read live, by pool:
[hos aal::LowPrioWorkerThread] vcores: 146913 takes, 24520 waits (mean 80 us, longest 10.2 ms, 0 over HOS_VCORE_LOG_MS), 11622 kicks, 11855 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 17660000 (10776300 pages from the captures)
[hos host] vk: program picks: 25388129 kept (10555272 by the last picks' textures, a uniform buffer bound since), 26745 new, 9185125 with other textures (of 34600000)
[hos host] vk: 7 textures retired (1238 left, 2272 MB): budget 7, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17680000 (10788473 pages from the captures)
[hos host] vk: 17 textures retired (1226 left, 2270 MB): budget 17, drawn over 0, dead targets 0
[hos host] vk: queues ordered by their events: 150000 submissions stopped at a wait, 149938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 440670, not met 0; captures at a CPU event's wait 13793, 0 of them at a wait met on arrival
[hos host] vk: sparse mirrors: 123, 1004 MB with memory of 1472 MB (467 MB spared); blocks given memory later 13843, given back 18946; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1476 binds (356.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 398 taken from the spares, 166 made new; 441 kept as spares, 0 destroyed; 43 spares now, 1368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2460 dirty pages from their submits' captures, 50474 from earlier ones', 45 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 910 dispatches, 39771 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8372 runs, 630.5 MB compared; 73 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19628 in captures, 1 only in the collection, 25 only by the replay itself
[hos host] vk: read live, by pool: 0x7fa8400000 (4 MB, captured, state 2): 45;
[hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos host] vk: replay reads of guest memory as submitted: 17700000 (10800966 pages from the captures)
[hos host] vk: 20 textures retired (1210 left, 2267 MB): budget 20, drawn over 0, dead targets 0
[hos host] vk: 1 textures retired (1209 left, 2267 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17720000 (10812909 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 564 ranges, 3332 MB), 280 pages written, 280 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.029; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (170 MB), 10 spare buffers, 5896 texture pages copied (23 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #325000: event 0x100d6c8004 mode 0 value 324999 (lr 0x80291504)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1043 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 15 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 851 kept their virtual core (nothing there to wait for), 2733 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 17740000 (10825151 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 799, 3550.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1852 dirty pages from their submits' captures, 50125 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 38811 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8140 runs, 594.9 MB compared; 85 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19605 in captures, 0 only in the collection, 22 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 8 textures retired (1204 left, 2261 MB): budget 8, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17760000 (10836493 pages from the captures)
[hos host] vk: 13 textures retired (1192 left, 2244 MB): budget 13, drawn over 0, dead targets 0
[hos host] vk: program picks: 25544529 kept (10613961 by the last picks' textures, a uniform buffer bound since), 26745 new, 9228725 with other textures (of 34800000)
[hos host] vk: 10 textures retired (1193 left, 2253 MB): budget 10, drawn over 0, dead targets 0
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15283 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 17780000 (10848295 pages from the captures)
[hos host] vk: sparse mirrors: 123, 1004 MB with memory of 1472 MB (468 MB spared); blocks given memory later 13845, given back 18953; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1480 binds (356.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 398 taken from the spares, 166 made new; 441 kept as spares, 0 destroyed; 43 spares now, 1368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 775, 3441.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1547 dirty pages from their submits' captures, 47350 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 834 dispatches, 36972 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 8152 runs, 616.5 MB compared; 62 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19097 in captures, 0 only in the collection, 19 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos host] vk: a frame: 345 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 12.64, present 0.17, draw 7.99, descriptors 3.27, texture lookups 0.14, uploads 0.02, pipelines 0.35, programs 1.43, commands 2.31, dispatch 1.04, memory sync 0.57, queue submit 0.00, slot wait 0.27
[hos host] vk: a frame: mirrors 564 (1004 MB): 6.32 MB synced in 1201.2 copies, 19.18 MB written back
[hos host] vk: 22440 presents, 58.0 fps; a frame: 2513 draws (0 skipped), 216 dispatches, 228 passes, 135 barriers, 6 submits, 0.5 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14233 pipelines (3752 ms), 0 programs (10 failed), 1193 textures (2253 MB)
[hos host] vk: host time in GPU resource waits: 5.04 ms a frame (fences 1.44, timelines 3.61, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 808.6 full barriers recorded, 317.4 left out (nothing recorded since the one before), 48.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 70.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 126.0, an op 1357.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.4 ms a frame (the longest submission 3.8 ms); 0.5 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.5 (0.4 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 17800000 (10859974 pages from the captures)
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 800000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040)
[hos host] vk: late pages found at the replay (600 takes): 20 pages in 20 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 17820000 (10873649 pages from the captures)
[hos host] vk: 11 textures retired (1186 left, 2221 MB): budget 11, drawn over 0, dead targets 0
[hos host] vk: 5 textures retired (1181 left, 2211 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2103 dirty pages from their submits' captures, 47461 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 852 dispatches, 37422 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7770 runs, 601.1 MB compared; 61 pages read into the shadows; 288 MB of shadows (55 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18819 in captures, 0 only in the collection, 23 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 17840000 (10885931 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 566 ranges, 3358 MB), 274 pages written, 274 copied (1.1 MB)
[hos host] vk: a capture taken (ms): all 0.036; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 24 kept (118 MB), 11 spare buffers, 5982 texture pages copied (23 MB)
[hos host] vk: 7 textures retired (1188 left, 2228 MB): budget 7, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 17860000 (10898607 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 17880000 (10910585 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 7438107 pages so far (0 held: may be written ahead in their submission)
[hos sead::MainThread] nvn: EventSignal (CPU) #45000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: program picks: 25701719 kept (10674154 by the last picks' textures, a uniform buffer bound since), 26746 new, 9271534 with other textures (of 35000000)
[hos host] vk: sparse mirrors: 127, 1014 MB with memory of 1496 MB (481 MB spared); blocks given memory later 13845, given back 18954; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1485 binds (358.2 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 402 taken from the spares, 166 made new; 441 kept as spares, 0 destroyed; 39 spares now, 1344 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 775, 3441.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 3353 dirty pages from their submits' captures, 50299 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 877 dispatches, 37835 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 7320 runs, 581.3 MB compared; 59 pages read into the shadows; 288 MB of shadows (55 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18659 in captures, 0 only in the collection, 30 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos host] vk: replay reads of guest memory as submitted: 17900000 (10923077 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 17920000 (10935147 pages from the captures)
[hos host] vk: 3 textures retired (1204 left, 2237 MB): budget 3, drawn over 0, dead targets 0
[hos sead::MainThread] nvn: GPU waits in 5 s: 278 sync waits (955 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1693 dirty pages from their submits' captures, 47888 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 892 dispatches, 37601 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6859 runs, 621.3 MB compared; 64 pages read into the shadows; 288 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18238 in captures, 0 only in the collection, 25 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: 7 textures retired (1207 left, 2236 MB): budget 7, drawn over 0, dead targets 0
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #445000: event 0x100d6c800c mode 0 value 327699 (now 327680) (lr 0x80660e38)
[hos host] vk: replay reads of guest memory as submitted: 17940000 (10947248 pages from the captures)
[hos Presentation Thread] vi: 22546 frames presented (56.5 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 12, 15.5-17.8 169, 17.8-22 91, 22-30 11, 30-40 0, over 40 0
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1034 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 944 kept their virtual core (nothing there to wait for), 3152 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 17960000 (10959457 pages from the captures)
[hos host] vk: 5 textures retired (1224 left, 2251 MB): budget 5, drawn over 0, dead targets 0
[hos host] vk: a frame: 363 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 13.66, present 0.17, draw 8.65, descriptors 3.54, texture lookups 0.24, uploads 0.07, pipelines 0.36, programs 1.57, commands 2.51, dispatch 1.00, memory sync 0.57, queue submit 0.00, slot wait 0.29
[hos host] vk: a frame: mirrors 570 (1017 MB): 6.33 MB synced in 1175.7 copies, 18.78 MB written back
[hos host] vk: 22560 presents, 54.6 fps; a frame: 2676 draws (0 skipped), 174 dispatches, 230 passes, 135 barriers, 6 submits, 0.7 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14241 pipelines (3757 ms), 0 programs (10 failed), 1224 textures (2251 MB)
[hos host] vk: host time in GPU resource waits: 5.12 ms a frame (fences 1.32, timelines 3.80, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 769.4 full barriers recorded, 317.7 left out (nothing recorded since the one before), 49.0 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 70.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 126.1, an op 1472.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.5 ms); 0.7 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.7 (0.8 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.11 MB unswizzled on the CPU in 0.01 ms
[hos Presentation Thread] vk: capture: a submit 0.20 ms (lock 0.00, scan 0.04 over 568 ranges, 3376 MB), 307 pages written, 307 copied (1.2 MB)
[hos host] vk: a capture taken (ms): all 0.035; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (88 MB), 12 spare buffers, 7123 texture pages copied (27 MB)
[hos sead::MainThread] nvn: 2950000 recordings made, 66726 live copies (25.3 MB), 6923 texture copies; 5 command buffers' segments (4.3 MB)
[hos host] vk: sparse mirrors: 129, 1017 MB with memory of 1504 MB (486 MB spared); blocks given memory later 13846, given back 18957; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1489 binds (358.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 404 taken from the spares, 166 made new; 441 kept as spares, 0 destroyed; 37 spares now, 1336 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2242 dirty pages from their submits' captures, 49540 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 956 dispatches, 38987 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6605 runs, 617.1 MB compared; 83 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17465 in captures, 0 only in the collection, 20 only by the replay itself
[hos host] vk: read live, by pool: 0x1042000000 (4 MB, captured, state 2): 1; 0x1041c00000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos host] vk: 5 textures retired (1247 left, 2266 MB): budget 0, drawn over 0, dead targets 5
[hos host] vk: replay reads of guest memory as submitted: 17980000 (10972355 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11948 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 25858579 kept (10736833 by the last picks' textures, a uniform buffer bound since), 26751 new, 9314669 with other textures (of 35200000)
[hos host] vk: replay reads of guest memory as submitted: 18000000 (10985186 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 13 pages in 13 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05
[hos host] vk: recorder: 7589637 packets run in 10.0 s (busy 2.02 s, spun idle 0.13 s, slept 12026 times, 4084 of them under 200 us), 0 calls made directly (no room); 19492 batches over 567 frames (34.4 a frame, 389 packets a batch), 10300 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 88 times, 8.64 ms (0.2 times, 0.02 ms a frame); every thread 869 times, 86.17 ms in all, the longest 1.03 ms; vkWaitForFences 515/45.41 ms (max 0.51), vmaDestroyBuffer 103/17.22 ms (max 1.03), the queue's lock 74/8.42 ms (max 0.47), vkDestroyImageView 50/5.41 ms (max 0.43), vmaFreeMemory 26/4.10 ms (max 0.82), vkQueueBindSparse 51/2.88 ms (max 0.30)
[hos host] vk: recorder: of its busy 2.02 s, 0.59 s in parts' ends and submits (the longest 1.50 ms), 0.24 s of that blocked on the queue's lock (373 times, the longest 1.33 ms), so 0.35 s in the ends, submits and holds; 1.43 s running the other packets; 13386 packets a frame
[hos host] vk: replay reads of guest memory as submitted: 18020000 (10996711 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1904 dirty pages from their submits' captures, 49693 from earlier ones', 1 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 943 dispatches, 39089 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6555 runs, 637.4 MB compared; 74 pages read into the shadows; 288 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18941 in captures, 1 only in the collection, 17 only by the replay itself
[hos host] vk: read live, by pool: 0x7fb6000000 (4 MB, captured, state 2): 1;
[hos host] vk: replay reads of guest memory as submitted: 18040000 (11008256 pages from the captures)
[hos host] vk: 6 textures retired (1265 left, 2274 MB): budget 0, drawn over 6, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 18060000 (11019795 pages from the captures)
[hos aal::LowPrioWorkerThread] vcores: 146564 takes, 24783 waits (mean 82 us, longest 21.7 ms, 0 over HOS_VCORE_LOG_MS), 11725 kicks, 11942 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 129, 1014 MB with memory of 1508 MB (493 MB spared); blocks given memory later 13846, given back 18970; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1492 binds (359.1 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 405 taken from the spares, 166 made new; 442 kept as spares, 0 destroyed; 37 spares now, 1332 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2204 dirty pages from their submits' captures, 49678 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 926 dispatches, 39393 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6475 runs, 627.2 MB compared; 70 pages read into the shadows; 288 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18339 in captures, 0 only in the collection, 18 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos host] vk: 7 textures retired (1258 left, 2268 MB): budget 6, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 18080000 (11031331 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 568 ranges, 3611 MB), 240 pages written, 240 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 24 kept (128 MB), 12 spare buffers, 7123 texture pages copied (27 MB)
[hos host] vk: program picks: 26015220 kept (10800824 by the last picks' textures, a uniform buffer bound since), 26751 new, 9358028 with other textures (of 35400000)
[hos host] vk: 7 textures retired (1275 left, 2254 MB): budget 7, drawn over 0, dead targets 0
[hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #330000: event 0x100d6c8008 mode 0 value 329997 (lr 0x80660f34)
[hos host] vk: replay reads of guest memory as submitted: 18100000 (11042877 pages from the captures)
[hos host] vk: 13 textures retired (1262 left, 2245 MB): budget 13, drawn over 0, dead targets 0
[hos host] vk: readback pages refreshed from guest memory: 7508580 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 1963 dirty pages from their submits' captures, 48468 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 38797 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6375 runs, 630.3 MB compared; 75 pages read into the shadows; 288 MB of shadows (55 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18061 in captures, 0 only in the collection, 14 only by the replay itself
[hos host] vk: read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 18120000 (11054422 pages from the captures)
[hos host] vk: 2 textures retired (1268 left, 2250 MB): budget 2, drawn over 0, dead targets 0
[hos host] vk: a frame: 350 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 11.85, present 0.16, draw 7.65, descriptors 3.11, texture lookups 0.18, uploads 0.05, pipelines 0.33, programs 1.37, commands 2.24, dispatch 0.89, memory sync 0.40, queue submit 0.00, slot wait 0.27
[hos host] vk: a frame: mirrors 571 (1014 MB): 5.73 MB synced in 1112.1 copies, 18.07 MB written back
[hos host] vk: 22680 presents, 58.9 fps; a frame: 2557 draws (0 skipped), 146 dispatches, 226 passes, 143 barriers, 6 submits, 1.1 uploads (0.1 MB), 81 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 14241 pipelines (3757 ms), 0 programs (10 failed), 1280 textures (2259 MB)
[hos host] vk: host time in GPU resource waits: 6.27 ms a frame (fences 1.67, timelines 4.60, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 730.8 full barriers recorded, 323.9 left out (nothing recorded since the one before), 48.9 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.6; 0x8: 3.0; 0x9: 1.0; 0x10: 0.6; 0x11: 39.6; 0x12: 70.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.2, an op 1439.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0;
[hos host] vk: GPU time 12.2 ms a frame (the longest submission 3.2 ms); 1.1 textures made a frame: colour targets 0.2 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.9 (0.7 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.08 MB unswizzled on the CPU in 0.01 ms
[hos host] vk: 1 textures retired (1291 left, 2269 MB): budget 1, drawn over 0, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 18140000 (11066488 pages from the captures)
[hos sead::MainThread] vk: readbacks of GPU results: 1020 of 3000 waited for their write-back, 0.09 ms each; 0 looked at the stopped queues first, 9 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 1039 kept their virtual core (nothing there to wait for), 3057 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04
[hos host] vk: a submission's parts ended: 116043 where the game had ended its render pass, 23957 at 512 draws
[hos host] vk: late pages' fix (600 recordings with one): of 1139 captures with late pages 1136 fixed (91.6 pages, 92.2 fixes a capture), waited as before: 0 no pool, 3 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 444 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 896 blocks (57344 KB); of the reads on the CPU, 5 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 163, of an earlier recording 276, not recorded yet 0, of this part 0; pages fixed from stamps 103080, from snapshots 1009; the shadows: 555761 pages synced from captures, 542355 made known by it, 13406 left unknown because the replay or the collection wrote them since
[hos host] vk: replay reads of guest memory as submitted: 18160000 (11079306 pages from the captures)
[hos host] vk: sparse mirrors: 130, 1019 MB with memory of 1516 MB (496 MB spared); blocks given memory later 13846, given back 18970; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1493 binds (359.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 406 taken from the spares, 166 made new; 442 kept as spares, 0 destroyed; 36 spares now, 1324 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 849, 3772.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far
[hos host] vk: mirror syncs (600 submits): 2356 dirty pages from their submits' captures, 46235 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 869 dispatches, 37332 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5948 runs, 604.1 MB compared; 96 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16838 in captures, 0 only in the collection, 13 only by the replay itself
[hos host] vk: read live, by pool: 0x7fd3800000 (4 MB, captured, state 2): 1;
[hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9978 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 18180000 (11092354 pages from the captures)
[hos host] vk: 10 textures retired (1369 left, 2417 MB): budget 10, drawn over 0, dead targets 0
[hos host] vk: program picks: 26170057 kept (10862153 by the last picks' textures, a uniform buffer bound since), 26761 new, 9403181 with other textures (of 35600000)
[hos host] vk: replay reads of guest memory as submitted: 18200000 (11105379 pages from the captures)
[hos host] vk: 41 textures retired (1396 left, 2501 MB): budget 3, drawn over 0, dead targets 38
[hos host] vk: late pages found at the replay (600 takes): 19 pages in 19 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: slow upload 15.0 ms: #27966 0x10c9e7e200 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 568 ranges, 3663 MB), 252 pages written, 252 copied (1.0 MB)
[hos PlacementMgr] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 42 made): 15235183 lookups, 123849 missed (0.81 %)
[hos host] vk: a capture taken (ms): all 0.030; GPU wait for late pages 0.009 (2 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (154 MB), 9 spare buffers, 7777 texture pages copied (30 MB)
[hos host] vk: program census (close): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 31 program records loaded (30.6 ms), 10 new graphics pipelines, pipelines from the driver's cache 14, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms)
[hos host] vk: program loads at create (close): 6566 queued, 6566 loaded (median 1.46 ms, 99th percentile 16.63 ms, worst 38.13 ms, 16513 ms and 351.7 MB read in all); 8876 programs (1634 of them compute variants without their code, 0 graphics programs no note names), 12373 graphics pipelines (1 not in a cache), 746 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2494 pipelines made at create taken by draws; 3157 create calls with nothing kept for their stages, 0 before the index; units 9723 (6566 ready, 0 loading), 17816 programs alive; 4 noted graphics pipeline probes deferred to first use
[hos host] vk: program loads at create (close): 14110 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them)
[hos host] vk: module identifiers (close): 792 known, 35 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before)
[hos host] vk: indexed descriptor arrays (close): 1343 in 1343 of 20151 program stages, 0 of them near a limit
[hos host] vk: indexed descriptor arrays (close), fragment stages: 1343 texture (17-64: 1343); the longest 64
[hos host] vk: page table census (close): no pool outside its windows, 64-72 GB and 508-512 GB (4456 pools in the table now)
[hos host] vk: pipeline cache: play file 11 saved (4 KB): 1 pipelines, 1 from the driver's cache, 0 compiled; 68 more from the start's cache
[hos host] vk: the window is closing: the swapchain is let go with its images (what the compositor needs back for its final frames)
[hos host] display: window closed, exiting
[hos host] stop: close work done (a file the game was writing waited for, at most 3 s)

Screenshots or video

No response

Before you send it

  • The problem is still there on the newest release or nightly.
  • I searched the open issues for it.
  • I'm not attaching game files, keys, the game's executable, its .bin caches or saves.
### Longshot version Nightly 680 9dd41326 ### Channel Nightly ### What kind of problem Graphics (missing, black or wrong-coloured things, lighting, flicker) ### What happened Newest Nightly release broke Hyrule field the field appears segmented with seams between chunks? seams disappear when you get close to them but return when far away. ### Where in the game Hyrule Field ### How often Every time ### Graphics card and driver version RX 6750XT ### CPU, RAM and operating system Ryzen 7 5700X 32GB ddr4, win11 ### Changes from the defaults I have the no fog mod installed but tested the bug with with it off. ### longshot_log.txt [hos host] longshot nightly-680-9dd41326 (source e287b3d8) [hos host] run_id: bf8f392313b2578d0aeadbc8c186b6e4 [a64] unimplemented import _ZN2nn2ro12ProtectRelroEPKvS2_S2_S2_S2_ (lr 0x81237908) [a64] unimplemented import _ZN2nn2vi36SetDisplay120HzRefreshRateCapabilityEPNS0_7DisplayEb (lr 0x8348bf64) [hos host] crash window: watching this run (HOS_CRASH_WATCH=0: off) [hos host] options: the settings file is D:\Fan games and Decomps\Longshot Ocarina of Time Remake\LongshotLauncher-data\user\longshot_options.ini [hos host] oe: output resolution 2560x1440 [hos host] boot: widescreen: a 1.7778 screen is 16:9: the game's own picture [hos host] boot: image 96 MB at 0x80000000, 9809 MB of memory for the game [hos host] boot: 32694 MB of memory, 17408 MB of swap [hos host] mods: the list (the order and the switches) is D:\Fan games and Decomps\Longshot Ocarina of Time Remake\LongshotLauncher-data\user\mods.json [hos host] mods: Remove Fog: off (the Mods tab) [hos host] mods: 1 mods from D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\mods: 0 RomFS entries [hos host] vk: reserved core selection: 8 cores, 16 logical, best class=0, fast cores=8, selected=0xc000 (of the logical processors 0xffff) [hos host] vk: core 0: mask=0x3, class=0, popcount=2 [hos host] vk: core 1: mask=0xc, class=0, popcount=2 [hos host] vk: core 2: mask=0x30, class=0, popcount=2 [hos host] vk: core 3: mask=0xc0, class=0, popcount=2 [hos host] vk: core 4: mask=0x300, class=0, popcount=2 [hos host] vk: core 5: mask=0xc00, class=0, popcount=2 [hos host] vk: core 6: mask=0x3000, class=0, popcount=2 [hos host] vk: core 7: mask=0xc000, class=0, popcount=2 [hos host] vk: reserved core mask (enhanced): 0xc000 (8 cores; a core the process may run on wholly, of 0xffff) [hos MainThread] mem: the working set's minimum is 4096 MB of the PC's 32694 MB: Windows keeps that much of the game in RAM before it trims below it (HOS_WS_FLOOR_MB=0: Windows' own) [hos MainThread] boot: _init_libc0 (0x85717e50) [hos MainThread] boot: nnosInitialize (0x8510f720) [hos MainThread] libc: memcpy, memmove, memset, memcmp, strlen and strcmp are the host's (HOS_GUEST_LIBC=1: the game's own) [hos MainThread] boot: _init_libc1 (0x85717e60) [hos MainThread] boot: nninitStartup (0x8262bb10) [hos MainThread] heap: 6894 MB at 0x1000000000 [hos MainThread] mem: 0x1000000040 committed on first touch (nothing handed it out) [hos MainThread] mem: 0x1000010000 committed on first touch (nothing handed it out) [hos MainThread] boot: _init_libc2 (0x85717e80) [hos MainThread] boot: main DT_INIT (0x812377e0) [hos MainThread] boot: subsdk0 DT_INIT (0x84ec0040) [hos MainThread] boot: sdk DT_INIT (0x84fa0060) [hos MainThread] boot: nnMain (0x81222cd0) [hos MainThread] mem: 0x1002400010 committed on first touch (nothing handed it out) [hos sead::MainThread] mem: 0x10024109c0 committed on first touch (nothing handed it out) [hos sead::MainThread] fs: mounted content: -> D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\romfs [hos sead::MainThread] mem: 0x11aecffff2 committed on first touch (nothing handed it out) [hos sead::MainThread] mem: 0x11aed00010 committed on first touch (nothing handed it out) [hos sead::MainThread] mem: 0x11aed27998 committed on first touch (nothing handed it out) [hos sead::MainThread] zstd: the game's whole-buffer decode on the host's libzstd 1.5.7 (HOS_NATIVE_ZSTD_FRAMES) [hos sead::MainThread] mem: 0x1003f01a88 committed on first touch (nothing handed it out) [hos sead::MainThread] mem: 0x11aedfff52 committed on first touch (nothing handed it out) [hos sead::MainThread] oe: language en-US (language.txt) [hos sead::MainThread] oe: GetCurrentGraphicsScale -> 1.00 [hos sead::MainThread] gs: BootGraphicsSystem(0, 0): no graphics server yet (headless) [hos host] display: fullscreen (F11 for a window) [hos sead::MainThread] nvn: native NVN layer: 429 procs [hos sead::MainThread] nvn: native NVN device (flags 0): no Switch GPU emulation [a64] unimplemented import _ZN2nn3aoc17CountAddOnContentEv (lr 0x83392a38) [hos sead::MainThread] nvn: window origin upper left [hos sead::MainThread] nvn: depth mode near = -w (GL) [hos sead::MainThread] mem: 0x1004032000 committed on first touch (nothing handed it out) [hos sead::MainThread] nvn: texture pool of 33068 descriptors [hos sead::MainThread] vi: GetNativeWindow -> 0x1102400 [hos sead::MainThread] mem: 0x1004133000 committed on first touch (nothing handed it out) [hos sead::MainThread] mem: 0x1006e37000 committed on first touch (nothing handed it out) [hos sead::MainThread] nvn: window of 3 textures (3 active), 2560x1440 format 0x25 [hos sead::MainThread] mem: 0x1006f3b000 committed on first touch (nothing handed it out) [hos sead::MainThread] mem: 0x10072b8020 committed on first touch (nothing handed it out) [hos sead::MainThread] mem: 0x10073b9000 committed on first touch (nothing handed it out) [hos sead::MainThread] mem: 0x10073ca000 committed on first touch (nothing handed it out) [hos sead::MainThread] sleep: at most 4 of the game's threads yield through a short sleep at a time, the others wait on the timer (16 logical processors; HOS_SHORT_SLEEP_POLLERS=0: all yield) [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 1 made): 65537 lookups, 37293 missed (56.90 %) [hos host] vk: Intel XeSS: no libxess.dll in D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\bin, D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326 (the readme says where Intel's file comes from) [hos ParallelDecompThread] zstd: 1 stream(s) decoding on the host at once [hos sead::MainThread] applet: render target pool 512 MB (the game's 305 MB) [hos Resource Memory] mem: the game's mapped pages 360 MB (peak 360); heap 6894 MB; the console's total 9809 MB [hos Resource Memory] mem: the game's mapped pages 512 MB (peak 512); heap 6894 MB; the console's total 9809 MB [hos ParallelDecompThread] zstd: 2 stream(s) decoding on the host at once [hos ARes Loading 1] texts: mals/usen.product.110.sarc.zs <- D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\texts/usen.product.110.sarc.zs (the console's wording made the PC's) [hos host] hid: 55 controls from longshot_options.ini [hos host] vk: the Output resolution's sizes: 31 (the screen's modes, with the common sizes of their shapes, and the game's own); shapes 5:4, 4:3, 16:10, 16:9 [hos host] vk: device 0: AMD Radeon RX 6750 XT, Vulkan 1.4, driver 0x800161, type 2, id 1002:73df [hos host] vk: float atomic add on storage buffers: compare-exchange loop, by default only with NVIDIA's driver and Intel's Windows driver on Meteor Lake, Arrow Lake, Panther Lake and Battlemage (RX_VK_NATIVE_FADD=1: native where the GPU has it) [hos host] vk: subgroups: default 64 lanes, 32-64, pinned to 32 on stages 0x20; RADV on GCN 1-4: no [hos host] vk: shader caps: subgroup stages 0x3fff; denorm independence 0, fp32 denorm preserve 1 flush 1, signed zero fp16 1 fp32 1; int64 buffer atomics 1 [hos host] vk: mesh-only and task + mesh programs use the compute fallback (RX_VK_TASK_MESH_FALLBACK=0: native) [hos host] vk: mesh shaders: Automatic, from the default: mesh-only and task + mesh draws on the compute path [hos host] vk: robust buffer access on [hos host] vk: vendor fixes for AMD proprietary driver (26.11.2 (AMD proprietary shader compiler)): out-of-bounds texel fetches 0 1, NaN tests by bits 1, gather offset 1 [hos host] vk: programs are loaded when the game makes them (create): the default, on the compute fallback too (RX_VK_LOAD=start: every listed program before the title) (RX_VK_LOAD=create or start says which) [hos host] vk: the presenter thread waits for its presents to be shown: no [hos host] vk: colour outputs NaN-guarded [hos host] vk: positions invariant (not an NVIDIA card) [hos host] vk: matrix multiplies in software (the cooperative matrices lack 16x8x16 f16) [hos host] vk: software matrix multiplies looped (not RADV or RDNA 4 on AMD's proprietary driver, RX_VK_MMA_UNROLL=1 tries the unrolled form) [hos host] vk: AMD Anti-Lag 2: available (VK_AMD_anti_lag) [hos host] vk: AMD Radeon RX 6750 XT, Vulkan 1.4, driver 0x800161; mesh shaders used, host memory import yes, pipeline libraries yes [hos host] vk: driver: AMD proprietary driver 26.11.2 (AMD proprietary shader compiler) [hos host] vk: policy: facts 1002:73df, driver id 1, version 0x800161 (26.11.2 (AMD proprietary shader compiler)), type 2, Windows; VRAM 12272 MB, allowed at the start 11474 MB, RAM 32694 MB, resizable BAR 1; CPUs 16 (16 usable, 8 cores); mesh shaders 1, float add 0, cooperative matrices 1, 16-bit stage inputs 1, lanes 64 to 64 (32 on stages 0x20) [hos host] vk: policy: driver fixes on driver version 0x800161: fetch_bounds=1 nan_bits=1 gather_offset=1 frag_shuffle_fallback=1 position_invariant=1 equal_depth_bias=0 mesh_no_optimizer=1 amd_retry_no_optimizer=1 amd_cloth_no_optimizer=1 fault_gate_serial=1 task_mesh_fallback=1(seen on 0x800161) mesh_mode=automatic fallback_phi_limit=0 intel_compile_gate=0 compute_builds_serial=0 pcache_fetch_lock=0 split_before_draw=0 astc_on_cpu=0 inline_deku_web_loads=0 networks_left_out=1 dma_buf_imports=0 nv_checkpoints=0 retired=after a wait idle_import_high=0% allowance_at_count_rule=1 [hos host] vk: policy: can and measured: mirrors=1 import_chunk=4 MB ub_snapshot=0 split_ub_flush=1 native_float_add=0 warp64=1 batch_submits=0 mma_unroll_try=0 global_table_loop=0 occ_copy=0 intel_side_by_side_auto=0 fsr4_tune=1 gpu_clock_peak=0 readback_late_newest=0 acquire_ahead=1 pcache_small_pc=0 [hos host] vk: policy: sizes by the allowance at the start (the sites read it later): textures 7158 MB, mirrors 2048, target pool 768, compaction from 512, shadows 1434, staging kept 256, spare captures no limit; blocks at the start 24, of 64 MB (targets 32); count line 97% room_in_blocks_rule=1 squeeze_damped=0 linux_hold_caps_budget=0 small_vram_textures=0 astc_bc_all=0 dlss_cnn=0 [hos host] vk: policy: auto settings and threads: auto_tier=4 auto_render_third=0 rtx=0 rdna4=0; workers 6, capture helpers 4, texture threads 8 (gather 6), program workers in play 4 threaded_auto_render=1 threaded_auto_recorder=1 drop_modules_small_ram=0 native_zstd_keep_core=0 [hos host] vk: policy: switches (* above: set by its switch; the other values are this PC's defaults): none of these is set [hos host] report: gpu AMD Radeon RX 6750 XT, AMD proprietary driver 26.11.2 (AMD proprietary shader compiler) [hos host] report: cpu AMD Ryzen 7 5700X 8-Core Processor (16 threads), 32 GB [hos host] report: os Windows 11 (build 26300) [hos host] report: renderer Vulkan [hos host] vk: switches set: none [hos host] vk: Auto (this PC) would be: 12272 MB of video memory: tier 4; the game's own upscaler (AMD Radeon RX 6750 XT, 12272 MB of video memory, dedicated, screen 3840x2160 at 165 Hz, 32694 MB of RAM, 16 threads) [hos host] vk: surface caps images 2..16 extent 3840x2160..3840x2160 usage 0x9f alpha 0x1 transform 0x1 [hos host] discord: connected (application 1570559772111933578) [hos host] discord: the boss heart containers: 8 of 8 found in the RomFS [hos host] discord: the status: N680 9dd41326 / - [hos host] discord: status set [hos host] vk: swapchain 3840x2160, 4 images, fifo (V-Sync On) [hos host] vk: guest memory imported in place (alignment 4096) [hos host] vk: replay ready (window; push descriptors yes (32), extended dynamic state 3 yes, custom border colours yes) [hos host] vk: driver shader cache at the start: 2968 MB (%USERPROFILE%\AppData\Local\AMD\VkCache 2968 MB) [hos host] vk: VRAM mirror of buffer pools on [hos host] vk: Threaded rendering: Auto: on, the render thread and the recorder thread (8 cores of 16 logical processors) [hos host] vk: Compile shaders: Most (the default): the predicted pipelines are left out; a pipeline no list names is made when it is first drawn [hos host] vk: handle tables as descriptor arrays: on [hos host] vk: pipeline cache file read: not NVIDIA's own driver [hos host] vk: acquire ahead: on by default: the next image asked for right after each present (RX_VK_ACQUIRE_AHEAD=0: off) [hos host] vk: RX_VK_RECORDER=thread: vkCmd* packed on the replay thread and run in order by a recorder thread, handed over by part (RX_VK_RECORDER_PARTS): a part's packets at its end, and its end and submit made by the recorder; the replay waits for it at a reset, begin, destruction or direct recording, and for a part's submit where a wait, a fence or the queue needs it; ring 8192 KB, the recorder waits on the ring after 10 us [hos host] vk: AMD FSR 4: D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\amd_fidelityfx_loader_vk.dll [hos host] vk: ui: the game's own menu font, from its files [hos host] vk: ui: Dear ImGui 1.92.9b on the window's images (format 44) [hos host] vk: fsr3: AMD FSR 3.1 can run on this GPU [hos host] vk: AMD FSR 4: upscalers 4.1.2, 3.1.6, 2.3.4; using 4.1.2 [hos host] menu: the upscalers asked in the background while the start-up screen shows (21 ms) [hos host] vk: images from a pool of 32 MB blocks (memory type 0), 0 made at the start [hos host] vk: presenter thread: each present made on a thread of its own once its copy is done [hos host] vk: AMD Anti-Lag 2: on [hos host] vk: the window's screen is 3840x2160 (16:9): 0 sizes added to the Output resolution's [hos host] vk: pipeline cache D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\cache/longshot_vk_pipelines.bin (1973875 KB) [hos host] vk: pipeline cache: 6 play files merged (968 KB) [hos host] vk: ASTC is decoded on the GPU (Eden's decoder, Switch 2 tiling) [hos host] vk: big ASTC textures are kept as BC3 (re-encoded on the GPU; ASTC as RGBA8 or RX_VK_ASTC_RGBA8=1: as RGBA8) [hos host] vk: big block-linear textures are unswizzled on the GPU from their pools [hos host] vk: texture pipelines made before the game in 1 ms (ASTC decoder yes, BC encoder yes, GPU unswizzle yes) [hos host] vk: program cache D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\cache/longshot_programs.bin: 47798 program keys kept, 147312 KB [hos host] vk: shader store %USERPROFILE%\AppData\Local\Longshot\shaders/: 50685 SPIR-V blobs, for every game folder [hos host] vk: one build of a program goes to the driver at a time (AMD; RX_VK_DRIVER_FAULTS=0: as many as ask) [hos host] vk: vertex format NVN 0x29 -> vk 97 [hos host] vk: vertex format NVN 0x22 -> vk 106 [hos host] vk: vertex format NVN 0x69 -> vk 65 [hos host] vk: vertex format NVN 0x25 -> vk 37 [hos host] vk: vertex format NVN 0x27 -> vk 41 [hos host] vk: vertex format NVN 0x12 -> vk 77 [hos host] vk: vertex format NVN 0x13 -> vk 78 [hos host] vk: vertex format NVN 0x11 -> vk 83 [hos host] vk: vertex format NVN 0x16 -> vk 103 [hos host] vk: vertex format NVN 0x2e -> vk 109 [hos host] vk: vertex format NVN 0x3 -> vk 13 [hos host] vk: the driver's shader cache has 51 of 51 noted pipelines asked (a sample, in 99 ms) [hos host] vk: start-up: pipelines 0 of 4528, 0% of all [hos host] vk: programs: load at create: 37641 notes of 23925 programs indexed, 4 type signatures, 3357 compute programs listed, in 14.5 ms (45798 kept program keys); 2 workers in play, 2 while the game loads; pipelines from: pipeline caches [hos host] vk: precompile: load at create (RX_VK_LOAD=create): 37641 noted pipelines found, none built; 3357 compute programs named (3357 in the shipped list, 325 dispatched in play here); 2264 mesh fallback programs kept ready, 1589 of their noted pipelines with them (167 in no cache); in 680 ms. No other program is kept ready by a list: each is loaded when the game makes it (the notes' file walked 1 times this start, in 11072 ms: start_note_refs) [hos host] vk: presenter thread: the copy's fence 7.33 ms, the present 0.44 ms a present [hos host] vk: start-up: the shaders took 20.9 s [hos host] vk: pipeline cache: what play makes goes into play files (RX_VK_PLAY_CACHE=0: into the one cache) [hos host] rx: renderer: Vulkan (longshot_vk.dll) [hos host] vk: the replay runs on a thread of its own, the render thread (RX_VK_ASYNC=0: on the submitting thread) [hos host] vk: the replay thread runs on logical processors 0xc000, kept from the game's threads (its ideal processor there; free to go to the other fastest cores when another program takes it) [hos host] vk: the replay thread may run on logical processors 0xffff [hos host] rx: 123 texture and sampler calls made before the renderer started (33 texel writes), given to it [hos sead::MainThread] nvn: QueueSubmitCommands #1: queue 00001000010FBF90 (its #1), 1 handles (lr 0x812c3cf4) [hos sead::MainThread] vk: submit-time capture of written pages on [hos sead::MainThread] mem: the write marks' walk: 256 marks a test, else 64, in 32-byte loads (AVX2) (HOS_COLLECT_VECTOR=0: word loads) [hos sead::MainThread] vk: RX_VK_READBACK_EARLY left off: the render thread is on (RX_VK_ASYNC, Threaded rendering), which it was not measured with; reads of GPU results wait for their whole submission [hos host] [debug] vk: 78 pools mapped [hos host] vk: texture storage in the submit-time capture on [hos host] vk: a 32 MB capture buffer made ahead in 7.8 ms, off the game's thread [hos host] vk: a 64 MB capture buffer made ahead in 13.9 ms, off the game's thread [hos host] vk: a 128 MB capture buffer made ahead in 24.9 ms, off the game's thread [hos host] vk: sparse VRAM mirrors (RX_VK_MIRROR_SPARSE): on [hos host] vk: occlusion queries on (RX_VK_OCC_QUERIES): the game's samples-passed reports get the GPU's counts; read from the pool [hos host] vk: render target pool: up to 768 MB (RX_VK_TARGET_POOL_MB) [hos host] vk: render targets' images from blocks of their own (32 MB, memory type 0) [hos host] vk: texture 0x103d650000 32x32x32 x1 levels 1 NVN 0x26 -> vk 38 (1 textures, 0 MB) [hos host] [debug] vk: a staging arena of 16 MB took 6.7 ms to make [hos host] vk: staging arenas: 1 (16 MB in this one) [hos host] vk: 1 pools imported [hos host] vk: sparse mirrors' buffers are kept for the next mirror of their size, up to 4096 MB of address space (RX_VK_SPARSE_SPARE_MB=0: destroyed at once, as before) [hos host] vk: pool 0x103d800000 (4 MB): mirror made at its first binding (1 so) [hos host] vk: batched submits off (RX_VK_BATCH) [a64] unimplemented import _ZN2nn5audio34GetAudioRendererVoiceDropParameterENS0_19AudioRendererHandleE (lr 0x81044370) [a64] unimplemented import _ZN2nn5audio39IsDeviceOutputVolumeAutoTuneModeEnabledEv (lr 0x81043560) [a64] unimplemented import _ZN2nn5audio36AcquireAudioOutputDeviceNotificationEPNS_2os11SystemEventE (lr 0x80f72978) [a64] unimplemented import _ZN2nn5audio40SetDeviceOutputVolumeAutoTuneModeEnabledEb (lr 0x80f729e0) [a64] unimplemented import _ZN2nn5audio26GetAudioDeviceOutputVolumeEPKNS0_15AudioDeviceNameE (lr 0x80f72e00) [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 1 made): 222210 lookups, 73443 missed (33.05 %) [hos GeneralFormatAsync_] vcores: 3363 takes, 472 waits (mean 1273 us, longest 14.8 ms, 0 over HOS_VCORE_LOG_MS), 184 kicks, 168 yields, 0 starved grants, 0 escapes in 22 s [hos ARes Loading 0] prompts: switch icons, 16 buttons on these pad's controls [hos ParallelDecompThread] zstd: the game's stream decode on the host's libzstd 1.5.7 (HOS_NATIVE_ZSTD=1, this GPU's default: the core given up while it decodes) [hos sead::MainThread] audio: renderer opened [hos host] audio: stereo for the game (the selected/default output has 2 channels) [hos sead::MainThread] audio: device sink on mix 0, 6 channels (0 1 2 3 4 5) [hos host] audio: output on the default device (3 devices), stereo [a64] unimplemented import _ZN2nn5prepo30GetOverSaveFrequencyLimitEventEv (lr 0x80e5ec00) [hos sead::MainThread] fs: mounted save: -> D:\Fan games and Decomps\Longshot Ocarina of Time Remake\LongshotLauncher-data\user\save\5a3c9e1f0b7d24680000000000000001 [hos BootupBackgroundWorkThread] fs: mounted savt: -> D:\Fan games and Decomps\Longshot Ocarina of Time Remake\LongshotLauncher-data\user\temp [hos GeneralFormatAsync_] vcores: GeneralFormatAsync_ (priority 24) waited 131.4 ms for a core: given by sead::MainThread (priority 16), which held it 131.4 ms and was never asked; the cores it may use now: core 0: it; [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 68.7 ms for a core: given by GeneralFormatAsync_ (priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 0: it; [hos PlayReporter] vcores: PlayReporter (priority 29) waited 117.3 ms for a core: given by PreActorModelCompose (priority 26), which held it 0.0 ms and was never asked; the cores it may use now: core 0: it; [a64] unimplemented import _ZN2nn5audio34SetAudioRendererVoiceDropParameterENS0_19AudioRendererHandleEf (lr 0x836f755c) [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 498.5 ms for a core: given by sead::MainThread (priority 16), which held it 580.6 ms and was never asked; the cores it may use now: core 0: it; [hos host] vk: module identifiers: 757 known (longshot_module_ids.bin): a learned compute variant's pipeline is made at its program's create call without its code [a64] unimplemented import _ZN2nn4nifm10InitializeEv (lr 0x8119aea8) [a64] unimplemented import _ZN2nn2vi30QueryDisplayCurrentRefreshRateEPNS0_11RefreshRateEPbPKNS0_7DisplayE (lr 0x83b6eab8) [a64] unimplemented import _ZN2nn3hid25SetGyroscopeZeroDriftModeERKNS0_19SixAxisSensorHandleERKNS0_22GyroscopeZeroDriftModeE (lr 0x834898d0) [a64] unimplemented import _ZN2nn6socket10InitializeERKNS0_6ConfigE (lr 0x83c55b24) [a64] unimplemented import _ZN2nn3ssl10InitializeEv (lr 0x83c55b30) [a64] unimplemented import _ZN2nn6socket13IsInitializedEv (lr 0x824d92f8) [a64] unimplemented import _ZN2nn5audio15AddBiquadFilterEPNS0_19AudioRendererConfigEPNS0_16BiquadFilterTypeEPNS0_10SubMixTypeE (lr 0x80dc7494) [a64] unimplemented import _ZN2nn5audio26SetBiquadFilterInputOutputEPNS0_16BiquadFilterTypeEPKaS4_i (lr 0x80dc740c) [a64] unimplemented import _ZN2nn5audio22SetBiquadFilterEnabledEPNS0_16BiquadFilterTypeEb (lr 0x80de9464) [a64] unimplemented import _ZN2nn6socket7EventFdEmNS0_12EventFdFlagsE (lr 0x82213be4) [a64] unimplemented import _ZN2nn6socket5CloseEi (lr 0x82213bf4) [a64] unimplemented import _ZN2nn5audio24SetBiquadFilterParameterEPNS0_16BiquadFilterTypeERKNS0_22TBiquadFilterParameterIfEE (lr 0x836fcdb4) [a64] unimplemented import _ZN2nn4nifm22GetCurrentProxySettingEPNS0_12ProxySettingE (lr 0x82378f6c) [a64] unimplemented import _ZN2nn5audio8AddDelayEPNS0_19AudioRendererConfigEPNS0_9DelayTypeEPvmPNS0_10SubMixTypeENS_8TimeSpanEi (lr 0x80dd53b8) [a64] unimplemented import _ZN2nn5audio19SetDelayInputOutputEPNS0_9DelayTypeEPKaS4_i (lr 0x80dd531c) [a64] unimplemented import _ZN2nn5audio15SetDelayEnabledEPNS0_9DelayTypeEb (lr 0x80de9464) [a64] unimplemented import _ZN2nn4nifm18IsNetworkAvailableEv (lr 0x82145cec) [a64] unimplemented import _ZN2nn5audio14IsDelayEnabledEPKNS0_9DelayTypeE (lr 0x8119f138) [a64] unimplemented import _ZN2nn5audio12GetDelayTimeEPKNS0_9DelayTypeE (lr 0x8119f164) [a64] unimplemented import _ZN2nn5audio12SetDelayTimeEPNS0_9DelayTypeENS_8TimeSpanE (lr 0x8119f178) [a64] unimplemented import _ZN2nn5audio15SetDelayDryGainEPNS0_9DelayTypeEf (lr 0x8119f184) [a64] unimplemented import _ZN2nn5audio20SetDelayFeedbackGainEPNS0_9DelayTypeEf (lr 0x8119f190) [a64] unimplemented import _ZN2nn5audio14SetDelayInGainEPNS0_9DelayTypeEf (lr 0x8119f19c) [a64] unimplemented import _ZN2nn5audio21SetDelayChannelSpreadEPNS0_9DelayTypeEf (lr 0x8119f1a8) [a64] unimplemented import _ZN2nn5audio21SetDelayLowPassAmountEPNS0_9DelayTypeEf (lr 0x8119f1b4) [a64] unimplemented import _ZN2nn4nifm27SubmitNetworkRequestAndWaitEv (lr 0x8120873c) [a64] unimplemented import _ZN2nn5audio9AddReverbEPNS0_19AudioRendererConfigEPNS0_10ReverbTypeEPvmPNS0_10SubMixTypeEi (lr 0x80de96ec) [a64] unimplemented import _ZN2nn5audio20SetReverbInputOutputEPNS0_10ReverbTypeEPKaS4_i (lr 0x80de965c) [a64] unimplemented import _ZN2nn5audio16SetReverbEnabledEPNS0_10ReverbTypeEb (lr 0x80de9464) [a64] unimplemented import _ZN2nn5audio15IsReverbEnabledEPKNS0_10ReverbTypeE (lr 0x80de91a0) [a64] unimplemented import _ZN2nn5audio18SetReverbEarlyModeEPNS0_10ReverbTypeENS1_9EarlyModeE (lr 0x80de91c4) [a64] unimplemented import _ZN2nn5audio18SetReverbEarlyGainEPNS0_10ReverbTypeEf (lr 0x80de91d0) [a64] unimplemented import _ZN2nn5audio21SetReverbPredelayTimeEPNS0_10ReverbTypeENS_8TimeSpanE (lr 0x80de91e4) [a64] unimplemented import _ZN2nn5audio17SetReverbLateModeEPNS0_10ReverbTypeENS1_8LateModeE (lr 0x80de91f0) [a64] unimplemented import _ZN2nn5audio17SetReverbLateGainEPNS0_10ReverbTypeEf (lr 0x80de91fc) [a64] unimplemented import _ZN2nn5audio18SetReverbDecayTimeEPNS0_10ReverbTypeENS_8TimeSpanE (lr 0x80de9208) [a64] unimplemented import _ZN2nn5audio32SetReverbHighFrequencyDecayRatioEPNS0_10ReverbTypeEf (lr 0x80de9214) [a64] unimplemented import _ZN2nn5audio19SetReverbColorationEPNS0_10ReverbTypeEf (lr 0x80de9220) [a64] unimplemented import _ZN2nn5audio19SetReverbReverbGainEPNS0_10ReverbTypeEf (lr 0x80de922c) [a64] unimplemented import _ZN2nn5audio16SetReverbOutGainEPNS0_10ReverbTypeEf (lr 0x80de9238) [a64] unimplemented import _ZN2nn5audio16SetReverbDryGainEPNS0_10ReverbTypeEf (lr 0x80de9244) [a64] unimplemented import _ZN2nn5audio20GetReverbInputOutputEPKNS0_10ReverbTypeEPaS4_Pii (lr 0x80de9260) [a64] unimplemented import _ZN2nn5prepo27AddInternetConnectionStatusEPNS0_10PlayReportEPKcS4_S4_ (lr 0x80e5bd60) [a64] unimplemented import _ZN2nn5prepo28RequestImmediateTransmissionEv (lr 0x80df26c0) [hos] guest backtrace: lr 0x804fc1ec 0x83b6f194 0x83b6f080 [hos] guest backtrace: lr 0x8348c5e0 0x83b6f08c [hos sead::MainThread] audio in: stream 0 opened: format 2, 1 channel, 48000 Hz, 4800 frames [hos sead::MainThread] audio in: stream 1 opened (raw): format 2, 1 channel, 48000 Hz, 4800 frames [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 54.2 ms for a core: given by sead::MainThread (priority 16), which held it 122.0 ms and was never asked; the cores it may use now: core 0: it; [hos sead::MainThread] vi: QueryDisplayHdrStatus -> SDR [hos sead::MainThread] hid: the game reads npad 0 (No.1, lr 0x83489670) [hos sead::MainThread] sdl: SDL3 [hos sead::MainThread] hid: game window not in front (foreground window: process 7404, class "MozillaWindowClass") [hos sead::MainThread] hid: six-axis handles: 2 asked for, 3 reported [hos ModuleSystemWorker5] tex: the game's texture heap is compacted at most every 1000 ms (HOS_TEX_COMPACT_MS; 0: at every frame, the game's own; off: never) [hos ModuleSystemWorker4] nvn: the game reports counter type 16 (lr 0x8074b278) [hos ModuleSystemWorker4] nvn: the game reports counter type 0 (lr 0x8074b290) [hos sead::MainThread] nvn: window crop 0,0 2560x1440 [hos sead::MainThread] nvn: EventSignal (CPU) #1: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #1: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #1: event 0x100d6c8000 mode 0 value 0 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #2: event 0x100d6c8000 mode 0 value 0 (now 0) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #3: event 0x100d6c8000 mode 0 value 0 (now 0) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #2: queue 000010100D695020 (its #1), 1 handles (lr 0x80666be0) [hos Porter] fs: mounted _npln: -> D:\Fan games and Decomps\Longshot Ocarina of Time Remake\nightly\nightly-680-9dd41326\romfs [hos sead::MainThread] nvn: CommandBufferWaitEvent #4: event 0x100d6c8000 mode 0 value 0 (now 0) (lr 0x806661f8) [hos host] [debug] vk: 1299 pools mapped [hos sead::MainThread] nvn: QueueSubmitCommands #3: queue 000010100D6BF020 (its #1), 1 handles (lr 0x80666be0) [hos Porter] fiber 0x100ac54230: function 0x8504b440 arg 0x100ac541f8 stack 0x100acc1000+0x10000 flags 1 [hos Porter] fiber 0x100acd1090: function 0x8504b440 arg 0x100acd1058 stack 0x100acd4000+0x10000 flags 1 [hos nn.npln.Worker] fiber 0x100ac543d0: function 0x8504b440 arg 0x100ac54398 stack 0x100ace6000+0x10000 flags 1 [hos Porter] fiber 0x100acf7090: function 0x8504b440 arg 0x100acf7058 stack 0x100acfa000+0x10000 flags 1 [hos nn.npln.Worker] fiber 0x100ac3c5d0: function 0x8504b440 arg 0x100ac3c598 stack 0x100ad0a000+0x10000 flags 1 [hos host] vk: a 128 MB capture buffer made ahead in 19.0 ms, off the game's thread [hos sead::MainThread] nvn: EventSignal (CPU) #2: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos Presentation Thread] nvn: QueueSubmitCommands #4: queue 000010100D5EB020 (its #1), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #5: queue 000010100D640020 (its #1), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #6: queue 000010100D6AA020 (its #1), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #7: queue 0000101004133020 (its #1), 1 handles (lr 0x80590358) [hos host] vk: mirror syncs' pages from the captures by a compute pass (RX_VK_MSYNC_PASS) [hos host] [debug] vk: 1699 pools mapped [hos sead::MainThread] nvn: EventSignal (CPU) #3: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #5: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #2: event 0x100d6c8000 mode 0 value 1 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #6: event 0x100d6c8000 mode 0 value 1 (now 0) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #7: event 0x100d6c8000 mode 0 value 1 (now 0) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #8: queue 000010100D695020 (its #2), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #8: event 0x100d6c8000 mode 0 value 1 (now 0) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #9: queue 000010100D6BF020 (its #2), 1 handles (lr 0x80666be0) [hos host] [debug] vk: 1819 pools mapped [hos Presentation Thread] nvn: QueueSubmitCommands #10: queue 000010100D5EB020 (its #2), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #11: queue 000010100D640020 (its #2), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #12: queue 000010100D6AA020 (its #2), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #13: queue 0000101004133020 (its #2), 1 handles (lr 0x80590358) [hos host] vk: 2 pools imported [hos host] vk: written bindings over 1024 KB are written back only where the CPU reads or writes them, after 8 submissions whole (RX_VK_WB_MAX_KB) [hos host] vk: texture 0x1001f0ffe00 2560x1440x1 x1 levels 1 NVN 0x3f -> vk 122 (2 textures, 14 MB) [hos host] vk: texture 0x1004137000 2560x1440x1 x1 levels 1 NVN 0x25 -> vk 37 (3 textures, 28 MB) [hos host] vk: upscaler: faded: while a temporal upscaler or frame generation draws, the depth draws of a skinned mesh that had colour and has had none for 3 frames are left out until its colour is back (RX_VK_DLSS_FADED=0 draws them) [hos host] vk: descriptor slotgen ready: draws keep their descriptors through the registrations of slots they do not sample (menu: Descriptor reuse) [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 5, mip filter linear, LOD bias 0; made with 0, an image of one level: not the row's) [hos host] vk: 3 pools imported [hos host] vk: import 2.5 ms: commit+properties 0.0, buffer 0.0, allocate 2.5, bind+address 0.0 [hos host] vk: pool 0x102dd34020+0x40 imported in 2.6 ms as 0x102dc00000..0x102e000000 (4 imports, 7 ms in all) [hos host] vk: 4 pools imported [hos host] vk: write-back diff on (RX_VK_WB_DIFF=0: off) [hos sead::MainThread] nvn: EventSignal (CPU) #4: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: present texture 0x1004137000 2560x1440 format 0x25: rendered [hos host] vk: present: 2560x1440, format 37 (its image's 37); the Insert menu's HDR 0, HDR10 swapchain no: SDR [hos host] wide ui: picture 0.0000: the HUD at 100% of its 16:9 size, its sides moved 0 and its top and bottom 0 canvas units; the menus at 100% [hos host] vk: GPU hang watch: work that finishes nothing for 30 s ends the game (RX_VK_GPU_HANG_SEC) [hos host] vk: acquire ahead: not now, the swapchain is FIFO (V-Sync On, or the window's only mode): each image asked for at its present [hos Presentation Thread] vi: real-time step on: ratio 3.035, game's frame rate 20 [hos sead::MainThread] nvn: EventSignal (CPU) #5: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #9: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #3: event 0x100d6c8000 mode 0 value 2 (lr 0x80666598) [hos host] vk: late pages' fix on (RX_VK_LATE_FIX=0: the wait, as before) [hos sead::MainThread] nvn: CommandBufferWaitEvent #10: event 0x100d6c8000 mode 0 value 2 (now 0) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #11: event 0x100d6c8000 mode 0 value 2 (now 0) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #14: queue 000010100D695020 (its #3), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #12: event 0x100d6c8000 mode 0 value 2 (now 0) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #15: queue 000010100D6BF020 (its #3), 1 handles (lr 0x80666be0) [hos Presentation Thread] nvn: QueueSubmitCommands #16: queue 000010100D5EB020 (its #3), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #17: queue 000010100D640020 (its #3), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #18: queue 000010100D6AA020 (its #3), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #19: queue 0000101004133020 (its #3), 1 handles (lr 0x80590358) [hos host] vk: submission 11 (queue 0x10100d5eb020) stopped at command buffer 0's wait for event 0x100d6c8000 == 1 (has 0) [hos host] vk: submission 12 (queue 0x10100d640020) stopped at command buffer 0's wait for event 0x100d6c8000 == 1 (has 0) [hos host] vk: submission 13 (queue 0x10100d6aa020) stopped at command buffer 0's wait for event 0x100d6c8000 == 1 (has 0) [hos host] vk: submission 14 (queue 0x101004133020) stopped at command buffer 0's wait for event 0x1006e38004 == 1 (has 0) [hos host] vk: texture 0x1005037000 2560x1440x1 x1 levels 1 NVN 0x25 -> vk 37 (4 textures, 42 MB) [hos sead::MainThread] nvn: EventSignal (CPU) #6: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: the game sets present interval 1 -> 3 [hos sead::MainThread] nvn: the game sets present interval 3 -> 1 [hos host] vk: present texture 0x1005037000 2560x1440 format 0x25: rendered [hos sead::MainThread] nvn: EventSignal (CPU) #7: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos host] [debug] vk: 1857 pools mapped [hos sead::MainThread] nvn: CommandBufferWaitEvent #13: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #4: event 0x100d6c8000 mode 0 value 3 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #14: event 0x100d6c8000 mode 0 value 3 (now 1) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #15: event 0x100d6c8000 mode 0 value 3 (now 1) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #20: queue 000010100D695020 (its #4), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #16: event 0x100d6c8000 mode 0 value 3 (now 1) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #21: queue 000010100D6BF020 (its #4), 1 handles (lr 0x80666be0) [hos Presentation Thread] nvn: QueueSubmitCommands #22: queue 000010100D5EB020 (its #4), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #23: queue 000010100D640020 (its #4), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #24: queue 000010100D6AA020 (its #4), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #25: queue 0000101004133020 (its #4), 1 handles (lr 0x80590358) [hos host] vk: submission 18 (queue 0x10100d5eb020) stopped at command buffer 0's wait for event 0x100d6c8000 == 2 (has 1) [hos host] vk: submission 19 (queue 0x10100d640020) stopped at command buffer 0's wait for event 0x100d6c8000 == 2 (has 1) [hos host] vk: submission 20 (queue 0x10100d6aa020) stopped at command buffer 0's wait for event 0x100d6c8000 == 2 (has 1) [hos host] vk: submission 21 (queue 0x101004133020) stopped at command buffer 0's wait for event 0x1006e38000 == 1 (has 0) [hos host] vk: texture 0x1005f37000 2560x1440x1 x1 levels 1 NVN 0x25 -> vk 37 (5 textures, 56 MB) [hos sead::MainThread] nvn: EventSignal (CPU) #8: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: present texture 0x1005f37000 2560x1440 format 0x25: rendered [hos host] [debug] vk: 1858 pools mapped [hos sead::MainThread] nvn: EventSignal (CPU) #9: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #17: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #5: event 0x100d6c8000 mode 0 value 4 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #18: event 0x100d6c8000 mode 0 value 4 (now 2) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #19: event 0x100d6c8000 mode 0 value 4 (now 2) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #26: queue 000010100D695020 (its #5), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #20: event 0x100d6c8000 mode 0 value 4 (now 2) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #27: queue 000010100D6BF020 (its #5), 1 handles (lr 0x80666be0) [hos Presentation Thread] nvn: QueueSubmitCommands #28: queue 000010100D5EB020 (its #5), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #29: queue 000010100D640020 (its #5), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #30: queue 000010100D6AA020 (its #5), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #31: queue 0000101004133020 (its #5), 1 handles (lr 0x80590358) [hos host] [debug] vk: 1862 pools mapped [hos host] vk: submission 25 (queue 0x10100d5eb020) stopped at command buffer 0's wait for event 0x100d6c8000 == 3 (has 2) [hos host] [debug] vk: 1927 pools mapped [hos host] vk: submission 26 (queue 0x10100d640020) stopped at command buffer 0's wait for event 0x100d6c8000 == 3 (has 2) [hos host] vk: submission 27 (queue 0x10100d6aa020) stopped at command buffer 0's wait for event 0x100d6c8000 == 3 (has 2) [hos host] vk: submission 28 (queue 0x101004133020) stopped at command buffer 0's wait for event 0x1006e38004 == 1 (has 0) [hos sead::MainThread] nvn: EventSignal (CPU) #10: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: present texture 0x1004137000 2560x1440 format 0x25: rendered [hos sead::MainThread] nvn: EventSignal (CPU) #11: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #21: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #6: event 0x100d6c8000 mode 0 value 5 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #22: event 0x100d6c8000 mode 0 value 5 (now 3) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #23: event 0x100d6c8000 mode 0 value 5 (now 3) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #32: queue 000010100D695020 (its #6), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #24: event 0x100d6c8000 mode 0 value 5 (now 3) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #33: queue 000010100D6BF020 (its #6), 1 handles (lr 0x80666be0) [hos Presentation Thread] nvn: QueueSubmitCommands #34: queue 000010100D5EB020 (its #6), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #35: queue 000010100D640020 (its #6), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #36: queue 000010100D6AA020 (its #6), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #37: queue 0000101004133020 (its #6), 1 handles (lr 0x80590358) [hos SceneLowPriority] nfp: the game uses NFC tags: the Insert menu's NFC tab holds one of %USERPROFILE%\AppData\Local\Longshot\nfc's to the reader [hos host] hid: SDL controller support ready (DualSense, DualShock 4, Switch Pro, Xbox and others) [hos host] hid: controller "XInput Controller #1" connected (SDL); guid=0300fa67c82d00000631000014017801; mapping=available; mapping-name="*" [hos host] hid: guid=0300fa67c82d00000631000014017801; mapping-string[1/1]="0300fa67c82d00000631000014017801,*,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b10,leftshoulder:b4,leftstick:b8,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b9,righttrigger:a5,rightx:a3,righty:a4,start:b7,x:b2,y:b3,platform:Windows," [hos host] vibration: RenHDrum 0.5.0 loaded [hos host] vibration: "XInput Controller #1": four motors (RenHDrum) [hos sead::MainThread] nvn: EventSignal (CPU) #12: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #13: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #25: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #7: event 0x100d6c8000 mode 0 value 6 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #26: event 0x100d6c8000 mode 0 value 6 (now 4) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #27: event 0x100d6c8000 mode 0 value 6 (now 4) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #38: queue 000010100D695020 (its #7), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #28: event 0x100d6c8000 mode 0 value 6 (now 4) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #39: queue 000010100D6BF020 (its #7), 1 handles (lr 0x80666be0) [hos Presentation Thread] nvn: QueueSubmitCommands #40: queue 000010100D5EB020 (its #7), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #41: queue 000010100D640020 (its #7), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #42: queue 000010100D6AA020 (its #7), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #43: queue 0000101004133020 (its #7), 1 handles (lr 0x80590358) [hos sead::MainThread] nvn: EventSignal (CPU) #14: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos ModuleSystemWorker1] ocb: a new neighbour to preload is taken once it stays the choice 2000 ms (HOS_OCB_HOLD_MS; 0: the game's own at every frame) [hos sead::MainThread] nvn: EventSignal (CPU) #15: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #29: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #8: event 0x100d6c8000 mode 0 value 7 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #30: event 0x100d6c8000 mode 0 value 7 (now 5) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #31: event 0x100d6c8000 mode 0 value 7 (now 5) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #44: queue 000010100D695020 (its #8), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #32: event 0x100d6c8000 mode 0 value 7 (now 5) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #45: queue 000010100D6BF020 (its #8), 1 handles (lr 0x80666be0) [hos Presentation Thread] nvn: QueueSubmitCommands #46: queue 000010100D5EB020 (its #8), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #47: queue 000010100D640020 (its #8), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #48: queue 000010100D6AA020 (its #8), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #49: queue 0000101004133020 (its #8), 1 handles (lr 0x80590358) [hos sead::MainThread] nvn: EventSignal (CPU) #16: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #17: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #33: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #9: event 0x100d6c8000 mode 0 value 8 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #34: event 0x100d6c8000 mode 0 value 8 (now 6) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #35: event 0x100d6c8000 mode 0 value 8 (now 6) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #50: queue 000010100D695020 (its #9), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #36: event 0x100d6c8000 mode 0 value 8 (now 6) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #51: queue 000010100D6BF020 (its #9), 1 handles (lr 0x80666be0) [hos Presentation Thread] nvn: QueueSubmitCommands #52: queue 000010100D5EB020 (its #9), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #53: queue 000010100D640020 (its #9), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #54: queue 000010100D6AA020 (its #9), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #55: queue 0000101004133020 (its #9), 1 handles (lr 0x80590358) [hos sead::MainThread] nvn: EventSignal (CPU) #18: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: texture budget: 7158 MB (driver allows 11474 MB, the card has 12272 MB, RAM 32694 MB) [hos sead::MainThread] nvn: EventSignal (CPU) #19: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferWaitEvent #37: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444) [hos sead::MainThread] nvn: CommandBufferSignalEvent #10: event 0x100d6c8000 mode 0 value 9 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferWaitEvent #38: event 0x100d6c8000 mode 0 value 9 (now 8) (lr 0x806661f8) [hos sead::MainThread] nvn: CommandBufferWaitEvent #39: event 0x100d6c8000 mode 0 value 9 (now 8) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #56: queue 000010100D695020 (its #10), 1 handles (lr 0x80666be0) [hos sead::MainThread] nvn: CommandBufferWaitEvent #40: event 0x100d6c8000 mode 0 value 9 (now 8) (lr 0x806661f8) [hos sead::MainThread] nvn: QueueSubmitCommands #57: queue 000010100D6BF020 (its #10), 1 handles (lr 0x80666be0) [hos Presentation Thread] nvn: QueueSubmitCommands #58: queue 000010100D5EB020 (its #10), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #59: queue 000010100D640020 (its #10), 1 handles (lr 0x805282f4) [hos Presentation Thread] nvn: QueueSubmitCommands #60: queue 000010100D6AA020 (its #10), 1 handles (lr 0x805282f4) [hos sead::MainThread] nvn: EventSignal (CPU) #20: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: RX_VK_RECORDER=thread: a command buffer or recorded call on a thread without the replay lock while the recorder was busy: made after the recorder caught up (counted in the stats line; logged once) [hos sead::MainThread] nvn: EventSignal (CPU) #21: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #11: event 0x100d6c8000 mode 0 value 10 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #22: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #23: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #12: event 0x100d6c8000 mode 0 value 11 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #24: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #25: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #13: event 0x100d6c8000 mode 0 value 12 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #26: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #27: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #14: event 0x100d6c8000 mode 0 value 13 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #28: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: vertex format NVN 0x8 -> vk 74 [hos sead::MainThread] nvn: EventSignal (CPU) #29: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #15: event 0x100d6c8000 mode 0 value 14 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #30: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos ModuleSystemWorker2] barrier: the game's job barriers (f_800998a0, f_800738e0, f_8007e200) wait until their masks move instead of sleeping 20 us (HOS_OOT_BARRIER_WAIT) [hos sead::MainThread] nvn: EventSignal (CPU) #31: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #16: event 0x100d6c8000 mode 0 value 15 (lr 0x80666598) [hos ModuleSystemWorker1] vk: invalidates with nothing to wait for keep their virtual core: on [hos sead::MainThread] nvn: EventSignal (CPU) #32: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #33: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #17: event 0x100d6c8000 mode 0 value 16 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #34: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #35: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #18: event 0x100d6c8000 mode 0 value 17 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #36: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #37: event 0x1006e38000 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #19: event 0x100d6c8000 mode 0 value 18 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #38: event 0x1006e38000 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: EventSignal (CPU) #39: event 0x1006e38004 mode 0 value 0 (lr 0x802eb420) [hos sead::MainThread] nvn: CommandBufferSignalEvent #20: event 0x100d6c8000 mode 0 value 19 (lr 0x80666598) [hos sead::MainThread] nvn: EventSignal (CPU) #40: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos sead::MainThread] nvn: CommandBufferSignalEvent #21: event 0x100d6c8000 mode 0 value 20 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #22: event 0x100d6c8000 mode 0 value 21 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #23: event 0x100d6c8000 mode 0 value 22 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #24: event 0x100d6c8000 mode 0 value 23 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #25: event 0x100d6c8000 mode 0 value 24 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #26: event 0x100d6c8000 mode 0 value 25 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #27: event 0x100d6c8000 mode 0 value 26 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #28: event 0x100d6c8000 mode 0 value 27 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #29: event 0x100d6c8000 mode 0 value 28 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #30: event 0x100d6c8000 mode 0 value 29 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #31: event 0x100d6c8000 mode 0 value 30 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #32: event 0x100d6c8000 mode 0 value 31 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #33: event 0x100d6c8000 mode 0 value 32 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #34: event 0x100d6c8000 mode 0 value 33 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #35: event 0x100d6c8000 mode 0 value 34 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #36: event 0x100d6c8000 mode 0 value 35 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #37: event 0x100d6c8000 mode 0 value 36 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #38: event 0x100d6c8000 mode 0 value 37 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #39: event 0x100d6c8000 mode 0 value 38 (lr 0x80666598) [hos sead::MainThread] nvn: CommandBufferSignalEvent #40: event 0x100d6c8000 mode 0 value 39 (lr 0x80666598) [hos host] vk: vertex format NVN 0xd -> vk 16 [hos host] vk: vertex format NVN 0xf -> vk 20 [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 260 dirty pages from their submits' captures, 1222 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 148 dispatches, 326 chunks; 0 syncs as copies (no table room) [hos host] vk: read live, by pool: [hos Resource Memory] mem: the game's mapped pages 951 MB (peak 951); heap 6894 MB; the console's total 9809 MB [hos ARes Loading 1] prompts: the Talk ring's A B X Y at places 0 1 2 3, the ocarina hint as is [hos ARes Loading 0] prompts: switch textures, 1373 runs for this pad's controls [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.03 over 1185 ranges, 702 MB), 122 pages written, 122 copied (0.5 MB) [hos host] vk: a capture taken (ms): all 0.008; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (58 MB), 12 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: late pages left to their reads off: the pools have mirrors [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07 [hos ModuleSystemWorker5] prompts: the textures' pool at 0x7fc3035000 (+0xe86c000): a new face or new controls show at once [hos sead::MainThread] prompts: the layouts at 0x7fe1a31000: the Talk ring and the ocarina hint follow at once [hos host] vk: import 2.1 ms: commit+properties 0.0, buffer 0.0, allocate 2.0, bind+address 0.0 [hos host] vk: pool 0x107db39100+0x40 imported in 2.1 ms as 0x107d800000..0x107e000000 (5 imports, 9 ms in all) [hos host] vk: 5 pools imported [hos host] vk: import 3.4 ms: commit+properties 0.0, buffer 0.0, allocate 3.4, bind+address 0.0 [hos host] vk: pool 0x107cd70000+0x120 imported in 3.5 ms as 0x107cc00000..0x107d800000 (6 imports, 12 ms in all) [hos host] vk: 6 pools imported [hos host] vk: write-back diff: a shadow for pool 0x100d400000 (4 MB; 4 MB of shadows) [a64] unimplemented import _ZN2nn5movie10InitializeEv (lr 0x83b7d25c) [a64] unimplemented import _ZN2nn5movie13IsInitializedEv (lr 0x85467d74) [hos host] vk: 3901 programs from the program cache [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 3.10, present 0.32, draw 0.08, descriptors 0.01, texture lookups 0.13, uploads 0.06, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.00, memory sync 0.48, queue submit 0.00, slot wait 0.21 [hos host] vk: a frame: mirrors 6 (15 MB): 0.16 MB synced in 3.4 copies, 0.00 MB written back [hos host] vk: 120 presents, 4.7 fps; a frame: 1 draws (0 skipped), 0 dispatches, 2 passes, 1 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 3426 pipelines (661 ms), 0 programs (0 failed), 5 textures (56 MB) [hos host] vk: host time in GPU resource waits: 45.03 ms a frame (fences 43.28, timelines 1.76, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 23.4 full barriers recorded, 13.9 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 0.0; 0x12: 1.0; 0x78: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 0.0, an op 1.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.5 ms a frame (the longest submission 2.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.5 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 108.2 ms for a core: given by GeneralFormatAsync_ (priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 53.4 ms for a core: given by tera::FileResourceMgr (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos ModuleSystemWorker2] oe: startup logo finished [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos Resource Memory] mem: the game's mapped pages 1027 MB (peak 1027); heap 6894 MB; the console's total 9809 MB [hos Resource Memory] mem: the game's mapped pages 1283 MB (peak 1283); heap 6894 MB; the console's total 9809 MB [hos host] vk: late pages found at the replay (600 takes): 7 pages in 7 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: vertex format NVN 0x2c -> vk 95 [hos OptimizeChangeBanc] [debug] nvn: 1 copies of recordings whose control memory was reused let go (304 live) [hos host] vk: vertex format NVN 0x26 -> vk 38 [hos MovieOperateThread] video: nn::codec::video on libavcodec [hos MovieOperateThread] video: AV1 on dav1d 1.5.4 (a thread per logical core) [hos MovieOperateThread] video: AV1 decoder ready on dav1d (1920x1080) [hos AudioFrameworkThread] audio: voice 0: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s [hos AudioFrameworkThread] audio: voice 1: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s [hos AudioFrameworkThread] audio: voice 2: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s [hos AudioFrameworkThread] audio: voice 3: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s [hos AudioFrameworkThread] audio: voice 4: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s [hos AudioFrameworkThread] audio: voice 5: 48000 Hz, 1 channels, format 2, parameter 0 (0 bytes) at 106891.0 s [hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x108964d000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2) [hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x1089655000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2) [hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x108965d000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2) [hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x1089665000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2) [hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x108966d000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2) [hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x1089675000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2) [hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x108964e000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2) [hos AudioFrameworkThread] audio: wave buffer 0x108acecd50: data 0x1089656000 size 4096 samples 0..2048 loop 0 context 0 (voice format 2) [hos AudioUserTaskThread] [debug] codec: Opus decoder 0: 48000 Hz, 1 channels [hos AudioUserTaskThread] [debug] codec: Opus decoder 1: 48000 Hz, 1 channels [hos AudioUserTaskThread] [debug] codec: Opus decoder 2: 48000 Hz, 1 channels [hos AudioUserTaskThread] [debug] codec: Opus decoder 3: 48000 Hz, 1 channels [hos AudioUserTaskThread] [debug] codec: Opus decoder 4: 48000 Hz, 1 channels [hos AudioUserTaskThread] [debug] codec: Opus decoder 5: 48000 Hz, 1 channels [hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe000, 49152): header 000000ae 07bf4b00 -> 182 [hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe0b6, 48970): header 00000094 036ed700 -> 156 [hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe152, 48814): header 000000a1 129a5e00 -> 169 [hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe1fb, 48645): header 000000a1 03898c00 -> 169 [hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe2a4, 48476): header 000000a1 66171600 -> 169 [hos AudioUserTaskThread] [trace] codec: GetOpusPacketSize(0x108acfe34d, 48307): header 000000a1 0322bb00 -> 169 [hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960) [hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960) [hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960) [hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960) [hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960) [hos OpusDecoder] codec: decode 161 bytes (of 169) -> 960 samples (room for 960) [hos OptimizeChangeBanc] [debug] nvn: 4 copies of recordings whose control memory was reused let go (596 live) [hos MovieOperateThread] video: the first buffer went out before the first picture was decoded (shown black) [hos Presentation Thread] vi: a movie is playing: the frame rate goes to 240 for it (HOS_MOVIE_PACE=0 leaves it) [hos host] vk: texture 0x10000000000 1920x1080x1 x1 levels 1 NVN 0x1 -> vk 9 (6 textures, 58 MB) [hos host] vk: import 3.5 ms: commit+properties 0.0, buffer 0.0, allocate 3.4, bind+address 0.0 [hos host] vk: pool 0x10766a2000+0x1fa400 imported in 3.5 ms as 0x1075c00000..0x1077000000 (7 imports, 16 ms in all) [hos host] vk: 7 pools imported [hos host] vk: texture 0x1000021c200 960x540x1 x1 levels 1 NVN 0xd -> vk 16 (7 textures, 59 MB) [hos host] vk: sampler 268 is nearest: min filter 4, mag filter 1, wrap 7/7/7, lod -1000..1000 bias 0, aniso 1, compare 0 [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 4, mip filter linear, LOD bias 0; made with 0, an image of one level: not the row's) [hos OptimizeChangeBanc] nvn: DrawMeshTasks: 1 so far (x1 0 x2 0x2 x3 0x400000 x4 0x83b70de0) [hos OptimizeChangeBanc] nvn: DrawMeshTasks: 2 so far (x1 0 x2 0x2 x3 0x400000 x4 0x83b70de0) [hos OptimizeChangeBanc] nvn: DrawMeshTasks: 3 so far (x1 0 x2 0x2 x3 0x400000 x4 0x83b70de0) [hos host] vk: textures from a pool of their own, 32 MB blocks (RX_VK_COMPACT), 24 made at the start [hos host] vk: texture 0x10755b4000 1920x1080x1 x1 levels 1 NVN 0x25 -> vk 37 (8 textures, 67 MB) [hos host] vk: import 10.4 ms: commit+properties 0.2, buffer 0.0, allocate 10.2, bind+address 0.0 [hos host] vk: pool 0x10755b4000+0x870000 imported in 10.4 ms as 0x1072c00000..0x1077000000 (8 imports, 26 ms in all) [hos host] vk: pool 0x1075c00000 remade larger (1 retired) [hos host] vk: 7 pools imported [hos host] vk: in play: compute e409733c45530d75 (e409733c45530d75 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program e409733c45530d75 not translated: a network left out on this card [hos host] vk: rejected compute e409733c45530d75 (control 0x1075e24768) grid 900x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x1072e26320+0xfd2000; ssbo2 0x1072dea000+0x1200; ssbo3 0x1072deb200+0x20; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute 8306877b7dc8784b (8306877b7dc8784b 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program 8306877b7dc8784b not translated: a network left out on this card [hos host] vk: rejected compute 8306877b7dc8784b (control 0x1075e260d8) grid 720x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x10745e1320+0xfd2000; ssbo1 0x1072e26320+0xfd2000; ssbo2 0x1072deb300+0x1200; ssbo3 0x1072dec500+0x20; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute de3c10dba4ff39d3 (de3c10dba4ff39d3 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program de3c10dba4ff39d3 not translated: a network left out on this card [hos host] vk: rejected compute de3c10dba4ff39d3 (control 0x1075e27948) grid 97x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x1072e26320+0x7e9000; ssbo1 0x10745e1320+0xfd2000; ssbo2 0x1072dec600+0x1000; ssbo3 0x1072ded600+0x40; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute a41c78faa56ccf7d (a41c78faa56ccf7d 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program a41c78faa56ccf7d not translated: a network left out on this card [hos host] vk: rejected compute a41c78faa56ccf7d (control 0x1075e28860) grid 120x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x107360f320+0x7e9000; ssbo1 0x1072e26320+0x7e9000; ssbo2 0x1072ded700+0x4800; ssbo3 0x1072df1f00+0x40; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute e959b9b1451491c0 (e959b9b1451491c0 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos OptimizeChangeBanc] [debug] nvn: 3 copies of recordings whose control memory was reused let go (2160 live) [hos host] [trace] vk: program e959b9b1451491c0 not translated: a network left out on this card [hos host] vk: rejected compute e959b9b1451491c0 (control 0x1075e2a100) grid 70x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x1072e26320+0x3f4800; ssbo1 0x107360f320+0x7e9000; ssbo2 0x1072df2000+0x4000; ssbo3 0x1072df6000+0x80; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute 5a4d7378db437fed (5a4d7378db437fed 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program 5a4d7378db437fed not translated: a network left out on this card [hos host] vk: rejected compute 5a4d7378db437fed (control 0x1075e2b038) grid 60x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x107321ab20+0x3f4800; ssbo1 0x1072e26320+0x3f4800; ssbo2 0x1072df6100+0x12000; ssbo3 0x1072e08100+0x80; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute 8b73eaee02efc5e4 (8b73eaee02efc5e4 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program 8b73eaee02efc5e4 not translated: a network left out on this card [hos host] vk: rejected compute 8b73eaee02efc5e4 (control 0x1075e2c8f0) grid 85x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x1073df8320+0x7e9000; ssbo1 0x1072e26320+0x3f4800; ssbo2 0x1072e1a300+0x4000; ssbo3 0x1072e1e300+0x40; ssbo4 0x107360f320+0x7e9000; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute 14140d4c1dbe754f (14140d4c1dbe754f 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program 14140d4c1dbe754f not translated: a network left out on this card [hos host] vk: rejected compute 14140d4c1dbe754f (control 0x1075e2d970) grid 8100x1x1: img 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x107360f320+0xfd2000; ssbo1 0x1072e26320+0x7e9000; ssbo2 0x1072e22d00+0x4000; ssbo3 0x1072e23d00+0x40; ssbo4 0x10745e1320+0xfd2000; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute 4c2c129fe950b041 (4c2c129fe950b041 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program 4c2c129fe950b041 not translated: a network left out on this card [hos host] vk: rejected compute 4c2c129fe950b041 (control 0x1075e2e9c8) grid 5400x1x1: img 0x1075e31000 1920x1080x1 fmt 0x3d; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x10745e1320+0xfd2000; ssbo1 0x10745e1320+0xfd2000; ssbo2 0x1072e25100+0x1200; ssbo3 0x1072e26300+0x20; ssbo4 0x10745e1320+0xfd2000; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; ubo7 0x100c91e000+0x100; [hos host] vk: in play: compute 3f6f2b1d671badcd (3f6f2b1d671badcd 0000000000000000), inputs 0000000000000000, outputs 00 00 00 00 00 00 00 00, 0 other variants [hos host] [trace] vk: program 3f6f2b1d671badcd not translated: a network left out on this card [hos host] vk: rejected compute 3f6f2b1d671badcd (control 0x1072da4fe0) grid 80x180x1: img 0x1075e31000 1920x1080x1 fmt 0x3d; img 0x10000348400 2560x1440x1 fmt 0x38; tex 0x1075e31000 1920x1080x1 fmt 0x3d; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x1000021c200 960x540x1 fmt 0xd; tex 0x1072ca0200 512x512x1 fmt 0x3d; tex 0x10755b4000 1920x1080x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; tex 0x100c922a00 4x4x1 fmt 0x25; ssbo0 0x10745e1320+0xfd2000; ssbo1 0x10745e1320+0xfd2000; ssbo2 0x1072da1f00+0x3000; ssbo3 0x1072e26300+0x20; ssbo4 0x10745e1320+0xfd2000; ubo0 0x100c91e000+0x100; ubo1 0x100c91e000+0x100; ubo2 0x100c91e000+0x100; ubo3 0x100c91e000+0x100; ubo4 0x100c91e000+0x100; ubo5 0x100c91e000+0x100; ubo6 0x100c91e000+0x100; [hos host] vk: texture 0x10000348400 2560x1440x1 x1 levels 1 NVN 0x38 -> vk 37 (9 textures, 81 MB) [hos host] vk: import 3.9 ms: commit+properties 0.0, buffer 0.0, allocate 3.9, bind+address 0.0 [hos host] vk: pool 0x100df5f400+0xf00000 imported in 3.9 ms as 0x100dc00000..0x100f000000 (9 imports, 30 ms in all) [hos host] vk: movie upscaling uses the linear compute path [hos Presentation Thread] vi: the movie is over: back to the game's frame rate cap 60 [hos host] vk: pool 0x1072c00000..0x1077000000: 6540 KB the CPU wrote since retired pool 0x1075c00000's last sync kept from guest memory, not its mirror [hos host] vk: pool 0x1072c00000..0x1077000000: mirror made with the GPU's contents of 1 retired pool(s) [hos host] vk: late pages' fix: stamps for 0x100d6c0000 (pool 0x100d400000; 1 blocks, 64 KB) [hos host] vk: late pages' fix: stamps for 0x100d400000 (pool 0x100d400000; 2 blocks, 128 KB) [hos host] vk: late pages' fix: stamps for 0x100d410000 (pool 0x100d400000; 3 blocks, 192 KB) [hos host] vk: late pages' fix: stamps for 0x100d420000 (pool 0x100d400000; 4 blocks, 256 KB) [hos host] vk: late pages' fix: stamps for 0x100d430000 (pool 0x100d400000; 5 blocks, 320 KB) [hos host] vk: late pages' fix: stamps for 0x100d440000 (pool 0x100d400000; 6 blocks, 384 KB) [hos host] vk: late pages' fix: stamps for 0x100d450000 (pool 0x100d400000; 7 blocks, 448 KB) [hos host] vk: late pages' fix: stamps for 0x100d460000 (pool 0x100d400000; 8 blocks, 512 KB) [hos host] vk: late pages' fix: stamps for 0x100d7f0000 (pool 0x100d400000; 64 blocks, 4096 KB) [hos Presentation Thread] vi: a movie is playing: the frame rate goes to 240 for it (HOS_MOVIE_PACE=0 leaves it) [hos host] vk: sparse mirrors: 7, 79 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 0; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 9 binds (1.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 0 kept as spares, 0 destroyed; 0 spares now, 0 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 136 dirty pages from their submits' captures, 3325 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 183 dispatches, 500 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 40 runs, 0.0 MB compared; 40 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 60 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: mirror of pool 0x1075c00000 (20 MB) given back: retired, 50 submissions since its sync [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 1265 ranges, 907 MB), 153 pages written, 153 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.019 (2 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (90 MB), 8 spare buffers, 0 texture pages copied (0 MB) [hos OptimizeChangeBanc] [debug] nvn: 6 copies of recordings whose control memory was reused let go (2974 live) [hos Resource Memory] mem: the game's mapped pages 1536 MB (peak 1536); heap 6894 MB; the console's total 9809 MB [hos sead::MainThread] nvn: GPU waits in 5 s: 46 sync waits (570 ms), 0 polls timed out (0 ms), 1 queue finishes (46 ms) [hos ParallelDecompThread] zstd: 3 stream(s) decoding on the host at once [hos sead::MainThread] [debug] nvn: 3 copies of recordings whose control memory was reused let go (3512 live) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.10 [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3602 dirty pages from their submits' captures, 32761 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2620 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 90 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 4427 programs from the program cache [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 3.54, present 0.05, draw 0.06, descriptors 0.01, texture lookups 1.18, uploads 0.14, pipelines 0.00, programs 0.00, commands 0.04, dispatch 2.10, memory sync 0.75, queue submit 0.00, slot wait 0.17 [hos host] vk: a frame: mirrors 8 (78 MB): 2.17 MB synced in 5.1 copies, 0.00 MB written back [hos host] vk: 240 presents, 91.4 fps; a frame: 2 draws (9 skipped), 2 dispatches, 3 passes, 11 barriers, 6 submits, 0.0 uploads (0.2 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 5049 pipelines (1030 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 7.22 ms a frame (fences 3.16, timelines 4.06, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 9.3 [hos host] vk: a frame: 25.3 full barriers recorded, 26.3 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 10.1; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 10.1, an op 11.8, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 2.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.2 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.02 ms [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.03 over 1296 ranges, 948 MB), 161 pages written, 161 copied (0.6 MB) [hos OptimizeChangeBanc] [debug] nvn: 3 copies of recordings whose control memory was reused let go (4005 live) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (84 MB), 8 spare buffers, 0 texture pages copied (0 MB) [hos BaseProcExecutor0] vcores: 123740 takes, 27998 waits (mean 83 us, longest 498.5 ms, 7 over HOS_VCORE_LOG_MS), 13303 kicks, 12240 yields, 0 starved grants, 0 escapes in 5 s [hos OptimizeChangeBanc] [debug] nvn: 3 copies of recordings whose control memory was reused let go (4541 live) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 120 pages in 120 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.11 [hos host] vk: vertex format NVN 0xe -> vk 17 [hos host] vk: sparse mirrors: 7, 78 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 11 binds (1.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 1 kept as spares, 0 destroyed; 1 spares now, 20 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2769 dirty pages from their submits' captures, 30218 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2414 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 95 in captures, 1 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos OptimizeChangeBanc] [debug] nvn: 4 copies of recordings whose control memory was reused let go (5437 live) [hos OptimizeChangeBanc] [debug] nvn: 3 copies of recordings whose control memory was reused let go (5723 live) [hos Resource Memory] mem: the game's mapped pages 1792 MB (peak 1792); heap 6894 MB; the console's total 9809 MB [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.03 over 1318 ranges, 1024 MB), 166 pages written, 166 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (130 MB), 11 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.16, present 0.05, draw 0.06, descriptors 0.01, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.04, dispatch 0.06, memory sync 0.68, queue submit 0.00, slot wait 0.17 [hos host] vk: a frame: mirrors 8 (78 MB): 1.68 MB synced in 5.0 copies, 0.00 MB written back [hos host] vk: 360 presents, 164.9 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 6231 pipelines (1331 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 5.41 ms a frame (fences 2.78, timelines 2.64, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 27.9 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.6 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the recording streams: 48 of 1 MB, the most one held 0 KB; 0 made on demand (0 MB kept now); c1 stages: at most 0 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6532 dirty pages from their submits' captures, 27295 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2439 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 85 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: vertex format NVN 0x1 -> vk 9 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.09 [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 1361 ranges, 1139 MB), 191 pages written, 191 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (104 MB), 8 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 147 pages in 147 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 7, 78 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 11 binds (1.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 1 kept as spares, 0 destroyed; 1 spares now, 20 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2984 dirty pages from their submits' captures, 29347 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2367 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 103 in captures, 1 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: 5240 programs from the program cache [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.25, present 0.05, draw 0.06, descriptors 0.01, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.04, dispatch 0.07, memory sync 0.70, queue submit 0.00, slot wait 0.19 [hos host] vk: a frame: mirrors 8 (78 MB): 1.73 MB synced in 5.1 copies, 0.00 MB written back [hos host] vk: 480 presents, 165.0 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 7234 pipelines (1629 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 5.35 ms a frame (fences 2.84, timelines 2.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 27.4 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.1 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.11 [hos Presentation Thread] vi: 503 frames presented (100.5 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 358, 15.5-17.8 134, 17.8-22 2, 22-30 0, 30-40 4, over 40 4 [hos OptimizeChangeBanc] [debug] nvn: 1 copies of recordings whose control memory was reused let go (13140 live) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 1396 ranges, 1222 MB), 177 pages written, 177 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (112 MB), 10 spare buffers, 0 texture pages copied (0 MB) [hos OptimizeChangeBanc] nvn: 50000 recordings made, 13368 live copies (1.3 MB), 7 texture copies; 2 command buffers' segments (0.1 MB) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 8529 dirty pages from their submits' captures, 23677 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2348 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 90 in captures, 1 only in the collection, 1 only by the replay itself [hos host] vk: read live, by pool: [hos Resource Memory] mem: the game's mapped pages 2048 MB (peak 2048); heap 6894 MB; the console's total 9809 MB [hos host] vk: presenter thread: the copy's fence 3.45 ms, the present 0.77 ms a present [hos host] vk: memory: working set 9754 MB (peak 9754), private 12033 MB, staging 16 MB in 1 arenas, 1 made so far (7 ms) [hos host] vk: memory heaps: heap 0 (system): 12 blocks 592 MB, 37 allocations 540 MB; used 756 of 15579 MB; heap 1 (device): 83 blocks 1136 MB, 1472 allocations 306 MB; used 2616 of 11474 MB; [hos host] vk: image blocks: 2 blocks 64 MB, 2 images 63 MB; 1 free ranges, the largest 0 MB [hos host] vk: image blocks (render targets): 3 blocks 96 MB, 8 images 73 MB; 3 free ranges, the largest 17 MB [hos host] vk: image blocks (textures, compacted): 24 blocks 768 MB, 1 images 7 MB; 24 free ranges, the largest 32 MB [hos host] vk: render target images: 0 reused from the pool, 8 made new, 0 destroyed from it, 0 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 0 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 8 with the pool empty (0 of them with targets still in the dead queue), 0 with none in it a present yet, 0 with none of a kind they can use, 0 with all smaller (the largest 0% short on average), 0 with all bigger (over 1.5 times), 0 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 0 idle [hos host] vk: VRAM: 2616 MB used of 11474 MB allowed (829 MB of it free in the blocks); textures 81 MB (budget 7158 MB), 9 textures [hos host] vk: budget: VRAM 1136 of 11474 MB (2616 MB by the driver, 1480 MB unaccounted): blocks free 829, textures 81, mirrors 78, other images 63, other buffers 62, staging 16, shadows 4; RAM 703 MB (the process's private memory 12033 MB): capture spares 420, imports 124, captures 106, retired imports 20, capture table 19, other buffers 14, others 0 [hos host] vk: textures: decoded ASTC 0 MB, rendered 81 MB (1 drawn in one frame, kept: 128 KB), other 0 MB; used in the last 8 presents 81 MB [hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 0 MB of blocks hashed in 0 ms [hos host] vk: 14 MB: 0x1004137000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 2 presents ago [hos host] vk: 14 MB: 0x1005f37000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 3 presents ago [hos host] vk: 14 MB: 0x1001f0ffe00 2560x1440x1 x1 levels 1 NVN 0x3f rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x10000348400 2560x1440x1 x1 levels 1 NVN 0x38 rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x1005037000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago [hos host] vk: 7 MB: 0x10755b4000 1920x1080x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago [hos host] vk: 1 MB: 0x10000000000 1920x1080x1 x1 levels 1 NVN 0x1 rendered, last used 1 presents ago [hos host] vk: 0 MB: 0x1000021c200 960x540x1 x1 levels 1 NVN 0xd rendered, last used 1 presents ago [hos Presentation Thread] vk: presents after the CPU's signal (600): 567 waited for it, 3.00 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.11 [hos host] vk: status: 600 frames, the slowest 275.6 ms; VRAM 2.6 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 0 program records loaded (0.0 ms), 0 new graphics pipelines, pipelines from the driver's cache 0, from the start's cache 0, compiled 0; 0 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 3114 queued, 3114 loaded (median 1.05 ms, 99th percentile 17.17 ms, worst 34.27 ms, 7049 ms and 171.6 MB read in all); 3836 programs (583 of them compute variants without their code, 0 graphics programs no note names), 6183 graphics pipelines (0 not in a cache), 365 compute pipelines (19 not in a cache); a first use ran 0 itself and waited for 0 (0.0 ms); 3 pipelines made at create taken by draws; 1295 create calls with nothing kept for their stages, 0 before the index; units 4409 (3114 ready, 0 loading), 8148 programs alive; 0 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 6422 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 757 known, 0 learned in this run; 365 compute pipelines made by them at a create call, 19 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.29, present 0.06, draw 0.06, descriptors 0.01, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.04, dispatch 0.07, memory sync 0.75, queue submit 0.00, slot wait 0.18 [hos host] vk: a frame: mirrors 8 (78 MB): 1.73 MB synced in 5.0 copies, 0.00 MB written back [hos host] vk: 600 presents, 165.2 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 7869 pipelines (1827 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 5.26 ms a frame (fences 2.80, timelines 2.47, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 27.6 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 0.00 ms a frame): its jobs 6.01 ms a frame (submits 1.46, presents 3.81, others 0.74), idle -6.01; the game waited for room at 431 presents, 1.79 ms each (1.28 a frame; 16.09 jobs queued at a present); 43 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 1.91 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 580 captures at a CPU signal; the submits' waits read in 0.002 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.05 ms after its hand-over (265.75 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 7.98 + 0.12 (the window's copy is in the idle), after it 0.09 + 0.00; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 89.92 (0.13 ms) [hos host] vk: frame ahead (600 presents): 1.00 parts a frame recorded in the signal's job, 17.8% of them took nothing that came with the signal; of the others, 0.82 a frame took pages of the signal's capture into a mirror (47.6 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 103 of 579 frames, its first three in 103 [hos host] vk: frame ahead, what came with the signal: 579 signals with a capture, each of 163.4 pages in 2.3 pools; between a frame's last submit and that signal 2.07 texture and sampler registrations, 2.78 releases and 0.00 other CPU signals a signal (295 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.09+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.04 over 1445 ranges, 1320 MB), 183 pages written, 183 copied (0.7 MB) [hos host] vk: sparse mirrors: 7, 78 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 11 binds (1.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 1 kept as spares, 0 destroyed; 1 spares now, 20 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2750 dirty pages from their submits' captures, 30326 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2400 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 86 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: a capture taken (ms): all 0.007; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (108 MB), 11 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 143 pages in 143 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 7531 dirty pages from their submits' captures, 26285 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2428 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 74 in captures, 1 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1445 ranges, 1320 MB), 142 pages written, 142 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (96 MB), 9 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 0.94, present 0.05, draw 0.04, descriptors 0.00, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.06, memory sync 0.43, queue submit 0.00, slot wait 0.21 [hos host] vk: a frame: mirrors 8 (78 MB): 1.70 MB synced in 4.7 copies, 0.00 MB written back [hos host] vk: 720 presents, 164.9 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 7869 pipelines (1827 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 5.46 ms a frame (fences 2.63, timelines 2.83, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 27.8 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.6 ms a frame (the longest submission 3.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the recording streams: 48 of 1 MB, the most one held 0 KB; 0 made on demand (0 MB kept now); c1 stages: at most 0 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 7, 78 MB with memory of 104 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 11 binds (1.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 1 kept as spares, 0 destroyed; 1 spares now, 20 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4908 dirty pages from their submits' captures, 28034 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2372 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: mirror of pool 0x103d800000 (4 MB) given back: idle, 6015 submissions since its sync [hos sead::MainThread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1445 ranges, 1320 MB), 142 pages written, 142 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (106 MB), 6 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 175 pages in 175 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos host] vk: late pages' fix (600 recordings with one): of 856 captures with late pages 854 fixed (0.7 pages, 0.7 fixes a capture), waited as before: 0 no pool, 2 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 0 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 64 blocks (4096 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 0, of an earlier recording 0, not recorded yet 0, of this part 0; pages fixed from stamps 599, from snapshots 0; the shadows: 853 pages synced from captures, 127 made known by it, 726 left unknown because the replay or the collection wrote them since [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2322 dirty pages from their submits' captures, 29974 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2332 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 74 in captures, 0 only in the collection, 1 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 0.86, present 0.04, draw 0.04, descriptors 0.00, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.06, memory sync 0.42, queue submit 0.00, slot wait 0.16 [hos host] vk: a frame: mirrors 8 (76 MB): 1.73 MB synced in 4.7 copies, 0.00 MB written back [hos host] vk: 840 presents, 164.3 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 7869 pipelines (1827 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 5.55 ms a frame (fences 2.63, timelines 2.92, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 28.1 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.8 ms a frame (the longest submission 3.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos sead::MainThread] vk: capture: a submit 0.09 ms (lock 0.00, scan 0.03 over 1445 ranges, 1320 MB), 140 pages written, 140 copied (0.5 MB) [hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (102 MB), 9 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: mirror of pool 0x107cc00000 (12 MB) given back: idle, 6039 submissions since its sync [hos host] vk: mirror of pool 0x107d800000 (8 MB) given back: idle, 6039 submissions since its sync [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07 [hos host] vk: sparse mirrors: 4, 64 MB with memory of 80 MB (15 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 13 binds (1.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 4 spares now, 44 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4489 dirty pages from their submits' captures, 29327 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2428 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: vertex format NVN 0x14 -> vk 81 [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 1464 ranges, 1335 MB), 150 pages written, 150 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (116 MB), 10 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.07, present 0.05, draw 0.03, descriptors 0.00, texture lookups 0.01, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.07, memory sync 0.58, queue submit 0.00, slot wait 0.14 [hos host] vk: a frame: mirrors 8 (64 MB): 1.73 MB synced in 4.9 copies, 0.00 MB written back [hos host] vk: 960 presents, 165.9 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 8106 pipelines (1903 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 5.01 ms a frame (fences 2.51, timelines 2.49, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 28.1 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 181 pages in 181 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6124 dirty pages from their submits' captures, 26843 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2397 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 91 in captures, 0 only in the collection, 1 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.11 [hos OptimizeChangeBanc] [debug] nvn: 2 copies of recordings whose control memory was reused let go (18707 live) [hos host] vk: recorder: 154659 packets run in 10.0 s (busy 0.64 s, spun idle 0.10 s, slept 9659 times, 4872 of them under 200 us), 0 calls made directly (no room); 16150 batches over 1011 frames (16.0 a frame, 10 packets a batch), 8066 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 19 times, 1.41 ms (0.0 times, 0.00 ms a frame); every thread 791 times, 103.32 ms in all, the longest 3.87 ms; the queue's lock 380/57.35 ms (max 1.65), vkWaitForFences 360/43.22 ms (max 3.87), vkQueueBindSparse 13/1.15 ms (max 0.35), vmaDestroyBuffer 27/0.76 ms (max 0.08), vkDestroyPipeline 5/0.58 ms (max 0.51), vmaFreeMemory 4/0.13 ms (max 0.06) [hos host] vk: recorder: of its busy 0.64 s, 0.49 s in parts' ends and submits (the longest 1.98 ms), 0.15 s of that blocked on the queue's lock (198 times, the longest 1.90 ms), so 0.35 s in the ends, submits and holds; 0.14 s running the other packets; 153 packets a frame [hos sead::MainThread] nvn: GPU waits in 5 s: 807 sync waits (2232 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos Resource Memory] mem: the game's mapped pages 2304 MB (peak 2304); heap 6894 MB; the console's total 9809 MB [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.02, scan 0.06 over 1588 ranges, 1375 MB), 159 pages written, 159 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (114 MB), 9 spare buffers, 0 texture pages copied (0 MB) [hos ModuleSystemWorker3] hid: game window in front (foreground window: process 32080, class "Longshot") [hos ModuleSystemWorker3] hid: 1 controller through SDL [hos host] vk: sparse mirrors: 4, 64 MB with memory of 80 MB (15 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 13 binds (1.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 4 spares now, 44 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4860 dirty pages from their submits' captures, 29882 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 228 dispatches, 2536 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 103 in captures, 0 only in the collection, 1 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: 6004 programs from the program cache [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.45, present 0.05, draw 0.03, descriptors 0.00, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.02, dispatch 0.06, memory sync 1.00, queue submit 0.00, slot wait 0.14 [hos host] vk: a frame: mirrors 8 (64 MB): 1.76 MB synced in 5.3 copies, 0.00 MB written back [hos host] vk: 1080 presents, 158.4 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 8991 pipelines (2077 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 5.40 ms a frame (fences 2.77, timelines 2.63, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 28.4 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 3.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos BaseProcExecutor1] vcores: 215849 takes, 39742 waits (mean 27 us, longest 12.6 ms, 0 over HOS_VCORE_LOG_MS), 20952 kicks, 20442 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.15 [hos ModuleSystemWorker2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 13 made): 7769385 lookups, 168816 missed (2.17 %) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.05 over 1661 ranges, 1392 MB), 153 pages written, 153 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (102 MB), 10 spare buffers, 0 texture pages copied (0 MB) [hos host] audio: 1001 blocks played in the first 10 s, loudest sample 408 of 32767 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 0 KB; 0 made on demand (0 MB kept now); c1 stages: at most 0 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3991 dirty pages from their submits' captures, 29839 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2442 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 83 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 193 pages in 193 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: presenter thread: the copy's fence 2.69 ms, the present 0.74 ms a present [hos host] vk: memory: working set 10388 MB (peak 10400), private 12789 MB, staging 16 MB in 1 arenas, 1 made so far (7 ms) [hos host] vk: memory heaps: heap 0 (system): 12 blocks 592 MB, 37 allocations 540 MB; used 756 of 15579 MB; heap 1 (device): 83 blocks 1136 MB, 1244 allocations 293 MB; used 2744 of 11474 MB; [hos host] vk: image blocks: 2 blocks 64 MB, 2 images 63 MB; 1 free ranges, the largest 0 MB [hos host] vk: image blocks (render targets): 3 blocks 96 MB, 8 images 73 MB; 3 free ranges, the largest 17 MB [hos host] vk: image blocks (textures, compacted): 24 blocks 768 MB, 1 images 7 MB; 24 free ranges, the largest 32 MB [hos host] vk: render target images: 0 reused from the pool, 8 made new, 0 destroyed from it, 0 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 0 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 8 with the pool empty (0 of them with targets still in the dead queue), 0 with none in it a present yet, 0 with none of a kind they can use, 0 with all smaller (the largest 0% short on average), 0 with all bigger (over 1.5 times), 0 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 0 idle [hos host] vk: VRAM: 2744 MB used of 11474 MB allowed (842 MB of it free in the blocks); textures 81 MB (budget 7158 MB), 9 textures [hos host] vk: budget: VRAM 1136 of 11474 MB (2744 MB by the driver, 1608 MB unaccounted): blocks free 842, textures 81, mirrors 64, other images 63, other buffers 63, staging 16, shadows 4; RAM 704 MB (the process's private memory 12789 MB): capture spares 424, imports 124, captures 112, retired imports 20, capture table 20, other buffers 4, others 0 [hos host] vk: textures: decoded ASTC 0 MB, rendered 81 MB (1 drawn in one frame, kept: 128 KB), other 0 MB; used in the last 8 presents 81 MB [hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 0 MB of blocks hashed in 0 ms [hos host] vk: 14 MB: 0x1004137000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 2 presents ago [hos host] vk: 14 MB: 0x1005f37000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 3 presents ago [hos host] vk: 14 MB: 0x1001f0ffe00 2560x1440x1 x1 levels 1 NVN 0x3f rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x10000348400 2560x1440x1 x1 levels 1 NVN 0x38 rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x1005037000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago [hos host] vk: 7 MB: 0x10755b4000 1920x1080x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago [hos host] vk: 1 MB: 0x10000000000 1920x1080x1 x1 levels 1 NVN 0x1 rendered, last used 1 presents ago [hos host] vk: 0 MB: 0x1000021c200 960x540x1 x1 levels 1 NVN 0xd rendered, last used 1 presents ago [hos Presentation Thread] vk: presents after the CPU's signal (600): 595 waited for it, 3.07 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07 [hos host] vk: status: 600 frames, the slowest 40.0 ms; VRAM 2.7 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 0 program records loaded (0.0 ms), 0 new graphics pipelines, pipelines from the driver's cache 0, from the start's cache 0, compiled 0; 0 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 3379 queued, 3379 loaded (median 1.15 ms, 99th percentile 17.66 ms, worst 34.27 ms, 8477 ms and 202.2 MB read in all); 4278 programs (583 of them compute variants without their code, 0 graphics programs no note names), 7257 graphics pipelines (0 not in a cache), 365 compute pipelines (19 not in a cache); a first use ran 0 itself and waited for 0 (0.0 ms); 3 pipelines made at create taken by draws; 1564 create calls with nothing kept for their stages, 0 before the index; units 4943 (3379 ready, 0 loading), 9293 programs alive; 0 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 7268 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 757 known, 0 learned in this run; 365 compute pipelines made by them at a create call, 19 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 0.85, present 0.03, draw 0.02, descriptors 0.00, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.00, commands 0.01, dispatch 0.04, memory sync 0.45, queue submit 0.00, slot wait 0.15 [hos host] vk: a frame: mirrors 8 (64 MB): 1.73 MB synced in 4.7 copies, 0.00 MB written back [hos host] vk: 1200 presents, 164.8 fps; a frame: 2 draws (12 skipped), 2 dispatches, 3 passes, 14 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 8991 pipelines (2077 ms), 0 programs (10 failed), 9 textures (81 MB) [hos host] vk: host time in GPU resource waits: 6.62 ms a frame (fences 3.06, timelines 3.56, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 29.1 full barriers recorded, 30.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.0; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.0, an op 15.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.8 ms a frame (the longest submission 0.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 6.11 ms a frame): its jobs 6.10 ms a frame (submits 0.93, presents 4.95, others 0.22), idle 0.01; the game waited for room at 593 presents, 1.90 ms each (1.88 a frame; 11.82 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 597 captures at a CPU signal; the submits' waits read in 0.002 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.97 ms after its hand-over (44.30 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 5.20 + 0.12 (the window's copy is in the idle), after it 0.05 + 0.00; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 1.00 parts a frame recorded in the signal's job, 0.8% of them took nothing that came with the signal; of the others, 0.99 a frame took pages of the signal's capture into a mirror (62.9 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 5 of 597 frames, its first three in 5 [hos host] vk: frame ahead, what came with the signal: 597 signals with a capture, each of 78.5 pages in 2.8 pools; between a frame's last submit and that signal 0.37 texture and sampler registrations, 0.85 releases and 0.00 other CPU signals a signal (112 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.05+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos sead::MainThread] vk: capture: a submit 0.09 ms (lock 0.00, scan 0.03 over 1661 ranges, 1392 MB), 142 pages written, 142 copied (0.6 MB) [hos host] vk: sparse mirrors: 4, 64 MB with memory of 80 MB (15 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 13 binds (1.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 0 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 4 spares now, 44 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4093 dirty pages from their submits' captures, 28849 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 225 dispatches, 2372 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 0 texture pages copied (0 MB) [hos sead::MainThread] nvn: CommandBufferWaitEvent #5000: event 0x100d6c8000 mode 0 value 1249 (now 1247) (lr 0x806661f8) [hos AudioFrameworkThread] audio: voice 6: 48000 Hz, 1 channels, format 6, parameter 0x1089637500 (32 bytes) at 106898.3 s [hos AudioFrameworkThread] audio: voice 7: 48000 Hz, 1 channels, format 6, parameter 0x1089637740 (32 bytes) at 106898.3 s [hos host] vk: ASTC not sampled by this GPU: blocks are decoded on upload [hos host] vk: texture 0x7fcd55d200 1664x1280x1 x1 levels 1 NVN 0x87 -> vk 37 (ASTC decoded) (10 textures, 89 MB) [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 3, mip filter nearest, LOD bias -0.5; made with 0, an image of one level: not the row's) [hos host] vk: texture 0x7fcd17d200 1464x640x1 x1 levels 1 NVN 0x87 -> vk 37 (ASTC decoded) (11 textures, 93 MB) [hos host] vk: texture 0x7fcbe9b200 683x683x1 x1 levels 1 NVN 0x49 -> vk 139 (12 textures, 93 MB) [hos host] vk: texture 0x7fdc8a7000 832x640x1 x1 levels 1 NVN 0x87 -> vk 37 (ASTC decoded) (13 textures, 95 MB) [hos host] vk: texture 0x7fdccb7000 832x640x1 x1 levels 1 NVN 0x87 -> vk 37 (ASTC decoded) (14 textures, 97 MB) [hos host] vk: texture 0x1001f0e1c00 240x94x1 x1 levels 1 NVN 0x25 -> vk 37 (15 textures, 97 MB) [hos host] vk: texture 0x1001f0d4a00 198x40x1 x1 levels 1 NVN 0x25 -> vk 37 (16 textures, 97 MB) [hos host] vk: texture 0x100c922a00 4x4x1 x1 levels 1 NVN 0x25 -> vk 37 (17 textures, 97 MB) [hos host] vk: texture 0x107cd71600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (18 textures, 97 MB) [hos host] vk: page table of guest memory at 0x39fc00000 (8 pools) [hos host] vk: texture #19 at 0x10000000000 beside #6: 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture 0x10000000000 262x104x1 x1 levels 1 NVN 0x1 -> vk 9 (19 textures, 97 MB) [hos host] vk: texture 0x1001f0ca800 262x104x1 x1 levels 1 NVN 0x1 -> vk 9 (20 textures, 97 MB) [hos host] vk: texture 0x1001f0bd600 198x40x1 x1 levels 1 NVN 0x25 -> vk 37 (21 textures, 97 MB) [hos host] vk: texture 0x107cd74600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (22 textures, 97 MB) [hos host] vk: texture 0x1001f0b3400 262x104x1 x1 levels 1 NVN 0x1 -> vk 9 (23 textures, 97 MB) [hos host] vk: texture 0x1001f0a6200 198x40x1 x1 levels 1 NVN 0x25 -> vk 37 (24 textures, 97 MB) [hos host] vk: texture 0x107cd77600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (25 textures, 97 MB) [hos host] vk: texture 0x1001f09c000 262x104x1 x1 levels 1 NVN 0x1 -> vk 9 (26 textures, 97 MB) [hos host] vk: texture 0x1001f077e00 285x94x1 x1 levels 1 NVN 0x25 -> vk 37 (27 textures, 97 MB) [hos host] vk: texture 0x1001f067c00 242x40x1 x1 levels 1 NVN 0x25 -> vk 37 (28 textures, 97 MB) [hos host] vk: texture 0x107cd7d600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (29 textures, 97 MB) [hos host] vk: texture #30 at 0x10000000000 beside #6: 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #30 at 0x10000000000 beside #19: 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture 0x10000000000 306x104x1 x1 levels 1 NVN 0x1 -> vk 9 (30 textures, 97 MB) [hos host] vk: texture 0x1001f05da00 306x104x1 x1 levels 1 NVN 0x1 -> vk 9 (31 textures, 98 MB) [hos host] vk: texture 0x1001f04d800 242x40x1 x1 levels 1 NVN 0x25 -> vk 37 (32 textures, 98 MB) [hos host] vk: texture 0x107cd80600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (33 textures, 98 MB) [hos host] vk: texture 0x1001f043600 306x104x1 x1 levels 1 NVN 0x1 -> vk 9 (34 textures, 98 MB) [hos host] vk: texture 0x1001f033400 242x40x1 x1 levels 1 NVN 0x25 -> vk 37 (35 textures, 98 MB) [hos host] vk: texture 0x107cd83600 256x46x1 x1 levels 1 NVN 0x1 -> vk 9 (36 textures, 98 MB) [hos host] vk: texture 0x1001f029200 306x104x1 x1 levels 1 NVN 0x1 -> vk 9 (37 textures, 98 MB) [hos host] vk: texture 0x1001efd9000 631x94x1 x1 levels 1 NVN 0x25 -> vk 37 (38 textures, 98 MB) [hos host] vk: texture 0x1001efb3e00 592x49x1 x1 levels 1 NVN 0x25 -> vk 37 (39 textures, 98 MB) [hos host] vk: texture 0x107cd89600 640x46x1 x1 levels 1 NVN 0x1 -> vk 9 (40 textures, 98 MB) [hos host] vk: texture #41 at 0x10000000000 beside #6: 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #41 at 0x10000000000 beside #19: 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #41 at 0x10000000000 beside #30: 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2681 dirty pages from their submits' captures, 29837 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 238 dispatches, 2358 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.09 ms (lock 0.00, scan 0.03 over 1662 ranges, 1624 MB), 141 pages written, 141 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (110 MB), 12 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 220 pages in 220 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos nn.npln.Worker] mutex 0x100ac110c8 (obj 000002DB48F189C0): nn.npln.Worker waiting 10 s; held by nn.npln.Worker (count 2, 1 waiting) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.00, present 0.05, draw 0.11, descriptors 0.04, texture lookups 0.03, uploads 0.02, pipelines 0.00, programs 0.02, commands 0.04, dispatch 0.05, memory sync 0.44, queue submit 0.00, slot wait 0.14 [hos host] vk: a frame: mirrors 10 (76 MB): 1.86 MB synced in 5.2 copies, 0.00 MB written back [hos host] vk: 1320 presents, 163.7 fps; a frame: 11 draws (12 skipped), 2 dispatches, 3 passes, 15 barriers, 6 submits, 0.4 uploads (0.1 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 8991 pipelines (2077 ms), 0 programs (10 failed), 59 textures (99 MB) [hos host] vk: host time in GPU resource waits: 6.26 ms a frame (fences 2.80, timelines 3.47, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 29.8 full barriers recorded, 30.7 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 13.3; 0x10: 0.3; 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 13.6, an op 15.3, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 1.0 ms a frame (the longest submission 2.7 ms); 0.4 textures made a frame: colour targets 0.3 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.01 ms [hos Presentation Thread] vi: 1323 frames presented (163.8 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 818, 15.5-17.8 0, 17.8-22 0, 22-30 1, 30-40 0, over 40 1 [hos host] vk: late pages' fix (600 recordings with one): of 705 captures with late pages 705 fixed (0.9 pages, 0.9 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 0 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 64 blocks (4096 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 0, of an earlier recording 0, not recorded yet 0, of this part 0; pages fixed from stamps 600, from snapshots 0; the shadows: 605 pages synced from captures, 55 made known by it, 550 left unknown because the replay or the collection wrote them since [hos AudioFrameworkThread] audio: voice 8: 48000 Hz, 1 channels, format 6, parameter 0x1089637980 (32 bytes) at 106898.7 s [hos AudioFrameworkThread] audio: voice 9: 48000 Hz, 1 channels, format 6, parameter 0x1089637bc0 (32 bytes) at 106898.7 s [hos host] vk: sparse mirrors: 6, 76 MB with memory of 100 MB (23 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 15 binds (2.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 2 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 2 spares now, 24 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 5610 dirty pages from their submits' captures, 29550 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 300 dispatches, 2578 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 8 live (124 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.03 over 1662 ranges, 1624 MB), 146 pages written, 146 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (112 MB), 10 spare buffers, 0 texture pages copied (0 MB) [hos ModuleSystemWorker1] prompts: the font at 0x7fd37800e8: the icons in texts follow at once [hos host] vk: texture storage read as submitted (1 times; 2 pages from the copies, 0 of them rewritten in guest memory since their submit; 6 pages kept, 6 spans, 317 MB) [hos host] vk: texture #65 at 0x10000000000 beside #6: 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #65 at 0x10000000000 beside #19: 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #65 at 0x10000000000 beside #30: 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #65 at 0x10000000000 beside #41: 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #69 at 0x10000000000 beside #6: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #69 at 0x10000000000 beside #19: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #69 at 0x10000000000 beside #30: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #69 at 0x10000000000 beside #41: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #69 at 0x10000000000 beside #65: 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #73 at 0x10000000000 beside #6: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #73 at 0x10000000000 beside #19: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #73 at 0x10000000000 beside #30: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #73 at 0x10000000000 beside #41: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #73 at 0x10000000000 beside #65: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #73 at 0x10000000000 beside #69: 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1, the other 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #75 at 0x10000000000 beside #6: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #75 at 0x10000000000 beside #19: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 262x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #75 at 0x10000000000 beside #30: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 306x104x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #75 at 0x10000000000 beside #41: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 656x113x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #75 at 0x10000000000 beside #65: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 154x98x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #75 at 0x10000000000 beside #69: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 138x96x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #75 at 0x10000000000 beside #73: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 106x105x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] [debug] vk: a staging arena of 16 MB took 9.2 ms to make [hos host] vk: staging arenas: 2 (16 MB in this one) [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 3, mip filter nearest, LOD bias 0; made with 0, an image of one level: not the row's) [hos host] vk: import 3.9 ms: commit+properties 0.0, buffer 0.0, allocate 3.9, bind+address 0.0 [hos host] vk: pool 0x102e645400+0x20 imported in 3.9 ms as 0x102e400000..0x102f800000 (10 imports, 34 ms in all) [hos host] vk: sampler 318 is nearest: min filter 2, mag filter 0, wrap 7/7/7, lod -1000..1000 bias -0.5, aniso 1, compare 0 [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 2, mip filter nearest, LOD bias -0.5; made with 0, an image of one level: not the row's) [hos host] vk: slow submit: 21.8 ms: draw 1.8 texture lookups 16.3 uploads 15.9 [hos host] vk: 3 textures retired (120 left, 139 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: images are destroyed on a thread of their own (RX_VK_REAPER=0: on the render thread) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07 [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 5946 dirty pages from their submits' captures, 30445 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 343 dispatches, 2758 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a frame: 5 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.31, present 0.05, draw 0.26, descriptors 0.12, texture lookups 0.18, uploads 0.16, pipelines 0.01, programs 0.03, commands 0.07, dispatch 0.06, memory sync 0.49, queue submit 0.00, slot wait 0.14 [hos host] vk: a frame: mirrors 11 (94 MB): 2.05 MB synced in 7.1 copies, 0.00 MB written back [hos host] vk: 1440 presents, 160.8 fps; a frame: 49 draws (12 skipped), 2 dispatches, 9 passes, 22 barriers, 6 submits, 0.5 uploads (0.2 MB), 10 ub updates (6 overlays), GPU-bound wait 0.0 ms (recording slots only); 9026 pipelines (2081 ms), 0 programs (10 failed), 124 textures (140 MB) [hos host] vk: host time in GPU resource waits: 5.69 ms a frame (fences 2.29, timelines 3.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 41.4 full barriers recorded, 38.1 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 14.0; 0x10: 1.5; 0x12: 6.5; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 21.1, an op 19.6, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 1.5 ms a frame (the longest submission 3.9 ms); 0.6 textures made a frame: colour targets 0.4 (0.2 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.15 MB unswizzled on the CPU in 0.08 ms [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 150 pages written, 150 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (146 MB), 10 spare buffers, 64 texture pages copied (0 MB) [hos AudioFrameworkThread] audio: voice 10: 48000 Hz, 1 channels, format 6, parameter 0x1089637e00 (32 bytes) at 106899.6 s [hos AudioFrameworkThread] audio: voice 11: 48000 Hz, 1 channels, format 6, parameter 0x1089638040 (32 bytes) at 106899.6 s [hos ModuleSystemWorker4] [debug] nvn: 1 copies of recordings whose control memory was reused let go (20830 live) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 64 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 199 pages in 199 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.08 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 57 KB; 0 made on demand (0 MB kept now); c1 stages: at most 3 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 7, 94 MB with memory of 120 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 16 binds (2.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6473 dirty pages from their submits' captures, 29758 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2747 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 74 in captures, 0 only in the collection, 1 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: queues ordered by their events: 6000 submissions stopped at a wait, 5997 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 6024, not met 0; captures at a CPU event's wait 1482, 0 of them at a wait met on arrival [hos host] vk: 3 textures retired (122 left, 142 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: texture #129 at 0x107cdcc600 beside #108: 256x32x1 x1 levels 1 layout 0x80000100 class 0x40001 NVN 0x1 target 1, the other 256x39x1 x1 levels 1 layout 0x80000100 class 0x40001 NVN 0x1 target 1 [hos host] vk: texture #131 at 0x107cdb9600 beside #104: 512x46x1 x1 levels 1 layout 0x80000200 class 0x40001 NVN 0x1 target 1, the other 256x39x1 x1 levels 1 layout 0x80000100 class 0x40001 NVN 0x1 target 1 [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 148 pages written, 148 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (142 MB), 9 spare buffers, 64 texture pages copied (0 MB) [hos host] vk: a frame: 9 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.22, present 0.05, draw 0.25, descriptors 0.11, texture lookups 0.02, uploads 0.00, pipelines 0.01, programs 0.03, commands 0.07, dispatch 0.07, memory sync 0.48, queue submit 0.00, slot wait 0.13 [hos host] vk: a frame: mirrors 11 (94 MB): 1.82 MB synced in 8.0 copies, 0.00 MB written back [hos host] vk: 1560 presents, 164.9 fps; a frame: 68 draws (12 skipped), 2 dispatches, 16 passes, 31 barriers, 6 submits, 0.1 uploads (0.0 MB), 21 ub updates (11 overlays), GPU-bound wait 0.0 ms (recording slots only); 9034 pipelines (2082 ms), 0 programs (10 failed), 131 textures (142 MB) [hos host] vk: host time in GPU resource waits: 5.13 ms a frame (fences 2.00, timelines 3.13, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 55.9 full barriers recorded, 47.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 15.0; 0x10: 3.0; 0x12: 13.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 30.0, an op 24.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 1.9 ms a frame (the longest submission 2.2 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6597 dirty pages from their submits' captures, 29211 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 378 dispatches, 2770 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07 [hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 143 pages written, 143 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (100 MB), 10 spare buffers, 64 texture pages copied (0 MB) [hos host] vk: sparse mirrors: 7, 94 MB with memory of 120 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 16 binds (2.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6606 dirty pages from their submits' captures, 27265 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2632 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 27 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 217 pages in 217 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.04, present 0.04, draw 0.17, descriptors 0.08, texture lookups 0.01, uploads 0.00, pipelines 0.01, programs 0.02, commands 0.04, dispatch 0.05, memory sync 0.45, queue submit 0.00, slot wait 0.14 [hos host] vk: a frame: mirrors 11 (94 MB): 1.81 MB synced in 7.7 copies, 0.00 MB written back [hos host] vk: 1680 presents, 165.4 fps; a frame: 47 draws (12 skipped), 2 dispatches, 16 passes, 31 barriers, 6 submits, 0.0 uploads (0.0 MB), 21 ub updates (11 overlays), GPU-bound wait 0.0 ms (recording slots only); 9034 pipelines (2082 ms), 0 programs (10 failed), 134 textures (143 MB) [hos host] vk: host time in GPU resource waits: 5.61 ms a frame (fences 2.19, timelines 3.42, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 56.1 full barriers recorded, 47.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 15.0; 0x10: 3.0; 0x12: 13.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 30.0, an op 25.7, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 1.7 ms a frame (the longest submission 1.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07 [hos sead::MainThread] barrier: 18848 looks: 2464 found the masks moved already, 16384 waited until they moved, 0 for the whole 1 ms (a wait 77 us on average); 19426 wakes [hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 147 pages written, 147 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (88 MB), 9 spare buffers, 64 texture pages copied (0 MB) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6153 dirty pages from their submits' captures, 28364 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2672 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: presenter thread: the copy's fence 2.32 ms, the present 0.56 ms a present [hos host] vk: memory: working set 10435 MB (peak 10435), private 12926 MB, staging 32 MB in 2 arenas, 2 made so far (16 ms) [hos host] vk: memory heaps: heap 0 (system): 12 blocks 592 MB, 36 allocations 532 MB; used 776 of 15579 MB; heap 1 (device): 85 blocks 1224 MB, 1849 allocations 433 MB; used 2833 of 11474 MB; [hos host] vk: image blocks: 2 blocks 64 MB, 2 images 63 MB; 1 free ranges, the largest 0 MB [hos host] vk: image blocks (render targets): 3 blocks 96 MB, 12 images 84 MB; 3 free ranges, the largest 10 MB [hos host] vk: image blocks (textures, compacted): 24 blocks 768 MB, 122 images 59 MB; 35 free ranges, the largest 32 MB [hos host] vk: render target images: 0 reused from the pool, 12 made new, 0 destroyed from it, 0 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 0 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 12 with the pool empty (0 of them with targets still in the dead queue), 0 with none in it a present yet, 0 with none of a kind they can use, 0 with all smaller (the largest 0% short on average), 0 with all bigger (over 1.5 times), 0 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 0 idle [hos host] vk: VRAM: 2833 MB used of 11474 MB allowed (790 MB of it free in the blocks); textures 143 MB (budget 7158 MB), 134 textures [hos host] vk: budget: VRAM 1224 of 11474 MB (2833 MB by the driver, 1609 MB unaccounted): blocks free 790, textures 143, other buffers 96, mirrors 94, other images 63, staging 32, shadows 4; RAM 716 MB (the process's private memory 12926 MB): capture spares 396, imports 144, captures 102, other buffers 34, retired imports 20, capture table 20, others 0 [hos host] vk: textures: decoded ASTC 30 MB, rendered 106 MB (39 drawn in one frame, kept: 26702 KB), other 6 MB; used in the last 8 presents 101 MB [hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 10 MB of blocks hashed in 1 ms [hos host] vk: 14 MB: 0x1004137000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 2 presents ago [hos host] vk: 14 MB: 0x1001f0ffe00 2560x1440x1 x1 levels 1 NVN 0x3f rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x10000348400 2560x1440x1 x1 levels 1 NVN 0x38 rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x1005f37000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 3 presents ago [hos host] vk: 14 MB: 0x1005037000 2560x1440x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago [hos host] vk: 8 MB: 0x7fcd55d200 1664x1280x1 x1 levels 1 NVN 0x87 ASTC, last used 417 presents ago [hos host] vk: 7 MB: 0x10755b4000 1920x1080x1 x1 levels 1 NVN 0x25 rendered, last used 1 presents ago [hos host] vk: 7 MB: 0x7fc9cd8400 1440x1440x1 x1 levels 1 NVN 0x87 ASTC, last used 280 presents ago [hos Presentation Thread] vk: presents after the CPU's signal (600): 599 waited for it, 3.49 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07 [hos host] vk: status: 600 frames, the slowest 27.6 ms; VRAM 2.8 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 0 program records loaded (0.0 ms), 0 new graphics pipelines, pipelines from the driver's cache 0, from the start's cache 0, compiled 0; 0 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 3398 queued, 3398 loaded (median 1.15 ms, 99th percentile 17.66 ms, worst 34.27 ms, 8522 ms and 203.4 MB read in all); 4300 programs (583 of them compute variants without their code, 0 graphics programs no note names), 7300 graphics pipelines (0 not in a cache), 365 compute pipelines (19 not in a cache); a first use ran 0 itself and waited for 0 (0.0 ms); 36 pipelines made at create taken by draws; 1564 create calls with nothing kept for their stages, 0 before the index; units 4962 (3398 ready, 0 loading), 9312 programs alive; 0 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 7312 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 757 known, 0 learned in this run; 365 compute pipelines made by them at a create call, 19 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.03, present 0.05, draw 0.16, descriptors 0.07, texture lookups 0.01, uploads 0.00, pipelines 0.01, programs 0.02, commands 0.04, dispatch 0.05, memory sync 0.46, queue submit 0.00, slot wait 0.14 [hos host] vk: a frame: mirrors 11 (94 MB): 1.81 MB synced in 7.7 copies, 0.00 MB written back [hos host] vk: 1800 presents, 164.8 fps; a frame: 46 draws (12 skipped), 2 dispatches, 16 passes, 31 barriers, 6 submits, 0.0 uploads (0.0 MB), 21 ub updates (11 overlays), GPU-bound wait 0.0 ms (recording slots only); 9034 pipelines (2082 ms), 0 programs (10 failed), 134 textures (143 MB) [hos host] vk: host time in GPU resource waits: 5.60 ms a frame (fences 2.15, timelines 3.46, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 56.0 full barriers recorded, 47.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 15.0; 0x10: 3.0; 0x12: 13.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 30.0, an op 25.8, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 1.7 ms a frame (the longest submission 1.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 6.10 ms a frame): its jobs 6.08 ms a frame (submits 0.78, presents 4.84, others 0.46), idle 0.02; the game waited for room at 595 presents, 1.63 ms each (1.62 a frame; 14.26 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 600 captures at a CPU signal; the submits' waits read in 0.002 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.69 ms after its hand-over (31.16 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.36 + 0.16 (the window's copy is in the idle), after it 0.03 + 0.00; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 1.00 parts a frame recorded in the signal's job, 0.2% of them took nothing that came with the signal; of the others, 1.00 a frame took pages of the signal's capture into a mirror (77.6 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 1 of 600 frames, its first three in 1 [hos host] vk: frame ahead, what came with the signal: 600 signals with a capture, each of 80.0 pages in 3.0 pools; between a frame's last submit and that signal 0.04 texture and sampler registrations, 0.01 releases and 0.00 other CPU signals a signal (16 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.03+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos sead::MainThread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 146 pages written, 146 copied (0.6 MB) [hos host] vk: sparse mirrors: 7, 94 MB with memory of 120 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 16 binds (2.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6490 dirty pages from their submits' captures, 28901 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2728 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (110 MB), 9 spare buffers, 64 texture pages copied (0 MB) [hos host] vk: late pages' fix (600 recordings with one): of 657 captures with late pages 657 fixed (0.9 pages, 0.9 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 0 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 64 blocks (4096 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 0, of an earlier recording 0, not recorded yet 0, of this part 0; pages fixed from stamps 600, from snapshots 0; the shadows: 493 pages synced from captures, 0 made known by it, 493 left unknown because the replay or the collection wrote them since [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 221 pages in 221 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] nvn: GPU waits in 5 s: 813 sync waits (2746 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 57 KB; 0 made on demand (0 MB kept now); c1 stages: at most 4 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6491 dirty pages from their submits' captures, 26506 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 375 dispatches, 2576 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.03 over 1663 ranges, 1627 MB), 142 pages written, 142 copied (0.6 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (124 MB), 9 spare buffers, 64 texture pages copied (0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07 [hos BaseProcExecutor1] vcores: 186020 takes, 20174 waits (mean 23 us, longest 5.2 ms, 0 over HOS_VCORE_LOG_MS), 8608 kicks, 8371 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.06, present 0.05, draw 0.19, descriptors 0.08, texture lookups 0.02, uploads 0.00, pipelines 0.01, programs 0.02, commands 0.06, dispatch 0.05, memory sync 0.47, queue submit 0.00, slot wait 0.13 [hos host] vk: a frame: mirrors 11 (94 MB): 1.73 MB synced in 7.7 copies, 0.00 MB written back [hos host] vk: 1920 presents, 165.0 fps; a frame: 46 draws (12 skipped), 2 dispatches, 16 passes, 31 barriers, 6 submits, 0.0 uploads (0.0 MB), 21 ub updates (11 overlays), GPU-bound wait 0.0 ms (recording slots only); 9034 pipelines (2082 ms), 0 programs (10 failed), 134 textures (143 MB) [hos host] vk: host time in GPU resource waits: 5.61 ms a frame (fences 2.15, timelines 3.46, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 12.0 [hos host] vk: a frame: 56.2 full barriers recorded, 47.0 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 15.0; 0x10: 3.0; 0x12: 13.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 30.0, an op 24.5, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 1.7 ms a frame (the longest submission 2.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: sparse mirrors: 7, 94 MB with memory of 120 MB (25 MB spared); blocks given memory later 37, given back 2; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 16 binds (2.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 9298 dirty pages from their submits' captures, 24574 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 376 dispatches, 2633 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [a64] unimplemented import _ZN2nn5movie8FinalizeEv (lr 0x80df26c0) [hos MovieOperateThread] video: a movie of 690 pictures on dav1d: 1.6 ms to decode each (worst 41.1), one shown every 16.7 ms; 15 skipped to keep up with its sound; 422 submissions put off while 16 waited [hos MovieOperateThread] video: the longest wait between two of its pictures 289 ms [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 1665 ranges, 1627 MB), 139 pages written, 139 copied (0.5 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (96 MB), 11 spare buffers, 64 texture pages copied (0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.09 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6717 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 207 pages in 207 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vi: the movie is over: back to the game's frame rate cap 60 [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1619 dirty pages from their submits' captures, 5244 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 251 dispatches, 738 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 75 runs, 0.0 MB compared; 75 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 75 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a frame: 1 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.08, present 0.06, draw 0.13, descriptors 0.05, texture lookups 0.01, uploads 0.00, pipelines 0.01, programs 0.01, commands 0.05, dispatch 0.03, memory sync 0.56, queue submit 0.00, slot wait 0.16 [hos host] vk: a frame: mirrors 11 (94 MB): 0.72 MB synced in 5.4 copies, 0.00 MB written back [hos host] vk: 2040 presents, 115.0 fps; a frame: 20 draws (5 skipped), 1 dispatches, 7 passes, 12 barriers, 6 submits, 0.0 uploads (0.0 MB), 8 ub updates (4 overlays), GPU-bound wait 0.0 ms (recording slots only); 9173 pipelines (2095 ms), 0 programs (10 failed), 138 textures (143 MB) [hos host] vk: host time in GPU resource waits: 5.26 ms a frame (fences 3.02, timelines 2.24, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: skips a frame by vk_draw.cpp line: 11682: 4.7 [hos host] vk: a frame: 32.9 full barriers recorded, 25.8 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.7; 0x10: 1.0; 0x12: 4.9; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 10.6, an op 9.9, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 1.1 ms a frame (the longest submission 1.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos Presentation Thread] vi: 2060 frames presented (147.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 685, 15.5-17.8 50, 17.8-22 1, 22-30 0, 30-40 1, over 40 0 [hos sead::MainThread] vk: capture: a submit 0.08 ms (lock 0.00, scan 0.04 over 1672 ranges, 1626 MB), 92 pages written, 92 copied (0.4 MB) [hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (108 MB), 8 spare buffers, 3531 texture pages copied (13 MB) [hos host] vk: 1 textures retired (137 left, 139 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.00, writes collected 0.05 [hos host] vk: sparse mirrors: 7, 36 MB with memory of 120 MB (83 MB spared); blocks given memory later 37, given back 930; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 17 binds (2.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1 dirty pages from their submits' captures, 303 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 228 dispatches, 304 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 77 runs, 0.0 MB compared; 77 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 77 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9807 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Resource Memory] mem: the game's mapped pages 2580 MB (peak 2580); heap 6894 MB; the console's total 9809 MB [hos bake::ModuleMgr] vcores: bake::ModuleMgr (priority 28) waited 54.1 ms for a core: given by aglTextureDataMgr::Compaction (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos GeneralFormatAsync_] vcores: GeneralFormatAsync_ (priority 24) waited 64.1 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos BancModule] vcores: BancModule (priority 24) waited 110.6 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos GeneralFormatAsync_] vcores: GeneralFormatAsync_ (priority 24) waited 133.7 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 1702 ranges, 1672 MB), 127 pages written, 127 copied (0.5 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (116 MB), 10 spare buffers, 6391 texture pages copied (24 MB) [hos BancModule] vcores: BancModule (priority 24) waited 183.0 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos GeneralFormatAsync_] vcores: GeneralFormatAsync_ (priority 24) waited 124.9 ms for a core: given by nn.npln.Callback (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos BancModule] vcores: BancModule (priority 24) waited 50.6 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 0.94, present 0.06, draw 0.07, descriptors 0.02, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.01, commands 0.03, dispatch 0.00, memory sync 0.47, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 11 (30 MB): 0.02 MB synced in 4.1 copies, 0.00 MB written back [hos host] vk: 2160 presents, 60.0 fps; a frame: 5 draws (0 skipped), 0 dispatches, 2 passes, 1 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 9410 pipelines (2131 ms), 0 programs (10 failed), 137 textures (139 MB) [hos host] vk: host time in GPU resource waits: 4.16 ms a frame (fences 4.16, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 18.9 full barriers recorded, 13.8 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 0.0, an op 1.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos GeneralFormatAsync_] vcores: GeneralFormatAsync_ (priority 24) waited 281.0 ms for a core: given by BaseProcCreateAndDeleteThread (priority 23), which held it 4.5 ms and was never asked; the cores it may use now: core 1: it; [hos BancModule] vcores: BancModule (priority 24) waited 256.0 ms for a core: given by BaseProcCreateAndDeleteThread (priority 23), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 634.5 ms for a core: given by BancModule (priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 578.7 ms for a core: given by tera::FileResourceMgr (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos bake::ModuleMgr] vcores: bake::ModuleMgr (priority 28) waited 603.1 ms for a core: given by PreActorModelCompose (priority 26), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos BaseProcCreateAndDeleteThread] nvn: 100000 recordings made, 21028 live copies (1.9 MB), 53 texture copies; 2 command buffers' segments (0.2 MB) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 5 dirty pages from their submits' captures, 325 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 231 dispatches, 330 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 76 runs, 0.0 MB compared; 76 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 93 in captures, 1 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.10 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 1495 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.02, scan 0.09 over 1710 ranges, 1618 MB), 97 pages written, 97 copied (0.4 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 6401 texture pages copied (25 MB) [hos host] vk: recorder: 348216 packets run in 10.0 s (busy 0.62 s, spun idle 0.12 s, slept 11151 times, 5850 of them under 200 us), 0 calls made directly (no room); 19639 batches over 1232 frames (15.9 a frame, 18 packets a batch), 9816 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 7 times, 0.86 ms (0.0 times, 0.00 ms a frame); every thread 650 times, 26.52 ms in all, the longest 0.97 ms; the queue's lock 365/13.60 ms (max 0.97), vkWaitForFences 267/11.80 ms (max 0.35), vkQueueBindSparse 5/0.84 ms (max 0.57), vkDestroyImageView 3/0.12 ms (max 0.08), vmaDestroyBuffer 6/0.11 ms (max 0.03), vmaFreeMemory 2/0.03 ms (max 0.02) [hos host] vk: recorder: of its busy 0.62 s, 0.43 s in parts' ends and submits (the longest 1.01 ms), 0.03 s of that blocked on the queue's lock (64 times, the longest 0.96 ms), so 0.40 s in the ends, submits and holds; 0.19 s running the other packets; 283 packets a frame [hos host] vk: the recording streams: 48 of 1 MB, the most one held 57 KB; 0 made on demand (0 MB kept now); c1 stages: at most 4 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 0 sets (ub 0, sb 0, cis 0, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 7, 30 MB with memory of 120 MB (89 MB spared); blocks given memory later 37, given back 1026; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 18 binds (2.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 3 taken from the spares, 8 made new; 4 kept as spares, 0 destroyed; 1 spares now, 4 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 0 bindings followed, 0 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 7 dirty pages from their submits' captures, 346 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 229 dispatches, 353 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 78 runs, 0.0 MB compared; 78 pages read into the shadows; 4 MB of shadows (0 MB of pages known, in 1 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 119 in captures, 0 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 9 live (144 MB), 1 retired still imported (20 MB), 0 retired given back (0 MB) [hos Resource Control] vcores: 180481 takes, 67693 waits (mean 116 us, longest 634.5 ms, 12 over HOS_VCORE_LOG_MS), 30073 kicks, 28790 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 6746 programs from the program cache [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.53, present 0.07, draw 0.09, descriptors 0.03, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.01, commands 0.04, dispatch 0.00, memory sync 1.02, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 11 (30 MB): 0.02 MB synced in 4.6 copies, 0.00 MB written back [hos host] vk: 2280 presents, 60.0 fps; a frame: 5 draws (0 skipped), 0 dispatches, 2 passes, 1 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 9972 pipelines (2219 ms), 0 programs (10 failed), 137 textures (139 MB) [hos host] vk: host time in GPU resource waits: 4.17 ms a frame (fences 4.17, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 18.7 full barriers recorded, 13.7 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 0.0, an op 1.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 0.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos Resource Control] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 31 made): 10195546 lookups, 134132 missed (1.32 %) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.14 [hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.02, scan 0.10 over 1866 ranges, 1622 MB), 108 pages written, 108 copied (0.4 MB) [hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 6515 texture pages copied (25 MB) [hos ModuleSystemWorker3] shadow: the sun's nearest cascade keeps the game's fit below Shadow resolution Highest (1600 texels) [hos sead::MainThread] nvn: the game reports counter type 1 (lr 0x800876dc) [hos host] vk: texture 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e -> vk 137 (ASTC decoded) (139 textures, 357 MB) [hos host] [debug] vk: a staging arena of 57 MB took 8.5 ms to make [hos host] vk: staging arenas: 3 (57 MB in this one) [hos host] vk: vertex format NVN 0xa -> vk 100 [hos host] vk: slow upload 179.1 ms: #146 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC (BC3), 57825 of 57825 KB [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 5, mip filter linear, LOD bias 0; made with 0) [hos host] vk: texture #149 at 0x10000000000 beside #6: 256x144x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: import 2.8 ms: commit+properties 0.1, buffer 0.0, allocate 2.7, bind+address 0.0 [hos host] vk: pool 0x113b681b00+0x90 imported in 2.8 ms as 0x113b400000..0x113cc00000 (11 imports, 37 ms in all) [hos host] vk: sampler 409 is nearest: min filter 2, mag filter 0, wrap 7/7/7, lod 0..15 bias 0, aniso 1, compare 0 [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 2, mip filter nearest, LOD bias 0; made with 0, an image of one level: not the row's) [hos host] vk: mesh draw kind 0: 0 3 512, indirect 0 count 0 max 0 stride 0 [hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256] [hos host] vk: import 3.5 ms: commit+properties 0.3, buffer 0.0, allocate 3.2, bind+address 0.0 [hos host] vk: pool 0x7fab0e5000+0x14300 imported in 3.5 ms as 0x7fab000000..0x7faf000000 (14 imports, 42 ms in all) [hos host] vk: mesh draw kind 0: 0 3 512, indirect 0 count 0 max 0 stride 0 [hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256] [hos host] vk: mesh draw kind 0: 0 3 512, indirect 0 count 0 max 0 stride 0 [hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256] [hos host] vk: import 4.0 ms: commit+properties 0.2, buffer 0.0, allocate 3.8, bind+address 0.0 [hos host] vk: pool 0x104eafe800+0x60 imported in 4.0 ms as 0x104e000000..0x104f800000 (15 imports, 46 ms in all) [hos host] vk: import 2.4 ms: commit+properties 0.1, buffer 0.0, allocate 2.3, bind+address 0.0 [hos host] vk: pool 0x7fd3c04000+0x69800 imported in 2.4 ms as 0x7fd3c00000..0x7fd4c00000 (16 imports, 49 ms in all) [hos host] vk: mesh draw kind 0: 0 2 512, indirect 0 count 0 max 0 stride 0 [hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256] [hos host] vk: import 4.8 ms: commit+properties 0.2, buffer 0.0, allocate 4.6, bind+address 0.0 [hos host] vk: pool 0x7fb7df5000+0x35900 imported in 4.8 ms as 0x7fb5c00000..0x7fb8800000 (17 imports, 53 ms in all) [hos host] vk: mesh draw kind 0: 0 4 512, indirect 0 count 0 max 0 stride 0 [hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256] [hos host] vk: mesh draw kind 0: 0 4 512, indirect 0 count 0 max 0 stride 0 [hos host] vk: mesh draw: program 1bd1, 3 stages: [0: stage 1, code 0x7ffb119000, sph type 5 bits24-25 0, isbe 0 payload 0] [1: stage 6, code 0x7ffae43c00, sph type 1 bits24-25 0, isbe 1 payload 256] [2: stage 7, code 0x7ffb119200, sph type 3 bits24-25 2, isbe 1 payload 256] [hos host] vk: mesh draw kind 0: 0 5 512, indirect 0 count 0 max 0 stride 0 [hos host] vk: mesh draw kind 0: 0 15 512, indirect 0 count 0 max 0 stride 0 [hos host] vk: pool 0x7fb8f04000+0x1b2800 imported in 5.9 ms as 0x7fb8800000..0x7fbbc00000 (18 imports, 59 ms in all) [hos host] vk: pool 0x7fd1fbe200+0x76c imported in 4.2 ms as 0x7fd1800000..0x7fd2400000 (19 imports, 64 ms in all) [hos host] vk: pool 0x7fb4943f00+0xa2ca8 imported in 2.4 ms as 0x7fb4400000..0x7fb5c00000 (20 imports, 66 ms in all) [hos host] vk: pool 0x7fd6367b00+0x13390 imported in 3.0 ms as 0x7fd4c00000..0x7fd6400000 (21 imports, 69 ms in all) [hos host] vk: pool 0x7fbe136600+0x7560 imported in 4.1 ms as 0x7fbd800000..0x7fbf800000 (22 imports, 73 ms in all) [hos host] vk: pool 0x7fb23dce00+0x1ad9c imported in 3.8 ms as 0x7fb2000000..0x7fb4400000 (23 imports, 77 ms in all) [hos host] vk: pool 0x7fb1e73900+0xb61c imported in 3.6 ms as 0x7fb0800000..0x7fb2000000 (24 imports, 80 ms in all) [hos host] vk: pool 0x7fbcda0b00+0x8e68 imported in 3.3 ms as 0x7fbcc00000..0x7fbd800000 (25 imports, 84 ms in all) [hos host] vk: pool 0x7fd694e300+0x150 imported in 3.8 ms as 0x7fd6400000..0x7fd9400000 (27 imports, 89 ms in all) [hos host] vk: pool 0x7fc20d8900+0x11288 imported in 8.0 ms as 0x7fc1400000..0x7fc3000000 (28 imports, 97 ms in all) [hos host] vk: pool 0x7fc120db00+0x1b8dc imported in 2.8 ms as 0x7fbf800000..0x7fc1400000 (29 imports, 100 ms in all) [hos host] vk: pool 0x7faf0a7f00+0x93e4 imported in 4.1 ms as 0x7faf000000..0x7faf400000 (30 imports, 104 ms in all) [hos host] vk: pool 0x7fbc4fd200+0x1f8a64 imported in 2.8 ms as 0x7fbc000000..0x7fbcc00000 (31 imports, 106 ms in all) [hos host] vk: pool 0x7fd254ff00+0x76c64 imported in 4.0 ms as 0x7fd2400000..0x7fd2800000 (32 imports, 110 ms in all) [hos host] vk: pool 0x7faefcb100+0x874c8 imported in 7.9 ms as 0x7fab000000..0x7faf400000 (33 imports, 118 ms in all) [hos host] vk: pool 0x7faf000000 remade larger (2 retired) [hos host] vk: pool 0x7fab000000 remade larger (3 retired) [hos host] vk: pool 0x7fab000000..0x7faf400000 took in 2 pools [hos host] vk: pool 0x7fab000000..0x7faf400000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: pool 0x7fdf775200+0x960 imported in 13.4 ms as 0x7fdf000000..0x7fdfc00000 (34 imports, 132 ms in all) [hos host] vk: pool 0x7fd3babd00+0x7b5c imported in 4.0 ms as 0x7fd3800000..0x7fd3c00000 (35 imports, 136 ms in all) [hos host] vk: sampler anisotropy: the game asks for 2 (min filter 5, mip filter linear, LOD bias -0.5; made with 0) [hos host] vk: sampler anisotropy: the game asks for 0 (min filter 1, mip filter nearest, LOD bias 0; made with 0) [hos host] vk: texture slot 0x10b4 (handle 0x18c010b4): no image for it; NVN target 1 format 0x8c 683x683x1 levels 10 gpu 0x10d58f2200, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc30 (handle 0x18c00c30): no image for it; NVN target 1 format 0x8b 512x512x1 levels 10 gpu 0x10c1776600, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xe5d (handle 0x18c00e5d): no image for it; NVN target 1 format 0x8b 512x512x1 levels 10 gpu 0x10cd3e7a00, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0x1001 (handle 0x18c01001): no image for it; NVN target 1 format 0x8a 341x341x1 levels 9 gpu 0x10d385da00, view yes (format 0x8a target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0x1011 (handle 0x18c01011): no image for it; NVN target 1 format 0x8b 512x512x1 levels 10 gpu 0x10d39c3600, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0x10e9 (handle 0x18c010e9): no image for it; NVN target 1 format 0x8d 171x171x1 levels 8 gpu 0x10d6b62a00, view yes (format 0x8d target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0x1060 (handle 0x18c01060): no image for it; NVN target 1 format 0x88 512x512x1 levels 10 gpu 0x10d4d6c800, view yes (format 0x88 target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc33 (handle 0x18c00c33): no image for it; NVN target 1 format 0x8d 1024x1024x1 levels 11 gpu 0x10c19fd600, view yes (format 0x8d target 1), wanted view type 1 use 0 [hos host] vk: texture #192 at 0x10000000000 beside #6: 128x128x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #192 at 0x10000000000 beside #149: 128x128x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 256x144x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: pool 0x108c6f2540+0x40 imported in 5.4 ms as 0x108c000000..0x108f000000 (36 imports, 141 ms in all) [hos host] vk: texture slot 0xe37 (handle 0x18c00e37): no image for it; NVN target 1 format 0x8d 85x341x1 levels 9 gpu 0x10cc819400, view yes (format 0x8d target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xe38 (handle 0x18c00e38): no image for it; NVN target 1 format 0x7b 128x512x1 levels 10 gpu 0x10cc81e600, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: pool 0x10e5be9600+0x2028000 imported in 6.6 ms as 0x10e5800000..0x10e8000000 (37 imports, 148 ms in all) [hos host] vk: texture #197 at 0x113755c000 beside #159: 4096x4096x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0xa target 1, the other 4096x4096x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: BC image 1120x1516 (vk 143) as a raw colour attachment: the driver says -11 [hos host] vk: BC image 2240x3032 (vk 141) as a raw colour attachment: the driver says -11 [hos host] vk: texture slot 0xcfd (handle 0x18c00cfd): no image for it; NVN target 1 format 0x8b 1024x1024x1 levels 11 gpu 0x10c6aaa000, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc3a (handle 0x18c00c3a): no image for it; NVN target 1 format 0x8b 1024x1024x1 levels 11 gpu 0x10c1b0b200, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: BC image 1528x920 (vk 143) as a raw colour attachment: the driver says -11 [hos host] vk: staging arenas: 4 (16 MB in this one) [hos host] vk: BC image 3056x1840 (vk 141) as a raw colour attachment: the driver says -11 [hos host] vk: texture slot 0xc70 (handle 0x18c00c70): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c350d200, view yes (format 0x8d target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc71 (handle 0x18c00c71): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c352c000, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xafa (handle 0x18c00afa): no image for it; NVN target 1 format 0x8b 1024x1024x1 levels 11 gpu 0x10bb07d600, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xafc (handle 0x18c00afc): no image for it; NVN target 1 format 0x7f 768x768x1 levels 10 gpu 0x10bb319c00, view yes (format 0x7f target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xad6 (handle 0x18c00ad6): no image for it; NVN target 1 format 0x8b 341x341x1 levels 9 gpu 0x10ba37e000, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xad7 (handle 0x18c00ad7): no image for it; NVN target 1 format 0x7b 512x512x1 levels 10 gpu 0x10ba393000, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc18 (handle 0x18c00c18): no image for it; NVN target 1 format 0x8c 512x512x1 levels 10 gpu 0x10c10f8a00, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc19 (handle 0x18c00c19): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c111d000, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd57 (handle 0x18c00d57): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c8380e00, view yes (format 0x8d target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd58 (handle 0x18c00d58): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c839fc00, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc84 (handle 0x18c00c84): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c3e51600, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc49 (handle 0x18c00c49): no image for it; NVN target 1 format 0x8c 683x683x1 levels 10 gpu 0x10c249ee00, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc4a (handle 0x18c00c4a): no image for it; NVN target 1 format 0x7b 1024x1024x1 levels 11 gpu 0x10c24e2e00, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc4b (handle 0x18c00c4b): no image for it; NVN target 1 format 0x7e 512x512x1 levels 10 gpu 0x10c25c7000, view yes (format 0x7e target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd83 (handle 0x18c00d83): no image for it; NVN target 1 format 0x8b 512x512x1 levels 10 gpu 0x10c939de00, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd84 (handle 0x18c00d84): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c93c8600, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xcee (handle 0x18c00cee): no image for it; NVN target 1 format 0x8b 683x683x1 levels 10 gpu 0x10c5e2aa00, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xcef (handle 0x18c00cef): no image for it; NVN target 1 format 0x7b 1024x1024x1 levels 11 gpu 0x10c5e76000, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xcf0 (handle 0x18c00cf0): no image for it; NVN target 1 format 0x7e 512x512x1 levels 10 gpu 0x10c5f5a200, view yes (format 0x7e target 1), wanted view type 1 use 0 [hos host] [debug] vk: a staging arena of 16 MB took 12.4 ms to make [hos host] vk: texture slot 0xcfa (handle 0x18c00cfa): no image for it; NVN target 1 format 0x8c 1024x1024x1 levels 11 gpu 0x10c67e1400, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xcfc (handle 0x18c00cfc): no image for it; NVN target 1 format 0x7f 768x768x1 levels 10 gpu 0x10c6a69400, view yes (format 0x7f target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0x919 (handle 0x18c00919): no image for it; NVN target 1 format 0x7b 512x512x1 levels 10 gpu 0x10b2e54200, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xa49 (handle 0x18900a49): no image for it; NVN target 1 format 0x7b 64x64x1 levels 7 gpu 0x10b7a23400, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xaf4 (handle 0x18c00af4): no image for it; NVN target 1 format 0x92 512x512x1 levels 10 gpu 0x10bae49a00, view yes (format 0x92 target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xaf5 (handle 0x18c00af5): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10bae5aa00, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xaf6 (handle 0x18c00af6): no image for it; NVN target 1 format 0x7d 768x768x1 levels 10 gpu 0x10baedee00, view yes (format 0x7d target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xcf4 (handle 0x18c00cf4): no image for it; NVN target 1 format 0x8c 1024x1024x1 levels 11 gpu 0x10c6267600, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: slow upload 32.3 ms: #244 0x10c62f3c00 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB [hos host] vk: texture slot 0xcf6 (handle 0x18c00cf6): no image for it; NVN target 1 format 0x83 768x768x1 levels 10 gpu 0x10c64ef600, view yes (format 0x83 target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xcf7 (handle 0x18c00cf7): no image for it; NVN target 1 format 0x8c 1024x1024x1 levels 11 gpu 0x10c6518800, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xcf9 (handle 0x18c00cf9): no image for it; NVN target 1 format 0x7f 768x768x1 levels 10 gpu 0x10c67a0800, view yes (format 0x7f target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc3c (handle 0x18c00c3c): no image for it; NVN target 1 format 0x7e 768x768x1 levels 10 gpu 0x10c1da7800, view yes (format 0x7e target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc0c (handle 0x18c00c0c): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c0b56400, view yes (format 0x8d target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc0e (handle 0x18c00c0e): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c0b9fa00, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc10 (handle 0x18c00c10): no image for it; NVN target 1 format 0x7f 512x512x1 levels 10 gpu 0x10c0ca8200, view yes (format 0x7f target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xdc1 (handle 0x18c00dc1): no image for it; NVN target 1 format 0x8c 512x512x1 levels 10 gpu 0x10ca474400, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd7f (handle 0x18c00d7f): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c9193a00, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xdc2 (handle 0x18900dc2): no image for it; NVN target 1 format 0x93 341x341x1 levels 9 gpu 0x10ca498a00, view yes (format 0x93 target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xdc3 (handle 0x18900dc3): no image for it; NVN target 1 format 0x7b 512x512x1 levels 10 gpu 0x10ca4a0800, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd81 (handle 0x18900d81): no image for it; NVN target 1 format 0x7e 512x512x1 levels 10 gpu 0x10c9278c00, view yes (format 0x7e target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xdc4 (handle 0x18c00dc4): no image for it; NVN target 1 format 0x7b 512x512x1 levels 10 gpu 0x10ca4dba00, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd7b (handle 0x18c00d7b): no image for it; NVN target 1 format 0x8b 1024x1024x1 levels 11 gpu 0x10c8dc7c00, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd7c (handle 0x18900d7c): no image for it; NVN target 1 format 0x8b 683x683x1 levels 10 gpu 0x10c8e68800, view yes (format 0x8b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xd7e (handle 0x18900d7e): no image for it; NVN target 1 format 0x7b 1024x1024x1 levels 11 gpu 0x10c90af800, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc1b (handle 0x18c00c1b): no image for it; NVN target 1 format 0x8e 512x512x1 levels 10 gpu 0x10c11cca00, view yes (format 0x8e target 1), wanted view type 1 use 0 [hos host] vk: slow upload 15.2 ms: #268 0x10c11e2c00 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB [hos host] vk: texture slot 0xc1f (handle 0x18c00c1f): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c1480a00, view yes (format 0x8d target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc20 (handle 0x18c00c20): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c149f800, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc60 (handle 0x18c00c60): no image for it; NVN target 1 format 0x8d 512x512x1 levels 10 gpu 0x10c2ae4e00, view yes (format 0x8d target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc61 (handle 0x18c00c61): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c2b03c00, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc37 (handle 0x18c00c37): no image for it; NVN target 1 format 0x8c 171x171x1 levels 8 gpu 0x10c1af0c00, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc38 (handle 0x18c00c38): no image for it; NVN target 1 format 0x7b 256x256x1 levels 9 gpu 0x10c1af6e00, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc29 (handle 0x18c00c29): no image for it; NVN target 1 format 0x8c 512x512x1 levels 10 gpu 0x10c1683c00, view yes (format 0x8c target 1), wanted view type 1 use 0 [hos host] vk: texture slot 0xc2a (handle 0x18c00c2a): no image for it; NVN target 1 format 0x7b 768x768x1 levels 10 gpu 0x10c16a8200, view yes (format 0x7b target 1), wanted view type 1 use 0 [hos host] vk: pool 0x104da60000+0xc0 imported in 4.1 ms as 0x104cc00000..0x104dc00000 (38 imports, 152 ms in all) [hos host] vk: slow upload 25.5 ms: #388 0x10c2c3fc00 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB [hos host] [debug] vk: a staging arena of 16 MB took 7.5 ms to make [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 3, mip filter nearest, LOD bias 0; made with 0) [hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 64x64) [hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 32x32) [hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 16x16) [hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 8x8) [hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 4x4) [hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 2x2) [hos host] vk: a draw (program 5aae8ca508400194) samples targets it writes (mask 0x3f, 1x1) [hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 64x64) [hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 32x32) [hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 16x16) [hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 8x8) [hos host] vk: a draw (program 3efdde30088538bf) samples targets it writes (mask 0x3f, 4x4) [hos host] vk: texture #456 at 0x10000000000 beside #6: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 1920x1080x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #456 at 0x10000000000 beside #149: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 256x144x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: texture #456 at 0x10000000000 beside #192: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 128x128x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: slow submit: 606.3 ms: draw 581.0 descriptors 418.7 texture lookups 363.0 uploads 318.6 programs 13.3 commands 129.1 dispatch 6.4 [hos host] vk: pool 0x117883d100+0x700 imported in 18.2 ms as 0x1178400000..0x117c400000 (42 imports, 173 ms in all) [hos host] vk: pool 0x117e724000+0x18000 imported in 4.3 ms as 0x117e400000..0x117e800000 (43 imports, 178 ms in all) [hos host] vk: pool 0x117d058000+0x44b040 imported in 2.2 ms as 0x117d000000..0x117d800000 (45 imports, 181 ms in all) [hos host] vk: pool 0x117cabc000+0x230040 imported in 2.8 ms as 0x117c800000..0x117d000000 (47 imports, 185 ms in all) [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 1, mip filter nearest, LOD bias 0; made with 0, an image of one level: not the row's) [hos host] vk: pool 0x117c4a0000+0xa0040 imported in 2.1 ms as 0x117c400000..0x117c800000 (48 imports, 187 ms in all) [hos host] vk: pool 0x117c800000 remade larger (4 retired) [hos host] vk: pool 0x117c400000 remade larger (5 retired) [hos host] vk: pool 0x117c400000..0x117d000000 took in 2 pools [hos host] vk: pool 0x117c400000..0x117d000000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: pool 0x117d800000 remade larger (6 retired) [hos host] vk: pool 0x117d800000..0x117e000000: mirror made with the GPU's contents of 1 retired pool(s) [hos host] vk: pool 0x117e3d8000+0x8b040 imported in 3.0 ms as 0x117e000000..0x117e800000 (51 imports, 193 ms in all) [hos host] vk: pool 0x117e400000 remade larger (7 retired) [hos host] vk: pool 0x117e000000..0x117e800000: mirror made with the GPU's contents of 1 retired pool(s) [hos host] vk: pool 0x1146b29e00+0x80 imported in 8.3 ms as 0x1145400000..0x1147c00000 (52 imports, 202 ms in all) [hos host] vk: pool 0x115bd19580+0x1b580 imported in 9.7 ms as 0x1159c00000..0x115c000000 (53 imports, 211 ms in all) [hos host] vk: sampler 410 is nearest: min filter 4, mag filter 0, wrap 7/7/7, lod 0..15 bias 0, aniso 1, compare 0 [hos host] vk: slow submit: 114.4 ms: texture lookups 1.7 uploads 1.7 dispatch 114.3 [hos host] vk: the indirect mesh draws' ring (4 MB) made and filled with 1s [hos host] vk: 6 background workers (shader translation, pipelines) [hos host] vk: quick pipelines only for draws that must wait (RX_VK_QUICK_MS=required) [hos host] vk: quick pipeline (no optimizer) for cb71f79f6129aab8 in 13.2 ms, drawn with until its full build is done (1 so far, 13 ms in all, each kept for the session) [hos host] vk: RX_VK_RECORDER=thread: vkCmdCopyBufferToImage did not fit in the room a packet is given (81920 bytes): made directly once the recorder had run what was before it (logged once) [hos host] vk: pool 0x117e800000 remade larger (8 retired) [hos host] vk: pool 0x117e000000..0x117ec00000 took in 2 pools [hos host] vk: pool 0x117e000000..0x117ec00000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: a clear after 1 draws of its pass (2560x1440, 1 colour targets + depth, last program 3bde059fb051af1f): a pass of its own [hos host] vk: sampler 404 is nearest: min filter 0, mag filter 0, wrap 7/7/7, lod 0..15 bias 0, aniso 1, compare 0 [hos host] vk: sampler anisotropy: the game asks for 1 (min filter 0, mip filter nearest, LOD bias 0; made with 0, an image of one level: not the row's) [hos host] vk: a clear after 3 draws of its pass (640x360, 1 colour targets + depth, last program 619664ed509bbd99): a pass of its own [hos host] vk: pool 0x102fde0a00+0x180 imported in 12.8 ms as 0x102fc00000..0x1033000000 (55 imports, 226 ms in all) [hos host] vk: a clear after 3 draws of its pass (640x360, 2 colour targets + depth, last program 2479abad219cb19a): a pass of its own [hos host] vk: sampler 414 is nearest: min filter 5, mag filter 0, wrap 7/7/7, lod 0..15 bias 0, aniso 1, compare 0 [hos host] vk: texture #479 at 0x10004dd1a00 beside #474: 1600x1600x1 x3 levels 1 layout 0x4 class 0x1007c NVN 0x32 target 4, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: table entry bound again at frame 2333: prog 507e946d64f2d55e ssbo 0 of 3 (c0+0x880) addr 0x108c702000 size 0x2550000 (before 0) [hos host] vk: pipeline from the caches for a draw ahead of its build (1 so far, each kept for the session) [hos host] vk: pool 0x7fe04e8a00+0x6888 imported in 5.3 ms as 0x7fdfc00000..0x7fe0800000 (56 imports, 231 ms in all) [hos host] vk: pool 0x7fd37e4400+0x760 imported in 3.3 ms as 0x7fd3400000..0x7fd3800000 (57 imports, 234 ms in all) [hos host] vk: write-back diff: a shadow for pool 0x113b400000 (24 MB; 28 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x1159c00000 (36 MB; 64 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117d000000 (8 MB; 72 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117d800000 (8 MB; 80 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117c400000 (4 MB; 84 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117c400000 (12 MB; 96 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117e400000 (4 MB; 100 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x1178400000 (64 MB; 164 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117d800000 (4 MB; 168 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117e800000 (4 MB; 172 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117c800000 (8 MB; 180 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x117e000000 (8 MB; 188 MB of shadows) [hos host] vk: slow submit: 134.5 ms: draw 14.0 descriptors 11.4 texture lookups 33.1 uploads 9.7 programs 1.3 [hos host] vk: depth-compare sampler: the game's compare mode 1 function 4 -> compare on op 3 (filters min 1 mag 1) [hos rmbl::VibrationThread] hid: 4096 vibration values, strongest 0.00 since the last count [hos host] vk: texture 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 -> vk 124 (499 textures, 1748 MB) [hos host] vk: upload: 12288x12288 NVN 0x32 (288 MB) is all zero: cleared on the GPU (checked in 59.0 ms, 288 MB read) [hos host] vk: slow upload 59.1 ms: #506 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32, 294912 of 294912 KB [hos host] vk: texture 0x10008ac2600 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (500 textures, 1751 MB) [hos host] vk: texture #509 at 0x10008342200 beside #492: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: texture #511 at 0x10004921600 beside #494: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #512 at 0x10004831400 beside #493: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #513 at 0x10004741200 beside #495: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #514 at 0x10009242400 beside #510: 1280x720x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #520 at 0x10004831400 beside #493: 160x90x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x3 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #520 at 0x10004831400 beside #512: 160x90x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x3 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: NVN2 program 3d33bb73d3deee34 (stage 7) lowered with extended registers (577 REGMAP words) [hos host] vk: NVN2 program 463f42630fce83a0 (stage 7) lowered with extended registers (1539 REGMAP words) [hos host] [debug] vk: a staging arena of 16 MB took 11.8 ms to make [hos host] vk: NVN2 program 66c494fe4cfa930b (stage 7) lowered with extended registers (1460 REGMAP words) [hos host] vk: NVN2 program 354dde838352c262 (stage 7) lowered with extended registers (320 REGMAP words) [hos host] vk: NVN2 program e3fc8fcf540f7a48 (stage 7) lowered with extended registers (1503 REGMAP words) [hos host] vk: staging arenas: 16 (16 MB in this one) [hos host] vk: NVN2 program d38b0d52a94ab5a8 (stage 7) lowered with extended registers (337 REGMAP words) [hos host] vk: NVN2 program b509ae330829ce02 (stage 7) lowered with extended registers (264 REGMAP words) [hos host] vk: texture #532 at 0x10001e00400 beside #466: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: NVN2 program bd0ee311e887f8e4 (stage 0) lowered with extended registers (1132 REGMAP words) [hos host] vk: depth-compare sampler: the game's compare mode 0 function 0 -> compare on op 3 (filters min 1 mag 1) [hos host] vk: upload: 4096x4096 NVN 0x32 (32 MB) is all zero: cleared on the GPU (checked in 1.7 ms, 32 MB read) [hos host] vk: NVN2 program 7d8136fe8a9978f0 (stage 0) lowered with extended registers (1148 REGMAP words) [hos host] vk: pool 0x7fdac97800+0xb40 imported in 6.1 ms as 0x7fd9800000..0x7fdc400000 (58 imports, 240 ms in all) [hos host] vk: NVN2 program b764b6275a5f4b66 (stage 0) lowered with extended registers (1129 REGMAP words) [hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own [hos host] vk: texture #537 at 0x100021c0600 beside #536: 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 960x540x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #542 at 0x100023caa00 beside #540: 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1, the other 240x135x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #544 at 0x100023caa00 beside #540: 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1, the other 240x135x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #544 at 0x100023caa00 beside #542: 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1, the other 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #546 at 0x100023caa00 beside #540: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 240x135x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #546 at 0x100023caa00 beside #542: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #546 at 0x100023caa00 beside #544: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: sampler 423 is nearest: min filter 0, mag filter 0, wrap 5/5/5, lod 0..15 bias 0, aniso 1, compare 0 [hos host] vk: sampler 424 is nearest: min filter 2, mag filter 0, wrap 1/1/1, lod 0..15 bias 0, aniso 1, compare 0 [hos host] vk: texture #556 at 0x10004380c00 beside #469: 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40002 NVN 0x5 target 1 (rendered) [hos host] vk: a clear after 11 draws of its pass (2560x1440, 1 colour targets + depth, last program 5986d2784af21135): a pass of its own [hos host] vk: texture #557 at 0x10001e00400 beside #466: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: texture #557 at 0x10001e00400 beside #532: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: pool 0x1012e97e00+0xf00000 imported in 5.4 ms as 0x1012c00000..0x1014000000 (59 imports, 246 ms in all) [hos host] vk: pool 0x100fb07600+0x1c2000 imported in 2.2 ms as 0x100f800000..0x1010000000 (60 imports, 248 ms in all) [hos host] vk: pool 0x7fff4bda00+0x10 imported in 3.6 ms as 0x7ffe000000..0x7fff800000 (61 imports, 251 ms in all) [hos host] vk: sampler 267 is nearest: min filter 0, mag filter 0, wrap 5/5/5, lod -1000..1000 bias 0, aniso 1, compare 0 [hos host] vk: network 9dc98871eb7fc101: dispatch 64x36x1, 1 textures 1 images: 2560x1440x1/122 2560x1440x1/122 [hos host] vk: upscaler: Off: network 9dc98871eb7fc101 left out, its input stretched linearly (2560x1440 -> 2560x1440), packed output [hos host] vk: slow submit: 194.3 ms: draw 118.2 descriptors 109.3 texture lookups 141.8 uploads 88.2 programs 8.4 dispatch 20.8 [hos host] vk: write-back diff: a shadow for pool 0x100f800000 (8 MB; 196 MB of shadows) [hos host] vk: a frame's uploads took 423 ms: ASTC on the GPU 27 (107.0 MB, 290 ms); unswizzled on the GPU 2 (47.2 MB, 12 ms); on the CPU 118 (503.5 MB, 121 ms); [1365x1365 x90 NVN 0x8e ASTC on the GPU 56.5 MB 179 ms] [12288x12288 x1 NVN 0x32 on the CPU 288.0 MB 59 ms] [1536x1536 x1 NVN 0x7b ASTC on the GPU 2.0 MB 32 ms] [hos host] vk: 204 new ASTC textures left for a later frame (this frame's uploads took their budget; 204 so far) [hos sead::MainThread] nvn: GPU waits in 5 s: 173 sync waits (1629 ms), 0 polls timed out (0 ms), 1 queue finishes (8 ms) [hos Presentation Thread] vi: 2334 frames presented (48.6 fps over the last 5.6 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 11, 15.5-17.8 247, 17.8-22 15, 22-30 0, 30-40 0, over 40 1 [hos host] vk: readback pages refreshed from guest memory: 7 pages so far (0 held: may be written ahead in their submission) [hos host] vk: table entry bound again at frame 2334: prog a03bdab95992c48e ssbo 0 of 3 (c0+0x880) addr 0x108c702000 size 0x2550000 (before 0) [hos host] vk: pool 0x117d000000..0x117e000000 took in 2 pools [hos host] vk: pool 0x117d000000..0x117e000000: 12 KB the CPU wrote since retired pool 0x117d800000's last sync kept from guest memory, not its mirror [hos host] vk: pool 0x117d000000..0x117e000000: 4 KB the CPU wrote since retired pool 0x117d000000's last sync kept from guest memory, not its mirror [hos host] vk: pool 0x117d000000..0x117e000000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: pool 0x117cfbb040+0x9b040 imported in 2.5 ms as 0x117c400000..0x117e000000 (63 imports, 255 ms in all) [hos host] vk: pool 0x117c400000..0x117e000000 took in 2 pools [hos host] vk: pool 0x117c400000..0x117e000000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: pool 0x117c3fe000+0xa0040 imported in 5.4 ms as 0x1178400000..0x117e000000 (64 imports, 261 ms in all) [hos host] vk: pool 0x1178400000..0x117e000000 took in 2 pools [hos host] vk: pool 0x1178400000..0x117e000000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: pool 0x117df8b040+0x44b040 imported in 7.0 ms as 0x1178400000..0x117ec00000 (65 imports, 268 ms in all) [hos host] vk: pool 0x1178400000..0x117ec00000 took in 2 pools [hos host] vk: pool 0x1178400000..0x117ec00000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: slow submit: 35.9 ms: dispatch 35.8 [hos host] vk: upscaler: NVIDIA DLAA after the game's bloom and tone map (DLAA before the bloom: off) [hos host] vk: a clear after 3 draws of its pass (640x360, 1 colour targets + depth, last program 619664ed509bbd99): a pass of its own [hos host] vk: a clear after 3 draws of its pass (640x360, 2 colour targets + depth, last program 2479abad219cb19a): a pass of its own [hos host] vk: write-back diff: a shadow for pool 0x117c400000 (12 MB; 116 MB of shadows) [hos host] vk: write-back diff: a shadow for pool 0x1178400000 (64 MB; 180 MB of shadows) [hos host] vk: slow submit: 20.4 ms: [hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own [hos host] vk: 92 new ASTC textures left for a later frame (this frame's uploads took their budget; 296 so far) [hos ModuleSystemWorker4] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 2.0 ms; 1 in all, 0 MB, 2 ms) [hos host] vk: slow submit: 16.9 ms: draw 16.2 descriptors 16.0 texture lookups 15.8 uploads 15.7 [hos] guest backtrace: lr 0x804fc1ec 0x83b6f194 0x83b6f080 [hos] guest backtrace: lr 0x8348c5e0 0x83b6f08c [hos] guest backtrace: lr 0x804fc1ec 0x83b6f194 0x83b6f080 [hos] guest backtrace: lr 0x8348c5e0 0x83b6f08c [hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own [hos host] vk: 62 new ASTC textures left for a later frame (this frame's uploads took their budget; 358 so far) [hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.8 ms, 1 frames after their copies were recorded, the first #191 0x113739f000 128x128 levels 8 NVN 0x3f (1 so far, 528 KB, 1 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos sead::MainThread] nvn: the game sets present interval 1 -> 4 [hos sead::MainThread] nvn: the game sets present interval 4 -> 1 [hos host] vk: slow submit: 17.0 ms: draw 16.8 descriptors 16.1 texture lookups 15.6 uploads 15.5 [hos host] vk: texture #466 0x10001e00400 2560x1440 NVN 0x34: the GPU wrote its storage (submission 16346): uploaded again [hos host] vk: upload over a drawn image: #466 0x10001e00400 2560x1440x1 x1 NVN 0x34 at present 2335 (texture frame 2336), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: texture 0x10001e00400 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since) [hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own [hos host] vk: 136 new ASTC textures left for a later frame (this frame's uploads took their budget; 494 so far) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 20 bindings followed, 32 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1843 dirty pages from their submits' captures, 2893 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 282 dispatches, 2687 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 363 runs, 247.8 MB compared; 52310 pages read into the shadows; 432 MB of shadows (108 MB of pages known, in 123 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 213 in captures, 0 only in the collection, 4 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.9 ms; 2 in all, 0 MB, 3 ms) [hos sead::MainThread] nvn: the game sets present interval 1 -> 4 [hos sead::MainThread] nvn: the game sets present interval 4 -> 1 [hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own [hos host] vk: 22 new ASTC textures left for a later frame (this frame's uploads took their budget; 516 so far) [hos host] vk: slow upload 15.6 ms: #282 0x10c6e4fe00 512x512x1 x1 levels 10 NVN 0x7f ASTC (BC3), 123 of 123 KB [hos host] vk: slow submit: 17.5 ms: draw 17.3 descriptors 17.0 texture lookups 16.5 uploads 16.4 [hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own [hos host] vk: 30 new ASTC textures left for a later frame (this frame's uploads took their budget; 546 so far) [hos ModuleSystemWorker1] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.8 ms; 3 in all, 0 MB, 4 ms) [hos host] vk: mirror of pool 0x7faf000000 (4 MB) given back: retired, 60 submissions since its sync [hos host] vk: mirror of pool 0x7fab000000 (64 MB) given back: retired, 60 submissions since its sync [hos host] vk: mirror of pool 0x117c800000 (8 MB) given back: retired, 56 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x117c400000 (4 MB) given back: retired, 56 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x117d800000 (4 MB) given back: retired, 56 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x117e400000 (4 MB) given back: retired, 56 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x117e800000 (4 MB) given back: retired, 56 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x117e000000 (8 MB) given back: retired, 56 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x117d800000 (8 MB) given back: retired, 54 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x117d000000 (8 MB) given back: retired, 54 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x117d000000 (16 MB) given back: retired, 45 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x117c400000 (12 MB) given back: retired, 45 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x117c400000 (28 MB) given back: retired, 45 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x1178400000 (64 MB) given back: retired, 45 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x117e000000 (12 MB) given back: retired, 45 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x1178400000 (92 MB) given back: retired, 45 submissions since its sync, GPU-only data, GPU-written [hos host] vk: 20 mirrors given back (353 MB in all); 874 MB of mirrors now, 0 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: slow submit: 15.2 ms: draw 15.2 descriptors 14.8 texture lookups 14.3 uploads 14.2 [hos host] vk: a draw (program 51b56e5cab37da1f) writes a target an earlier draw sampled, after 7 draws of the pass (1280x720): a pass of its own [hos host] vk: 17 new ASTC textures left for a later frame (this frame's uploads took their budget; 563 so far) [hos host] vk: 1 textures retired (549 left, 1901 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: NVN2 program 5599b891dfe20fdf (stage 0) lowered with extended registers (1132 REGMAP words) [hos host] vk: NVN2 program e5f2865bdce56993 (stage 0) lowered with extended registers (1129 REGMAP words) [hos host] vk: NVN2 program 57b9a80d47da5078 (stage 0) lowered with extended registers (1148 REGMAP words) [hos ModuleSystemWorker4] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 4.0 ms; 4 in all, 0 MB, 8 ms) [hos ModuleSystemWorker3] vk: readbacks read without their wait: the culling camera block (0x3980) by default (newest 0, 1 sizes) [hos host] vk: reaper: 8 images (7 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 3 (10 MB) [hos ModuleSystemWorker5] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.2 ms; 5 in all, 0 MB, 8 ms) [hos BaseProcExecutor3] map: Market_Child (child) [hos BaseProcExecutor3] cheats: inventory probe: Weapon_Sword_Kokiri=0 Weapon_Sword_Master=0 Weapon_Sword_Giant=0 Weapon_Sword_Biggoron=0 Weapon_Shield_Deku=0 Weapon_Shield_Hylian=0 Weapon_Shield_Mirror=0 Armor_Adult_Goron=0 Armor_Adult_Zora=0 Armor_Adult_HeavyBoot=0 Armor_Adult_HoverBoot=0 Item_DekuStick=0 Item_DekuNut=0 Item_Bomb=0 Item_Bombchu=0 [hos BaseProcExecutor3] cheats: inventory probe: Weapon_Range_SlingShot=0 Item_DekuSeed=0 Weapon_Range_Bow=0 Item_Arrow_Normal=0 Item_Arrow_Fire=0 Item_Arrow_Ice=0 Item_Arrow_Light=0 Item_DinFire=0 Item_FaroreWind=0 Item_NayruLove=0 Item_Ocarina_Fairy=0 Item_Ocarina_Time=0 Item_Hookshot=0 Item_Longshot=0 Item_Boomerang=0 Item_GlassTruth=0 Item_MagicBean=0 [hos BaseProcExecutor3] cheats: inventory probe: Item_Hammer=0 Item_Bottle=0 Armor_Child_Mask_Keaton=0 Armor_Child_Mask_Skull=0 Armor_Child_Mask_Spooky=0 Armor_Child_Mask_Bunny=0 Armor_Child_Mask_Goron=0 Armor_Child_Mask_Zora=0 Armor_Child_Mask_Gerudo=0 Armor_Child_Mask_Truth=0 Item_Gauntlet_Goron=0 Armor_Adult_Gauntlet_Silver=0 Armor_Adult_Gauntlet_Gold=0 [hos BaseProcExecutor3] cheats: inventory probe: Item_Scale_Silver=0 Item_Scale_Gold=0 Item_BombBag_Small=0 Item_BombBag_Medium=0 Item_BombBag_Large=0 Item_Quiver_Small=0 Item_Quiver_Medium=0 Item_Quiver_Large=0 Item_BulletBag_Small=0 Item_BulletBag_Medium=0 Item_BulletBag_Large=0 Item_Wallet_Medium=0 Item_Wallet_Large=0 Item_FairyBless_MP=0 Item_FairyBless_MPDouble=0 [hos BaseProcExecutor3] cheats: inventory probe: Item_FairyBless_DefenseDouble=0 Item_Stone_Forest=0 Item_Stone_Fire=0 Item_Stone_Water=0 Item_Medal_Forest=0 Item_Medal_Fire=0 Item_Medal_Water=0 Item_Medal_Shadow=0 Item_Medal_Spirit=0 Item_Medal_Light=0 Item_Melody_ZeldaLullaby=0 Item_Melody_EponaSong=0 Item_Melody_SariaSong=0 Item_Melody_SunSong=0 [hos BaseProcExecutor3] cheats: inventory probe: Item_Melody_TimeSong=0 Item_Melody_StormSong=0 Item_Melody_ForestMinuet=0 Item_Melody_FireBolero=0 Item_Melody_WaterSerenade=0 Item_Melody_ShadowNocturne=0 Item_Melody_SpiritRequiem=0 Item_Melody_LightPrelude=0 Item_AgonyStone=0 Item_Token_Gerudo=0 Item_Token_Gold=0 Item_Sword_Kokiri=0 Item_Sword_Giant=0 [hos BaseProcExecutor3] cheats: inventory probe: Item_Sword_Biggoron=0 Item_Shield_Mirror=0 Item_Range_Bow=0 Item_Range_SlingShot=0 Weapon_Range_Hookshot=0 Weapon_Range_Longshot=0 Weapon_Boomerang=0 Weapon_Hammer=0 Item_HeavyBoot=0 Item_HoverBoot=0 Item_Mask_Keaton=0 Item_Gauntlet_Silver=0 Armor_GlassTruth=0 Armor_Child_GlassTruth=0 Armor_Adult_GlassTruth=0 [hos BaseProcExecutor3] cheats: inventory probe: Armor_Child_Gauntlet_Goron=0 Item_Wallet_Small=1 [hos host] vk: slow upload 27.8 ms: #294 0x10bad3a200 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB [hos host] vk: slow submit: 32.1 ms: draw 31.7 descriptors 30.2 texture lookups 28.8 uploads 28.6 [hos host] vk: 94 new ASTC textures left for a later frame (this frame's uploads took their budget; 657 so far) [hos ModuleSystemWorker1] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.9 ms; 6 in all, 0 MB, 9 ms) [hos host] vk: texture #809 at 0x10001e00400 beside #466: 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: texture #809 at 0x10001e00400 beside #532: 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #809 at 0x10001e00400 beside #557: 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1 (rendered) [hos host] vk: slow submit: 72.7 ms: draw 68.6 descriptors 48.9 texture lookups 46.3 uploads 18.6 programs 15.5 commands 2.8 [hos host] vk: texture #831 at 0x10001fe0800 beside #811: 1280x720x1 x1 levels 1 layout 0x4 class 0x40002 NVN 0x5 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: texture #833 at 0x10004a11800 beside #478: 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x40010 NVN 0x2e target 1 (rendered) [hos host] vk: texture #840 at 0x10005192200 beside #835: 1600x1600x1 x3 levels 1 layout 0x4 class 0x1007c NVN 0x32 target 4, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: slow submit: 18.1 ms: draw 2.9 texture lookups 11.7 programs 1.4 [hos host] vk: texture #874 at 0x100092d4400 beside #868: 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: texture #876 at 0x10004ce1e00 beside #870: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #877 at 0x10004bf1c00 beside #869: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #878 at 0x10004b01a00 beside #871: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #879 at 0x1000a1d4600 beside #875: 1280x720x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #884 at 0x10004bf1c00 beside #869: 160x90x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x3 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: texture #884 at 0x10004bf1c00 beside #877: 160x90x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x3 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: NVN2 program b81f83fe05a2a32a (stage 0) lowered with extended registers (939 REGMAP words) [hos host] vk: NVN2 program 2d6b101e5868b4ea (stage 0) lowered with extended registers (1132 REGMAP words) [hos host] vk: NVN2 program 40d0c0768e874600 (stage 7) lowered with extended registers (331 REGMAP words) [hos host] vk: NVN2 program c567dec57b3bbad1 (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: NVN2 program 56cc6c3a14ce827a (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: NVN2 program 73b819fbe71d2814 (stage 0) lowered with extended registers (1128 REGMAP words) [hos host] discord: the status: Hyrule Market - Child / ❤️ 0.5/3 · 🕷️ 0 · 💗 0 [hos host] vk: NVN2 program 167110f0f411da2e (stage 0) lowered with extended registers (1122 REGMAP words) [hos host] vk: NVN2 program c5f227a6efbe5b41 (stage 0) lowered with extended registers (1141 REGMAP words) [hos host] vk: NVN2 program a19c1ffa6e31ecd4 (stage 0) lowered with extended registers (1145 REGMAP words) [hos host] vk: NVN2 program 2de89066ce43a6c1 (stage 0) lowered with extended registers (1134 REGMAP words) [hos host] vk: NVN2 program 9039474cf8f30ef4 (stage 0) lowered with extended registers (940 REGMAP words) [hos host] vk: NVN2 program 5c55ab4d22261ede (stage 0) lowered with extended registers (1132 REGMAP words) [hos host] vk: NVN2 program a19c1ffa6e31ecd4 (stage 0) lowered with extended registers (1145 REGMAP words) [hos host] vk: NVN2 program 690e54aa085d7bd5 (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: NVN2 program f33fc74589ea6ff0 (stage 0) lowered with extended registers (1131 REGMAP words) [hos host] vk: NVN2 program 196c6bfc68efb1dd (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: texture 0x10d14b8e00 171x341x1 x1 levels 9 NVN 0x8d -> vk 37 (ASTC decoded) (1000 textures, 2606 MB) [hos host] vk: NVN2 program 73b819fbe71d2814 (stage 0) lowered with extended registers (1128 REGMAP words) [hos host] vk: NVN2 program 2ef7974fd3b630c7 (stage 0) lowered with extended registers (943 REGMAP words) [hos host] vk: NVN2 program 809483568f9bbfc9 (stage 0) lowered with extended registers (1137 REGMAP words) [hos host] vk: NVN2 program a19c1ffa6e31ecd4 (stage 0) lowered with extended registers (1145 REGMAP words) [hos host] vk: NVN2 program 2855427a5b00e4d4 (stage 0) lowered with extended registers (1142 REGMAP words) [hos host] vk: NVN2 program 2d6b101e5868b4ea (stage 0) lowered with extended registers (1132 REGMAP words) [hos host] vk: NVN2 program f14336ce5b132a86 (stage 0) lowered with extended registers (1122 REGMAP words) [hos host] vk: NVN2 program b07021ee80078f45 (stage 0) lowered with extended registers (1117 REGMAP words) [hos host] vk: NVN2 program 73b819fbe71d2814 (stage 0) lowered with extended registers (1128 REGMAP words) [hos host] vk: NVN2 program 167110f0f411da2e (stage 0) lowered with extended registers (1122 REGMAP words) [hos host] vk: NVN2 program 78d7ec0f190748f5 (stage 0) lowered with extended registers (1117 REGMAP words) [hos host] vk: NVN2 program 56cc6c3a14ce827a (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: NVN2 program a6d021d3dce2b9f9 (stage 0) lowered with extended registers (1128 REGMAP words) [hos host] vk: NVN2 program 389c379053aab5fe (stage 0) lowered with extended registers (1129 REGMAP words) [hos host] vk: NVN2 program 13a1f5b177ab2712 (stage 0) lowered with extended registers (1130 REGMAP words) [hos host] vk: NVN2 program 2d6b101e5868b4ea (stage 0) lowered with extended registers (1132 REGMAP words) [hos host] vk: NVN2 program 2ef7974fd3b630c7 (stage 0) lowered with extended registers (943 REGMAP words) [hos host] vk: NVN2 program b1fc459b591f7806 (stage 0) lowered with extended registers (939 REGMAP words) [hos host] vk: NVN2 program 9039474cf8f30ef4 (stage 0) lowered with extended registers (940 REGMAP words) [hos host] vk: texture #1169 at 0x100023a0e00 beside #828: 1280x720x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 2560x1440x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: NVN2 program 57b9a80d47da5078 (stage 0) lowered with extended registers (1148 REGMAP words) [hos host] vk: NVN2 program a9cad530cd625cc4 (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: NVN2 program 196c6bfc68efb1dd (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: a draw (program b426f821e80124dd) samples a target the pass wrote, after 1 draws of the pass (2560x1440): a pass of its own [hos host] vk: texture #1201 at 0x10001fe0800 beside #811: 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: texture #1201 at 0x10001fe0800 beside #831: 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40002 NVN 0x5 target 1 (rendered) [hos host] vk: texture #1208 at 0x100021aea00 beside #1206: 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1, the other 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #1210 at 0x100021aea00 beside #1206: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 120x67x1 x1 levels 1 layout 0x3 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #1210 at 0x100021aea00 beside #1208: 30x16x1 x1 levels 1 layout 0x1 class 0x40004 NVN 0x3f target 1, the other 60x33x1 x1 levels 1 layout 0x2 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #1211 at 0x100020e5a00 beside #1202: 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1, the other 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #1212 at 0x10001fe0800 beside #811: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 640x360x1 x1 levels 1 layout 0x4 class 0x1007e NVN 0x34 target 1 (rendered) [hos host] vk: texture #1212 at 0x10001fe0800 beside #831: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40002 NVN 0x5 target 1 (rendered) [hos host] vk: texture #1212 at 0x10001fe0800 beside #1201: 640x360x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x14 target 1, the other 480x270x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x3f target 1 (rendered) [hos host] vk: texture #1213 at 0x10004921600 beside #494: 2560x1440x1 x1 levels 1 layout 0x4 class 0x40004 NVN 0x25 target 1, the other 1280x720x1 x1 levels 1 layout 0x4 class 0x40001 NVN 0x1 target 1 (rendered) [hos host] vk: pool 0x1012538600+0xf00000 imported in 2.7 ms as 0x1012400000..0x1014000000 (79 imports, 286 ms in all) [hos host] vk: pool 0x1010bfcc00+0x1c2000 imported in 2.2 ms as 0x1010800000..0x1011000000 (80 imports, 289 ms in all) [hos host] vk: slow submit: 122.7 ms: draw 89.6 descriptors 45.1 texture lookups 60.6 uploads 8.3 programs 42.5 commands 1.2 dispatch 2.7 [hos host] vk: 391 new ASTC textures left for a later frame (this frame's uploads took their budget; 1048 so far) [hos host] vk: slow upload 25.2 ms: #587 0x10c4fe3e00 1024x1024x1 x1 levels 11 NVN 0x8b ASTC (BC3), 642 of 642 KB [hos host] vk: slow submit: 40.2 ms: draw 37.1 descriptors 31.0 texture lookups 27.4 uploads 26.9 programs 2.3 commands 2.8 [hos host] vk: NVN2 program 875fa04074350384 (stage 0) lowered with extended registers (1142 REGMAP words) [hos host] vk: NVN2 program cb9e651e1a2900cc (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: NVN2 program 196c6bfc68efb1dd (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: NVN2 program 2e41b85d34167b3b (stage 0) lowered with extended registers (1113 REGMAP words) [hos host] vk: NVN2 program d88eb83da249de23 (stage 0) lowered with extended registers (940 REGMAP words) [hos host] vk: NVN2 program 42be3be9f8a7413b (stage 0) lowered with extended registers (944 REGMAP words) [hos host] vk: NVN2 program 500b27d6180ac176 (stage 0) lowered with extended registers (1134 REGMAP words) [hos host] vk: NVN2 program 9039474cf8f30ef4 (stage 0) lowered with extended registers (940 REGMAP words) [hos host] vk: slow submit: 21.0 ms: draw 18.7 descriptors 7.3 texture lookups 6.1 uploads 1.4 programs 10.4 [hos host] vk: 395 new ASTC textures left for a later frame (this frame's uploads took their budget; 1443 so far) [hos ModuleSystemWorker1] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 3.4 ms; 7 in all, 0 MB, 12 ms) [hos host] vk: slow upload 20.9 ms: #593 0x10b7b09c00 512x512x1 x1 levels 10 NVN 0x8d ASTC (BC3), 123 of 123 KB [hos host] vk: slow submit: 31.7 ms: draw 30.1 descriptors 25.0 texture lookups 22.0 uploads 21.7 programs 1.8 commands 2.4 [hos host] vk: 332 new ASTC textures left for a later frame (this frame's uploads took their budget; 1775 so far) [hos host] vk: slow submit: 30.6 ms: draw 24.2 descriptors 17.7 texture lookups 14.0 uploads 13.4 programs 2.5 commands 2.7 [hos host] vk: NVN2 program 2251fa487a08a403 (stage 0) lowered with extended registers (1136 REGMAP words) [hos host] vk: NVN2 program c30cc196df83447f (stage 0) lowered with extended registers (1128 REGMAP words) [hos host] vk: NVN2 program c2a34f499fd72642 (stage 0) lowered with extended registers (1127 REGMAP words) [hos host] vk: NVN2 program 00528550fc06c30c (stage 0) lowered with extended registers (1093 REGMAP words) [hos host] vk: NVN2 program 847f64de27c1dc0c (stage 0) lowered with extended registers (1129 REGMAP words) [hos host] vk: slow submit: 38.9 ms: draw 37.7 descriptors 19.7 texture lookups 17.5 uploads 4.2 programs 16.1 commands 1.3 [hos host] vk: 433 new ASTC textures left for a later frame (this frame's uploads took their budget; 2208 so far) [hos ModuleSystemWorker3] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 26.6 ms; 8 in all, 0 MB, 39 ms) [hos host] vk: slow upload 18.2 ms: #375 0x10d5789a00 341x341x1 x1 levels 9 NVN 0x8c ASTC, 73 of 73 KB [hos host] vk: slow submit: 37.2 ms: draw 31.9 descriptors 24.9 texture lookups 20.7 uploads 19.9 programs 2.4 commands 3.4 [hos host] vk: texture 0x10afe8cc00 576x576x1 x1 levels 10 NVN 0x7b -> vk 37 (ASTC decoded) (1500 textures, 2876 MB) [hos host] vk: replay reads of guest memory as submitted: 20000 (12380 pages from the captures) [hos host] vk: slow submit: 23.2 ms: draw 22.1 descriptors 14.7 texture lookups 12.7 uploads 2.7 programs 5.7 commands 1.3 [hos host] vk: 522 new ASTC textures left for a later frame (this frame's uploads took their budget; 2730 so far) [hos host] vk: texture #8 0x10755b4000 1920x1080x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 383 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 85 textures retired (1545 left, 2925 MB): budget 85, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 7 MB [hos host] vk: late pages' fix: stamps for 0x10387f0000 (pool 0x1038400000; 128 blocks, 8192 KB) [hos host] vk: late pages' fix: stamps for 0x103d7f0000 (pool 0x103d400000; 192 blocks, 12288 KB) [hos host] vk: late pages' fix: stamps for 0x117d5f0000 (pool 0x1178400000; 256 blocks, 16384 KB) [hos host] vk: late pages' fix: stamps for 0x117dbf0000 (pool 0x1178400000; 320 blocks, 20480 KB) [hos host] vk: late pages' fix: stamps for 0x117e5f0000 (pool 0x1178400000; 384 blocks, 24576 KB) [hos host] vk: slow submit: 25.3 ms: draw 22.7 descriptors 17.7 texture lookups 16.1 uploads 15.7 programs 1.2 commands 2.7 [hos host] vk: slow submit: 16.7 ms: draw 12.4 descriptors 5.9 texture lookups 5.6 programs 4.6 commands 1.3 [hos host] vk: 431 new ASTC textures left for a later frame (this frame's uploads took their budget; 3161 so far) [hos ModuleSystemWorker2] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.5 ms; 9 in all, 0 MB, 39 ms) [hos host] vk: slow submit: 19.8 ms: draw 17.8 descriptors 14.7 texture lookups 13.2 uploads 12.9 commands 1.6 [hos host] vk: slow submit: 17.7 ms: draw 11.9 descriptors 5.2 texture lookups 4.6 programs 4.0 commands 2.0 [hos host] vk: 478 new ASTC textures left for a later frame (this frame's uploads took their budget; 3639 so far) [hos host] vk: late pages' fix: stamps for 0x10383f0000 (pool 0x1038000000; 448 blocks, 28672 KB) [hos host] vk: slow upload 23.6 ms: #1535 0x10bd32f400 576x576x1 x1 levels 10 NVN 0x7e ASTC (BC3), 187 of 187 KB [hos host] vk: slow submit: 34.9 ms: draw 32.6 descriptors 27.3 texture lookups 25.0 uploads 24.6 programs 4.0 [hos host] vk: 359 new ASTC textures left for a later frame (this frame's uploads took their budget; 3998 so far) [hos ModuleSystemWorker1] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 5.0 ms; 10 in all, 0 MB, 44 ms) [hos host] vk: slow upload 20.8 ms: #1808 0x10b3bbcc00 576x576x1 x1 levels 10 NVN 0x7e ASTC (BC3), 187 of 187 KB [hos host] vk: slow submit: 30.2 ms: draw 28.8 descriptors 25.9 texture lookups 23.2 uploads 22.7 programs 1.1 commands 1.3 [hos host] vk: 344 new ASTC textures left for a later frame (this frame's uploads took their budget; 4342 so far) [hos host] vk: slow submit: 24.4 ms: draw 22.4 descriptors 16.3 texture lookups 12.5 uploads 12.2 programs 1.9 commands 3.0 [hos host] vk: 430 new ASTC textures left for a later frame (this frame's uploads took their budget; 4772 so far) [hos ModuleSystemWorker2] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 0.3 ms; 11 in all, 0 MB, 45 ms) [hos host] vk: slow upload 21.5 ms: #1560 0x10b2853c00 576x576x1 x1 levels 10 NVN 0x7e ASTC (BC3), 187 of 187 KB [hos host] vk: slow submit: 33.1 ms: draw 29.9 descriptors 26.3 texture lookups 22.9 uploads 22.3 programs 1.1 commands 1.7 [hos host] vk: 323 new ASTC textures left for a later frame (this frame's uploads took their budget; 5095 so far) [hos host] vk: slow submit: 19.4 ms: draw 18.7 descriptors 15.6 texture lookups 13.9 uploads 13.5 commands 1.6 [hos host] vk: 356 new ASTC textures left for a later frame (this frame's uploads took their budget; 5451 so far) [hos ModuleSystemWorker3] vk: virtual pages 0x1165e17000..0x1165e17080 remapped: 1 texture(s) the GPU wrote there written back into their memory (0 KB, 2.6 ms; 12 in all, 0 MB, 47 ms) [hos host] vk: slow upload 17.1 ms: #1361 0x10cab48200 576x576x1 x1 levels 10 NVN 0x7b ASTC, 301 of 301 KB [hos host] vk: slow submit: 27.3 ms: draw 23.5 descriptors 20.4 texture lookups 18.5 uploads 18.1 programs 1.5 commands 1.1 [hos host] vk: 302 new ASTC textures left for a later frame (this frame's uploads took their budget; 5753 so far) [hos host] vk: slow upload 19.8 ms: #1378 0x10af4d8a00 768x768x1 x1 levels 10 NVN 0x7e ASTC (BC3), 324 of 324 KB [hos host] vk: slow submit: 26.9 ms: draw 26.0 descriptors 23.2 texture lookups 21.0 uploads 20.5 programs 1.3 commands 1.1 [hos host] vk: 281 new ASTC textures left for a later frame (this frame's uploads took their budget; 6034 so far) [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #10000: event 0x100d6c8028 mode 0 value 2855 (now 2817) (lr 0x806674b0) [hos host] vk: slow submit: 22.2 ms: draw 20.5 descriptors 16.4 texture lookups 14.3 uploads 13.9 programs 1.1 commands 2.1 [hos host] vk: 323 new ASTC textures left for a later frame (this frame's uploads took their budget; 6357 so far) [hos host] vk: replay reads of guest memory as submitted: 40000 (26086 pages from the captures) [hos host] vk: slow upload 21.2 ms: #1780 0x10afcdb200 768x768x1 x1 levels 10 NVN 0x7e ASTC (BC3), 324 of 324 KB [hos host] vk: slow submit: 27.7 ms: draw 26.6 descriptors 24.4 texture lookups 22.4 uploads 22.0 commands 1.0 [hos host] vk: 269 new ASTC textures left for a later frame (this frame's uploads took their budget; 6626 so far) [hos host] vk: slow submit: 20.2 ms: draw 19.0 descriptors 16.5 texture lookups 14.0 uploads 13.5 commands 1.2 [hos host] vk: 244 new ASTC textures left for a later frame (this frame's uploads took their budget; 6870 so far) [hos host] vk: slow upload 15.2 ms: #1436 0x10b5369400 576x576x1 x1 levels 10 NVN 0x7b ASTC, 301 of 301 KB [hos host] vk: slow submit: 23.6 ms: draw 23.4 descriptors 20.1 texture lookups 17.0 uploads 16.3 programs 1.0 commands 1.8 [hos host] vk: 223 new ASTC textures left for a later frame (this frame's uploads took their budget; 7093 so far) [hos host] vk: presenter thread: the copy's fence 3.34 ms, the present 0.60 ms a present [hos host] vk: slow submit: 27.1 ms: draw 26.7 descriptors 21.2 texture lookups 18.4 uploads 15.8 programs 3.5 commands 1.4 [hos host] vk: 215 new ASTC textures left for a later frame (this frame's uploads took their budget; 7308 so far) [hos host] vk: slow submit: 18.1 ms: draw 17.0 descriptors 14.6 texture lookups 12.5 uploads 12.0 commands 1.2 [hos host] vk: 187 new ASTC textures left for a later frame (this frame's uploads took their budget; 7495 so far) [hos host] vk: slow submit: 22.0 ms: draw 20.2 descriptors 16.9 texture lookups 14.9 uploads 14.4 programs 1.7 commands 1.1 [hos host] vk: 172 new ASTC textures left for a later frame (this frame's uploads took their budget; 7667 so far) [hos host] vk: slow submit: 19.4 ms: draw 17.2 descriptors 14.9 texture lookups 13.1 uploads 12.6 [hos host] vk: 150 new ASTC textures left for a later frame (this frame's uploads took their budget; 7817 so far) [hos host] vk: slow submit: 21.7 ms: draw 19.2 descriptors 16.4 texture lookups 14.3 uploads 13.8 programs 1.1 commands 1.2 [hos host] vk: 124 new ASTC textures left for a later frame (this frame's uploads took their budget; 7941 so far) [hos host] vk: slow submit: 26.7 ms: draw 25.5 descriptors 22.9 texture lookups 20.6 uploads 20.2 commands 1.3 [hos host] vk: 90 new ASTC textures left for a later frame (this frame's uploads took their budget; 8031 so far) [hos host] vk: slow submit: 22.7 ms: draw 21.4 descriptors 18.6 texture lookups 16.3 uploads 15.9 commands 1.2 [hos host] vk: 71 new ASTC textures left for a later frame (this frame's uploads took their budget; 8102 so far) [hos host] vk: slow submit: 18.8 ms: draw 19.0 descriptors 15.7 texture lookups 13.9 uploads 13.5 commands 1.6 [hos host] vk: 98 new ASTC textures left for a later frame (this frame's uploads took their budget; 8200 so far) [hos host] vk: sparse mirrors: 67, 933 MB with memory of 1184 MB (250 MB spared); blocks given memory later 71, given back 1026; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 103 binds (16.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 13 taken from the spares, 74 made new; 20 kept as spares, 0 destroyed; 7 spares now, 288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 537, 2529.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3543 dirty pages from their submits' captures, 50023 from earlier ones', 14 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 834 dispatches, 35231 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5135 runs, 1068.8 MB compared; 1183 pages read into the shadows; 196 MB of shadows (112 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8811 in captures, 174 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: 0x7fdf000000 (12 MB, captured, state 2): 7; 0x7fd4c00000 (24 MB, captured, state 2): 7; [hos host] vk: imports: 70 live (1272 MB), 18 retired still imported (380 MB), 0 retired given back (0 MB) [hos host] vk: slow submit: 20.4 ms: draw 18.8 descriptors 15.8 texture lookups 13.2 uploads 12.8 programs 1.1 commands 1.4 [hos host] vk: 51 new ASTC textures left for a later frame (this frame's uploads took their budget; 8251 so far) [hos host] vk: replay reads of guest memory as submitted: 60000 (40507 pages from the captures) [hos host] vk: slow submit: 23.4 ms: draw 22.9 descriptors 19.0 texture lookups 15.0 uploads 13.9 programs 1.5 commands 1.8 [hos host] vk: 52 new ASTC textures left for a later frame (this frame's uploads took their budget; 8303 so far) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5796 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos OptimizeChangeBanc] nvn: 150000 recordings made, 53206 live copies (26.3 MB), 259 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: slow submit: 23.1 ms: draw 21.0 descriptors 17.2 texture lookups 15.2 uploads 14.8 commands 2.1 [hos host] vk: 75 new ASTC textures left for a later frame (this frame's uploads took their budget; 8378 so far) [hos host] vk: late pages found at the replay (600 takes): 695 pages in 10 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: slow submit: 21.4 ms: draw 19.5 descriptors 17.0 texture lookups 14.6 uploads 14.3 commands 1.2 [hos host] vk: 32 new ASTC textures left for a later frame (this frame's uploads took their budget; 8410 so far) [hos host] vk: slow submit: 20.0 ms: draw 19.1 descriptors 16.5 texture lookups 15.1 uploads 15.0 commands 1.4 [hos host] vk: 12 new ASTC textures left for a later frame (this frame's uploads took their budget; 8422 so far) [hos host] vk: slow upload 29.1 ms: #1979 0x10d026cc00 341x341x1 x1 levels 9 NVN 0x8d ASTC, 62 of 62 KB [hos host] vk: slow submit: 46.9 ms: draw 46.2 descriptors 38.1 texture lookups 34.7 uploads 31.7 programs 5.8 commands 1.7 [hos host] vk: 4 new ASTC textures left for a later frame (this frame's uploads took their budget; 8426 so far) [hos host] vk: program picks: 115897 kept (42840 by the last picks' textures, a uniform buffer bound since), 24784 new, 59318 with other textures (of 200000) [hos host] vk: replay reads of guest memory as submitted: 80000 (54583 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 100000 (68581 pages from the captures) [hos host] vk: memory: working set 12370 MB (peak 12387), private 19295 MB, staging 256 MB in 16 arenas, 17 made so far (56 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 624 MB, 36 allocations 558 MB; used 2296 of 15579 MB; heap 1 (device): 168 blocks 5104 MB, 17371 allocations 4829 MB; used 6915 of 7823 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 16 blocks 512 MB, 108 images 405 MB; 71 free ranges, the largest 16 MB [hos host] vk: image blocks (textures, compacted): 62 blocks 1984 MB, 1685 images 1979 MB; 82 free ranges, the largest 4 MB [hos host] vk: render target images: 57 reused from the pool, 174 made new, 64 destroyed from it, 10 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 1 missed after their image was taken by size; it holds 12 (64 MB) [hos host] vk: render targets not reused by size: 154 with the pool empty (0 of them with targets still in the dead queue), 2 with none in it a present yet, 0 with none of a kind they can use, 0 with all smaller (the largest 0% short on average), 10 with all bigger (over 1.5 times), 8 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 64 idle [hos host] vk: VRAM: 6915 MB used of 7823 MB allowed (275 MB of it free in the blocks); textures 3062 MB (budget 2900 MB), 1787 textures [hos host] vk: budget: VRAM 5104 of 7823 MB (6915 MB by the driver, 1810 MB unaccounted): textures 3062, mirrors 933, blocks free 275, staging 256, other buffers 205, shadows 196, other images 111, target pool 64; RAM 2236 MB (the process's private memory 19295 MB): imports 1272, capture spares 416, retired imports 380, captures 118, capture table 26, other buffers 24, others 0 [hos host] vk: textures: decoded ASTC 1796 MB, rendered 861 MB (73 drawn in one frame, kept: 51753 KB), other 365 MB; used in the last 8 presents 2682 MB [hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 358 MB of blocks hashed in 38 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0xa, last used 67 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 13 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 1 (2 KB): 8x8x1 NVN 0x3f: 1 (2 KB); [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.16 [hos Presentation Thread] vk: presents after the CPU's signal (600): 429 waited for it, 4.94 ms each, 1 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 1079.0 ms; VRAM 6.8 of 7.6 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 23 program records loaded (8.3 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 4810 queued, 4810 loaded (median 1.25 ms, 99th percentile 17.14 ms, worst 34.27 ms, 11581 ms and 267.7 MB read in all); 6641 programs (1430 of them compute variants without their code, 0 graphics programs no note names), 9515 graphics pipelines (1 not in a cache), 576 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 985 pipelines made at create taken by draws; 2185 create calls with nothing kept for their stages, 0 before the index; units 6995 (4810 ready, 0 loading), 12768 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 10218 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 574 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before) [hos host] vk: 7535 programs from the program cache [hos host] vk: a frame: 221 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 27.09, present 0.13, draw 18.69, descriptors 13.27, texture lookups 11.88, uploads 9.46, pipelines 0.26, programs 1.95, commands 2.63, dispatch 2.37, memory sync 1.04, queue submit 0.02, slot wait 0.30 [hos host] vk: a frame: mirrors 87 (933 MB): 13.41 MB synced in 559.7 copies, 15.36 MB written back [hos host] vk: 2400 presents, 25.8 fps; a frame: 1579 draws (0 skipped), 79 dispatches, 94 passes, 98 barriers, 6 submits, 13.2 uploads (6.2 MB), 59 ub updates (39 overlays), GPU-bound wait 0.0 ms (recording slots only); 11303 pipelines (2506 ms), 0 programs (10 failed), 1787 textures (3062 MB) [hos host] vk: host time in GPU resource waits: 23.88 ms a frame (fences 11.48, timelines 12.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 426.4 full barriers recorded, 173.2 left out (nothing recorded since the one before), 10.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.5; 0x8: 1.8; 0x9: 0.6; 0x10: 1.4; 0x11: 27.8; 0x12: 48.4; 0x13: 1.1; 0x21: 2.2; 0x28: 9.1; 0x92: 1.7; 0x97: 1.6; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 19.7 from 13.6 batches at the game's barriers (21.5 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 92.6, an op 1255.2, the end 4.4, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.1; 5: 11.3; 48: 3.4; [hos host] vk: GPU time 8.0 ms a frame (the longest submission 43.8 ms); 15.8 textures made a frame: colour targets 1.7 (6.5 MB), depth 0.8 (6.3 MB), ASTC or BC 13.3 (17.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 2.77 MB unswizzled on the CPU in 0.47 ms [hos host] vk: the render thread (600 presents, 17.37 ms a frame): its jobs 9.45 ms a frame (submits 4.71, presents 2.85, others 1.89), idle 7.92; the game waited for room at 213 presents, 1.91 ms each (0.68 a frame; 10.90 jobs queued at a present); 35 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 27.83 ms each; 1 presents held for the CPU's signal, 72.11 ms each (0 let go by RX_VK_HOLD_MS); 488 captures at a CPU signal; the submits' waits read in 0.012 ms a frame; readbacks: 313 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 9.47 ms after its hand-over (1082.38 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 10.71 + 0.27 (the window's copy is in the idle), after it 0.92 + 1.40; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 876.78 (0.44 ms) [hos host] vk: frame ahead (600 presents): 1.73 parts a frame recorded in the signal's job, 56.8% of them took nothing that came with the signal; of the others, 0.75 a frame took pages of the signal's capture into a mirror (51.9 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 234 of 489 frames, its first three in 234 [hos host] vk: frame ahead, what came with the signal: 489 signals with a capture, each of 71.7 pages in 2.4 pools; between a frame's last submit and that signal 0.59 texture and sampler registrations, 1.05 releases and 0.00 other CPU signals a signal (37 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.33+0.47, 0.06+0.03, 0.08+0.02, 0.06+0.13, 0.05+0.01, 0.02+0.03, 0.02+0.40, 0.11+0.09, 0.09+0.15, 0.01+0.06, 0.03+0.00, 0.07+0.01 [hos host] vk: pipeline cache: play file 7 saved (363 KB): 8 pipelines, 6 from the driver's cache, 2 compiled; 10231 more from the start's cache [hos host] vk: write-backs left out (600 submits): 682, 3247.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2063 dirty pages from their submits' captures, 41181 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 742 dispatches, 31867 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5485 runs, 540.0 MB compared; 174 pages read into the shadows; 196 MB of shadows (113 MB of pages known, in 133 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 9478 in captures, 1 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 120000 (82210 pages from the captures) [hos BaseProcExecutor5] vcores: 124237 takes, 54276 waits (mean 113 us, longest 37.0 ms, 0 over HOS_VCORE_LOG_MS), 25760 kicks, 24949 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: render targets drawn over from another start, let go: 209 so far (968 MB), 199 of them (907 MB) between the passes [hos host] vk: 13 textures retired (1842 left, 3100 MB): budget 0, drawn over 3, dead targets 10 [hos host] vk: pool 0x103d000000..0x103d800000: mirror made with the GPU's contents of 1 retired pool(s) [hos host] vk: late pages' fix: stamps for 0x103d3e0000 (pool 0x103d000000; 512 blocks, 32768 KB) [hos host] vk: mirror of pool 0x103d400000 (4 MB) given back: retired, 58 submissions since its sync, GPU-written [hos host] vk: 1 textures retired (1862 left, 3111 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: replay reads of guest memory as submitted: 140000 (94738 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.29 ms (lock 0.05, scan 0.10 over 1061 ranges, 3248 MB), 308 pages written, 308 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.010 (4 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (166 MB), 10 spare buffers, 6553 texture pages copied (25 MB) [hos host] vk: late pages' fix: stamps for 0x103d7f0000 (pool 0x103d000000; 512 blocks, 32768 KB) [hos host] vk: 2 textures retired (1912 left, 3191 MB): budget 0, drawn over 1, dead targets 1 [hos host] vk: sparse mirrors: 68, 937 MB with memory of 1192 MB (254 MB spared); blocks given memory later 71, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 107 binds (17.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 13 taken from the spares, 76 made new; 21 kept as spares, 0 destroyed; 8 spares now, 292 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 682, 3247.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2224 dirty pages from their submits' captures, 41922 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 744 dispatches, 32151 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5930 runs, 579.0 MB compared; 961 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 10162 in captures, 3 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 71 live (1280 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB) [hos host] vk: slow submit: 15.5 ms: draw 15.0 descriptors 3.3 programs 9.4 commands 1.7 [hos host] vk: replay reads of guest memory as submitted: 160000 (106714 pages from the captures) [hos host] vk: budget: gave back 256 MB from staging at 90% of the allowance [hos host] vk: 2 textures retired (1974 left, 3300 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: program picks: 267399 kept (99085 by the last picks' textures, a uniform buffer bound since), 24880 new, 107720 with other textures (of 400000) [hos host] vk: texture 0x1000b9a3800 1280x720x1 x1 levels 1 NVN 0x29 -> vk 97 (2000 textures, 3422 MB) [hos host] vk: texture 0x10be4dca00 1024x1024x1 x1 levels 11 NVN 0x8e -> vk 137 (ASTC decoded) (2000 textures, 3451 MB) [hos host] vk: replay reads of guest memory as submitted: 180000 (118167 pages from the captures) [hos host] vk: 12 textures retired (2036 left, 3500 MB): budget 1, drawn over 0, dead targets 11 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 72 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1334, sb 559, cis 10086, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 682, 3247.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1970 dirty pages from their submits' captures, 41215 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 737 dispatches, 31833 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5855 runs, 561.1 MB compared; 124 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 10261 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 200000 (130074 pages from the captures) [hos host] vk: texture #558 0x10005280e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 126 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #495 0x10004741200 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 126 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #493 0x10004831400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 126 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #479 0x10004dd1a00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 126 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 118 textures retired (1937 left, 3483 MB): budget 117, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 30 MB [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #5000: event 0x100d6c802c mode 0 value 5003 (lr 0x806616b0) [hos OptimizeChangeBanc] nvn: 200000 recordings made, 65901 live copies (24.6 MB), 533 texture copies; 4 command buffers' segments (4.2 MB) [hos sead::MainThread] vk: readbacks of GPU results: 816 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 18919 kept their virtual core (nothing there to wait for), 2841 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 42 textures retired (1920 left, 3124 MB): budget 0, drawn over 8, dead targets 34 [hos host] vk: replay reads of guest memory as submitted: 220000 (142585 pages from the captures) [hos host] vk: texture #2308 0x100091e2e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2494 (drawn at 2433) [hos Presentation Thread] vi: 2495 frames presented (32.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 2, 15.5-17.8 3, 17.8-22 39, 22-30 67, 30-40 26, over 40 24 [hos host] vk: readback pages refreshed from guest memory: 73115 pages so far (0 held: may be written ahead in their submission) [hos sead::MainThread] nvn: GPU waits in 5 s: 78 sync waits (640 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: sparse mirrors: 70, 938 MB with memory of 1200 MB (261 MB spared); blocks given memory later 71, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 109 binds (17.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 14 taken from the spares, 77 made new; 21 kept as spares, 0 destroyed; 7 spares now, 288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 727, 3458.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1340 dirty pages from their submits' captures, 44461 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 809 dispatches, 33950 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6139 runs, 596.3 MB compared; 116 pages read into the shadows; 200 MB of shadows (112 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11382 in captures, 2 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: 0x7fe2800000 (4 MB, captured, state 2): 2; [hos host] vk: imports: 73 live (1288 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 612 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 134 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] nvn: EventSignal (CPU) #5000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.03, writes collected 0.21 [hos host] vk: replay reads of guest memory as submitted: 240000 (154302 pages from the captures) [hos host] vk: mirror of pool 0x1072c00000 (68 MB) given back: idle, 6048 submissions since its sync [hos host] vk: slow submit: 23.3 ms: draw 21.7 descriptors 18.6 texture lookups 17.2 uploads 14.9 programs 1.3 commands 1.3 [hos sead::MainThread] vk: capture: a submit 0.24 ms (lock 0.04, scan 0.08 over 1328 ranges, 3419 MB), 267 pages written, 267 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.003 (1 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (106 MB), 11 spare buffers, 6552 texture pages copied (25 MB) [hos host] vk: program picks: 416049 kept (153866 by the last picks' textures, a uniform buffer bound since), 24943 new, 159007 with other textures (of 600000) [hos host] vk: slow submit: 15.9 ms: memory sync 15.8 [memory sync: maps 0.0 (lock waits 0.0), capture 15.6, collect 0.2] [hos host] vk: replay reads of guest memory as submitted: 260000 (166568 pages from the captures) [hos host] vk: a frame: 503 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 17.40, present 0.22, draw 10.55, descriptors 4.40, texture lookups 1.07, uploads 0.28, pipelines 0.46, programs 1.93, commands 2.94, dispatch 1.10, memory sync 1.40, queue submit 0.00, slot wait 0.28 [hos host] vk: a frame: mirrors 91 (938 MB): 5.54 MB synced in 1029.5 copies, 18.09 MB written back [hos host] vk: 2520 presents, 42.1 fps; a frame: 2864 draws (0 skipped), 162 dispatches, 164 passes, 171 barriers, 6 submits, 2.8 uploads (0.8 MB), 88 ub updates (60 overlays), GPU-bound wait 0.1 ms (recording slots only); 11914 pipelines (2757 ms), 0 programs (10 failed), 1991 textures (3159 MB) [hos host] vk: host time in GPU resource waits: 6.88 ms a frame (fences 4.03, timelines 2.85, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 592.8 full barriers recorded, 294.6 left out (nothing recorded since the one before), 17.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.8; 0x8: 3.5; 0x9: 1.0; 0x10: 4.8; 0x11: 46.1; 0x12: 82.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 34.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 161.8, an op 2320.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 13.5 ms a frame (the longest submission 3.6 ms); 5.2 textures made a frame: colour targets 1.2 (7.2 MB), depth 1.4 (6.5 MB), ASTC or BC 2.6 (2.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.33 MB unswizzled on the CPU in 0.07 ms [hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2213 dirty pages from their submits' captures, 45673 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 873 dispatches, 35034 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6291 runs, 612.6 MB compared; 76 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11556 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 280000 (178165 pages from the captures) [hos host] vk: 19 textures retired (1973 left, 3145 MB): budget 19, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 300000 (189793 pages from the captures) [hos host] vk: 16 textures retired (1957 left, 3139 MB): budget 16, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: CommandBufferWaitEvent #15000: event 0x100d6c8028 mode 0 value 6655 (now 6636) (lr 0x806674b0) [hos host] vk: sparse mirrors: 70, 938 MB with memory of 1136 MB (197 MB spared); blocks given memory later 71, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 110 binds (17.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 14 taken from the spares, 78 made new; 22 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 726, 3457.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2897 dirty pages from their submits' captures, 44391 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 867 dispatches, 34537 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5788 runs, 588.2 MB compared; 73 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12184 in captures, 1 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 74 live (1292 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB) [hos host] vk: replay reads of guest memory as submitted: 320000 (201516 pages from the captures) [hos host] vk: 17 textures retired (1952 left, 3138 MB): budget 7, drawn over 0, dead targets 10 [hos host] vk: 16 textures retired (1936 left, 3133 MB): budget 16, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 340000 (213968 pages from the captures) [hos host] vk: texture #2123 0x10008b23400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2587 (drawn at 2413) [hos host] vk: 22 textures retired (1915 left, 3122 MB): budget 21, drawn over 0, dead targets 1 [hos host] vk: program picks: 559264 kept (206757 by the last picks' textures, a uniform buffer bound since), 24959 new, 215776 with other textures (of 800000) [hos host] vk: texture #2124 0x10008bb3600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2594 (drawn at 2413) [hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4475 dirty pages from their submits' captures, 44014 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 969 dispatches, 35291 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5705 runs, 601.8 MB compared; 74 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11964 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 54 textures retired (1878 left, 3093 MB): budget 44, drawn over 0, dead targets 10 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05 [hos host] vk: texture #2148 0x10008c43800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2600 (drawn at 2413) [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.04 over 1312 ranges, 3448 MB), 227 pages written, 227 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.043; GPU wait for late pages 0.026 (1 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (106 MB), 8 spare buffers, 6552 texture pages copied (25 MB) [hos host] vk: replay reads of guest memory as submitted: 360000 (226547 pages from the captures) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 849 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 23 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 7339 kept their virtual core (nothing there to wait for), 2901 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 18 textures retired (1860 left, 3083 MB): budget 18, drawn over 0, dead targets 0 [hos host] vk: 55 textures retired (1809 left, 3035 MB): budget 55, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 380000 (238983 pages from the captures) [hos host] vk: handle table: element 25 of program 2fd8c7ca964e5a1d names texture #822 (format 83) for a depth comparison: bound as nothing [hos host] vk: handle table: element 26 of program 2fd8c7ca964e5a1d names texture #823 (format 83) for a depth comparison: bound as nothing [hos host] vk: handle table: element 28 of program 2fd8c7ca964e5a1d names texture #823 (format 83) for a depth comparison: bound as nothing [hos host] vk: handle table: element 25 of program b1b048510819ac68 names texture #822 (format 83) for a depth comparison: bound as nothing [hos host] vk: handle table: element 26 of program b1b048510819ac68 names texture #823 (format 83) for a depth comparison: bound as nothing [hos host] vk: handle table: element 28 of program b1b048510819ac68 names texture #823 (format 83) for a depth comparison: bound as nothing [hos host] vk: handle table: element 25 of program b1b048510819ac68 names texture #822 (format 83) for a depth comparison: bound as nothing [hos host] vk: handle table: element 26 of program b1b048510819ac68 names texture #823 (format 83) for a depth comparison: bound as nothing [hos host] vk: recorder: 4441759 packets run in 10.0 s (busy 1.26 s, spun idle 0.10 s, slept 9008 times, 2340 of them under 200 us), 4 calls made directly (no room); 11921 batches over 384 frames (31.0 a frame, 373 packets a batch), 6025 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 297 times, 34.33 ms (0.8 times, 0.09 ms a frame); every thread 734 times, 79.15 ms in all, the longest 4.52 ms; vkWaitForFences 329/31.96 ms (max 4.52), vkEndCommandBuffer 152/14.88 ms (max 0.74), vkDestroyImageView 84/14.70 ms (max 0.64), vkQueueBindSparse 102/8.12 ms (max 0.94), the queue's lock 19/2.91 ms (max 0.56), vkDestroyPipeline 15/2.48 ms (max 0.40) [hos host] vk: recorder: of its busy 1.26 s, 0.30 s in parts' ends and submits (the longest 4.57 ms), 0.06 s of that blocked on the queue's lock (108 times, the longest 2.68 ms), so 0.24 s in the ends, submits and holds; 0.96 s running the other packets; 11567 packets a frame [hos host] vk: 6 textures retired (1832 left, 3048 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 83 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1334, sb 559, cis 10086, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 78, 940 MB with memory of 1168 MB (227 MB spared); blocks given memory later 84, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 121 binds (18.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 14 taken from the spares, 86 made new; 22 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3883 dirty pages from their submits' captures, 46412 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1078 dispatches, 35828 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5814 runs, 603.3 MB compared; 90 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12353 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 82 live (1324 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB) [hos host] vk: replay reads of guest memory as submitted: 400000 (251457 pages from the captures) [hos host] vk: 16 textures retired (1828 left, 3048 MB): budget 6, drawn over 0, dead targets 10 [hos host] vk: a frame: 340 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 13.19, present 0.17, draw 8.79, descriptors 3.82, texture lookups 0.25, uploads 0.02, pipelines 0.38, programs 1.42, commands 2.48, dispatch 1.05, memory sync 0.44, queue submit 0.00, slot wait 0.27 [hos host] vk: a frame: mirrors 103 (941 MB): 5.68 MB synced in 1044.0 copies, 17.77 MB written back [hos host] vk: 2640 presents, 56.0 fps; a frame: 2425 draws (0 skipped), 149 dispatches, 165 passes, 167 barriers, 6 submits, 0.4 uploads (0.1 MB), 86 ub updates (59 overlays), GPU-bound wait 0.1 ms (recording slots only); 11914 pipelines (2757 ms), 0 programs (10 failed), 1829 textures (3049 MB) [hos host] vk: host time in GPU resource waits: 4.86 ms a frame (fences 2.24, timelines 2.62, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 545.7 full barriers recorded, 288.0 left out (nothing recorded since the one before), 21.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.8; 0x8: 3.0; 0x9: 1.0; 0x10: 4.8; 0x11: 41.8; 0x12: 84.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.6 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 158.4, an op 1850.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 11.5 ms a frame (the longest submission 3.5 ms); 1.8 textures made a frame: colour targets 1.0 (6.1 MB), depth 0.5 (2.7 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms [hos host] vk: 16 textures retired (1827 left, 3045 MB): budget 16, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 420000 (263799 pages from the captures) [hos AudioUserTaskThread] vcores: 135485 takes, 33541 waits (mean 103 us, longest 23.5 ms, 0 over HOS_VCORE_LOG_MS), 16686 kicks, 16782 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 666 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 15 pages in 15 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 60 textures retired (1796 left, 3031 MB): budget 49, drawn over 0, dead targets 11 [hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 35 made): 25122194 lookups, 191012 missed (0.76 %) [hos host] vk: readback pages refreshed from guest memory: 164054 pages so far (0 held: may be written ahead in their submission) [hos host] vk: mirror of pool 0x7fd3400000 (4 MB) given back: idle, 6007 submissions since its sync [hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3708 dirty pages from their submits' captures, 48557 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1197 dispatches, 36190 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5813 runs, 604.4 MB compared; 78 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12683 in captures, 0 only in the collection, 27 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 440000 (276171 pages from the captures) [hos host] vk: 18 textures retired (1791 left, 3024 MB): budget 8, drawn over 0, dead targets 10 [hos host] vk: program picks: 699778 kept (256124 by the last picks' textures, a uniform buffer bound since), 24979 new, 275242 with other textures (of 1000000) [hos host] vk: 8 textures retired (1794 left, 3029 MB): budget 8, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 460000 (289091 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 1310 ranges, 3479 MB), 246 pages written, 246 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (142 MB), 12 spare buffers, 6555 texture pages copied (25 MB) [hos host] vk: 3 textures retired (1822 left, 3054 MB): budget 3, drawn over 0, dead targets 0 [hos ModuleSystemWorker4] vk: readbacks of GPU results: 600 of 3000 waited for their write-back, 0.08 ms each; 0 looked at the stopped queues first, 24 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 4327 kept their virtual core (nothing there to wait for), 3097 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: queues ordered by their events: 12000 submissions stopped at a wait, 11999 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 18341, not met 0; captures at a CPU event's wait 2479, 0 of them at a wait met on arrival [hos host] vk: 3 textures retired (1846 left, 2867 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: replay reads of guest memory as submitted: 480000 (301643 pages from the captures) [hos host] vk: sparse mirrors: 83, 945 MB with memory of 1188 MB (242 MB spared); blocks given memory later 85, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 129 binds (18.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 15 taken from the spares, 91 made new; 23 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 747, 3560.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4254 dirty pages from their submits' captures, 48890 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1221 dispatches, 36545 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5552 runs, 599.2 MB compared; 67 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12578 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 88 live (1348 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB) [hos host] vk: 32 textures retired (1851 left, 2861 MB): budget 0, drawn over 0, dead targets 32 [hos host] vk: replay reads of guest memory as submitted: 500000 (314201 pages from the captures) [hos host] vk: 20 textures retired (1859 left, 2862 MB): budget 0, drawn over 0, dead targets 20 [hos host] vk: 25 textures retired (1881 left, 2899 MB): budget 25, drawn over 0, dead targets 0 [hos host] vk: mirror of pool 0x1042c00000 (4 MB) given back: idle, 6003 submissions since its sync [hos host] vk: write-backs left out (600 submits): 727, 3458.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4420 dirty pages from their submits' captures, 48370 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1280 dispatches, 35945 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5313 runs, 580.3 MB compared; 80 pages read into the shadows; 200 MB of shadows (112 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12469 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 520000 (326814 pages from the captures) [hos sead::MainThread] nvn: CommandBufferSignalEvent #10000: event 0x100d6c8028 mode 0 value 9999 (lr 0x80667410) [hos host] vk: replay reads of guest memory as submitted: 540000 (339417 pages from the captures) [hos host] vk: program picks: 837584 kept (306999 by the last picks' textures, a uniform buffer bound since), 25017 new, 337398 with other textures (of 1200000) [hos sead::MainThread] nvn: CommandBufferWaitEvent #20000: event 0x100d6c8028 mode 0 value 10455 (now 10436) (lr 0x806674b0) [hos host] vk: a frame: 273 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 13.90, present 0.20, draw 9.00, descriptors 3.84, texture lookups 0.60, uploads 0.08, pipelines 0.39, programs 1.52, commands 2.58, dispatch 0.97, memory sync 0.45, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 110 (945 MB): 6.23 MB synced in 1080.3 copies, 17.64 MB written back [hos host] vk: 2760 presents, 53.4 fps; a frame: 2438 draws (0 skipped), 149 dispatches, 184 passes, 175 barriers, 6 submits, 1.8 uploads (0.4 MB), 85 ub updates (58 overlays), GPU-bound wait 0.1 ms (recording slots only); 11914 pipelines (2757 ms), 0 programs (10 failed), 1928 textures (2908 MB) [hos host] vk: host time in GPU resource waits: 5.43 ms a frame (fences 2.36, timelines 3.07, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 568.9 full barriers recorded, 293.5 left out (nothing recorded since the one before), 30.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.3; 0x8: 3.0; 0x9: 1.0; 0x10: 4.3; 0x11: 42.4; 0x12: 92.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 166.0, an op 1822.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 11.7 ms a frame (the longest submission 3.2 ms); 6.3 textures made a frame: colour targets 2.6 (14.7 MB), depth 2.2 (7.8 MB), ASTC or BC 1.6 (1.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.17 MB unswizzled on the CPU in 0.03 ms [hos host] vk: texture #866 0x10009154000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2763 (drawn at 2351) [hos Presentation Thread] vi: 2766 frames presented (54.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 29, 15.5-17.8 111, 17.8-22 107, 22-30 20, 30-40 3, over 40 1 [hos host] vk: sparse mirrors: 86, 945 MB with memory of 1200 MB (254 MB spared); blocks given memory later 86, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 135 binds (19.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 15 taken from the spares, 95 made new; 24 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: replay reads of guest memory as submitted: 560000 (351936 pages from the captures) [hos host] vk: write-backs left out (600 submits): 747, 3560.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4440 dirty pages from their submits' captures, 51312 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1385 dispatches, 37467 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5359 runs, 596.5 MB compared; 70 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12287 in captures, 0 only in the collection, 20 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 92 live (1364 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 229 sync waits (614 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos sead::MainThread] vk: readbacks of GPU results: 510 of 3000 waited for their write-back, 0.09 ms each; 0 looked at the stopped queues first, 20 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 2934 kept their virtual core (nothing there to wait for), 2698 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 18 textures retired (1942 left, 2952 MB): budget 0, drawn over 0, dead targets 18 [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 1276 ranges, 3527 MB), 259 pages written, 259 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (128 MB), 10 spare buffers, 6564 texture pages copied (25 MB) [hos host] vk: texture 0x10aec87200 576x576x1 x1 levels 10 NVN 0x7b -> vk 37 (ASTC decoded) (2000 textures, 3049 MB) [hos host] vk: reaper: 817 images (1760 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 116 (258 MB) [hos host] vk: texture #38 0x1001efd9000 631x94x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1514 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #80 0x1001eccba00 1466x282x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 821 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #7 0x1000021c200 960x540x1 x1 NVN 0xd the GPU drew goes with no copy of its texels (finalized, unused for 813 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #9 0x10000348400 2560x1440x1 x1 NVN 0x38 the GPU drew goes with no copy of its texels (finalized, unused for 813 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 34 textures retired (1995 left, 2964 MB): budget 33, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 16 MB [hos host] vk: texture 0x1000a535400 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 2987 MB) [hos host] vk: texture 0x1000a5c5600 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 3000 MB) [hos host] vk: texture 0x1000a655800 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 3012 MB) [hos host] vk: replay reads of guest memory as submitted: 580000 (364547 pages from the captures) [hos host] vk: texture 0x1000a985800 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 3037 MB) [hos host] vk: texture 0x1000a295600 640x360x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3062 MB) [hos host] vk: 13 textures retired (1999 left, 3088 MB): budget 11, drawn over 2, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 600000 (377139 pages from the captures) [hos host] vk: texture 0x10d18f6800 768x768x1 x1 levels 10 NVN 0x49 -> vk 139 (2000 textures, 3089 MB) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 86 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1334, sb 559, cis 10086, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 704, 3352.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4518 dirty pages from their submits' captures, 51415 from earlier ones', 12 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1400 dispatches, 36553 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 4814 runs, 541.8 MB compared; 73 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 133 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11861 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: 0x1054000000 (4 MB, captured, state 2): 6; 0x1053800000 (4 MB, captured, state 2): 6; [hos host] vk: 3 textures retired (1998 left, 3088 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: texture 0x10008c43800 1920x1920x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3102 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 632 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 74 pages in 20 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture 0x1000a4d5e00 640x360x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3019 MB) [hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 20000 so far (x1 0x117daa3070 x2 0 x3 0x117daa3040 x4 0x9b040) [hos host] vk: texture 0x10008bb3600 1693x1693x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3045 MB) [hos host] vk: texture 0x1000a775c00 1280x720x1 x1 levels 1 NVN 0x25 -> vk 37 (2000 textures, 2977 MB) [hos host] vk: 16 textures retired (1990 left, 2961 MB): budget 5, drawn over 0, dead targets 11 [hos host] vk: texture 0x100099c4a00 384x384x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 2979 MB) [hos OptimizeChangeBanc] nvn: 250000 recordings made, 68472 live copies (23.7 MB), 1083 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: replay reads of guest memory as submitted: 620000 (389759 pages from the captures) [hos host] vk: program picks: 979481 kept (364878 by the last picks' textures, a uniform buffer bound since), 25038 new, 395480 with other textures (of 1400000) [hos host] vk: texture 0x10009273e00 384x384x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3025 MB) [hos host] vk: 26 textures retired (1973 left, 3014 MB): budget 26, drawn over 0, dead targets 0 [hos host] vk: mirror of pool 0x7f54000000 (4 MB) given back: idle, 6006 submissions since its sync [hos host] vk: texture 0x10d2b5de00 1536x1536x1 x1 levels 11 NVN 0x8d -> vk 137 (ASTC decoded) (2000 textures, 3044 MB) [hos host] vk: slow submit: 24.4 ms: draw 22.2 descriptors 16.1 texture lookups 13.3 uploads 12.8 programs 4.1 commands 1.5 [hos host] vk: replay reads of guest memory as submitted: 640000 (402940 pages from the captures) [hos host] vk: 2 textures retired (2030 left, 3066 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 298110 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 95, 950 MB with memory of 1236 MB (285 MB spared); blocks given memory later 89, given back 1027; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 148 binds (21.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 17 taken from the spares, 103 made new; 25 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 705, 3354.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 5403 dirty pages from their submits' captures, 53888 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1544 dispatches, 37776 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5279 runs, 579.2 MB compared; 142 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12154 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 102 live (1404 MB), 19 retired still imported (384 MB), 0 retired given back (0 MB) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 437 of 3000 waited for their write-back, 0.10 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 2372 kept their virtual core (nothing there to wait for), 3004 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: (its written part: 4 of 4 KB readable, 1014 bytes not zero; levels L0) [hos host] vk: texture 0x107cd82600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it [hos host] vk: 10 textures retired (2031 left, 3062 MB): budget 9, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 660000 (415527 pages from the captures) [hos host] vk: 32 textures retired (1999 left, 3047 MB): budget 31, drawn over 1, dead targets 0 [hos host] vk: texture 0x10cde3fe00 171x171x1 x1 levels 8 NVN 0x8e -> vk 37 (ASTC decoded) (2000 textures, 3047 MB) [hos host] vk: 26 textures retired (1976 left, 3030 MB): budget 26, drawn over 0, dead targets 0 [hos Presentation Thread] vk: capture: a submit 0.22 ms (lock 0.02, scan 0.05 over 688 ranges, 3337 MB), 332 pages written, 332 copied (1.3 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (126 MB), 9 spare buffers, 6576 texture pages copied (25 MB) [hos host] vk: replay reads of guest memory as submitted: 680000 (428168 pages from the captures) [hos host] vk: texture 0x10009875c00 2560x1440x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3065 MB) [hos host] vk: write-backs left out (600 submits): 704, 3352.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3537 dirty pages from their submits' captures, 52433 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1471 dispatches, 36828 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5269 runs, 561.9 MB compared; 96 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12180 in captures, 1 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 5 textures retired (1974 left, 2977 MB): budget 0, drawn over 3, dead targets 2 [hos host] vk: texture 0x1000a9b5c00 1280x720x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3034 MB) [hos host] vk: replay reads of guest memory as submitted: 700000 (440582 pages from the captures) [hos host] vk: texture 0x10009e15800 2560x1440x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3019 MB) [hos host] vk: 21 textures retired (1984 left, 2989 MB): budget 0, drawn over 0, dead targets 21 [hos host] vk: 8179 programs from the program cache [hos host] vk: a frame: 366 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 17.33, present 0.24, draw 10.58, descriptors 4.54, texture lookups 0.80, uploads 0.17, pipelines 0.46, programs 1.74, commands 3.03, dispatch 1.20, memory sync 1.00, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 120 (950 MB): 6.94 MB synced in 1149.0 copies, 17.54 MB written back [hos host] vk: 2880 presents, 45.4 fps; a frame: 2851 draws (0 skipped), 157 dispatches, 202 passes, 182 barriers, 6 submits, 1.6 uploads (0.3 MB), 82 ub updates (57 overlays), GPU-bound wait 0.1 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1984 textures (2995 MB) [hos host] vk: host time in GPU resource waits: 6.49 ms a frame (fences 2.55, timelines 3.94, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 595.6 full barriers recorded, 305.3 left out (nothing recorded since the one before), 37.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 44.8; 0x12: 99.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 34.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 173.1, an op 2099.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 13.3 ms a frame (the longest submission 3.9 ms); 6.4 textures made a frame: colour targets 2.1 (11.8 MB), depth 2.9 (9.9 MB), ASTC or BC 1.4 (1.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.12 MB unswizzled on the CPU in 0.02 ms [hos host] vk: texture 0x10009b74a00 384x384x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3044 MB) [hos host] vk: program picks: 1118828 kept (416106 by the last picks' textures, a uniform buffer bound since), 25070 new, 456101 with other textures (of 1600000) [hos host] vk: texture #4177 0x10009815c00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2888 (drawn at 2827) [hos host] vk: 23 textures retired (1972 left, 3052 MB): budget 23, drawn over 0, dead targets 0 [hos host] vk: texture #4209 0x10009665c00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2891 (drawn at 2830) [hos host] vk: texture 0x1000b2b5600 1280x720x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 3139 MB) [hos sead::MainThread] vk: readbacks of GPU results: 406 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1845 kept their virtual core (nothing there to wait for), 3019 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: sparse mirrors: 95, 950 MB with memory of 1236 MB (285 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 158 binds (21.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 17 taken from the spares, 103 made new; 25 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 747, 3560.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3706 dirty pages from their submits' captures, 53527 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1448 dispatches, 38187 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5531 runs, 580.7 MB compared; 86 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13259 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 102 live (1404 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB) [hos host] vk: replay reads of guest memory as submitted: 720000 (452502 pages from the captures) [hos tera::FileResourceMgr] vcores: 145002 takes, 37873 waits (mean 121 us, longest 48.0 ms, 0 over HOS_VCORE_LOG_MS), 18392 kicks, 18543 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.15 [hos host] vk: late pages' fix (600 recordings with one): of 803 captures with late pages 797 fixed (47.0 pages, 47.0 fixes a capture), waited as before: 0 no pool, 6 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 242 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 2 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 130, of an earlier recording 110, not recorded yet 0, of this part 0; pages fixed from stamps 36833, from snapshots 587; the shadows: 210447 pages synced from captures, 203319 made known by it, 7069 left unknown because the replay or the collection wrote them since [hos host] vk: replay reads of guest memory as submitted: 740000 (464479 pages from the captures) [hos host] vk: texture #1870 0x1001f023c00 870x133x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 196 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 13 textures retired (1943 left, 3041 MB): budget 11, drawn over 2, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: texture #4208 0x100095d5a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2921 (drawn at 2860) [hos host] vk: texture #4305 0x10009785a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2925 (drawn at 2864) [hos host] vk: 8 textures retired (1963 left, 3234 MB): budget 8, drawn over 0, dead targets 0 [hos host] vk: write-backs left out (600 submits): 749, 3563.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3370 dirty pages from their submits' captures, 49692 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1309 dispatches, 36374 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5621 runs, 612.4 MB compared; 71 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13029 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: texture #4351 0x10009935600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2935 (drawn at 2874) [hos host] vk: replay reads of guest memory as submitted: 760000 (476403 pages from the captures) [hos host] vk: 4 textures retired (1971 left, 3354 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: texture #4397 0x10009c95400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2945 (drawn at 2884) [hos host] vk: mirror of pool 0x7fe1800000 (4 MB) given back: idle, 6015 submissions since its sync [hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 764 ranges, 3369 MB), 269 pages written, 269 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (112 MB), 11 spare buffers, 6576 texture pages copied (25 MB) [hos host] vk: 84 textures retired (1899 left, 3382 MB): budget 84, drawn over 0, dead targets 0 [hos host] vk: texture #4441 0x10009db5200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2952 (drawn at 2891) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 409 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1928 kept their virtual core (nothing there to wait for), 3192 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 637 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 5 pages in 5 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #4512 0x1000a3e5000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2957 (drawn at 2896) [hos host] vk: replay reads of guest memory as submitted: 780000 (488380 pages from the captures) [hos host] vk: commit: 2476 MB left under Windows' limit as it is now, and the page files can grow by 80896 MB more: counted as room [hos sead::MainThread] nvn: CommandBufferWaitEvent #25000: event 0x100d6c8028 mode 0 value 14255 (now 14236) (lr 0x806674b0) [hos host] vk: presenter thread: the copy's fence 2.34 ms, the present 0.84 ms a present [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 86 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1334, sb 559, cis 10086, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 95, 950 MB with memory of 1236 MB (285 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 160 binds (21.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 17 taken from the spares, 104 made new; 26 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3174 dirty pages from their submits' captures, 51226 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1327 dispatches, 37697 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5683 runs, 626.1 MB compared; 85 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13427 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 103 live (1408 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB) [hos host] vk: 168 textures retired (1754 left, 3461 MB): budget 168, drawn over 0, dead targets 0 [hos host] vk: texture #4588 0x1000a354a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 2973 (drawn at 2912) [hos host] vk: mirror of pool 0x7fe2800000 (4 MB) given back: idle, 6063 submissions since its sync [hos host] vk: program picks: 1259712 kept (464759 by the last picks' textures, a uniform buffer bound since), 25070 new, 515217 with other textures (of 1800000) [hos host] vk: replay reads of guest memory as submitted: 800000 (500296 pages from the captures) [hos host] vk: 17 textures retired (1748 left, 3384 MB): budget 0, drawn over 17, dead targets 0 [hos host] vk: 29 textures retired (1725 left, 3328 MB): budget 18, drawn over 11, dead targets 0 [hos host] vk: mirror of pool 0x7fe2400000 (4 MB) given back: idle, 6018 submissions since its sync [hos sead::MainThread] nvn: CommandBufferSignalEvent #15000: event 0x100d6c8034 mode 0 value 14999 (lr 0x806616b0) [hos host] vk: replay reads of guest memory as submitted: 820000 (512282 pages from the captures) [hos host] vk: memory: working set 13072 MB (peak 13092), private 20589 MB, staging 32 MB in 2 arenas, 19 made so far (56 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 624 MB, 36 allocations 548 MB; used 2416 of 15579 MB; heap 1 (device): 196 blocks 5687 MB, 17640 allocations 4987 MB; used 7504 of 7868 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 42 blocks 1344 MB, 191 images 1068 MB; 171 free ranges, the largest 25 MB [hos host] vk: image blocks (textures, compacted): 69 blocks 2208 MB, 1545 images 1955 MB; 320 free ranges, the largest 11 MB [hos host] vk: render target images: 1416 reused from the pool, 688 made new, 495 destroyed from it, 138 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 25 missed after their image was taken by size; it holds 0 (0 MB), VRAM short: halved [hos host] vk: render targets not reused by size: 324 with the pool empty (0 of them with targets still in the dead queue), 50 with none in it a present yet, 13 with none of a kind they can use, 79 with all smaller (the largest 57% short on average), 86 with all bigger (over 1.5 times), 136 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 10 over its bound and 485 idle [hos host] vk: VRAM: 7504 MB used of 7868 MB allowed (699 MB of it free in the blocks); textures 3483 MB (budget 3145 MB), 1740 textures [hos host] vk: budget: VRAM 5687 of 7868 MB (7504 MB by the driver, 1817 MB unaccounted): textures 3483, mirrors 949, blocks free 699, other buffers 211, shadows 200, other images 111, staging 32; RAM 2346 MB (the process's private memory 20589 MB): imports 1408, capture spares 422, retired imports 364, captures 108, capture table 26, other buffers 18, others 0 [hos host] vk: textures: decoded ASTC 1600 MB, rendered 1457 MB (65 drawn in one frame, kept: 43444 KB), other 278 MB; used in the last 8 presents 2571 MB [hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 457 MB of blocks hashed in 49 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 1 presents ago [hos host] vk: 21 MB: 0x10006fce600 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x10005192200 1600x1600x1 x3 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 24 (8683 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 64x62x1 NVN 0x25: 1 (15 KB); 73x33x1 NVN 0x25: 1 (9 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 128x126x1 NVN 0x1: 1 (15 KB); 137x97x1 NVN 0x1: 1 (12 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 232x33x1 NVN 0x25: 1 (29 KB); 262x104x1 NVN 0x1: 1 (26 KB); 296x97x1 NVN 0x1: 1 (28 KB); 592x49x1 NVN 0x25: 2 (226 KB); [hos host] vk: 34 textures retired (1706 left, 3376 MB): budget 22, drawn over 12, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos Presentation Thread] vk: presents after the CPU's signal (600): 488 waited for it, 2.99 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 48.7 ms; VRAM 7.3 of 7.6 GB [hos host] vk: readback pages refreshed from guest memory: 461078 pages so far (0 held: may be written ahead in their submission) [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 24 program records loaded (9.5 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5253 queued, 5253 loaded (median 1.38 ms, 99th percentile 16.77 ms, worst 34.27 ms, 13012 ms and 296.2 MB read in all); 7285 programs (1431 of them compute variants without their code, 0 graphics programs no note names), 10490 graphics pipelines (1 not in a cache), 577 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1275 pipelines made at create taken by draws; 2569 create calls with nothing kept for their stages, 0 before the index; units 7822 (5253 ready, 0 loading), 14420 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 11458 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 575 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 307 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.51, present 0.23, draw 7.52, descriptors 3.19, texture lookups 0.88, uploads 0.01, pipelines 0.34, programs 1.23, commands 2.18, dispatch 0.93, memory sync 0.36, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 121 (949 MB): 6.05 MB synced in 1061.9 copies, 17.44 MB written back [hos host] vk: 3000 presents, 59.0 fps; a frame: 2152 draws (0 skipped), 149 dispatches, 186 passes, 173 barriers, 6 submits, 0.1 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1703 textures (3389 MB) [hos host] vk: host time in GPU resource waits: 3.43 ms a frame (fences 1.82, timelines 1.61, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 561.9 full barriers recorded, 295.0 left out (nothing recorded since the one before), 29.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 41.3; 0x12: 93.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 34.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 163.8, an op 1434.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 11.1 ms a frame (the longest submission 3.2 ms); 3.6 textures made a frame: colour targets 0.8 (4.7 MB), depth 2.8 (13.1 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.01 ms [hos host] vk: the render thread (600 presents, 19.86 ms a frame): its jobs 18.75 ms a frame (submits 4.02, presents 3.53, others 11.21), idle 1.10; the game waited for room at 14 presents, 0.21 ms each (0.00 a frame; 11.04 jobs queued at a present); 58 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 8.12 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 492 captures at a CPU signal; the submits' waits read in 0.062 ms a frame; readbacks: 2768 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 17.58 ms after its hand-over (40.94 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.20 + 0.21 (the window's copy is in the idle), after it 4.01 + 1.24; after a frame's last recording that reads or writes the game's memory come 1.04 recordings (0.00 ms of GPU time), after its last that reads it 1.04 (0.00 ms) [hos host] vk: frame ahead (600 presents): 11.08 parts a frame recorded in the signal's job, 62.3% of them took nothing that came with the signal; of the others, 4.18 a frame took pages of the signal's capture into a mirror (127.2 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 20 of 492 frames, its first three in 14 [hos host] vk: frame ahead, what came with the signal: 492 signals with a capture, each of 209.2 pages in 23.4 pools; between a frame's last submit and that signal 0.53 texture and sampler registrations, 0.74 releases and 0.00 other CPU signals a signal (71 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.72+0.25, 0.29+0.06, 0.44+0.07, 0.64+0.12, 0.33+0.10, 0.18+0.18, 0.14+0.16, 0.08+0.09, 0.09+0.13, 0.05+0.07, 0.03+0.02, 0.01+0.01 [hos host] vk: VRAM compaction 1 begins: the textures' blocks hold 2208 MB, 262 MB of it free space [hos host] vk: write-backs left out (600 submits): 769, 3664.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3445 dirty pages from their submits' captures, 49704 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1277 dispatches, 36678 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5279 runs, 589.6 MB compared; 76 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13250 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vi: real-time step on: ratio 1.051, game's frame rate 60 [hos host] vk: render targets drawn over from another start, let go: 2078 so far (9928 MB), 1865 of them (8920 MB) between the passes [hos host] vk: 111 textures retired (1609 left, 3397 MB): budget 104, drawn over 7, dead targets 0 [hos ModuleSystemWorker2] vk: readbacks of GPU results: 413 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1637 kept their virtual core (nothing there to wait for), 2715 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 840000 (524297 pages from the captures) [hos host] vk: 49 textures retired (1606 left, 3433 MB): budget 41, drawn over 8, dead targets 0 [hos Presentation Thread] vi: 3025 frames presented (51.7 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 22, 15.5-17.8 95, 17.8-22 97, 22-30 37, 30-40 6, over 40 2 [hos sead::MainThread] nvn: GPU waits in 5 s: 219 sync waits (652 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: 5 textures retired (1629 left, 3056 MB): budget 0, drawn over 5, dead targets 0 [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 763 ranges, 3369 MB), 257 pages written, 257 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (104 MB), 10 spare buffers, 6579 texture pages copied (25 MB) [hos host] vk: sparse mirrors: 93, 949 MB with memory of 1228 MB (278 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 162 binds (22.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 17 taken from the spares, 104 made new; 28 kept as spares, 0 destroyed; 11 spares now, 368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3947 dirty pages from their submits' captures, 48885 from earlier ones', 8 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1310 dispatches, 36181 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5532 runs, 609.1 MB compared; 75 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13016 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: 0x1053800000 (4 MB, captured, state 2): 7; 0x1051000000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 103 live (1408 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB) [hos host] vk: replay reads of guest memory as submitted: 860000 (536382 pages from the captures) [hos host] vk: 1 textures retired (1614 left, 2857 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: VRAM compaction 1 over: 274 textures moved (146 MB) in 23 passes, 0 left where they were; 8 blocks (256 MB) given back to the driver, 256 MB in all so far [hos host] vk: 40 textures retired (1627 left, 2920 MB): budget 0, drawn over 0, dead targets 40 [hos host] vk: replay reads of guest memory as submitted: 880000 (548381 pages from the captures) [hos host] vk: program picks: 1398101 kept (506854 by the last picks' textures, a uniform buffer bound since), 25081 new, 576817 with other textures (of 2000000) [hos sead::MainThread] vk: readbacks of GPU results: 401 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1802 kept their virtual core (nothing there to wait for), 3062 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2899 dirty pages from their submits' captures, 52014 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1367 dispatches, 38066 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5668 runs, 632.2 MB compared; 84 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13619 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 900000 (560340 pages from the captures) [hos ModuleSystemWorker4] nvn: 300000 recordings made, 65514 live copies (21.2 MB), 1886 texture copies; 3 command buffers' segments (4.1 MB) [hos host] vk: slow submit: 27.9 ms: draw 27.1 descriptors 19.3 texture lookups 17.0 uploads 14.2 programs 6.4 [hos host] vk: 27 textures retired (1742 left, 2804 MB): budget 0, drawn over 0, dead targets 27 [hos host] vk: texture #4560 0x1000a1a4e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3090 (drawn at 2902) [hos host] vk: replay reads of guest memory as submitted: 920000 (572989 pages from the captures) [hos host] vk: 50 textures retired (1731 left, 2766 MB): budget 0, drawn over 0, dead targets 50 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: mirror of pool 0x7f55400000 (4 MB) given back: idle, 6041 submissions since its sync [hos host] vk: sparse mirrors: 95, 948 MB with memory of 1236 MB (287 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 166 binds (23.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 20 taken from the spares, 104 made new; 29 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 793, 3773.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3165 dirty pages from their submits' captures, 51703 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1359 dispatches, 37629 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5549 runs, 627.4 MB compared; 91 pages read into the shadows; 200 MB of shadows (112 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13988 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: 0x1053800000 (4 MB, captured, state 2): 2; [hos host] vk: imports: 106 live (1420 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB) [hos host] vk: replay reads of guest memory as submitted: 940000 (584974 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 674 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 4 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #5286 0x1001f0d3800 382x97 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 3118 (drawn at 3053) [hos host] vk: 19 textures retired (1749 left, 2800 MB): budget 0, drawn over 0, dead targets 19 [hos host] vk: a frame: 330 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.63, present 0.43, draw 7.83, descriptors 3.60, texture lookups 1.14, uploads 0.22, pipelines 0.33, programs 1.30, commands 2.05, dispatch 0.88, memory sync 0.39, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 124 (948 MB): 5.97 MB synced in 1050.1 copies, 17.44 MB written back [hos host] vk: 3120 presents, 56.6 fps; a frame: 2049 draws (0 skipped), 143 dispatches, 183 passes, 172 barriers, 6 submits, 2.2 uploads (0.3 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1750 textures (2809 MB) [hos host] vk: host time in GPU resource waits: 4.54 ms a frame (fences 2.04, timelines 2.50, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 560.6 full barriers recorded, 291.0 left out (nothing recorded since the one before), 26.3 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 41.6; 0x12: 92.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 162.9, an op 1369.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 11.2 ms a frame (the longest submission 3.7 ms); 9.4 textures made a frame: colour targets 2.2 (11.8 MB), depth 5.4 (17.2 MB), ASTC or BC 1.9 (0.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.18 MB unswizzled on the CPU in 0.05 ms [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 756 ranges, 3381 MB), 252 pages written, 252 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (114 MB), 10 spare buffers, 6593 texture pages copied (25 MB) [hos sead::MainThread] vk: readbacks of GPU results: 396 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1623 kept their virtual core (nothing there to wait for), 2985 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 960000 (596871 pages from the captures) [hos host] vk: 33 textures retired (1779 left, 3008 MB): budget 0, drawn over 4, dead targets 29 [hos host] vk: 1 textures retired (1795 left, 3133 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 86 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 924, 4400.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2756 dirty pages from their submits' captures, 57920 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1375 dispatches, 42506 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6166 runs, 721.5 MB compared; 95 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15754 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 980000 (608764 pages from the captures) [hos host] vk: 1 textures retired (1778 left, 3012 MB): budget 1, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: CommandBufferWaitEvent #30000: event 0x100d6c8028 mode 0 value 18055 (now 18036) (lr 0x806674b0) [hos host] vk: recorder: 7455220 packets run in 10.0 s (busy 2.14 s, spun idle 0.14 s, slept 12323 times, 4210 of them under 200 us), 0 calls made directly (no room); 18677 batches over 537 frames (34.8 a frame, 399 packets a batch), 9288 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 387 times, 83.16 ms (0.7 times, 0.15 ms a frame); every thread 1085 times, 150.76 ms in all, the longest 0.93 ms; vkDestroyImageView 488/97.65 ms (max 0.93), vkWaitForFences 481/41.69 ms (max 0.70), the queue's lock 28/3.50 ms (max 0.47), vmaFreeMemory 8/2.75 ms (max 0.67), vkQueueBindSparse 44/2.42 ms (max 0.28), a part's submit 8/0.95 ms (max 0.29) [hos host] vk: recorder: of its busy 2.14 s, 0.51 s in parts' ends and submits (the longest 1.29 ms), 0.18 s of that blocked on the queue's lock (295 times, the longest 1.26 ms), so 0.33 s in the ends, submits and holds; 1.63 s running the other packets; 13883 packets a frame [hos host] vk: mirror of pool 0x7fbbc00000 (4 MB) given back: idle, 6064 submissions since its sync [hos host] vk: 6 textures retired (1773 left, 3008 MB): budget 0, drawn over 6, dead targets 0 [hos host] vk: texture #5802 0x10009c04800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3171 (drawn at 3110) [hos host] vk: program picks: 1530028 kept (535367 by the last picks' textures, a uniform buffer bound since), 25108 new, 644863 with other textures (of 2200000) [hos host] vk: mirror of pool 0x105dc00000 (4 MB) given back: idle, 6063 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 1000000 (620678 pages from the captures) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 406 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1549 kept their virtual core (nothing there to wait for), 3059 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #5962 0x10009e44400 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 3186 (drawn at 3125) [hos host] vk: readback pages refreshed from guest memory: 628224 pages so far (0 held: may be written ahead in their submission) [hos OptimizeChangeBanc] vcores: 152668 takes, 27567 waits (mean 81 us, longest 18.1 ms, 0 over HOS_VCORE_LOG_MS), 13588 kicks, 13803 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: sparse mirrors: 93, 944 MB with memory of 1228 MB (283 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 168 binds (23.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 20 taken from the spares, 104 made new; 31 kept as spares, 0 destroyed; 11 spares now, 368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 968, 4609.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 780 dirty pages from their submits' captures, 60672 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1403 dispatches, 43961 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7314 runs, 769.2 MB compared; 94 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17608 in captures, 0 only in the collection, 44 only by the replay itself [hos host] vk: read live, by pool: 0x1053800000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 106 live (1420 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB) [hos host] vk: replay reads of guest memory as submitted: 1020000 (632705 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: (its written part: 4 of 4 KB readable, 1512 bytes not zero; levels L0) [hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it [hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 35 made): 24285364 lookups, 146793 missed (0.60 %) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 828 ranges, 3367 MB), 254 pages written, 254 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 143 texture pages copied (0 MB) [hos host] vk: replay reads of guest memory as submitted: 1040000 (644927 pages from the captures) [hos host] vk: texture #5736 0x10009814a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3223 (drawn at 3104) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 269 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 5 textures retired (1740 left, 2767 MB): budget 0, drawn over 5, dead targets 0 [hos host] vk: write-backs left out (600 submits): 990, 4714.3 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 343 dirty pages from their submits' captures, 60303 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1375 dispatches, 44042 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7610 runs, 788.4 MB compared; 106 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17813 in captures, 0 only in the collection, 45 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 1060000 (657510 pages from the captures) [hos host] vk: a frame: 251 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.36, present 0.20, draw 5.69, descriptors 2.45, texture lookups 0.42, uploads 0.02, pipelines 0.25, programs 1.00, commands 1.57, dispatch 0.82, memory sync 0.38, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 124 (944 MB): 5.41 MB synced in 992.0 copies, 17.42 MB written back [hos host] vk: 3240 presents, 61.1 fps; a frame: 1546 draws (0 skipped), 144 dispatches, 166 passes, 164 barriers, 6 submits, 0.2 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1740 textures (2767 MB) [hos host] vk: host time in GPU resource waits: 2.15 ms a frame (fences 1.74, timelines 0.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 521.9 full barriers recorded, 280.3 left out (nothing recorded since the one before), 23.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.0; 0x12: 85.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 35.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 155.3, an op 1045.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 10.2 ms a frame (the longest submission 5.2 ms); 2.9 textures made a frame: colour targets 0.9 (5.0 MB), depth 1.9 (6.5 MB), ASTC or BC 0.1 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms [hos ModuleSystemWorker2] vk: readbacks of GPU results: 397 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1736 kept their virtual core (nothing there to wait for), 3128 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #6271 0x10008cd3a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3246 (drawn at 3185) [hos host] vk: 1 textures retired (1737 left, 2757 MB): budget 0, drawn over 1, dead targets 0 [hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #20000: event 0x100d6c8040 mode 0 value 19999 (lr 0x806621e8) [hos host] vk: replay reads of guest memory as submitted: 1080000 (670024 pages from the captures) [hos host] vk: texture #6432 0x10008cd3000 384x384 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (1 so far) [hos host] vk: replay reads of guest memory as submitted: 1100000 (682464 pages from the captures) [hos host] vk: mirror of pool 0x1042400000 (4 MB) given back: idle, 6058 submissions since its sync [hos host] vk: 32 mirrors given back (366 MB in all); 942 MB of mirrors now, 84 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: sparse mirrors: 92, 942 MB with memory of 1224 MB (281 MB spared); blocks given memory later 113, given back 1054; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 169 binds (23.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 20 taken from the spares, 104 made new; 32 kept as spares, 0 destroyed; 12 spares now, 372 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1078, 5133.3 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 616 dirty pages from their submits' captures, 61812 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1342 dispatches, 46427 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7948 runs, 853.0 MB compared; 138 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18655 in captures, 1 only in the collection, 49 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 106 live (1420 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB) [hos host] vk: texture #6270 0x10008c43800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3295 (drawn at 3234) [hos host] vk: 13 textures retired (1762 left, 2798 MB): budget 0, drawn over 0, dead targets 13 [hos ModuleSystemWorker3] vk: readbacks of GPU results: 393 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 9 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1592 kept their virtual core (nothing there to wait for), 3016 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: program picks: 1660735 kept (563714 by the last picks' textures, a uniform buffer bound since), 25123 new, 714141 with other textures (of 2400000) [hos host] vk: mirror of pool 0x105cc00000 (4 MB) given back: idle, 6029 submissions since its sync [hos host] vk: texture #6269 0x10008bb3600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3305 (drawn at 3244) [hos host] vk: replay reads of guest memory as submitted: 1120000 (695021 pages from the captures) [hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 40000 so far (x1 0x117cfbb040 x2 0 x3 0x117cfbb040 x4 0x9b040) [hos host] vk: 9 textures retired (1791 left, 2860 MB): budget 0, drawn over 0, dead targets 9 [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 823 ranges, 3373 MB), 244 pages written, 244 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (108 MB), 11 spare buffers, 147 texture pages copied (0 MB) [hos Presentation Thread] vi: 3322 frames presented (59.2 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 41, 15.5-17.8 207, 17.8-22 45, 22-30 3, 30-40 1, over 40 0 [hos host] vk: replay reads of guest memory as submitted: 1140000 (707637 pages from the captures) [hos host] vk: 1 textures retired (1795 left, 2911 MB): budget 0, drawn over 1, dead targets 0 [hos sead::MainThread] nvn: GPU waits in 5 s: 138 sync waits (277 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos Presentation Thread] nvn: QueueSubmitCommands #20000: queue 0000101004133020 (its #3334), 1 handles (lr 0x80590358) [hos host] vk: write-backs left out (600 submits): 989, 4712.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1569 dirty pages from their submits' captures, 57195 from earlier ones', 16 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1326 dispatches, 43414 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6974 runs, 765.2 MB compared; 102 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17154 in captures, 1 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: 0x1054c00000 (4 MB, captured, state 2): 10; 0x7fa2800000 (8 MB, captured, state 2): 4; 0x1055c00000 (4 MB, captured, state 2): 2; [hos rmbl::VibrationThread] hid: 8192 vibration values, strongest 0.00 since the last count [hos host] vk: reaper: 1415 images (4027 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 167 (248 MB) [hos host] vk: 25 textures retired (1862 left, 2960 MB): budget 0, drawn over 3, dead targets 22 [hos host] vk: texture #6105 0x10009b74000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3350 (drawn at 3143) [hos host] vk: (its written part: 4 of 4 KB readable, 1502 bytes not zero; levels L0) [hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it [hos host] vk: replay reads of guest memory as submitted: 1160000 (719829 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 196 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 383 of 3000 waited for their write-back, 0.10 ms each; 0 looked at the stopped queues first, 26 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1472 kept their virtual core (nothing there to wait for), 2880 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 1 textures retired (1879 left, 2969 MB): budget 0, drawn over 0, dead targets 1 [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #35000: event 0x100d6c8028 mode 0 value 21855 (now 21836) (lr 0x806674b0) [hos host] vk: a frame: 236 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.57, present 0.20, draw 5.82, descriptors 2.65, texture lookups 0.80, uploads 0.17, pipelines 0.24, programs 1.00, commands 1.51, dispatch 0.81, memory sync 0.41, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 127 (943 MB): 5.10 MB synced in 963.1 copies, 17.44 MB written back [hos host] vk: 3360 presents, 59.0 fps; a frame: 1506 draws (0 skipped), 137 dispatches, 157 passes, 159 barriers, 6 submits, 1.1 uploads (0.6 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1882 textures (2974 MB) [hos host] vk: host time in GPU resource waits: 2.84 ms a frame (fences 1.19, timelines 1.65, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 514.6 full barriers recorded, 273.3 left out (nothing recorded since the one before), 18.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.0; 0x12: 80.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 36.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 150.3, an op 982.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 10.5 ms a frame (the longest submission 5.9 ms); 5.4 textures made a frame: colour targets 2.0 (10.6 MB), depth 2.6 (8.1 MB), ASTC or BC 0.8 (0.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.52 MB unswizzled on the CPU in 0.12 ms [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 111 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 95, 943 MB with memory of 1236 MB (292 MB spared); blocks given memory later 113, given back 1056; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 175 binds (23.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 24 taken from the spares, 104 made new; 33 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 859, 4087.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2320 dirty pages from their submits' captures, 50554 from earlier ones', 67 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1195 dispatches, 38998 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6487 runs, 701.5 MB compared; 76 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15292 in captures, 1 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: 0x105e800000 (4 MB, captured, state 2): 67; [hos host] vk: imports: 108 live (1428 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB) [hos host] vk: replay reads of guest memory as submitted: 1180000 (731927 pages from the captures) [hos host] vk: 33 textures retired (1864 left, 2927 MB): budget 0, drawn over 20, dead targets 13 [hos OptimizeChangeBanc] nvn: 350000 recordings made, 67513 live copies (21.4 MB), 3065 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: replay reads of guest memory as submitted: 1200000 (743993 pages from the captures) [hos host] vk: 24 textures retired (1960 left, 2987 MB): budget 0, drawn over 8, dead targets 16 [hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.02, scan 0.04 over 902 ranges, 3384 MB), 280 pages written, 280 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (148 MB), 9 spare buffers, 147 texture pages copied (0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.12 [hos host] vk: (its written part: 14 of 14 KB readable, 4280 bytes not zero; levels L0) [hos host] vk: texture 0x107cd9f600 384x39x1 x1 levels 1 NVN 0x1 (14 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 14 KB of it [hos host] vk: readback pages refreshed from guest memory: 800005 pages so far (0 held: may be written ahead in their submission) [hos host] vk: program picks: 1806500 kept (624498 by the last picks' textures, a uniform buffer bound since), 25135 new, 768364 with other textures (of 2600000) [hos host] vk: texture 0x1000b4f4a00 1280x720x1 x1 levels 1 NVN 0x29 -> vk 97 (2000 textures, 3070 MB) [hos host] vk: write-backs left out (600 submits): 813, 3874.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2486 dirty pages from their submits' captures, 50880 from earlier ones', 23 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1300 dispatches, 38172 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6374 runs, 639.7 MB compared; 109 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14402 in captures, 1 only in the collection, 26 only by the replay itself [hos host] vk: read live, by pool: 0x1057c00000 (4 MB, captured, state 2): 10; 0x1059c00000 (4 MB, captured, state 2): 8; 0x1054400000 (4 MB, captured, state 2): 4; 0x105e400000 (4 MB, captured, state 2): 1; [hos host] vk: texture 0x1000ae04a00 1280x720x1 x1 levels 1 NVN 0x29 -> vk 97 (2000 textures, 3069 MB) [hos host] vk: 13 textures retired (1988 left, 3002 MB): budget 0, drawn over 0, dead targets 13 [hos ModuleSystemWorker2] vk: readbacks of GPU results: 333 of 3000 waited for their write-back, 0.18 ms each; 0 looked at the stopped queues first, 22 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1458 kept their virtual core (nothing there to wait for), 3150 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture 0x10d3801200 341x341x1 x1 levels 9 NVN 0x46 -> vk 133 (2000 textures, 3013 MB) [hos host] vk: 97 textures retired (1965 left, 3028 MB): budget 81, drawn over 2, dead targets 14 [hos host] vk: replay reads of guest memory as submitted: 1220000 (756005 pages from the captures) [hos host] vk: 75 textures retired (1900 left, 3028 MB): budget 73, drawn over 0, dead targets 2 [hos sead::MainThread] nvn: 400000 recordings made, 71335 live copies (22.0 MB), 3305 texture copies; 5 command buffers' segments (4.3 MB) [hos host] vk: 68 textures retired (1834 left, 3028 MB): budget 68, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 1240000 (768100 pages from the captures) [hos host] vk: sparse mirrors: 99, 948 MB with memory of 1252 MB (303 MB spared); blocks given memory later 117, given back 1059; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 184 binds (25.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 25 taken from the spares, 107 made new; 33 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 837, 3982.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2764 dirty pages from their submits' captures, 54070 from earlier ones', 38 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1401 dispatches, 40270 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7017 runs, 693.2 MB compared; 129 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15285 in captures, 2 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: 0x1057000000 (4 MB, captured, state 2): 18; 0x1053c00000 (4 MB, captured, state 2): 11; 0x1054800000 (4 MB, captured, state 2): 6; 0x1054000000 (4 MB, captured, state 2): 2; 0x1059800000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 108 live (1428 MB), 18 retired still imported (364 MB), 1 retired given back (20 MB) [hos host] vk: texture #5525 0x1001f077c00 64x62x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 205 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #5528 0x1001f034600 128x126x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 205 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 39 textures retired (1811 left, 3115 MB): budget 38, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 2 finalized, 0 MB [hos host] vk: 39 textures retired (1791 left, 3043 MB): budget 24, drawn over 0, dead targets 15 [hos ModuleSystemWorker2] vk: readbacks of GPU results: 338 of 3000 waited for their write-back, 0.14 ms each; 0 looked at the stopped queues first, 22 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1164 kept their virtual core (nothing there to wait for), 2676 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: queues ordered by their events: 18000 submissions stopped at a wait, 17996 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 37717, not met 0; captures at a CPU event's wait 3051, 0 of them at a wait met on arrival [hos host] vk: 18 textures retired (1797 left, 3088 MB): budget 0, drawn over 1, dead targets 17 [hos host] vk: replay reads of guest memory as submitted: 1260000 (780185 pages from the captures) [hos nn.npln.Shared] vcores: 161771 takes, 37763 waits (mean 83 us, longest 49.3 ms, 0 over HOS_VCORE_LOG_MS), 19294 kicks, 19315 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 5 textures retired (1814 left, 3348 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: a frame: 194 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.87, present 0.26, draw 7.00, descriptors 3.38, texture lookups 1.57, uploads 0.49, pipelines 0.29, programs 1.08, commands 1.78, dispatch 0.96, memory sync 1.30, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 133 (949 MB): 5.76 MB synced in 1040.3 copies, 17.71 MB written back [hos host] vk: 3480 presents, 52.7 fps; a frame: 1674 draws (0 skipped), 158 dispatches, 162 passes, 166 barriers, 6 submits, 2.0 uploads (1.6 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1795 textures (3080 MB) [hos host] vk: host time in GPU resource waits: 5.38 ms a frame (fences 1.92, timelines 3.46, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 546.9 full barriers recorded, 277.7 left out (nothing recorded since the one before), 22.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 42.8; 0x12: 85.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 30.8 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 157.2, an op 1169.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.8 ms); 8.7 textures made a frame: colour targets 3.3 (18.2 MB), depth 3.6 (22.5 MB), ASTC or BC 1.8 (1.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.44 MB unswizzled on the CPU in 0.41 ms [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 178 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 22 pages in 12 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 924, 4400.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1193 dirty pages from their submits' captures, 61091 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1498 dispatches, 44013 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7122 runs, 739.6 MB compared; 97 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15994 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 91 textures retired (1719 left, 3125 MB): budget 91, drawn over 0, dead targets 0 [hos sead::MainThread] vk: capture: a submit 0.21 ms (lock 0.03, scan 0.06 over 1319 ranges, 3589 MB), 303 pages written, 303 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (120 MB), 11 spare buffers, 146 texture pages copied (0 MB) [hos host] vk: pool 0x1018e0bc00+0xf00000 imported in 5.9 ms as 0x1018c00000..0x101a000000 (128 imports, 335 ms in all) [hos host] vk: replay reads of guest memory as submitted: 1280000 (792306 pages from the captures) [hos host] vk: pool 0x10184ab800+0xf00000 imported in 3.4 ms as 0x1018400000..0x101a000000 (129 imports, 338 ms in all) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.15 [hos host] vk: mirror of pool 0x1051000000 (4 MB) given back: idle, 6033 submissions since its sync [hos host] vk: mirror of pool 0x1053000000 (4 MB) given back: idle, 6048 submissions since its sync [hos host] vk: (its written part: 4 of 4 KB readable, 1495 bytes not zero; levels L0) [hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it [hos ModuleSystemWorker3] vk: readbacks of GPU results: 364 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1308 kept their virtual core (nothing there to wait for), 3044 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 10 textures retired (1774 left, 3020 MB): budget 0, drawn over 0, dead targets 10 [hos host] vk: replay reads of guest memory as submitted: 1300000 (805048 pages from the captures) [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #25000: event 0x100d6c8000 mode 0 value 24999 (lr 0x80291440) [hos host] vk: VRAM compaction 2 begins: the textures' blocks hold 2464 MB, 203 MB of it free space [hos host] vk: 15 textures retired (1786 left, 3028 MB): budget 3, drawn over 0, dead targets 12 [hos host] vk: slow upload 20.4 ms: #8318 0x10eead2a00 576x576x1 x1 levels 10 NVN 0x8d ASTC (BC3), 149 of 149 KB [hos host] vk: slow submit: 25.5 ms: draw 25.5 descriptors 24.0 texture lookups 22.6 uploads 22.0 [hos host] vk: sparse mirrors: 98, 947 MB with memory of 1248 MB (300 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 188 binds (25.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 25 taken from the spares, 108 made new; 35 kept as spares, 0 destroyed; 10 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 902, 4295.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 680 dirty pages from their submits' captures, 59745 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1437 dispatches, 42707 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6685 runs, 718.0 MB compared; 98 pages read into the shadows; 200 MB of shadows (113 MB of pages known, in 132 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16324 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 109 live (1456 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB) [hos host] vk: mirror of pool 0x1050800000 (4 MB) given back: idle, 6009 submissions since its sync [hos host] vk: program picks: 1938293 kept (658368 by the last picks' textures, a uniform buffer bound since), 25165 new, 836541 with other textures (of 2800000) [hos host] vk: 65 textures retired (1776 left, 3065 MB): budget 51, drawn over 0, dead targets 14 [hos host] vk: replay reads of guest memory as submitted: 1320000 (816929 pages from the captures) [hos host] vk: (its written part: 4 of 4 KB readable, 1495 bytes not zero; levels L0) [hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it [hos host] vk: 4 textures retired (1795 left, 3157 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: texture #8089 0x1000a894800 1280x720 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3557 (drawn at 3496) [hos host] vk: 4 textures retired (1808 left, 3288 MB): budget 4, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: CommandBufferWaitEvent #40000: event 0x100d6c8028 mode 0 value 25655 (now 25636) (lr 0x806674b0) [hos host] vk: presenter thread: the copy's fence 1.73 ms, the present 0.75 ms a present [hos host] vk: budget: gave back 34 MB from target pool at 90% of the allowance [hos host] vk: 34 textures retired (1745 left, 3031 MB): budget 33, drawn over 1, dead targets 0 [hos ModuleSystemWorker3] vk: readbacks of GPU results: 374 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1293 kept their virtual core (nothing there to wait for), 3059 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: mirror of pool 0x7fd9400000 (4 MB) given back: idle, 6023 submissions since its sync [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 111 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 924, 4400.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 784 dirty pages from their submits' captures, 59160 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1378 dispatches, 42934 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6847 runs, 735.1 MB compared; 92 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15617 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 1340000 (828851 pages from the captures) [hos host] vk: texture #8032 0x10008e84000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3579 (drawn at 3518) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 1319 ranges, 3589 MB), 270 pages written, 270 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (122 MB), 10 spare buffers, 147 texture pages copied (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 198 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 35 textures retired (1753 left, 3026 MB): budget 30, drawn over 5, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: memory: working set 13243 MB (peak 13245), private 20169 MB, staging 48 MB in 3 arenas, 20 made so far (56 ms) [hos host] vk: memory heaps: heap 0 (system): 15 blocks 640 MB, 36 allocations 572 MB; used 2500 of 15579 MB; heap 1 (device): 175 blocks 5015 MB, 17652 allocations 4722 MB; used 6833 of 7823 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 21 blocks 672 MB, 152 images 557 MB; 122 free ranges, the largest 28 MB [hos host] vk: image blocks (textures, compacted): 69 blocks 2208 MB, 1657 images 2189 MB; 310 free ranges, the largest 4 MB [hos host] vk: render target images: 3177 reused from the pool, 1530 made new, 1376 destroyed from it, 660 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 85 missed after their image was taken by size; it holds 33 (135 MB) [hos host] vk: render targets not reused by size: 418 with the pool empty (0 of them with targets still in the dead queue), 123 with none in it a present yet, 16 with none of a kind they can use, 263 with all smaller (the largest 48% short on average), 233 with all bigger (over 1.5 times), 477 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 65 over its bound and 1311 idle [hos host] vk: VRAM: 6833 MB used of 7823 MB allowed (292 MB of it free in the blocks); textures 3066 MB (budget 3028 MB), 1768 textures [hos host] vk: budget: VRAM 5011 of 7823 MB (6833 MB by the driver, 1821 MB unaccounted): textures 3066, mirrors 945, blocks free 292, other buffers 211, shadows 200, target pool 135, other images 111, staging 48; RAM 2438 MB (the process's private memory 20169 MB): imports 1456, capture spares 440, retired imports 384, captures 114, capture table 26, other buffers 18, others 0 [hos host] vk: textures: decoded ASTC 1608 MB, rendered 914 MB (104 drawn in one frame, kept: 94765 KB), other 480 MB; used in the last 8 presents 2536 MB [hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 528 MB of blocks hashed in 65 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 4 presents ago [hos host] vk: 21 MB: 0x10006fce600 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x10005192200 1600x1600x1 x3 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 17 (8407 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 208x105x1 NVN 0x1: 1 (21 KB); 262x104x1 NVN 0x1: 1 (26 KB); 592x49x1 NVN 0x25: 1 (113 KB); 656x113x1 NVN 0x1: 1 (72 KB); 1466x282x1 NVN 0x25: 5 (8074 KB); [hos host] vk: 28 textures retired (1740 left, 2971 MB): budget 0, drawn over 3, dead targets 25 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos Presentation Thread] vk: presents after the CPU's signal (600): 324 waited for it, 2.64 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: replay reads of guest memory as submitted: 1360000 (840734 pages from the captures) [hos host] vk: status: 600 frames, the slowest 43.5 ms; VRAM 6.7 of 7.6 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5253 queued, 5253 loaded (median 1.38 ms, 99th percentile 16.77 ms, worst 34.27 ms, 13012 ms and 296.2 MB read in all); 7285 programs (1431 of them compute variants without their code, 0 graphics programs no note names), 10490 graphics pipelines (1 not in a cache), 577 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1369 pipelines made at create taken by draws; 2569 create calls with nothing kept for their stages, 0 before the index; units 7822 (5253 ready, 0 loading), 14420 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 11458 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 575 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 176 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.97, present 0.54, draw 6.00, descriptors 2.85, texture lookups 1.16, uploads 0.38, pipelines 0.25, programs 0.98, commands 1.52, dispatch 0.86, memory sync 0.38, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 134 (945 MB): 5.68 MB synced in 1030.8 copies, 17.50 MB written back [hos host] vk: 3600 presents, 60.0 fps; a frame: 1402 draws (0 skipped), 146 dispatches, 162 passes, 166 barriers, 6 submits, 1.5 uploads (0.7 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1742 textures (2979 MB) [hos host] vk: host time in GPU resource waits: 2.82 ms a frame (fences 2.04, timelines 0.79, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 518.4 full barriers recorded, 275.6 left out (nothing recorded since the one before), 24.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 41.0; 0x12: 87.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 30.1 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 157.2, an op 920.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 9.5 ms a frame (the longest submission 3.0 ms); 6.5 textures made a frame: colour targets 2.1 (11.7 MB), depth 3.2 (13.2 MB), ASTC or BC 1.2 (0.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.37 MB unswizzled on the CPU in 0.08 ms [hos host] vk: queues ordered by their events: 18952 submissions stopped at a wait, 18952 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 41016, not met 0; captures at a CPU event's wait 3083, 0 of them at a wait met on arrival [hos host] vk: the render thread (600 presents, 17.33 ms a frame): its jobs 16.03 ms a frame (submits 5.34, presents 4.83, others 5.86), idle 1.30; the game waited for room at 50 presents, 0.19 ms each (0.02 a frame; 21.55 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 326 captures at a CPU signal; the submits' waits read in 0.044 ms a frame; readbacks: 1938 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.21 ms after its hand-over (36.00 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.68 + 0.35 (the window's copy is in the idle), after it 1.84 + 1.06; after a frame's last recording that reads or writes the game's memory come 1.02 recordings (0.00 ms of GPU time), after its last that reads it 1.02 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.60 parts a frame recorded in the signal's job, 66.9% of them took nothing that came with the signal; of the others, 2.84 a frame took pages of the signal's capture into a mirror (79.2 pages a frame) and 0.05 pages read as guest memory had them (0.4 pages a frame); a frame's first part took nothing in 37 of 326 frames, its first three in 21 [hos host] vk: frame ahead, what came with the signal: 326 signals with a capture, each of 201.5 pages in 32.3 pools; between a frame's last submit and that signal 0.03 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (8 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.60+0.09, 0.33+0.12, 0.23+0.10, 0.33+0.22, 0.11+0.21, 0.10+0.16, 0.09+0.11, 0.03+0.03, 0.02+0.02, 0.00+0.01, 0.00+0.00, 0.00+0.00 [hos ModuleSystemWorker3] tex: the game's texture heap: 45 compaction rounds asked for in 3600 frames without one running [hos Presentation Thread] vi: 3604 frames presented (56.3 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 64, 15.5-17.8 145, 17.8-22 45, 22-30 19, 30-40 7, over 40 2 [hos host] vk: render targets drawn over from another start, let go: 5303 so far (25283 MB), 4728 of them (22703 MB) between the passes [hos host] vk: readback pages refreshed from guest memory: 982054 pages so far (0 held: may be written ahead in their submission) [hos host] vk: VRAM compaction 2 over: 454 textures moved (337 MB) in 42 passes, 1 left where they were; 8 blocks (256 MB) given back to the driver, 512 MB in all so far [hos sead::MainThread] nvn: GPU waits in 5 s: 132 sync waits (486 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: sparse mirrors: 97, 945 MB with memory of 1244 MB (298 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 191 binds (25.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 26 taken from the spares, 108 made new; 37 kept as spares, 0 destroyed; 11 spares now, 368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 858, 4085.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1041 dirty pages from their submits' captures, 55813 from earlier ones', 97 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1314 dispatches, 40467 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6374 runs, 681.2 MB compared; 102 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14606 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 85; 0x1057800000 (4 MB, captured, state 2): 10; 0x1055800000 (4 MB, captured, state 2): 2; [hos host] vk: imports: 109 live (1456 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB) [hos host] vk: 87 textures retired (1724 left, 3027 MB): budget 75, drawn over 1, dead targets 11 [hos ModuleSystemWorker4] vk: readbacks of GPU results: 354 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1250 kept their virtual core (nothing there to wait for), 3102 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 1380000 (852769 pages from the captures) [hos host] vk: 2 textures retired (1733 left, 2996 MB): budget 0, drawn over 0, dead targets 2 [hos host] vk: 68 textures retired (1754 left, 3025 MB): budget 68, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 1400000 (865183 pages from the captures) [hos host] vk: texture #8536 0x1000a2c4200 1280x720 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 3647 (drawn at 3586) [hos host] vk: 16 textures retired (1775 left, 3024 MB): budget 13, drawn over 1, dead targets 2 [hos host] vk: write-backs left out (600 submits): 836, 3980.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2182 dirty pages from their submits' captures, 54215 from earlier ones', 16 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1339 dispatches, 39406 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6173 runs, 663.6 MB compared; 81 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14256 in captures, 0 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: 0x1055000000 (4 MB, captured, state 2): 11; 0x1055400000 (4 MB, captured, state 2): 5; [hos host] vk: program picks: 2066726 kept (685887 by the last picks' textures, a uniform buffer bound since), 25172 new, 908101 with other textures (of 3000000) [hos host] vk: slow submit: 18.6 ms: draw 16.9 descriptors 14.3 texture lookups 12.8 uploads 12.4 commands 1.3 [hos host] vk: 80 textures retired (1776 left, 3035 MB): budget 51, drawn over 3, dead targets 26 [hos host] vk: replay reads of guest memory as submitted: 1420000 (879032 pages from the captures) [hos host] vk: mirror of pool 0x105c000000 (4 MB) given back: idle, 6046 submissions since its sync [hos host] vk: 9 textures retired (1779 left, 3020 MB): budget 9, drawn over 0, dead targets 0 [hos ModuleSystemWorker2] barrier: 47779 looks: 15011 found the masks moved already, 31435 waited until they moved, 1333 for the whole 1 ms (a wait 302 us on average); 39649 wakes [hos ModuleSystemWorker3] vk: readbacks of GPU results: 351 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1180 kept their virtual core (nothing there to wait for), 2916 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 1319 ranges, 3589 MB), 258 pages written, 258 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (132 MB), 11 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: 22 textures retired (1768 left, 3026 MB): budget 21, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 1440000 (892706 pages from the captures) [hos host] vk: sparse mirrors: 98, 946 MB with memory of 1248 MB (301 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 194 binds (26.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 28 taken from the spares, 108 made new; 38 kept as spares, 0 destroyed; 10 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2733 dirty pages from their submits' captures, 50487 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1274 dispatches, 37156 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5864 runs, 631.1 MB compared; 91 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13046 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 109 live (1456 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB) [hos host] vk: 7 textures retired (1768 left, 3023 MB): budget 1, drawn over 6, dead targets 0 [hos host] vk: mirror of pool 0x7fb0000000 (8 MB) given back: idle, 6003 submissions since its sync [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 1460000 (906372 pages from the captures) [hos host] vk: 21 textures retired (1759 left, 2994 MB): budget 0, drawn over 21, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 1480000 (920106 pages from the captures) [hos host] vk: 1 textures retired (1760 left, 2996 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: a frame: 219 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.26, present 0.28, draw 6.90, descriptors 3.08, texture lookups 0.73, uploads 0.23, pipelines 0.31, programs 1.10, commands 1.91, dispatch 0.89, memory sync 0.39, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 136 (939 MB): 5.85 MB synced in 1040.8 copies, 17.53 MB written back [hos host] vk: 3720 presents, 55.6 fps; a frame: 1779 draws (0 skipped), 150 dispatches, 175 passes, 170 barriers, 6 submits, 1.8 uploads (0.5 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1761 textures (3000 MB) [hos host] vk: host time in GPU resource waits: 4.10 ms a frame (fences 1.68, timelines 2.42, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 547.3 full barriers recorded, 286.0 left out (nothing recorded since the one before), 29.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.8; 0x12: 91.1; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.2 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 161.4, an op 1223.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 11.2 ms a frame (the longest submission 5.6 ms); 5.5 textures made a frame: colour targets 1.3 (7.9 MB), depth 2.4 (7.0 MB), ASTC or BC 1.7 (1.5 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.22 MB unswizzled on the CPU in 0.04 ms [hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2898 dirty pages from their submits' captures, 51512 from earlier ones', 2 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1297 dispatches, 37759 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5976 runs, 633.1 MB compared; 72 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13465 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x7fd3800000 (4 MB, captured, state 2): 2; [hos ModuleSystemWorker1] vk: readbacks of GPU results: 360 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 10 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1253 kept their virtual core (nothing there to wait for), 3099 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: recorder: 5917301 packets run in 10.0 s (busy 1.86 s, spun idle 0.13 s, slept 11712 times, 4139 of them under 200 us), 0 calls made directly (no room); 17669 batches over 572 frames (30.9 a frame, 335 packets a batch), 8455 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 476 times, 86.65 ms (0.8 times, 0.15 ms a frame); every thread 1175 times, 139.25 ms in all, the longest 0.83 ms; vkDestroyImageView 596/98.90 ms (max 0.83), vkWaitForFences 505/35.06 ms (max 0.44), vkQueueBindSparse 28/1.82 ms (max 0.35), vmaFreeMemory 13/1.16 ms (max 0.37), a part's submit 5/0.77 ms (max 0.21), the queue's lock 9/0.62 ms (max 0.22) [hos host] vk: recorder: of its busy 1.86 s, 0.49 s in parts' ends and submits (the longest 1.66 ms), 0.19 s of that blocked on the queue's lock (325 times, the longest 1.62 ms), so 0.30 s in the ends, submits and holds; 1.37 s running the other packets; 10345 packets a frame [hos host] vk: replay reads of guest memory as submitted: 1500000 (933825 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 189 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 102 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 57 textures retired (1766 left, 2976 MB): budget 0, drawn over 25, dead targets 32 [hos host] vk: texture #9520 0x10009304600 384x384 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (2 so far) [hos host] vk: replay reads of guest memory as submitted: 1520000 (947496 pages from the captures) [hos host] vk: program picks: 2195084 kept (718626 by the last picks' textures, a uniform buffer bound since), 25172 new, 979743 with other textures (of 3200000) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 117 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 98, 946 MB with memory of 1248 MB (301 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 196 binds (26.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 29 taken from the spares, 108 made new; 39 kept as spares, 0 destroyed; 10 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2583 dirty pages from their submits' captures, 53540 from earlier ones', 5 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1320 dispatches, 38765 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6344 runs, 638.3 MB compared; 81 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14263 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: 0x1057400000 (4 MB, captured, state 2): 5; [hos host] vk: imports: 109 live (1456 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB) [hos sead::MainThread] nvn: CommandBufferWaitEvent #45000: event 0x100d6c8028 mode 0 value 29455 (now 29436) (lr 0x806674b0) [hos aal::LowPrioWorkerThread] vcores: 151722 takes, 27390 waits (mean 78 us, longest 18.3 ms, 0 over HOS_VCORE_LOG_MS), 13317 kicks, 13500 yields, 0 starved grants, 0 escapes in 5 s [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 1320 ranges, 3589 MB), 256 pages written, 256 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (128 MB), 11 spare buffers, 149 texture pages copied (0 MB) [hos host] vk: replay reads of guest memory as submitted: 1540000 (961071 pages from the captures) [hos BaseProcExecutor3] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 35 made): 21031640 lookups, 148040 missed (0.70 %) [hos host] vk: mirror of pool 0x1041c00000 (4 MB) given back: idle, 6063 submissions since its sync [hos ModuleSystemWorker1] vk: readbacks of GPU results: 358 of 3000 waited for their write-back, 0.08 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1152 kept their virtual core (nothing there to wait for), 2944 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos sead::MainThread] nvn: CommandBufferSignalEvent #30000: event 0x100d6c8010 mode 0 value 30000 (lr 0x80660adc) [hos host] vk: (its written part: 4 of 4 KB readable, 1495 bytes not zero; levels L0) [hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it [hos host] vk: 11 textures retired (1817 left, 3021 MB): budget 0, drawn over 0, dead targets 11 [hos host] vk: replay reads of guest memory as submitted: 1560000 (974468 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 1166032 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 836, 3980.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2646 dirty pages from their submits' captures, 55823 from earlier ones', 84 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1308 dispatches, 40669 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6946 runs, 676.5 MB compared; 79 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14690 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 84; [hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 60000 so far (x1 0x117b427040 x2 0x6001d40 x3 0x117b427000 x4 0xf33080) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: mirror of pool 0x105e800000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 1580000 (987723 pages from the captures) [hos host] vk: pool 0x7f9d35d600+0x2cb8 imported in 2.1 ms as 0x7f9c400000..0x7f9e400000 (130 imports, 340 ms in all) [hos ModuleSystemWorker2] nvn: 450000 recordings made, 72154 live copies (23.3 MB), 4043 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: slow submit: 30.4 ms: draw 29.7 descriptors 20.2 texture lookups 18.2 uploads 15.6 programs 8.1 [hos host] vk: mirror of pool 0x1050c00000 (4 MB) given back: idle, 6030 submissions since its sync [hos host] vk: 37 textures retired (1957 left, 3029 MB): budget 0, drawn over 37, dead targets 0 [hos host] vk: texture 0x10bf605200 768x768x1 x1 levels 10 NVN 0x49 -> vk 139 (2000 textures, 3055 MB) [hos host] vk: replay reads of guest memory as submitted: 1600000 (1000547 pages from the captures) [hos host] vk: 1 textures retired (2014 left, 3059 MB): budget 0, drawn over 0, dead targets 1 [hos ModuleSystemWorker2] vk: readbacks of GPU results: 344 of 3000 waited for their write-back, 0.13 ms each; 0 looked at the stopped queues first, 8 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1133 kept their virtual core (nothing there to wait for), 2707 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: sparse mirrors: 101, 978 MB with memory of 1292 MB (313 MB spared); blocks given memory later 117, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 205 binds (28.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 32 taken from the spares, 111 made new; 42 kept as spares, 0 destroyed; 10 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 814, 3876.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2358 dirty pages from their submits' captures, 52550 from earlier ones', 35 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1191 dispatches, 39168 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6344 runs, 651.8 MB compared; 74 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14862 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x1053800000 (4 MB, captured, state 2): 13; 0x1056800000 (4 MB, captured, state 2): 12; 0x1053400000 (4 MB, captured, state 2): 10; [hos host] vk: imports: 111 live (1496 MB), 19 retired still imported (384 MB), 1 retired given back (20 MB) [hos host] vk: 16 textures retired (2103 left, 3107 MB): budget 0, drawn over 0, dead targets 16 [hos host] vk: a frame: 218 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.73, present 0.20, draw 7.42, descriptors 3.25, texture lookups 0.68, uploads 0.32, pipelines 0.33, programs 1.26, commands 2.06, dispatch 0.94, memory sync 0.44, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 144 (978 MB): 6.28 MB synced in 1063.5 copies, 17.74 MB written back [hos host] vk: 3840 presents, 52.8 fps; a frame: 1839 draws (0 skipped), 164 dispatches, 176 passes, 171 barriers, 6 submits, 3.3 uploads (0.7 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2118 textures (3112 MB) [hos host] vk: host time in GPU resource waits: 5.10 ms a frame (fences 1.99, timelines 3.11, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 569.8 full barriers recorded, 286.1 left out (nothing recorded since the one before), 31.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 42.6; 0x12: 90.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.5 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 161.9, an op 1292.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 11.3 ms a frame (the longest submission 5.3 ms); 6.0 textures made a frame: colour targets 1.6 (8.8 MB), depth 1.1 (4.2 MB), ASTC or BC 3.3 (2.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.33 MB unswizzled on the CPU in 0.06 ms [hos host] vk: replay reads of guest memory as submitted: 1620000 (1012068 pages from the captures) [hos host] vk: 33 textures retired (2099 left, 3107 MB): budget 32, drawn over 1, dead targets 0 [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 1300 ranges, 3628 MB), 265 pages written, 265 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (142 MB), 11 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: program picks: 2327693 kept (751825 by the last picks' textures, a uniform buffer bound since), 25226 new, 1047080 with other textures (of 3400000) [hos host] vk: 11 textures retired (2096 left, 3106 MB): budget 11, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 1640000 (1023369 pages from the captures) [hos host] vk: write-backs left out (600 submits): 814, 3876.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1924 dirty pages from their submits' captures, 52272 from earlier ones', 16 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1272 dispatches, 38694 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6568 runs, 661.0 MB compared; 83 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13902 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x1056c00000 (4 MB, captured, state 2): 9; 0x105ac00000 (4 MB, captured, state 2): 6; 0x1042c00000 (4 MB, captured, state 2): 1; [hos Presentation Thread] vi: 3878 frames presented (54.7 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 36, 15.5-17.8 131, 17.8-22 79, 22-30 21, 30-40 5, over 40 2 [hos host] vk: 11 textures retired (2082 left, 3089 MB): budget 0, drawn over 0, dead targets 11 [hos host] vk: mirror of pool 0x7fa2800000 (8 MB) given back: idle, 6026 submissions since its sync [hos ModuleSystemWorker3] vk: readbacks of GPU results: 328 of 3000 waited for their write-back, 0.11 ms each; 0 looked at the stopped queues first, 20 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1229 kept their virtual core (nothing there to wait for), 3379 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 1660000 (1034773 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 172 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 113 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] nvn: GPU waits in 5 s: 207 sync waits (585 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 1680000 (1046288 pages from the captures) [hos host] vk: sparse mirrors: 103, 974 MB with memory of 1296 MB (321 MB spared); blocks given memory later 118, given back 1060; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 210 binds (28.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 34 taken from the spares, 112 made new; 43 kept as spares, 0 destroyed; 9 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1886 dirty pages from their submits' captures, 52109 from earlier ones', 1042 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1322 dispatches, 38315 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6316 runs, 637.3 MB compared; 93 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13763 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: 0x7f7dc00000 (8 MB, captured, state 2): 1014; 0x1053800000 (4 MB, captured, state 2): 15; 0x105dc00000 (4 MB, captured, state 2): 7; 0x105a800000 (4 MB, captured, state 2): 6; [hos host] vk: imports: 112 live (1500 MB), 3 retired still imported (44 MB), 17 retired given back (360 MB) [hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 67.2 ms for a core: given by NavMeshCAStepDepWorker1 (priority 22), which held it 0.6 ms and was never asked; the cores it may use now: core 1: it; [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 79.9 ms for a core: given by tera::FileResourceMgr (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos host] vk: reaper: 1531 images (3611 MB) destroyed in 10 s, 15 of them where they were handed over (its cap or VRAM pressure); the queue at most 90 (229 MB) [hos host] vk: replay reads of guest memory as submitted: 1700000 (1057771 pages from the captures) [hos host] vk: 17 textures retired (2098 left, 3101 MB): budget 7, drawn over 0, dead targets 10 [hos ModuleSystemWorker1] vk: readbacks of GPU results: 325 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1107 kept their virtual core (nothing there to wait for), 2989 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 34 textures retired (2079 left, 3107 MB): budget 34, drawn over 0, dead targets 0 [hos Presentation Thread] vk: capture: a submit 0.22 ms (lock 0.03, scan 0.05 over 801 ranges, 3436 MB), 342 pages written, 342 copied (1.3 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 24 kept (108 MB), 11 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 118 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 836, 3980.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2865 dirty pages from their submits' captures, 55429 from earlier ones', 8132 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1392 dispatches, 40439 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6657 runs, 673.3 MB compared; 90 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14464 in captures, 1 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: 0x7f9c400000 (32 MB, captured, state 2): 6892; 0x7f7dc00000 (8 MB, captured, state 2): 1236; 0x1051000000 (4 MB, captured, state 2): 2; 0x1053000000 (4 MB, captured, state 2): 1; 0x1042c00000 (4 MB, captured, state 2): 1; [hos host] vk: replay reads of guest memory as submitted: 1720000 (1070124 pages from the captures) [hos host] vk: 29 textures retired (2058 left, 3107 MB): budget 29, drawn over 0, dead targets 0 [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #50000: event 0x100d6c8028 mode 0 value 33255 (now 33236) (lr 0x806674b0) [hos host] vk: a frame: 244 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.47, present 0.23, draw 7.14, descriptors 2.98, texture lookups 0.34, uploads 0.03, pipelines 0.33, programs 1.26, commands 2.03, dispatch 0.99, memory sync 1.02, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 146 (973 MB): 6.17 MB synced in 1059.2 copies, 17.73 MB written back [hos host] vk: 3960 presents, 55.1 fps; a frame: 1814 draws (0 skipped), 162 dispatches, 181 passes, 171 barriers, 6 submits, 0.4 uploads (0.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2049 textures (3075 MB) [hos host] vk: host time in GPU resource waits: 3.02 ms a frame (fences 1.50, timelines 1.52, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 566.0 full barriers recorded, 290.6 left out (nothing recorded since the one before), 29.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 42.2; 0x12: 91.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.2 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 162.1, an op 1323.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 10.7 ms a frame (the longest submission 4.7 ms); 3.3 textures made a frame: colour targets 1.6 (9.4 MB), depth 1.4 (5.0 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.06 MB unswizzled on the CPU in 0.01 ms [hos host] vk: program picks: 2463396 kept (785830 by the last picks' textures, a uniform buffer bound since), 25247 new, 1111356 with other textures (of 3600000) [hos host] vk: replay reads of guest memory as submitted: 1740000 (1081484 pages from the captures) [hos host] vk: texture #10330 0x10009154a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 3980 (drawn at 3919) [hos host] vk: readback pages refreshed from guest memory: 1353050 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 103, 973 MB with memory of 1296 MB (322 MB spared); blocks given memory later 145, given back 1095; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 222 binds (29.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 34 taken from the spares, 112 made new; 43 kept as spares, 0 destroyed; 9 spares now, 364 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 901, 4293.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1050 dirty pages from their submits' captures, 57601 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1367 dispatches, 42249 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6827 runs, 720.7 MB compared; 103 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15344 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 2 retired still imported (24 MB), 18 retired given back (380 MB) [hos sead::MainThread] vk: readbacks of GPU results: 363 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1050 kept their virtual core (nothing there to wait for), 2790 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 1760000 (1092733 pages from the captures) [hos host] vk: mirror of pool 0x7fe2800000 (4 MB) given back: idle, 6028 submissions since its sync [hos host] vk: 10 textures retired (2052 left, 3076 MB): budget 0, drawn over 0, dead targets 10 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.16 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 164 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 1780000 (1103957 pages from the captures) [hos host] vk: mirror of pool 0x7fe2400000 (4 MB) given back: idle, 6055 submissions since its sync [hos host] vk: write-backs left out (600 submits): 903, 4297.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 518 dirty pages from their submits' captures, 57403 from earlier ones', 19 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1401 dispatches, 42064 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6652 runs, 716.1 MB compared; 103 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15551 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x1057400000 (4 MB, captured, state 2): 12; 0x1053400000 (4 MB, captured, state 2): 5; 0x7fe1800000 (4 MB, captured, state 2): 2; [hos host] vk: replay reads of guest memory as submitted: 1800000 (1115448 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 802 ranges, 3436 MB), 257 pages written, 257 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (110 MB), 12 spare buffers, 148 texture pages copied (0 MB) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 347 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1206 kept their virtual core (nothing there to wait for), 3146 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos NavMeshDepWorker2] vcores: 151037 takes, 34875 waits (mean 88 us, longest 79.9 ms, 2 over HOS_VCORE_LOG_MS), 17510 kicks, 17630 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 11 textures retired (2058 left, 3075 MB): budget 0, drawn over 1, dead targets 10 [hos host] vk: replay reads of guest memory as submitted: 1820000 (1127805 pages from the captures) [hos host] vk: 3 textures retired (2062 left, 3094 MB): budget 0, drawn over 3, dead targets 0 [hos sead::MainThread] nvn: CommandBufferSignalEvent #35000: event 0x100d6c8008 mode 0 value 34995 (lr 0x80660f34) [hos host] vk: 10 textures retired (2055 left, 3082 MB): budget 0, drawn over 10, dead targets 0 [hos host] vk: sparse mirrors: 101, 972 MB with memory of 1288 MB (315 MB spared); blocks given memory later 145, given back 1095; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 224 binds (29.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 34 taken from the spares, 112 made new; 45 kept as spares, 0 destroyed; 11 spares now, 372 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 879, 4188.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1461 dirty pages from their submits' captures, 56887 from earlier ones', 120 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1409 dispatches, 41871 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6367 runs, 687.6 MB compared; 93 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14643 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x1055800000 (4 MB, captured, state 2): 23; 0x105e400000 (4 MB, captured, state 2): 21; 0x1054c00000 (4 MB, captured, state 2): 15; 0x1057000000 (4 MB, captured, state 2): 11; 0x1053c00000 (4 MB, captured, state 2): 11; 0x1057c00000 (4 MB, captured, state 2): 10; 0x1057800000 (4 MB, captured, state 2): 10; 0x1054800000 (4 MB, captured, state 2): 6; [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: replay reads of guest memory as submitted: 1840000 (1140358 pages from the captures) [hos host] vk: 13 textures retired (2060 left, 3116 MB): budget 0, drawn over 3, dead targets 10 [hos host] vk: program picks: 2598639 kept (817264 by the last picks' textures, a uniform buffer bound since), 25257 new, 1176103 with other textures (of 3800000) [hos host] vk: texture #9520 0x10009304600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4075 (drawn at 3949) [hos sead::MainThread] nvn: 500000 recordings made, 70835 live copies (21.1 MB), 4274 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: a frame: 265 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.42, present 0.22, draw 6.42, descriptors 2.68, texture lookups 0.43, uploads 0.03, pipelines 0.29, programs 1.11, commands 1.82, dispatch 0.87, memory sync 0.36, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 148 (978 MB): 5.77 MB synced in 1046.6 copies, 17.57 MB written back [hos host] vk: 4080 presents, 56.8 fps; a frame: 1775 draws (0 skipped), 150 dispatches, 177 passes, 170 barriers, 6 submits, 0.6 uploads (0.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2101 textures (3058 MB) [hos host] vk: host time in GPU resource waits: 3.48 ms a frame (fences 1.97, timelines 1.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 548.0 full barriers recorded, 287.8 left out (nothing recorded since the one before), 29.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 41.5; 0x12: 90.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.8 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 160.9, an op 1247.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 10.4 ms a frame (the longest submission 4.1 ms); 3.9 textures made a frame: colour targets 1.9 (11.5 MB), depth 1.5 (7.7 MB), ASTC or BC 0.5 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.08 MB unswizzled on the CPU in 0.01 ms [hos sead::MainThread] vk: readbacks of GPU results: 309 of 3000 waited for their write-back, 0.31 ms each; 0 looked at the stopped queues first, 38 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1064 kept their virtual core (nothing there to wait for), 2776 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 1860000 (1152873 pages from the captures) [hos host] vk: write-backs left out (600 submits): 749, 3563.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4295 dirty pages from their submits' captures, 53565 from earlier ones', 34 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1521 dispatches, 38774 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5877 runs, 619.7 MB compared; 78 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12746 in captures, 1 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x105b400000 (4 MB, captured, state 2): 13; 0x1059c00000 (4 MB, captured, state 2): 11; 0x1054000000 (4 MB, captured, state 2): 10; [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 1880000 (1165097 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 802 ranges, 3436 MB), 276 pages written, 276 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (116 MB), 12 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: replay reads of guest memory as submitted: 1900000 (1177437 pages from the captures) [hos host] vk: 10 textures retired (2130 left, 3120 MB): budget 0, drawn over 0, dead targets 10 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 118 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 103, 978 MB with memory of 1300 MB (321 MB spared); blocks given memory later 145, given back 1095; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 226 binds (29.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 36 taken from the spares, 112 made new; 45 kept as spares, 0 destroyed; 9 spares now, 360 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4640 dirty pages from their submits' captures, 53533 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1517 dispatches, 38831 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6100 runs, 606.6 MB compared; 82 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13138 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 176 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 4 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] vk: readbacks of GPU results: 353 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1227 kept their virtual core (nothing there to wait for), 3125 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 1 textures retired (2149 left, 3136 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: replay reads of guest memory as submitted: 1920000 (1189679 pages from the captures) [hos host] vk: 13 textures retired (2176 left, 3132 MB): budget 0, drawn over 0, dead targets 13 [hos host] vk: program picks: 2729518 kept (853812 by the last picks' textures, a uniform buffer bound since), 25258 new, 1245223 with other textures (of 4000000) [hos Presentation Thread] vi: 4154 frames presented (55.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 26, 15.5-17.8 153, 17.8-22 61, 22-30 34, 30-40 2, over 40 0 [hos sead::MainThread] nvn: CommandBufferWaitEvent #55000: event 0x100d6c8028 mode 0 value 37055 (now 37036) (lr 0x806674b0) [hos host] vk: 144 textures retired (2074 left, 3171 MB): budget 142, drawn over 0, dead targets 2 [hos host] vk: replay reads of guest memory as submitted: 1940000 (1202143 pages from the captures) [hos host] vk: presenter thread: the copy's fence 1.75 ms, the present 0.78 ms a present [hos host] vk: readback pages refreshed from guest memory: 1540072 pages so far (0 held: may be written ahead in their submission) [hos sead::MainThread] nvn: GPU waits in 5 s: 154 sync waits (400 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: write-backs left out (600 submits): 726, 3457.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4758 dirty pages from their submits' captures, 52798 from earlier ones', 24 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1500 dispatches, 38131 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6351 runs, 594.1 MB compared; 64 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13440 in captures, 0 only in the collection, 24 only by the replay itself [hos host] vk: read live, by pool: 0x105a400000 (4 MB, captured, state 2): 24; [hos host] vk: 2 textures retired (2056 left, 3052 MB): budget 0, drawn over 0, dead targets 2 [hos host] vk: replay reads of guest memory as submitted: 1960000 (1214975 pages from the captures) [hos host] vk: 15 textures retired (2075 left, 3040 MB): budget 0, drawn over 0, dead targets 15 [hos ModuleSystemWorker4] vk: readbacks of GPU results: 348 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1008 kept their virtual core (nothing there to wait for), 2832 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 1980000 (1227742 pages from the captures) [hos host] vk: memory: working set 13219 MB (peak 13255), private 20304 MB, staging 48 MB in 3 arenas, 20 made so far (56 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 640 MB, 37 allocations 562 MB; used 2180 of 15579 MB; heap 1 (device): 181 blocks 5207 MB, 18473 allocations 4750 MB; used 7026 of 7823 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 22 blocks 704 MB, 139 images 523 MB; 113 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 74 blocks 2368 MB, 1955 images 2217 MB; 438 free ranges, the largest 14 MB [hos host] vk: render target images: 4634 reused from the pool, 2122 made new, 1981 destroyed from it, 866 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 118 missed after their image was taken by size; it holds 25 (128 MB) [hos host] vk: render targets not reused by size: 619 with the pool empty (0 of them with targets still in the dead queue), 161 with none in it a present yet, 22 with none of a kind they can use, 313 with all smaller (the largest 47% short on average), 355 with all bigger (over 1.5 times), 652 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 65 over its bound and 1916 idle [hos host] vk: VRAM: 7026 MB used of 7823 MB allowed (456 MB of it free in the blocks); textures 3072 MB (budget 3171 MB), 2073 textures [hos host] vk: budget: VRAM 5207 of 7823 MB (7026 MB by the driver, 1819 MB unaccounted): textures 3072, mirrors 978, blocks free 456, other buffers 211, shadows 200, target pool 128, other images 111, staging 48; RAM 2108 MB (the process's private memory 20304 MB): imports 1500, capture spares 434, captures 112, capture table 26, retired imports 20, other buffers 16, others 0 [hos host] vk: textures: decoded ASTC 1808 MB, rendered 884 MB (71 drawn in one frame, kept: 53518 KB), other 331 MB; used in the last 8 presents 2654 MB [hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 629 MB of blocks hashed in 86 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 4 presents ago [hos host] vk: 21 MB: 0x10006fce600 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x10005192200 1600x1600x1 x3 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 25 (8617 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 64x62x1 NVN 0x25: 1 (15 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 128x126x1 NVN 0x1: 1 (15 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 208x105x1 NVN 0x1: 1 (21 KB); 262x104x1 NVN 0x1: 1 (26 KB); 310x33x1 NVN 0x25: 1 (39 KB); 312x42x1 NVN 0x25: 1 (51 KB); 328x33x1 NVN 0x25: 1 (42 KB); 374x97x1 NVN 0x1: 1 (35 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 441 waited for it, 2.52 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: status: 600 frames, the slowest 53.4 ms; VRAM 6.9 of 7.6 GB [hos host] vk: sparse mirrors: 104, 978 MB with memory of 1304 MB (325 MB spared); blocks given memory later 145, given back 1095; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 227 binds (30.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 45 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 748, 3561.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3861 dirty pages from their submits' captures, 55695 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1559 dispatches, 39602 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6549 runs, 611.4 MB compared; 65 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13964 in captures, 0 only in the collection, 27 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5253 queued, 5253 loaded (median 1.38 ms, 99th percentile 16.77 ms, worst 34.27 ms, 13012 ms and 296.2 MB read in all); 7285 programs (1431 of them compute variants without their code, 0 graphics programs no note names), 10490 graphics pipelines (1 not in a cache), 577 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1463 pipelines made at create taken by draws; 2569 create calls with nothing kept for their stages, 0 before the index; units 7822 (5253 ready, 0 loading), 14420 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 11458 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 575 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 276 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 14.42, present 0.26, draw 8.84, descriptors 3.82, texture lookups 0.97, uploads 0.05, pipelines 0.40, programs 1.41, commands 2.55, dispatch 1.05, memory sync 0.44, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 149 (978 MB): 6.75 MB synced in 1150.1 copies, 17.91 MB written back [hos host] vk: 4200 presents, 51.7 fps; a frame: 2366 draws (0 skipped), 179 dispatches, 196 passes, 180 barriers, 6 submits, 0.9 uploads (0.2 MB), 82 ub updates (57 overlays), GPU-bound wait 0.1 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2075 textures (3090 MB) [hos host] vk: host time in GPU resource waits: 4.87 ms a frame (fences 2.01, timelines 2.86, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 598.0 full barriers recorded, 299.1 left out (nothing recorded since the one before), 34.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 42.9; 0x12: 99.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.9 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 170.9, an op 1708.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 12.2 ms a frame (the longest submission 3.6 ms); 7.1 textures made a frame: colour targets 1.6 (9.4 MB), depth 4.7 (13.8 MB), ASTC or BC 0.8 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.15 MB unswizzled on the CPU in 0.02 ms [hos host] vk: the render thread (600 presents, 18.40 ms a frame): its jobs 17.28 ms a frame (submits 3.75, presents 4.87, others 8.67), idle 1.12; the game waited for room at 67 presents, 0.26 ms each (0.03 a frame; 33.09 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 449 captures at a CPU signal; the submits' waits read in 0.050 ms a frame; readbacks: 2389 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 16.63 ms after its hand-over (44.61 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.48 + 0.31 (the window's copy is in the idle), after it 3.11 + 1.32; after a frame's last recording that reads or writes the game's memory come 1.01 recordings (0.00 ms of GPU time), after its last that reads it 1.01 (0.00 ms) [hos host] vk: frame ahead (600 presents): 9.73 parts a frame recorded in the signal's job, 68.5% of them took nothing that came with the signal; of the others, 3.05 a frame took pages of the signal's capture into a mirror (94.7 pages a frame) and 0.06 pages read as guest memory had them (18.2 pages a frame); a frame's first part took nothing in 39 of 449 frames, its first three in 28 [hos host] vk: frame ahead, what came with the signal: 449 signals with a capture, each of 208.0 pages in 33.1 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.01 releases and 0.00 other CPU signals a signal (2 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.64+0.21, 0.25+0.07, 0.40+0.12, 0.36+0.10, 0.16+0.19, 0.07+0.30, 0.11+0.20, 0.05+0.06, 0.05+0.04, 0.01+0.02, 0.00+0.01, 0.00+0.00 [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: render targets drawn over from another start, let go: 7457 so far (35065 MB), 6658 of them (31380 MB) between the passes [hos host] vk: 30 textures retired (2068 left, 3007 MB): budget 0, drawn over 2, dead targets 28 [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.03 over 815 ranges, 2842 MB), 281 pages written, 281 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (92 MB), 11 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: replay reads of guest memory as submitted: 2000000 (1240256 pages from the captures) [hos host] vk: 1 textures retired (2087 left, 3073 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: replay reads of guest memory as submitted: 2020000 (1252712 pages from the captures) [hos host] vk: write-backs left out (600 submits): 769, 3664.9 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4461 dirty pages from their submits' captures, 53629 from earlier ones', 10 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1441 dispatches, 39443 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6734 runs, 629.2 MB compared; 81 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14446 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x105a000000 (4 MB, captured, state 2): 10; [hos host] vk: program picks: 2862222 kept (892622 by the last picks' textures, a uniform buffer bound since), 25271 new, 1312506 with other textures (of 4200000) [hos host] vk: texture #11558 0x1000b1f5800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4235 (drawn at 4167) [hos host] vk: texture #11517 0x1000b4c5800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4237 (drawn at 4170) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 359 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1173 kept their virtual core (nothing there to wait for), 3179 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 2040000 (1265683 pages from the captures) [hos host] vk: queues ordered by their events: 24000 submissions stopped at a wait, 23998 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 57281, not met 0; captures at a CPU event's wait 3575, 0 of them at a wait met on arrival [hos host] vk: texture #11860 0x1000a355400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4266 (drawn at 4205) [hos host] vk: replay reads of guest memory as submitted: 2060000 (1278917 pages from the captures) [hos host] vk: 11 textures retired (2076 left, 3012 MB): budget 0, drawn over 0, dead targets 11 [hos host] vk: sparse mirrors: 104, 978 MB with memory of 1304 MB (325 MB spared); blocks given memory later 145, given back 1097; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 229 binds (30.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 45 kept as spares, 0 destroyed; 8 spares now, 356 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 749, 3563.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4136 dirty pages from their submits' captures, 49907 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1277 dispatches, 37351 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6537 runs, 611.0 MB compared; 77 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14061 in captures, 0 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: recorder: 6932152 packets run in 10.0 s (busy 1.98 s, spun idle 0.13 s, slept 11938 times, 4113 of them under 200 us), 0 calls made directly (no room); 18295 batches over 542 frames (33.8 a frame, 379 packets a batch), 8915 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 217 times, 46.58 ms (0.4 times, 0.09 ms a frame); every thread 871 times, 95.42 ms in all, the longest 0.94 ms; vkDestroyImageView 287/51.84 ms (max 0.94), vkWaitForFences 473/33.09 ms (max 0.54), vkQueueBindSparse 34/2.47 ms (max 0.36), vmaDestroyBuffer 24/1.93 ms (max 0.41), vmaFreeMemory 10/1.72 ms (max 0.48), vkDestroyBuffer 10/1.39 ms (max 0.58) [hos host] vk: recorder: of its busy 1.98 s, 0.48 s in parts' ends and submits (the longest 1.49 ms), 0.17 s of that blocked on the queue's lock (286 times, the longest 1.43 ms), so 0.31 s in the ends, submits and holds; 1.49 s running the other packets; 12790 packets a frame [hos host] vk: 4 textures retired (2076 left, 3000 MB): budget 0, drawn over 4, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 2080000 (1292531 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 7 pages in 7 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] vk: readbacks of GPU results: 350 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1151 kept their virtual core (nothing there to wait for), 2945 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 80000 so far (x1 0x117b427000 x2 0 x3 0x117b427000 x4 0xf33080) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 815 ranges, 2842 MB), 272 pages written, 272 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (94 MB), 12 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 2100000 (1306304 pages from the captures) [hos aal::LowPrioWorkerThread] vcores: 135512 takes, 26311 waits (mean 97 us, longest 34.1 ms, 0 over HOS_VCORE_LOG_MS), 12745 kicks, 13030 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 118 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 770, 3666.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3645 dirty pages from their submits' captures, 50859 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1318 dispatches, 37876 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6275 runs, 621.1 MB compared; 67 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14053 in captures, 0 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: [hos rmbl::VibrationThread] hid: vibration #11732: handle 00000003, low 0.04 @ 160 Hz, high 0.02 @ 252 Hz [hos rmbl::VibrationThread] hid: vibration #11733: handle 00010003, low 0.04 @ 160 Hz, high 0.02 @ 252 Hz [hos rmbl::VibrationThread] hid: vibration #11734: handle 00000003, low 0.05 @ 182 Hz, high 0.03 @ 182 Hz [hos rmbl::VibrationThread] hid: vibration #11735: handle 00010003, low 0.05 @ 182 Hz, high 0.03 @ 182 Hz [hos rmbl::VibrationThread] hid: vibration #11736: handle 00000003, low 0.06 @ 55 Hz, high 0.02 @ 231 Hz [hos rmbl::VibrationThread] hid: vibration #11737: handle 00010003, low 0.06 @ 55 Hz, high 0.02 @ 231 Hz [hos host] vk: mirror of pool 0x7f7dc00000 (8 MB) given back: idle, 6031 submissions since its sync [hos host] vk: 10 textures retired (2081 left, 3002 MB): budget 0, drawn over 0, dead targets 10 [hos host] vk: mirror of pool 0x7f9c400000 (32 MB) given back: idle, 6063 submissions since its sync [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #40000: event 0x100d6c801c mode 0 value 39997 (lr 0x802911b8) [hos BaseProcExecutor0] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 35 made): 24913091 lookups, 138131 missed (0.55 %) [hos host] vk: replay reads of guest memory as submitted: 2120000 (1320208 pages from the captures) [hos host] vk: a frame: 282 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.37, present 0.23, draw 7.65, descriptors 3.18, texture lookups 0.31, uploads 0.01, pipelines 0.36, programs 1.28, commands 2.23, dispatch 1.00, memory sync 0.41, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 149 (952 MB): 6.23 MB synced in 1098.0 copies, 17.87 MB written back [hos host] vk: 4320 presents, 54.9 fps; a frame: 2088 draws (0 skipped), 173 dispatches, 180 passes, 172 barriers, 6 submits, 0.2 uploads (0.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 2081 textures (3002 MB) [hos host] vk: host time in GPU resource waits: 4.17 ms a frame (fences 2.17, timelines 1.99, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 567.7 full barriers recorded, 288.6 left out (nothing recorded since the one before), 29.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 42.1; 0x12: 92.1; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 162.8, an op 1491.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 12.3 ms a frame (the longest submission 5.8 ms); 3.0 textures made a frame: colour targets 1.4 (8.0 MB), depth 1.5 (5.2 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms [hos host] vk: program picks: 2993533 kept (928942 by the last picks' textures, a uniform buffer bound since), 25278 new, 1381188 with other textures (of 4400000) [hos host] vk: replay reads of guest memory as submitted: 2140000 (1333903 pages from the captures) [hos host] vk: texture #11831 0x1000a6b5600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4337 (drawn at 4195) [hos host] vk: readback pages refreshed from guest memory: 1727578 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 102, 952 MB with memory of 1264 MB (311 MB spared); blocks given memory later 145, given back 1097; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 231 binds (30.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 47 kept as spares, 0 destroyed; 10 spares now, 396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 770, 3666.7 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3292 dirty pages from their submits' captures, 51029 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1313 dispatches, 38110 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6067 runs, 618.2 MB compared; 89 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13978 in captures, 0 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 344 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1203 kept their virtual core (nothing there to wait for), 3149 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 2160000 (1347732 pages from the captures) [hos host] vk: 10 textures retired (2118 left, 3074 MB): budget 0, drawn over 0, dead targets 10 [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #60000: event 0x100d6c8028 mode 0 value 40855 (now 40836) (lr 0x806674b0) [hos host] vk: late pages' fix (600 recordings with one): of 1196 captures with late pages 1196 fixed (54.8 pages, 55.0 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 386 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 3 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 164, of an earlier recording 219, not recorded yet 0, of this part 0; pages fixed from stamps 64424, from snapshots 1094; the shadows: 573707 pages synced from captures, 559260 made known by it, 14447 left unknown because the replay or the collection wrote them since [hos host] vk: replay reads of guest memory as submitted: 2180000 (1361410 pages from the captures) [hos host] vk: texture #12104 0x10009034600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4369 (drawn at 4308) [hos host] vk: texture #11080 0x100090c4800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4369 (drawn at 4308) [hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3115 dirty pages from their submits' captures, 50979 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1307 dispatches, 38112 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6017 runs, 632.1 MB compared; 89 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 134 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13744 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 12 textures retired (2115 left, 3071 MB): budget 0, drawn over 0, dead targets 12 [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 815 ranges, 2842 MB), 252 pages written, 252 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (110 MB), 12 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: replay reads of guest memory as submitted: 2200000 (1375073 pages from the captures) [hos host] vk: mirror of pool 0x105c000000 (4 MB) given back: idle, 6050 submissions since its sync [hos host] vk: 48 mirrors given back (413 MB in all); 952 MB of mirrors now, 68 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: 24 textures retired (2105 left, 3018 MB): budget 0, drawn over 0, dead targets 24 [hos host] vk: texture #11935 0x10009b75000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4394 (drawn at 4212) [hos sead::MainThread] vk: readbacks of GPU results: 357 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1155 kept their virtual core (nothing there to wait for), 3197 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #11903 0x10009e45400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4397 (drawn at 4207) [hos host] vk: texture #11900 0x10009ae4e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4398 (drawn at 4214) [hos host] vk: texture #11902 0x10009db5200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4398 (drawn at 4208) [hos host] vk: 20 textures retired (2106 left, 3036 MB): budget 0, drawn over 0, dead targets 20 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 2220000 (1388777 pages from the captures) [hos host] vk: mirror of pool 0x105d400000 (4 MB) given back: idle, 6058 submissions since its sync [hos host] vk: 27 textures retired (2094 left, 3171 MB): budget 26, drawn over 1, dead targets 0 [hos host] vk: sparse mirrors: 100, 952 MB with memory of 1256 MB (303 MB spared); blocks given memory later 145, given back 1097; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 233 binds (30.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 49 kept as spares, 0 destroyed; 12 spares now, 404 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 792, 3771.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3439 dirty pages from their submits' captures, 50858 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1313 dispatches, 37899 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5836 runs, 634.3 MB compared; 62 pages read into the shadows; 200 MB of shadows (114 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14016 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: budget: gave back 48 MB from staging at 93% of the allowance [hos host] vk: 395 textures retired (1719 left, 3186 MB): budget 394, drawn over 1, dead targets 0 [hos host] vk: program picks: 3123531 kept (963626 by the last picks' textures, a uniform buffer bound since), 25278 new, 1451190 with other textures (of 4600000) [hos Presentation Thread] vi: 4421 frames presented (53.2 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 5, 15.5-17.8 89, 17.8-22 155, 22-30 18, 30-40 0, over 40 0 [hos host] vk: replay reads of guest memory as submitted: 2240000 (1402313 pages from the captures) [hos host] vk: 20 textures retired (1696 left, 2837 MB): budget 0, drawn over 0, dead targets 20 [hos host] vk: texture #12029 0x10009304a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4430 (drawn at 4233) [hos sead::MainThread] nvn: GPU waits in 5 s: 233 sync waits (652 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: texture #12028 0x10009274800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4439 (drawn at 4234) [hos host] vk: 2 textures retired (1695 left, 2827 MB): budget 0, drawn over 1, dead targets 1 [hos host] vk: a frame: 283 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.45, present 0.24, draw 7.69, descriptors 3.37, texture lookups 0.97, uploads 0.14, pipelines 0.35, programs 1.23, commands 2.17, dispatch 0.92, memory sync 0.39, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 149 (952 MB): 5.91 MB synced in 1059.6 copies, 17.58 MB written back [hos host] vk: 4440 presents, 53.6 fps; a frame: 1993 draws (0 skipped), 153 dispatches, 174 passes, 170 barriers, 6 submits, 0.4 uploads (0.5 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1697 textures (2841 MB) [hos host] vk: host time in GPU resource waits: 4.43 ms a frame (fences 1.31, timelines 3.12, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 545.2 full barriers recorded, 287.3 left out (nothing recorded since the one before), 26.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 42.5; 0x12: 89.9; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 33.9 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 161.0, an op 1401.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 12.4 ms a frame (the longest submission 4.7 ms); 3.7 textures made a frame: colour targets 1.4 (8.3 MB), depth 2.0 (11.6 MB), ASTC or BC 0.2 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.44 MB unswizzled on the CPU in 0.13 ms [hos host] vk: replay reads of guest memory as submitted: 2260000 (1416192 pages from the captures) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 348 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1034 kept their virtual core (nothing there to wait for), 2550 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: write-backs left out (600 submits): 814, 3876.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3123 dirty pages from their submits' captures, 52914 from earlier ones', 8 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1361 dispatches, 39066 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6043 runs, 649.2 MB compared; 78 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14343 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: 0x1055000000 (4 MB, captured, state 2): 6; 0x1057400000 (4 MB, captured, state 2): 1; 0x1055400000 (4 MB, captured, state 2): 1; [hos rmbl::VibrationThread] hid: 12288 vibration values, strongest 0.06 since the last count [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 2280000 (1429977 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: reaper: 1243 images (3107 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 363 (189 MB) [hos OptimizeChangeBanc] nvn: 550000 recordings made, 70620 live copies (22.4 MB), 5131 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: 1 textures retired (1708 left, 2848 MB): budget 0, drawn over 1, dead targets 0 [hos BaseProcCreateAndDeleteThread] vcores: BaseProcCreateAndDeleteThread (priority 23) waited 69.0 ms for a core: given by NavMeshCAStepDepWorker1 (priority 22), which held it 1.4 ms and was never asked; the cores it may use now: core 1: it; [hos GeneralFormatAsync_] vcores: GeneralFormatAsync_ (priority 24) waited 67.7 ms for a core: given by BaseProcCreateAndDeleteThread (priority 23), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 75.5 ms for a core: given by GeneralFormatAsync_ (priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 54.2 ms for a core: given by tera::FileResourceMgr (priority 25), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.03 over 712 ranges, 2719 MB), 298 pages written, 298 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (224 MB), 8 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: replay reads of guest memory as submitted: 2300000 (1444306 pages from the captures) [hos host] vk: texture #1212 0x10001fe0800 640x360 NVN 0x14: the GPU wrote its storage (submission 31172): uploaded again [hos host] vk: upload over a drawn image: #1212 0x10001fe0800 640x360x1 x1 NVN 0x14 at present 4478 (texture frame 4479), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: texture 0x10001fe0800 640x360x1 x1 levels 1 NVN 0x14 (960 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 960 KB of it, over what the GPU wrote (finalized since) [hos host] vk: 56 textures retired (1732 left, 2877 MB): budget 0, drawn over 0, dead targets 56 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 118 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 100, 952 MB with memory of 1256 MB (303 MB spared); blocks given memory later 149, given back 1097; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 235 binds (30.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 49 kept as spares, 0 destroyed; 12 spares now, 404 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 814, 3876.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2950 dirty pages from their submits' captures, 56089 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1363 dispatches, 40024 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6298 runs, 649.6 MB compared; 84 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14568 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: (its written part: 4 of 4 KB readable, 1495 bytes not zero; levels L0) [hos host] vk: texture 0x107cd77600 128x38x1 x1 levels 1 NVN 0x1 (4 KB of storage) re-uploaded (its memory was written; flags 0x44): what the CPU wrote, 4 KB of it [hos host] vk: replay reads of guest memory as submitted: 2320000 (1458472 pages from the captures) [hos ModuleSystemWorker5] vk: readbacks of GPU results: 356 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1040 kept their virtual core (nothing there to wait for), 3056 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.18 [hos host] vk: 11 textures retired (1763 left, 2904 MB): budget 0, drawn over 0, dead targets 11 [hos host] vk: texture #12707 0x10009034600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4511 (drawn at 4450) [hos host] vk: replay reads of guest memory as submitted: 2340000 (1472083 pages from the captures) [hos OptimizeChangeBanc] nvn: 600000 recordings made, 72275 live copies (23.6 MB), 5174 texture copies; 4 command buffers' segments (4.2 MB) [hos host] vk: 11 textures retired (1769 left, 2911 MB): budget 0, drawn over 0, dead targets 11 [hos host] vk: readback pages refreshed from guest memory: 1914680 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 835, 3979.2 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2522 dirty pages from their submits' captures, 52803 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1229 dispatches, 39279 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6480 runs, 666.9 MB compared; 125 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14659 in captures, 2 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: program picks: 3253887 kept (996799 by the last picks' textures, a uniform buffer bound since), 25279 new, 1520833 with other textures (of 4800000) [hos host] vk: replay reads of guest memory as submitted: 2360000 (1485841 pages from the captures) [hos host] vk: texture #12803 0x1000a7a4600 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 4533 (drawn at 4472) [hos host] vk: texture slots registered again at another size: 200000 (of 276207 registrations; slot 10968 480x270 -> 128x128) [hos sead::MainThread] vk: readbacks of GPU results: 351 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1100 kept their virtual core (nothing there to wait for), 2996 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #12828 0x1001f0f5c00 285x98 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 4548 (drawn at 4462) [hos host] vk: replay reads of guest memory as submitted: 2380000 (1499301 pages from the captures) [hos host] vk: 1 textures retired (1836 left, 3119 MB): budget 0, drawn over 1, dead targets 0 [hos sead::MainThread] vk: capture: a submit 0.24 ms (lock 0.05, scan 0.06 over 1315 ranges, 3034 MB), 316 pages written, 316 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (114 MB), 9 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: texture #12900 0x10008f14200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4558 (drawn at 4497) [hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #65000: event 0x100d6c8028 mode 0 value 44655 (now 44636) (lr 0x806674b0) [hos host] vk: a frame: 228 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.16, present 0.21, draw 6.72, descriptors 2.89, texture lookups 0.45, uploads 0.07, pipelines 0.31, programs 1.11, commands 1.91, dispatch 0.98, memory sync 1.53, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 149 (952 MB): 5.78 MB synced in 1037.3 copies, 17.58 MB written back [hos host] vk: 4560 presents, 52.6 fps; a frame: 1683 draws (0 skipped), 156 dispatches, 167 passes, 167 barriers, 6 submits, 0.8 uploads (0.4 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1785 textures (2995 MB) [hos host] vk: host time in GPU resource waits: 3.72 ms a frame (fences 1.38, timelines 2.33, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 539.4 full barriers recorded, 282.9 left out (nothing recorded since the one before), 24.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 42.7; 0x12: 86.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 35.2 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 158.2, an op 1112.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 11.6 ms a frame (the longest submission 6.7 ms); 4.5 textures made a frame: colour targets 2.1 (9.6 MB), depth 1.9 (6.7 MB), ASTC or BC 0.5 (0.5 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.18 MB unswizzled on the CPU in 0.04 ms [hos host] vk: mirror of pool 0x1053400000 (4 MB) given back: idle, 6009 submissions since its sync [hos host] vk: sparse mirrors: 99, 951 MB with memory of 1252 MB (300 MB spared); blocks given memory later 149, given back 1100; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 237 binds (30.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 37 taken from the spares, 112 made new; 50 kept as spares, 0 destroyed; 13 spares now, 408 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 947, 4506.5 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2330 dirty pages from their submits' captures, 56041 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1188 dispatches, 42471 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7590 runs, 757.4 MB compared; 87 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17085 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: replay reads of guest memory as submitted: 2400000 (1512676 pages from the captures) [hos host] vk: texture #12898 0x10008df3e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4569 (drawn at 4508) [hos tera::FileResourceMgr] vcores: 155254 takes, 41666 waits (mean 97 us, longest 75.5 ms, 4 over HOS_VCORE_LOG_MS), 21743 kicks, 21974 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: texture #12897 0x10008d63c00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4572 (drawn at 4511) [hos sead::MainThread] nvn: CommandBufferSignalEvent #45000: event 0x100d6c8030 mode 0 value 44999 (lr 0x806612a8) [hos host] vk: mirror of pool 0x1053000000 (4 MB) given back: idle, 6008 submissions since its sync [hos host] vk: texture #13015 0x10008e83a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4579 (drawn at 4518) [hos host] vk: 1 textures retired (1777 left, 2939 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: mirror of pool 0x1051000000 (4 MB) given back: idle, 6018 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 2420000 (1526032 pages from the captures) [hos host] vk: mirror of pool 0x1042400000 (4 MB) given back: idle, 6063 submissions since its sync [hos host] vk: 2 textures retired (1777 left, 2941 MB): budget 0, drawn over 2, dead targets 0 [hos ModuleSystemWorker1] vk: readbacks of GPU results: 364 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1321 kept their virtual core (nothing there to wait for), 3287 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.13 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 2440000 (1539596 pages from the captures) [hos host] vk: texture #13114 0x10009783600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4607 (drawn at 4546) [hos host] vk: 2 textures retired (1778 left, 2945 MB): budget 0, drawn over 1, dead targets 1 [hos host] vk: texture #13112 0x100094b3200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4612 (drawn at 4551) [hos host] vk: write-backs left out (600 submits): 1034, 4923.8 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1363 dirty pages from their submits' captures, 57564 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1169 dispatches, 44248 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8090 runs, 824.2 MB compared; 105 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18212 in captures, 0 only in the collection, 47 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 1 textures retired (1777 left, 2945 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 2460000 (1553041 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 2480000 (1566538 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 1315 ranges, 3030 MB), 237 pages written, 237 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (98 MB), 11 spare buffers, 148 texture pages copied (0 MB) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 357 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1069 kept their virtual core (nothing there to wait for), 2771 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: program picks: 3382464 kept (1021580 by the last picks' textures, a uniform buffer bound since), 25280 new, 1592255 with other textures (of 5000000) [hos host] vk: sparse mirrors: 97, 948 MB with memory of 1244 MB (295 MB spared); blocks given memory later 149, given back 1100; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 241 binds (31.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 38 taken from the spares, 112 made new; 53 kept as spares, 0 destroyed; 15 spares now, 416 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1056, 5028.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 199 dirty pages from their submits' captures, 59576 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1224 dispatches, 45057 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7892 runs, 836.2 MB compared; 111 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17894 in captures, 0 only in the collection, 48 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: replay reads of guest memory as submitted: 2500000 (1580051 pages from the captures) [hos host] vk: a frame: 191 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 8.23, present 0.16, draw 5.03, descriptors 2.06, texture lookups 0.17, uploads 0.03, pipelines 0.23, programs 0.97, commands 1.39, dispatch 0.76, memory sync 0.42, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 150 (948 MB): 4.89 MB synced in 940.7 copies, 17.47 MB written back [hos host] vk: 4680 presents, 60.8 fps; a frame: 1328 draws (0 skipped), 139 dispatches, 157 passes, 157 barriers, 6 submits, 0.3 uploads (0.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1791 textures (2957 MB) [hos host] vk: host time in GPU resource waits: 1.73 ms a frame (fences 1.07, timelines 0.66, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 513.1 full barriers recorded, 275.1 left out (nothing recorded since the one before), 20.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.0; 0x12: 78.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 38.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 148.1, an op 888.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 10.5 ms a frame (the longest submission 6.9 ms); 0.6 textures made a frame: colour targets 0.2 (1.1 MB), depth 0.1 (0.5 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.01 ms [hos host] vk: a submission's parts ended: 16201 where the game had ended its render pass, 3799 at 512 draws [hos host] vk: replay reads of guest memory as submitted: 2520000 (1593577 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos ModuleSystemWorker1] vk: readbacks of GPU results: 357 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1085 kept their virtual core (nothing there to wait for), 3011 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos Presentation Thread] vi: 4703 frames presented (56.4 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 29, 15.5-17.8 163, 17.8-22 69, 22-30 21, 30-40 0, over 40 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 8 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1011, 4817.3 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 309 dirty pages from their submits' captures, 57340 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1169 dispatches, 43453 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7346 runs, 783.0 MB compared; 123 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16594 in captures, 0 only in the collection, 45 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 2540000 (1607044 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 156 sync waits (339 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 2560000 (1620597 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.05 over 1312 ranges, 3026 MB), 239 pages written, 239 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (108 MB), 10 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: replay reads of guest memory as submitted: 2580000 (1634219 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 2102340 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 97, 948 MB with memory of 1244 MB (295 MB spared); blocks given memory later 149, given back 1101; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 242 binds (31.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 38 taken from the spares, 112 made new; 53 kept as spares, 0 destroyed; 15 spares now, 416 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 991, 4716.0 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 418 dirty pages from their submits' captures, 56377 from earlier ones', 281 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1172 dispatches, 42732 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7358 runs, 798.6 MB compared; 112 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16651 in captures, 0 only in the collection, 46 only by the replay itself [hos host] vk: read live, by pool: 0x1052c00000 (4 MB, captured, state 2): 281; [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 360 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1103 kept their virtual core (nothing there to wait for), 2993 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #70000: event 0x100d6c8028 mode 0 value 48455 (now 48436) (lr 0x806674b0) [hos host] vk: presenter thread: the copy's fence 1.40 ms, the present 0.77 ms a present [hos host] vk: texture #13267 0x100092d3600 1280x720 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 4766 (drawn at 4705) [hos host] vk: replay reads of guest memory as submitted: 2600000 (1648440 pages from the captures) [hos host] vk: texture #12787 0x10008b53a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4771 (drawn at 4478) [hos host] vk: texture #13446 0x100091e2e00 384x384 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (3 so far) [hos host] vk: 5 textures retired (1833 left, 3170 MB): budget 5, drawn over 0, dead targets 0 [hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 100000 so far (x1 0x117d97b040 x2 0x2 x3 0x117d97b040 x4 0x8b040) [hos host] vk: replay reads of guest memory as submitted: 2620000 (1662674 pages from the captures) [hos host] vk: 148 textures retired (1695 left, 3166 MB): budget 148, drawn over 0, dead targets 0 [hos host] vk: memory: working set 13261 MB (peak 13264), private 20408 MB, staging 32 MB in 2 arenas, 22 made so far (56 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 640 MB, 36 allocations 564 MB; used 2180 of 15579 MB; heap 1 (device): 189 blocks 5463 MB, 17594 allocations 4826 MB; used 7282 of 7823 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 30 blocks 960 MB, 155 images 819 MB; 123 free ranges, the largest 10 MB [hos host] vk: image blocks (textures, compacted): 74 blocks 2368 MB, 1553 images 2044 MB; 434 free ranges, the largest 32 MB [hos host] vk: render target images: 5323 reused from the pool, 2734 made new, 2577 destroyed from it, 970 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 140 missed after their image was taken by size; it holds 11 (63 MB) [hos host] vk: render targets not reused by size: 832 with the pool empty (0 of them with targets still in the dead queue), 194 with none in it a present yet, 24 with none of a kind they can use, 410 with all smaller (the largest 49% short on average), 476 with all bigger (over 1.5 times), 798 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 76 over its bound and 2501 idle [hos host] vk: VRAM: 7282 MB used of 7823 MB allowed (636 MB of it free in the blocks); textures 3259 MB (budget 3171 MB), 1701 textures [hos host] vk: budget: VRAM 5463 of 7823 MB (7282 MB by the driver, 1819 MB unaccounted): textures 3259, mirrors 948, blocks free 636, other buffers 211, shadows 200, other images 111, target pool 63, staging 32; RAM 2110 MB (the process's private memory 20408 MB): imports 1500, capture spares 462, captures 102, capture table 26, retired imports 20, others 0 [hos host] vk: textures: decoded ASTC 1638 MB, rendered 1205 MB (81 drawn in one frame, kept: 68153 KB), other 309 MB; used in the last 8 presents 2306 MB [hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 0 listed; 650 MB of blocks hashed in 89 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: 31 MB: 0x10e5be9600 1775x1175x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 23 MB: 0x10d639e800 768x1536x1 x4 levels 11 NVN 0x7b ASTC, last used 4 presents ago [hos host] vk: 21 MB: 0x10006fce600 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 14 MB: 0x10004dd1a00 1600x1600x1 x3 levels 1 NVN 0x32 rendered, last used 322 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 23 (8517 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 73x33x1 NVN 0x25: 1 (9 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 137x97x1 NVN 0x1: 1 (12 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 208x105x1 NVN 0x1: 1 (21 KB); 262x104x1 NVN 0x1: 1 (26 KB); 285x98x1 NVN 0x1: 1 (27 KB); 328x33x1 NVN 0x25: 1 (42 KB); 374x97x1 NVN 0x1: 1 (35 KB); 452x106x1 NVN 0x1: 1 (46 KB); [hos host] vk: texture #12771 0x10004dd1a00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 322 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #12820 0x10005280e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 322 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #12793 0x10004741200 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 322 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #12791 0x10004831400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 322 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 87 textures retired (1614 left, 3171 MB): budget 86, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 30 MB [hos Presentation Thread] vk: presents after the CPU's signal (600): 405 waited for it, 2.42 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos host] vk: status: 600 frames, the slowest 31.4 ms; VRAM 7.1 of 7.7 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 4 new graphics pipelines, pipelines from the driver's cache 6, from the start's cache 0, compiled 2; 3 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5253 queued, 5253 loaded (median 1.38 ms, 99th percentile 16.77 ms, worst 34.27 ms, 13012 ms and 296.2 MB read in all); 7285 programs (1431 of them compute variants without their code, 0 graphics programs no note names), 10490 graphics pipelines (1 not in a cache), 577 compute pipelines (24 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1472 pipelines made at create taken by draws; 2569 create calls with nothing kept for their stages, 0 before the index; units 7822 (5253 ready, 0 loading), 14420 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 11458 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 575 compute pipelines made by them at a create call, 24 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 150 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 7.99, present 0.19, draw 4.57, descriptors 1.93, texture lookups 0.41, uploads 0.00, pipelines 0.21, programs 0.83, commands 1.26, dispatch 0.76, memory sync 0.42, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 150 (948 MB): 4.84 MB synced in 933.9 copies, 17.30 MB written back [hos host] vk: 4800 presents, 59.9 fps; a frame: 1169 draws (0 skipped), 129 dispatches, 163 passes, 158 barriers, 6 submits, 0.1 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1615 textures (3186 MB) [hos host] vk: host time in GPU resource waits: 2.01 ms a frame (fences 1.34, timelines 0.67, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 520.4 full barriers recorded, 280.1 left out (nothing recorded since the one before), 20.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 41.0; 0x12: 79.3; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 38.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 149.3, an op 726.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 10.4 ms a frame (the longest submission 6.0 ms); 1.6 textures made a frame: colour targets 0.7 (4.0 MB), depth 0.8 (4.9 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 17.80 ms a frame): its jobs 16.97 ms a frame (submits 3.91, presents 5.98, others 7.08), idle 0.84; the game waited for room at 106 presents, 0.16 ms each (0.03 a frame; 38.73 jobs queued at a present); 3 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 7.87 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 407 captures at a CPU signal; the submits' waits read in 0.043 ms a frame; readbacks: 2260 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 16.28 ms after its hand-over (28.67 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.65 + 0.33 (the window's copy is in the idle), after it 2.67 + 0.71; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.86 parts a frame recorded in the signal's job, 66.9% of them took nothing that came with the signal; of the others, 2.93 a frame took pages of the signal's capture into a mirror (88.3 pages a frame) and 0.01 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 15 of 406 frames, its first three in 13 [hos host] vk: frame ahead, what came with the signal: 406 signals with a capture, each of 185.9 pages in 35.7 pools; between a frame's last submit and that signal 0.03 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (9 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.39+0.15, 0.26+0.09, 0.31+0.10, 0.37+0.09, 0.11+0.07, 0.08+0.07, 0.08+0.06, 0.04+0.03, 0.03+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00 [hos host] vk: write-backs left out (600 submits): 1077, 5131.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1025 dirty pages from their submits' captures, 57908 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1238 dispatches, 44872 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7286 runs, 842.2 MB compared; 110 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17015 in captures, 0 only in the collection, 48 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: readbacks of GPU results: 343 of 3000 waited for their write-back, 0.09 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1207 kept their virtual core (nothing there to wait for), 3145 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 2640000 (1677050 pages from the captures) [hos host] vk: render targets drawn over from another start, let go: 8813 so far (41733 MB), 7826 of them (37136 MB) between the passes [hos host] vk: 67 textures retired (1553 left, 3171 MB): budget 57, drawn over 10, dead targets 0 [hos host] vk: mirror of pool 0x7fe2800000 (4 MB) given back: idle, 6043 submissions since its sync [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 3509204 kept (1049807 by the last picks' textures, a uniform buffer bound since), 25280 new, 1665515 with other textures (of 5200000) [hos host] vk: mirror of pool 0x7fe2400000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 2660000 (1691473 pages from the captures) [hos ModuleSystemWorker4] nvn: 650000 recordings made, 73843 live copies (21.6 MB), 5456 texture copies; 3 command buffers' segments (4.1 MB) [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #50000: event 0x100d6c803c mode 0 value 49999 (lr 0x80661bfc) [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 1312 ranges, 3008 MB), 226 pages written, 226 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (108 MB), 12 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: recorder: 5746095 packets run in 10.0 s (busy 1.85 s, spun idle 0.13 s, slept 11371 times, 3885 of them under 200 us), 0 calls made directly (no room); 17338 batches over 564 frames (30.7 a frame, 331 packets a batch), 8137 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 120 times, 24.37 ms (0.2 times, 0.04 ms a frame); every thread 806 times, 77.07 ms in all, the longest 0.93 ms; vkWaitForFences 500/35.69 ms (max 0.58), vkDestroyImageView 246/33.67 ms (max 0.71), vmaDestroyBuffer 25/3.80 ms (max 0.93), vmaFreeMemory 13/3.10 ms (max 0.70), the queue's lock 4/0.50 ms (max 0.39), vkQueueBindSparse 16/0.27 ms (max 0.07) [hos host] vk: recorder: of its busy 1.85 s, 0.51 s in parts' ends and submits (the longest 1.35 ms), 0.21 s of that blocked on the queue's lock (360 times, the longest 1.31 ms), so 0.29 s in the ends, submits and holds; 1.35 s running the other packets; 10188 packets a frame [hos host] vk: mirror of pool 0x1042800000 (4 MB) given back: idle, 6047 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 2680000 (1706231 pages from the captures) [hos host] vk: sparse mirrors: 94, 945 MB with memory of 1232 MB (286 MB spared); blocks given memory later 149, given back 1102; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 246 binds (31.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 38 taken from the spares, 112 made new; 56 kept as spares, 0 destroyed; 18 spares now, 428 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1057, 5030.3 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 880 dirty pages from their submits' captures, 57007 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1207 dispatches, 44268 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6939 runs, 821.8 MB compared; 118 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17316 in captures, 1 only in the collection, 49 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 356 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1104 kept their virtual core (nothing there to wait for), 2992 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos rmbl::VibrationThread] vcores: 160569 takes, 25082 waits (mean 70 us, longest 15.0 ms, 0 over HOS_VCORE_LOG_MS), 12588 kicks, 12577 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 2700000 (1721011 pages from the captures) [hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 36 made): 21384318 lookups, 150505 missed (0.70 %) [hos host] vk: mirror of pool 0x1050800000 (4 MB) given back: idle, 6028 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 2720000 (1735760 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: mirror of pool 0x7fa2800000 (8 MB) given back: idle, 6064 submissions since its sync [hos host] vk: mirror of pool 0x7fd9400000 (4 MB) given back: idle, 6048 submissions since its sync [hos host] vk: write-backs left out (600 submits): 1122, 5342.8 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1094 dirty pages from their submits' captures, 58913 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1220 dispatches, 45679 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7138 runs, 869.8 MB compared; 123 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17333 in captures, 0 only in the collection, 51 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: mirror of pool 0x7fbbc00000 (4 MB) given back: idle, 6032 submissions since its sync [hos ModuleSystemWorker2] vk: readbacks of GPU results: 358 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1117 kept their virtual core (nothing there to wait for), 2979 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 2 textures retired (1514 left, 2723 MB): budget 0, drawn over 1, dead targets 1 [hos host] vk: a frame: 112 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 7.45, present 0.22, draw 4.33, descriptors 1.92, texture lookups 0.27, uploads 0.03, pipelines 0.20, programs 0.74, commands 1.18, dispatch 0.79, memory sync 0.43, queue submit 0.00, slot wait 0.19 [hos host] vk: a frame: mirrors 151 (934 MB): 4.64 MB synced in 903.0 copies, 17.07 MB written back [hos host] vk: 4920 presents, 60.1 fps; a frame: 921 draws (0 skipped), 114 dispatches, 161 passes, 161 barriers, 6 submits, 0.1 uploads (0.2 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1513 textures (2722 MB) [hos host] vk: host time in GPU resource waits: 2.79 ms a frame (fences 1.93, timelines 0.87, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 479.3 full barriers recorded, 273.8 left out (nothing recorded since the one before), 23.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 41.0; 0x12: 82.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 38.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 152.2, an op 555.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 10.0 ms a frame (the longest submission 8.7 ms); 2.5 textures made a frame: colour targets 1.6 (9.7 MB), depth 0.8 (5.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.14 MB unswizzled on the CPU in 0.03 ms [hos host] vk: replay reads of guest memory as submitted: 2740000 (1750080 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 1307 ranges, 3008 MB), 222 pages written, 222 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.010; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (124 MB), 11 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 4 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: mirror of pool 0x7fd3800000 (4 MB) given back: idle, 6041 submissions since its sync [hos host] vk: mirror of pool 0x105e400000 (4 MB) given back: idle, 6013 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 2760000 (1763383 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 89, 932 MB with memory of 1208 MB (275 MB spared); blocks given memory later 149, given back 1103; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 252 binds (31.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 39 taken from the spares, 112 made new; 62 kept as spares, 0 destroyed; 23 spares now, 452 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1165, 5550.6 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 458 dirty pages from their submits' captures, 59778 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1285 dispatches, 46743 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7005 runs, 900.7 MB compared; 154 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17772 in captures, 0 only in the collection, 52 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: texture #13642 0x10008cd3a00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4956 (drawn at 4895) [hos sead::MainThread] nvn: CommandBufferWaitEvent #75000: event 0x100d6c8028 mode 0 value 52255 (now 52236) (lr 0x806674b0) [hos ModuleSystemWorker5] vk: readbacks of GPU results: 364 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1056 kept their virtual core (nothing there to wait for), 2784 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #13631 0x10008c43800 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4962 (drawn at 4901) [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34770): uploaded again [hos host] vk: upload over a drawn image: #12876 0x10001fe0800 1280x720x1 x1 NVN 0x5 at present 4967 (texture frame 4968), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it, over what the GPU wrote (finalized since) [hos host] vk: (its written part: 842 of 842 KB readable, 0 bytes not zero; levels L0) [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): what the CPU wrote, 842 KB of it [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34777): uploaded again [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34784): uploaded again [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34791): uploaded again [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it [hos host] vk: texture #13616 0x10008bb3600 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4973 (drawn at 4912) [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34805): uploaded again [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it [hos host] vk: (its written part: 842 of 842 KB readable, 0 bytes not zero; levels L0) [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): what the CPU wrote, 842 KB of it [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34812): uploaded again [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it [hos host] vk: texture #13573 0x10008a93200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4976 (drawn at 4915) [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34826): uploaded again [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 1920 KB of it [hos host] vk: (its written part: 842 of 842 KB readable, 0 bytes not zero; levels L0) [hos host] vk: texture 0x10001fe0800 1280x720x1 x1 levels 1 NVN 0x5 (1920 KB of storage) re-uploaded (its memory was written; flags 0x8): what the CPU wrote, 842 KB of it [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34833): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34847): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34854): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34868): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34875): uploaded again [hos host] vk: texture #13562 0x10008a03000 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4985 (drawn at 4924) [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34889): uploaded again [hos host] vk: texture #13356 0x10008972e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 4986 (drawn at 4925) [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34896): uploaded again [hos host] vk: replay reads of guest memory as submitted: 2780000 (1776037 pages from the captures) [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34910): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34917): uploaded again [hos host] vk: mirror of pool 0x7fd3c00000 (16 MB) given back: idle, 6056 submissions since its sync [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34924): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34931): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34945): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34952): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34966): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34973): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34987): uploaded again [hos sead::MainThread] nvn: EventSignal (CPU) #10000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 34994): uploaded again [hos host] vk: mirror of pool 0x105dc00000 (4 MB) given back: idle, 6011 submissions since its sync [hos host] vk: 64 mirrors given back (450 MB in all); 917 MB of mirrors now, 104 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35008): uploaded again [hos Presentation Thread] vi: 5003 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 29, 15.5-17.8 249, 17.8-22 20, 22-30 1, 30-40 1, over 40 0 [hos host] vk: readback pages refreshed from guest memory: 2289945 pages so far (0 held: may be written ahead in their submission) [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35022): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35029): uploaded again [hos host] vk: write-backs left out (600 submits): 1167, 5554.1 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 132 dirty pages from their submits' captures, 58485 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1236 dispatches, 46047 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6698 runs, 892.7 MB compared; 136 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17502 in captures, 0 only in the collection, 54 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35043): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35050): uploaded again [hos host] vk: 10 textures retired (1542 left, 2779 MB): budget 0, drawn over 0, dead targets 10 [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35064): uploaded again [hos ModuleSystemWorker4] vk: readbacks of GPU results: 354 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1229 kept their virtual core (nothing there to wait for), 3123 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35071): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35085): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35092): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35106): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35113): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35127): uploaded again [hos host] vk: 1 textures retired (1551 left, 2795 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35141): uploaded again [hos host] vk: texture #12876 0x10001fe0800 1280x720 NVN 0x5: the GPU wrote its storage (submission 35148): uploaded again [hos host] vk: texture #869 0x10004bf1c00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5027 (drawn at 4966) [hos host] vk: texture #12878 0x10004a11800 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 5027 (drawn at 4966) [hos host] vk: texture #870 0x10004ce1e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5027 (drawn at 4966) [hos host] vk: replay reads of guest memory as submitted: 2800000 (1788751 pages from the captures) [hos host] vk: mirror of pool 0x7fe1800000 (4 MB) given back: idle, 6057 submissions since its sync [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.05 over 1306 ranges, 3008 MB), 216 pages written, 216 copied (0.8 MB) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a frame: 49 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 5.97, present 0.16, draw 3.20, descriptors 1.42, texture lookups 0.17, uploads 0.13, pipelines 0.16, programs 0.57, commands 0.84, dispatch 0.72, memory sync 0.41, queue submit 0.00, slot wait 0.19 [hos host] vk: a frame: mirrors 151 (917 MB): 4.33 MB synced in 868.1 copies, 16.84 MB written back [hos host] vk: 5040 presents, 60.0 fps; a frame: 636 draws (0 skipped), 101 dispatches, 144 passes, 156 barriers, 6 submits, 0.9 uploads (1.1 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1572 textures (2904 MB) [hos host] vk: host time in GPU resource waits: 2.01 ms a frame (fences 1.86, timelines 0.15, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 449.1 full barriers recorded, 266.4 left out (nothing recorded since the one before), 14.3 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 41.0; 0x12: 77.1; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 38.0 from 24.0 batches at the game's barriers (38.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 146.5, an op 389.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 20.0; 48: 6.0; [hos host] vk: GPU time 9.3 ms a frame (the longest submission 5.9 ms); 1.4 textures made a frame: colour targets 0.7 (3.2 MB), depth 0.5 (2.5 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.96 MB unswizzled on the CPU in 0.11 ms [hos sead::MainThread] nvn: GPU waits in 5 s: 67 sync waits (97 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: queues ordered by their events: 30000 submissions stopped at a wait, 29996 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 77142, not met 0; captures at a CPU event's wait 4012, 0 of them at a wait met on arrival [hos host] vk: sparse mirrors: 86, 917 MB with memory of 1184 MB (266 MB spared); blocks given memory later 149, given back 1103; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 255 binds (32.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 39 taken from the spares, 112 made new; 65 kept as spares, 0 destroyed; 26 spares now, 476 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1166, 5552.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 141 dirty pages from their submits' captures, 57370 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1231 dispatches, 45440 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5967 runs, 881.5 MB compared; 139 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 136 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16882 in captures, 0 only in the collection, 53 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 112 live (1500 MB), 1 retired still imported (20 MB), 19 retired given back (384 MB) [hos host] vk: reaper: 800 images (2518 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 147 (156 MB) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 351 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1080 kept their virtual core (nothing there to wait for), 3016 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 2820000 (1801600 pages from the captures) [hos host] vk: 15 textures retired (1585 left, 3170 MB): budget 11, drawn over 4, dead targets 0 [hos host] vk: program picks: 3625076 kept (1071696 by the last picks' textures, a uniform buffer bound since), 25283 new, 1749640 with other textures (of 5400000) [hos host] vk: 164 textures retired (1441 left, 3170 MB): budget 146, drawn over 8, dead targets 10 [hos host] vk: 55 textures retired (1397 left, 3171 MB): budget 49, drawn over 6, dead targets 0 [hos host] vk: texture #13862 0x10008492200 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 5098 (drawn at 5008) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos sead::MainThread] nvn: CommandBufferSignalEvent #55000: event 0x100d6c8048 mode 0 value 54999 (lr 0x80666598) [hos host] vk: replay reads of guest memory as submitted: 2840000 (1814546 pages from the captures) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 356 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1088 kept their virtual core (nothing there to wait for), 3008 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: write-backs left out (600 submits): 1254, 5971.4 MB; 21 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 183 dirty pages from their submits' captures, 60591 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1296 dispatches, 48486 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6264 runs, 945.8 MB compared; 144 pages read into the shadows; 200 MB of shadows (115 MB of pages known, in 135 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17964 in captures, 1 only in the collection, 57 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.05 over 1299 ranges, 2998 MB), 215 pages written, 215 copied (0.8 MB) [hos host] vk: a capture taken (ms): all 0.007; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (122 MB), 11 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: pool 0x102bc88200+0xf00000 imported in 2.7 ms as 0x102bc00000..0x102cc00000 (133 imports, 345 ms in all) [hos host] vk: pool 0x1012087e00+0xf00000 imported in 2.5 ms as 0x1012000000..0x1014000000 (134 imports, 348 ms in all) [hos host] vk: a 128 MB capture buffer made ahead in 82.5 ms, off the game's thread [hos host] vk: a frame: 23 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 5.46, present 0.15, draw 2.41, descriptors 1.15, texture lookups 0.60, uploads 0.24, pipelines 0.10, programs 0.41, commands 0.59, dispatch 0.61, memory sync 0.62, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 152 (949 MB): 3.97 MB synced in 756.6 copies, 14.81 MB written back [hos host] vk: 5160 presents, 58.3 fps; a frame: 430 draws (0 skipped), 81 dispatches, 120 passes, 135 barriers, 6 submits, 0.8 uploads (1.6 MB), 74 ub updates (52 overlays), GPU-bound wait 0.0 ms (recording slots only); 12319 pipelines (2890 ms), 0 programs (10 failed), 1338 textures (2674 MB) [hos host] vk: host time in GPU resource waits: 1.80 ms a frame (fences 1.62, timelines 0.18, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 380.8 full barriers recorded, 231.9 left out (nothing recorded since the one before), 9.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.8; 0x8: 2.7; 0x9: 0.9; 0x10: 1.0; 0x11: 36.6; 0x12: 67.7; 0x13: 1.8; 0x21: 3.6; 0x28: 14.3; 0x92: 2.7; 0x97: 0.9; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 32.8 from 21.4 batches at the game's barriers (33.9 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 126.7, an op 273.0, the end 6.3, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.8; 5: 17.8; 48: 5.3; [hos host] vk: GPU time 8.4 ms a frame (the longest submission 6.7 ms); 0.9 textures made a frame: colour targets 0.2 (1.4 MB), depth 0.6 (4.6 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.33 MB unswizzled on the CPU in 0.19 ms [hos sead::MainThread] vcores: 161566 takes, 26774 waits (mean 61 us, longest 14.0 ms, 0 over HOS_VCORE_LOG_MS), 13412 kicks, 13306 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: sparse mirrors: 87, 949 MB with memory of 1216 MB (266 MB spared); blocks given memory later 149, given back 1103; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 256 binds (32.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 40 taken from the spares, 112 made new; 65 kept as spares, 0 destroyed; 25 spares now, 444 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 681, 3245.9 MB; 22 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 72 dirty pages from their submits' captures, 32581 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 26270 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 3410 runs, 515.4 MB compared; 572 pages read into the shadows; 232 MB of shadows (116 MB of pages known, in 137 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 9855 in captures, 2 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 113 live (1520 MB), 2 retired still imported (48 MB), 19 retired given back (384 MB) [hos Resource Control] vcores: Resource Control (priority 20) waited 57.2 ms for a core: given by ModuleSystemWorker1 (priority 16), which held it 7.2 ms and was never asked; the cores it may use now: core 1: it; [hos tera::FileResourceMgr] vcores: tera::FileResourceMgr (priority 25) waited 65.1 ms for a core: given by Resource Memory (priority 21), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos bake::ModuleMgr] vcores: bake::ModuleMgr (priority 28) waited 65.3 ms for a core: given by PreActorModelCompose (priority 26), which held it 0.0 ms and was never asked; the cores it may use now: core 1: it; [hos host] vk: 1 textures retired (1338 left, 2673 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.13 [hos host] vk: a 128 MB capture buffer made ahead in 28.3 ms, off the game's thread [hos Presentation Thread] vk: capture: a submit 0.61 ms (lock 0.02, scan 0.09 over 1157 ranges, 2993 MB), 755 pages written, 755 copied (2.9 MB) [hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (330 MB), 8 spare buffers, 372 texture pages copied (1 MB) [hos sead::MainThread] nvn: CommandBufferWaitEvent #80000: event 0x1006e38000 mode 0 value 1 (now 0) (lr 0x802eb444) [hos host] vk: 1 textures retired (1337 left, 2673 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 459 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 5 pages in 5 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 186 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 127 sets (ub 1415, sb 596, cis 10086, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 0, 0.0 MB; 22 bindings followed, 34 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 13 dirty pages from their submits' captures, 400 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 311 dispatches, 413 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 78 runs, 0.0 MB compared; 78 pages read into the shadows; 232 MB of shadows (112 MB of pages known, in 131 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 1053 in captures, 19 only in the collection, 0 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: mirror of pool 0x7fde000000 (4 MB) given back: idle, 6028 submissions since its sync [hos host] vk: 1 textures retired (1336 left, 2673 MB): budget 0, drawn over 1, dead targets 0 [hos ModuleSystemWorker1] vk: readbacks of GPU results: 238 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1365 kept their virtual core (nothing there to wait for), 3243 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 1.82, present 0.15, draw 0.09, descriptors 0.03, texture lookups 0.00, uploads 0.00, pipelines 0.00, programs 0.01, commands 0.04, dispatch 0.00, memory sync 0.97, queue submit 0.00, slot wait 0.27 [hos host] vk: a frame: mirrors 152 (698 MB): 0.02 MB synced in 5.4 copies, 0.00 MB written back [hos host] vk: 5280 presents, 57.4 fps; a frame: 5 draws (0 skipped), 0 dispatches, 2 passes, 1 barriers, 6 submits, 0.0 uploads (0.0 MB), 1 ub updates (1 overlays), GPU-bound wait 0.0 ms (recording slots only); 12784 pipelines (3060 ms), 0 programs (10 failed), 1336 textures (2673 MB) [hos host] vk: host time in GPU resource waits: 5.80 ms a frame (fences 4.96, timelines 0.84, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 33.7 full barriers recorded, 13.9 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x12: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 0.0, an op 1.0, the end 1.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 0.7 ms a frame (the longest submission 1.1 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: pool 0x112a982c00+0x2685000 imported in 3.7 ms as 0x112a800000..0x112d400000 (135 imports, 351 ms in all) [hos host] vk: pool 0x7f9933c100+0x5550 imported in 4.5 ms as 0x7f99000000..0x7f9a000000 (136 imports, 356 ms in all) [hos host] vk: pool 0x7f8b1de000+0x22ea3c imported in 2.3 ms as 0x7f8ac00000..0x7f8b800000 (137 imports, 358 ms in all) [hos host] vk: slow upload 30.4 ms: #14101 0x10fff15400 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB [hos host] vk: pool 0x108c711000+0xd5b6000 imported in 13.1 ms as 0x108c000000..0x109a000000 (138 imports, 371 ms in all) [hos host] vk: pool 0x108c000000..0x109a000000: mirror made with the GPU's contents of 1 retired pool(s) [hos host] vk: pool 0x7f88551b00+0x22f74 imported in 18.7 ms as 0x7f88000000..0x7f88c00000 (139 imports, 390 ms in all) [hos host] vk: pool 0x7fa16b0600+0x8420 imported in 3.4 ms as 0x7fa1400000..0x7fa1800000 (140 imports, 393 ms in all) [hos host] vk: pool 0x7f9bc33800+0x30510 imported in 3.2 ms as 0x7f9ac00000..0x7f9c000000 (141 imports, 396 ms in all) [hos host] vk: pool 0x7f87f0d300+0x19ea0 imported in 10.0 ms as 0x7f86c00000..0x7f88000000 (142 imports, 406 ms in all) [hos host] vk: pool 0x7f83f34000+0xca00 imported in 7.4 ms as 0x7f83c00000..0x7f84000000 (143 imports, 414 ms in all) [hos host] vk: slow submit: 136.2 ms: draw 220.7 descriptors 50.4 texture lookups 49.2 uploads 41.9 programs 3.9 commands 165.8 [hos host] vk: pool 0x117efc9ac0+0x540 imported in 2.3 ms as 0x117ec00000..0x117f000000 (146 imports, 419 ms in all) [hos host] vk: pool 0x1181000000..0x1181800000 took in 2 pools [hos host] vk: pool 0x1181000000..0x1181800000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: pool 0x11807c7000+0x595000 imported in 4.0 ms as 0x1180400000..0x1181000000 (150 imports, 426 ms in all) [hos host] vk: pool 0x117e8ea000+0x6d3000 imported in 9.3 ms as 0x1178400000..0x117f000000 (151 imports, 435 ms in all) [hos host] vk: pool 0x1178400000..0x117f000000 took in 2 pools [hos host] vk: pool 0x1178400000..0x117f000000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: pool 0x118163c000+0x215a00 imported in 5.2 ms as 0x1181000000..0x1181c00000 (152 imports, 440 ms in all) [hos host] vk: pool 0x1181000000..0x1181c00000 took in 2 pools [hos host] vk: pool 0x1181000000..0x1181c00000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: slow submit: 121.6 ms: dispatch 50.1 [hos host] vk: pool 0x1180d5d000+0x430000 imported in 8.8 ms as 0x1180400000..0x1181c00000 (153 imports, 449 ms in all) [hos host] vk: pool 0x1180400000..0x1181c00000 took in 2 pools [hos host] vk: pool 0x1180400000..0x1181c00000: mirror made with the GPU's contents of 2 retired pool(s) [hos host] vk: pool 0x117efcb000+0x17fb000 imported in 12.7 ms as 0x1178400000..0x1181c00000 (154 imports, 462 ms in all) [hos host] vk: pool 0x1178400000..0x1181c00000 took in 2 pools [hos host] vk: pool 0x7f96677700+0x84 imported in 23.3 ms as 0x7f95800000..0x7f96800000 (155 imports, 485 ms in all) [hos host] vk: pool 0x7f97ad9100+0xb970 imported in 17.8 ms as 0x7f97800000..0x7f98400000 (156 imports, 503 ms in all) [hos host] vk: pool 0x7f8b934900+0x350dc imported in 7.1 ms as 0x7f8b800000..0x7f8d000000 (157 imports, 510 ms in all) [hos host] vk: pool 0x7fffdd5000+0x8fe00 imported in 11.7 ms as 0x7fff800000..0x8000000000 (158 imports, 522 ms in all) [hos host] vk: pool 0x7f9a67e200+0xe0400 imported in 3.5 ms as 0x7f9a400000..0x7f9a800000 (159 imports, 525 ms in all) [hos host] vk: pool 0x7f9ee14000+0xedd00 imported in 6.2 ms as 0x7f9ec00000..0x7f9f000000 (161 imports, 532 ms in all) [hos host] vk: pool 0x7f98a95e00+0x14a0 imported in 2.0 ms as 0x7f98800000..0x7f99000000 (162 imports, 534 ms in all) [hos host] vk: pool 0x7f96982800+0xc844 imported in 5.1 ms as 0x7f96800000..0x7f96c00000 (163 imports, 539 ms in all) [hos host] vk: pool 0x7f96f8d500+0xc0 imported in 2.0 ms as 0x7f96c00000..0x7f97400000 (165 imports, 543 ms in all) [hos host] vk: pool 0x7f95038000+0xc79c imported in 3.6 ms as 0x7f95000000..0x7f95400000 (166 imports, 547 ms in all) [hos host] vk: pool 0x7f94c3ea00+0x1b6a8 imported in 14.7 ms as 0x7f94000000..0x7f95000000 (167 imports, 561 ms in all) [hos host] vk: pool 0x7f86079c00+0x1b404 imported in 4.7 ms as 0x7f86000000..0x7f86400000 (168 imports, 566 ms in all) [hos host] vk: pool 0x7f91aad400+0x23040 imported in 4.2 ms as 0x7f91000000..0x7f93000000 (169 imports, 570 ms in all) [hos host] vk: pool 0x7f8577ef00+0x42a54 imported in 7.8 ms as 0x7f85400000..0x7f85800000 (170 imports, 578 ms in all) [hos host] vk: pool 0x7f896f9a00+0xb340 imported in 2.2 ms as 0x7f89000000..0x7f8a000000 (172 imports, 582 ms in all) [hos host] vk: pool 0x7fa92b5600+0x1490 imported in 7.2 ms as 0x7fa9000000..0x7fa9400000 (173 imports, 589 ms in all) [hos host] vk: pool 0x7f86a2a100+0x86ef4 imported in 2.5 ms as 0x7f86400000..0x7f86c00000 (174 imports, 592 ms in all) [hos host] vk: pool 0x7f9ff8a500+0x9d68 imported in 5.0 ms as 0x7f9fc00000..0x7fa0000000 (175 imports, 597 ms in all) [hos host] vk: slow submit: 417.5 ms: draw 62.0 texture lookups 1.0 commands 60.6 [hos host] vk: pool 0x7f8f940700+0x9354 imported in 3.1 ms as 0x7f8f800000..0x7f8fc00000 (180 imports, 604 ms in all) [hos host] vk: pool 0x7f9f800000 (4 MB): mirror made at its first binding (201 so) [hos host] vk: texture #13584 0x10008b23400 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 4915) [hos host] vk: texture #13858 0x10004dd1a00 1600x1600 x3 NVN 0x32 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 5147) [hos host] vk: upload over a drawn image: #13858 0x10004dd1a00 1600x1600x1 x3 NVN 0x32 at present 5290 (texture frame 5291), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: texture 0x10004dd1a00 1600x1600x1 x3 levels 1 NVN 0x32 (15600 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15600 KB of it, over what the GPU wrote (finalized since) [hos host] vk: texture #13865 0x10004831400 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 5147) [hos host] vk: texture #13866 0x10004921600 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 5147) [hos host] vk: texture #13893 0x10005280e00 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 5291 (drawn at 5146) [hos host] vk: upload over a drawn image: #13893 0x10005280e00 2560x1440x1 x1 NVN 0x25 at present 5290 (texture frame 5291), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: texture 0x10005280e00 2560x1440x1 x1 levels 1 NVN 0x25 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since) [hos host] vk: texture 0x10f69cfc00 512x512x1 x1 levels 10 NVN 0x49 -> vk 139 (1500 textures, 3140 MB) [hos host] vk: the driver's allowance is the count itself (7842 MB used of 7842): past it, taken as 7058 MB [hos host] vk: slow submit: 104.5 ms: draw 70.1 descriptors 54.1 texture lookups 53.9 uploads 39.1 pipelines 3.5 programs 12.1 [hos host] vk: a frame's uploads took 82 ms: ASTC on the GPU 20 (15.2 MB, 38 ms); unswizzled on the GPU 2 (53.5 MB, 4 ms); on the CPU 73 (163.0 MB, 40 ms); [1024x1024 x1 NVN 0x7b ASTC on the GPU 0.9 MB 30 ms] [4096x4096 x1 NVN 0x4b on the CPU 16.0 MB 6 ms] [2048x2048 x3 NVN 0x50 on the CPU 12.0 MB 5 ms] [hos sead::MainThread] nvn: GPU waits in 5 s: 11 sync waits (904 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos Presentation Thread] vi: 5292 frames presented (50.1 fps over the last 5.8 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 37, 15.5-17.8 216, 17.8-22 22, 22-30 4, 30-40 6, over 40 4 [hos host] vk: slow submit: 25.3 ms: draw 7.9 descriptors 5.7 texture lookups 4.5 uploads 4.0 programs 1.1 [hos host] vk: slow upload 24.5 ms: #14230 0x10f88e4c00 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB [hos host] vk: slow submit: 28.8 ms: draw 28.4 descriptors 27.3 texture lookups 26.1 uploads 26.0 [hos host] vk: the driver's allowance is the count itself (7797 MB used of 7797): past it, taken as 7017 MB [hos host] vk: replay reads of guest memory as submitted: 2860000 (1827086 pages from the captures) [hos host] vk: pool 0x7faf5d5200+0x38fc0 imported in 2.4 ms as 0x7faf400000..0x7faf800000 (193 imports, 622 ms in all) [hos host] vk: slow upload 16.4 ms: #14471 0x10fa125800 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB [hos host] vk: slow upload 38.1 ms: #14677 0x11177bc400 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB [hos host] vk: texture 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e -> vk 137 (ASTC decoded) (1941 textures, 4061 MB) [hos host] [debug] vk: a staging arena of 57 MB took 55.8 ms to make [hos host] vk: slow upload 195.5 ms: #14681 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC (BC3), 57825 of 57825 KB [hos host] vk: texture 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 -> vk 137 (ASTC decoded) (1942 textures, 4279 MB) [hos host] [debug] vk: a staging arena of 57 MB took 60.7 ms to make [hos host] vk: slow upload 217.5 ms: #14682 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC (BC3), 57825 of 57825 KB [hos host] vk: slow submit: 940.6 ms: draw 972.5 descriptors 869.6 texture lookups 852.6 uploads 512.7 programs 44.7 commands 56.7 queue submit 1.0 [hos host] vk: a frame's uploads took 514 ms: ASTC on the GPU 41 (166.9 MB, 509 ms); on the CPU 152 (25.2 MB, 5 ms); [1365x1365 x90 NVN 0x80 ASTC on the GPU 56.5 MB 218 ms] [1365x1365 x90 NVN 0x8e ASTC on the GPU 56.5 MB 195 ms] [1536x1536 x1 NVN 0x7b ASTC on the GPU 2.0 MB 38 ms] [hos host] vk: 8 of 8 small textures the GPU drew written into their memory in 2.1 ms, 1 frames after their copies were recorded, the first #14061 0x1168866000 260x260 levels 9 NVN 0xd (9 so far, 2204 KB, 3 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: slow submit: 16.2 ms: draw 14.7 descriptors 14.5 texture lookups 14.4 uploads 14.4 [hos host] vk: the driver's allowance is the count itself (8424 MB used of 8424): past it, taken as 7581 MB [hos host] vk: ASTC texture #14738 0x1132250e00 1536x1536 x1 levels 11: the image of #14200 0x10e174b600 (the same blocks, finalized by the game) copied, not decoded again (1 so far, 11 MB) [hos host] vk: slow submit: 25.2 ms: draw 20.5 descriptors 13.9 programs 2.3 commands 3.1 [hos host] vk: 8 of 8 small textures the GPU drew written into their memory in 1.9 ms, 1 frames after their copies were recorded, the first #14252 0x1169e3e000 260x260 levels 9 NVN 0xd (17 so far, 3880 KB, 5 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: the driver's allowance is the count itself (8379 MB used of 8379): past it, taken as 7541 MB [hos host] vk: mirror of pool 0x108c000000 (48 MB) given back: retired, 69 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x1181400000 (4 MB) given back: retired, 68 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x1181000000 (4 MB) given back: retired, 68 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x117ec00000 (4 MB) given back: retired, 68 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x1178400000 (104 MB) given back: retired, 68 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x1181800000 (4 MB) given back: retired, 68 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x1181000000 (8 MB) given back: retired, 68 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x1181000000 (12 MB) given back: retired, 67 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x1180400000 (12 MB) given back: retired, 68 submissions since its sync, GPU-only data [hos host] vk: mirror of pool 0x1180400000 (24 MB) given back: retired, 67 submissions since its sync, GPU-only data, GPU-written [hos host] vk: mirror of pool 0x1178400000 (108 MB) given back: retired, 67 submissions since its sync, GPU-only data, GPU-written [hos host] vk: replay reads of guest memory as submitted: 2880000 (1845318 pages from the captures) [hos host] vk: slow submit: 37.4 ms: draw 29.7 descriptors 24.2 texture lookups 12.6 uploads 12.4 programs 2.0 commands 2.6 [hos host] vk: 7 of 7 small textures the GPU drew written into their memory in 1.5 ms, 1 frames after their copies were recorded, the first #14268 0x116b416000 260x260 levels 9 NVN 0xd (24 so far, 5346 KB, 6 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: slow upload 31.7 ms: #14361 0x10f9e0b800 1024x1024x1 x1 levels 11 NVN 0x8d ASTC (BC3), 475 of 475 KB [hos host] vk: slow submit: 55.8 ms: draw 50.9 descriptors 44.7 texture lookups 33.4 uploads 33.2 programs 2.5 commands 2.7 [hos host] vk: the driver's allowance is the count itself (8379 MB used of 8379): past it, taken as 7541 MB [hos host] vk: (its written part: 2 of 2 KB readable, 320 bytes not zero; levels L0) [hos host] vk: texture 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 2 KB of it [hos host] vk: (its written part: 1 of 1 KB readable, 864 bytes not zero; levels L0) [hos host] vk: texture 0x10d93e1600 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 1 KB of it [hos host] vk: replay reads of guest memory as submitted: 2900000 (1863408 pages from the captures) [hos host] vk: slow submit: 39.0 ms: draw 35.0 descriptors 28.6 texture lookups 15.8 uploads 15.6 programs 2.6 commands 2.8 [hos host] vk: texture 0x10001e00400 640x360x1 x1 levels 1 NVN 0x34 -> vk 126 (2000 textures, 4330 MB) [hos host] vk: the driver's allowance is the count itself (8424 MB used of 8424): past it, taken as 7581 MB [hos host] vk: (its written part: 2 of 2 KB readable, 320 bytes not zero; levels L0) [hos host] vk: texture 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 2 KB of it [hos host] vk: (its written part: 1 of 1 KB readable, 864 bytes not zero; levels L0) [hos host] vk: texture 0x10d93e1600 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 1 KB of it [hos host] vk: ASTC texture #14785 0x10da19b000 1536x1536 x1 levels 11: the image of #14738 0x1132250e00 (the same blocks, finalized by the game) copied, not decoded again (2 so far, 23 MB) [hos host] vk: texture #1212 0x10001fe0800 640x360 x1 NVN 0x14 the GPU drew, finalized by the game, sampled again at frame 5298 (drawn at 4966) [hos host] vk: upload over a drawn image: #1212 0x10001fe0800 640x360x1 x1 NVN 0x14 at present 5297 (texture frame 5298), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: slow submit: 46.9 ms: draw 49.0 descriptors 17.2 texture lookups 17.6 uploads 11.9 programs 30.1 commands 1.0 [hos host] vk: 191 new ASTC textures left for a later frame (this frame's uploads took their budget; 10072 so far) [hos host] vk: the driver's allowance is the count itself (8424 MB used of 8424): past it, taken as 7581 MB [hos host] vk: the driver's allowance is the count itself (8424 MB used of 8424): past it, taken as 7581 MB [hos host] vk: slow submit: 21.5 ms: [hos host] vk: slow upload 71.8 ms: #14697 0x10f5406c00 768x768x1 x1 levels 10 NVN 0x7b ASTC, 528 of 528 KB [hos host] vk: slow submit: 81.3 ms: draw 78.2 descriptors 75.5 texture lookups 73.5 uploads 73.3 programs 1.4 [hos host] vk: a frame's uploads took 73 ms: ASTC on the GPU 13 (5.3 MB, 73 ms); [768x768 x1 NVN 0x7b ASTC on the GPU 0.5 MB 72 ms] [hos host] vk: budget: gave back 192 MB from staging at 111% of the allowance [hos host] vk: budget: gave back 213 MB from target pool at 111% of the allowance [hos host] vk: write-back of #550 0x1137207e00: 0 KB the CPU wrote since the GPU kept [hos host] vk: write-back of #549 0x1137207d80: 0 KB the CPU wrote since the GPU kept [hos host] vk: textures the GPU drew, written back into their memory before eviction: 1 so far [hos host] vk: texture #14057 0x10004380c00 640x360x1 x1 NVN 0x14 the GPU drew goes with no copy of its texels (finalized, unused for 153 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 1017 textures retired (995 left, 2865 MB): budget 1013, drawn over 4, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: slow present: 15.2 ms: [hos BaseProcExecutor3] map: HyruleField_Child (child) [hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.02, writes collected 0.16 [hos host] vk: texture 0x1100866400 455x455x1 x1 levels 9 NVN 0x49 -> vk 139 (1000 textures, 2872 MB) [hos host] vk: slow submit: 47.2 ms: draw 45.3 descriptors 37.5 texture lookups 34.5 uploads 20.4 programs 6.0 commands 1.3 [hos host] vk: pool 0x1010917600+0xf00000 imported in 3.0 ms as 0x1010800000..0x1011c00000 (197 imports, 627 ms in all) [hos host] vk: pool 0x7f93912f00+0x138ea0 imported in 2.4 ms as 0x7f93800000..0x7f94000000 (201 imports, 633 ms in all) [hos ModuleSystemWorker3] nvn: 700000 recordings made, 66490 live copies (21.5 MB), 5678 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: slow submit: 15.9 ms: draw 4.9 descriptors 3.8 texture lookups 3.7 uploads 3.3 [hos host] vk: pool 0x1167269000+0x225d58 imported in 2.1 ms as 0x1167000000..0x1167800000 (204 imports, 638 ms in all) [hos host] vk: recorder: 1883653 packets run in 10.0 s (busy 0.81 s, spun idle 0.10 s, slept 9189 times, 3478 of them under 200 us), 0 calls made directly (no room); 10302 batches over 459 frames (22.4 a frame, 183 packets a batch), 4709 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 199 times, 26.98 ms (0.4 times, 0.06 ms a frame); every thread 711 times, 64.68 ms in all, the longest 14.19 ms; vkWaitForFences 403/28.29 ms (max 0.44), the queue's lock 4/14.27 ms (max 14.19), vkDestroyImageView 70/7.36 ms (max 0.40), vkQueueBindSparse 112/5.24 ms (max 0.49), vmaDestroyBuffer 48/3.92 ms (max 0.97), vkEndCommandBuffer 46/3.81 ms (max 0.88) [hos host] vk: recorder: of its busy 0.81 s, 0.23 s in parts' ends and submits (the longest 15.81 ms), 0.03 s of that blocked on the queue's lock (55 times, the longest 1.22 ms), so 0.20 s in the ends, submits and holds; 0.57 s running the other packets; 4104 packets a frame [hos host] vk: slow upload 26.1 ms: #15090 0x10d39ede00 512x512x1 x1 levels 10 NVN 0x7b ASTC, 236 of 236 KB [hos host] vk: slow submit: 57.9 ms: draw 46.7 descriptors 38.9 texture lookups 36.2 uploads 34.6 programs 5.3 commands 1.6 [hos host] vk: upload over a drawn image: #1212 0x10001fe0800 640x360x1 x1 NVN 0x14 at present 5301 (texture frame 5302), finalized, 842 KB dirty in 1 ranges [hos host] vk: slow submit: 16.8 ms: draw 14.3 descriptors 7.4 texture lookups 4.8 uploads 3.6 programs 5.0 commands 1.4 dispatch 1.5 [hos host] vk: slow submit: 25.8 ms: draw 26.6 descriptors 18.5 texture lookups 15.4 uploads 14.9 programs 5.9 commands 1.6 [hos host] vk: mirror of pool 0x1010800000 (8 MB) given back: retired, 99 submissions since its sync, GPU-written [hos host] vk: write-back diff: VRAM 8493 of 7643 MB: the shadows given back (120 MB) [hos host] vk: replay reads of guest memory as submitted: 2920000 (1877798 pages from the captures) [hos host] vk: slow upload 17.0 ms: #14628 0x11053b6000 341x341x1 x1 levels 9 NVN 0x7d ASTC, 83 of 83 KB [hos host] vk: (its written part: 0 of 0 KB readable, 192 bytes not zero; levels L0) [hos host] vk: texture 0x10debdc600 1240x1240x1 x3 levels 1 NVN 0x4d (4563 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 0 KB of it [hos sead::MainThread] vk: capture: a submit 0.36 ms (lock 0.03, scan 0.09 over 712 ranges, 4221 MB), 871 pages written, 587 copied (2.3 MB) [hos host] vk: (its written part: 0 of 0 KB readable, 288 bytes not zero; levels L0) [hos host] vk: texture 0x10dff20200 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 0 KB of it [hos host] vk: 35 textures retired (1390 left, 3132 MB): budget 35, drawn over 0, dead targets 0 [hos host] vk: a capture taken (ms): all 0.043; GPU wait for late pages 0.009 (3 of 600 takes), pools lock 0.000, retires and stamps 0.025, bytes copied 0.000; 25 kept (254 MB), 7 spare buffers, 7528 texture pages copied (29 MB) [hos host] vk: readback pages refreshed from guest memory: 2438387 pages so far (0 held: may be written ahead in their submission) [hos OptimizeChangeBanc] nvn: 750000 recordings made, 70160 live copies (26.0 MB), 5697 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: slow upload 17.6 ms: #14556 0x1105b1b800 1024x1024x1 x1 levels 11 NVN 0x8b ASTC (BC3), 642 of 642 KB [hos host] vk: sparse mirrors: 147, 1382 MB with memory of 1932 MB (549 MB spared); blocks given memory later 6023, given back 7934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 377 binds (95.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 66 taken from the spares, 159 made new; 78 kept as spares, 0 destroyed; 12 spares now, 632 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 501, 2281.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 10373 dirty pages from their submits' captures, 35718 from earlier ones', 71598 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 783 dispatches, 24732 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 2807 runs, 78.7 MB compared; 6682 pages read into the shadows; 0 MB of shadows (0 MB of pages known, in 0 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 4596 in captures, 388 only in the collection, 12 only by the replay itself [hos host] vk: read live, by pool: 0x7fab000000 (68 MB, captured, state 2): 32999; 0x7fb5c00000 (44 MB, captured, state 2): 10652; 0x108c000000 (224 MB, captured, state 2): 10605; 0x7fb8800000 (52 MB, captured, state 2): 8899; 0x1178400000 (152 MB, captured, state 2): 1257; 0x104e000000 (24 MB, captured, state 2): 1110; 0x1042800000 (4 MB, captured, state 2): 1002; 0x1042c00000 (4 MB, captured, state 2): 986; [hos host] vk: imports: 172 live (2232 MB), 13 retired still imported (368 MB), 20 retired given back (404 MB) [hos BaseProcExecutor1] vcores: 99428 takes, 42600 waits (mean 95 us, longest 65.3 ms, 3 over HOS_VCORE_LOG_MS), 19595 kicks, 18989 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 2940000 (1888903 pages from the captures) [hos Resource Control] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 91310689 lookups, 167903 missed (0.18 %) [hos host] vk: pool 0x11637ca000+0xa2438 imported in 2.7 ms as 0x1163400000..0x1163c00000 (206 imports, 642 ms in all) [hos host] vk: mirror of pool 0x1163400000 (4 MB) given back: retired, 16 submissions since its sync [hos host] vk: mirror of pool 0x7f55400000 (4 MB) given back: idle, 6037 submissions since its sync [hos host] vk: 80 mirrors given back (778 MB in all); 1384 MB of mirrors now, 144 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: 3 textures retired (1415 left, 3149 MB): budget 1, drawn over 0, dead targets 2 [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 68.4 ms for a core: given by GeneralFormatAsync_ (priority 24), which held it 0.0 ms and was never asked; the cores it may use now: core 3: it; [hos host] vk: replay reads of guest memory as submitted: 2960000 (1900796 pages from the captures) [hos host] vk: over the driver's video-memory allowance (8.2 of 7.4 GB) [hos host] discord: the status: Hyrule Field - Child / ❤️ 0.5/3 · 🕷️ 0 · 💗 0 [hos host] vk: pool 0x7fd9800000..0x7fdc400000 (44 MB) trimmed to its 2 live part(s), 8 MB: the rest outside every game pool for 1856 submissions (1 trims, 36 MB in all) [hos host] vk: pool 0x7fdac00000..0x7fdb000000: 3900 KB the CPU wrote since retired pool 0x7fd9800000's last sync kept from guest memory, not its mirror [hos host] vk: pool 0x7fdc000000..0x7fdc400000: 3948 KB the CPU wrote since retired pool 0x7fd9800000's last sync kept from guest memory, not its mirror [hos host] vk: 228 pools mirrored in VRAM (1385 MB) [hos host] vk: write-backs left out (600 submits): 799, 3550.5 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3815 dirty pages from their submits' captures, 47889 from earlier ones', 403 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 782 dispatches, 36653 chunks; 0 syncs as copies (no table room) [hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 400; 0x7f90000000 (4 MB, captured, state 2): 2; 0x7fd3c00000 (16 MB, captured, state 2): 1; [hos host] vk: program picks: 3771454 kept (1132652 by the last picks' textures, a uniform buffer bound since), 25971 new, 1802574 with other textures (of 5600000) [hos host] vk: replay reads of guest memory as submitted: 2980000 (1911627 pages from the captures) [hos host] vk: mirror of pool 0x7fd9800000 (44 MB) given back: retired, 2139 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 3000000 (1922573 pages from the captures) [hos host] vk: texture #14163 0x10005191c00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 63 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 85 textures retired (1342 left, 3028 MB): budget 85, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16527 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 1987 pages in 10 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: presenter thread: the copy's fence 2.62 ms, the present 0.72 ms a present [hos host] vk: texture #14173 0x10004ce1800 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 5363 (drawn at 5301) [hos host] vk: texture #14172 0x10004bf1600 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 69 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14174 0x10004b01400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 69 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14171 0x10004741200 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 69 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 169 textures retired (1174 left, 2661 MB): budget 167, drawn over 2, dead targets 0; GPU-drawn with no copy kept: 0 live, 3 finalized, 5 MB [hos host] vk: replay reads of guest memory as submitted: 3020000 (1933689 pages from the captures) [hos host] vk: sparse mirrors: 147, 1385 MB with memory of 1896 MB (510 MB spared); blocks given memory later 6124, given back 7970; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 393 binds (96.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 69 taken from the spares, 159 made new; 81 kept as spares, 0 destroyed; 12 spares now, 668 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1581 dirty pages from their submits' captures, 44268 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 695 dispatches, 35219 chunks; 0 syncs as copies (no table room) [hos host] vk: read live, by pool: [hos host] vk: imports: 173 live (2200 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos host] vk: 33 textures retired (1141 left, 2626 MB): budget 33, drawn over 0, dead targets 0 [hos host] vk: ASTC texture #15398 0x1132c9c600 768x768 x1 levels 10: the image of #14678 0x11179b7e00 (the same blocks, finalized by the game) copied, not decoded again (3 so far, 24 MB) [hos host] vk: ASTC texture #15399 0x1132391200 683x683 x1 levels 10: the image of #14919 0x11182db400 (the same blocks, finalized by the game) copied, not decoded again (4 so far, 25 MB) [hos host] vk: ASTC texture #15400 0x11322ad000 1024x1024 x1 levels 11: the image of #14920 0x1118314400 (the same blocks, finalized by the game) copied, not decoded again (5 so far, 30 MB) [hos host] vk: texture #14276 0x116c342000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 5384 (drawn at 5293) [hos host] vk: ASTC texture #15413 0x113248fc00 512x512 x1 levels 10: the image of #14139 0x11181f6c00 (the same blocks, finalized by the game) copied, not decoded again (6 so far, 31 MB) [hos host] vk: ASTC texture #15414 0x113240b800 768x768 x1 levels 10: the image of #14140 0x1118215a00 (the same blocks, finalized by the game) copied, not decoded again (7 so far, 34 MB) [hos host] vk: replay reads of guest memory as submitted: 3040000 (1946610 pages from the captures) [hos host] vk: ASTC texture #15416 0x10dd585200 768x768 x1 levels 10: the image of #15398 0x1132c9c600 (the same blocks, finalized by the game) copied, not decoded again (8 so far, 34 MB) [hos host] vk: 75 textures retired (1079 left, 2533 MB): budget 73, drawn over 2, dead targets 0 [hos Presentation Thread] vk: capture: a submit 0.23 ms (lock 0.03, scan 0.06 over 779 ranges, 4298 MB), 308 pages written, 308 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.297; GPU wait for late pages 0.249 (65 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (80 MB), 12 spare buffers, 3429 texture pages copied (13 MB) [hos host] vk: replay reads of guest memory as submitted: 3060000 (1958408 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.32, writes collected 0.12 [hos host] vk: memory: working set 15103 MB (peak 15142), private 22834 MB, staging 80 MB in 5 arenas, 37 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 752 MB, 37 allocations 710 MB; used 3388 of 15579 MB; heap 1 (device): 210 blocks 6628 MB, 23415 allocations 4288 MB; used 8493 of 8493 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 19 blocks 608 MB, 101 images 310 MB; 68 free ranges, the largest 29 MB [hos host] vk: image blocks (textures, compacted): 99 blocks 3168 MB, 972 images 1290 MB; 469 free ranges, the largest 31 MB [hos host] vk: render target images: 5657 reused from the pool, 3102 made new, 3000 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB), VRAM critical: given back [hos host] vk: render targets not reused by size: 1017 with the pool empty (0 of them with targets still in the dead queue), 252 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 189 over its bound and 2811 idle [hos host] vk: VRAM: 8493 MB used of 7643 MB allowed (2339 MB of it free in the blocks); textures 2533 MB (budget 1800 MB), 1079 textures [hos host] vk: budget: VRAM 6628 of 7643 MB (8493 MB by the driver, 1864 MB unaccounted): textures 2533, blocks free 2339, mirrors 1385, other buffers 177, other images 111, staging 80; RAM 3355 MB (the process's private memory 22834 MB): imports 2200, capture spares 612, retired imports 416, captures 82, capture table 29, other buffers 16, others 0 [hos host] vk: textures: decoded ASTC 1301 MB, rendered 859 MB (80 drawn in one frame, kept: 46554 KB), other 330 MB; used in the last 8 presents 2091 MB [hos host] vk: decoded ASTC used again: 12 textures (44 MB) copied from ones the game finalized, not decoded; 12 of those retired, 0 listed; 980 MB of blocks hashed in 160 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 253 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 253 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 20 (9459 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 138x96x1 NVN 0x1: 1 (12 KB); 154x98x1 NVN 0x1: 1 (14 KB); 198x40x1 NVN 0x25: 1 (30 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 262x104x1 NVN 0x1: 1 (26 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 656x113x1 NVN 0x1: 1 (72 KB); 1280x720x1 NVN 0x1: 1 (900 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 266 waited for it, 12.34 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 950.9 ms; VRAM 8.2 of 7.4 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5932 queued, 5932 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14130 ms and 314.7 MB read in all); 8028 programs (1613 of them compute variants without their code, 0 graphics programs no note names), 11132 graphics pipelines (1 not in a cache), 704 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1959 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8554 (5932 ready, 0 loading), 15572 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 12564 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 702 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: 8733 programs from the program cache [hos host] vk: a frame: 255 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.1 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 35.47, present 0.37, draw 24.04, descriptors 15.78, texture lookups 11.95, uploads 8.48, pipelines 0.37, programs 2.48, commands 4.75, dispatch 1.53, memory sync 2.52, queue submit 0.01, slot wait 0.32 [hos host] vk: a frame: mirrors 228 (1385 MB): 15.04 MB synced in 1034.6 copies, 17.79 MB written back [hos host] vk: 5400 presents, 24.2 fps; a frame: 2515 draws (0 skipped), 220 dispatches, 191 passes, 149 barriers, 6 submits, 8.9 uploads (5.7 MB), 84 ub updates (59 overlays), GPU-bound wait 0.1 ms (recording slots only); 12966 pipelines (3200 ms), 0 programs (10 failed), 1084 textures (2536 MB) [hos host] vk: host time in GPU resource waits: 36.44 ms a frame (fences 9.06, timelines 27.38, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 863.6 full barriers recorded, 324.7 left out (nothing recorded since the one before), 15.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 2.7; 0x9: 0.9; 0x10: 3.2; 0x11: 44.1; 0x12: 69.6; 0x13: 1.8; 0x21: 3.6; 0x28: 14.7; 0x92: 2.8; 0x97: 0.9; 0x9b: 0.2; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 42.6 from 27.5 batches at the game's barriers (44.9 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 141.0, an op 1818.6, the end 6.5, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.8; 5: 23.8; 48: 7.3; [hos host] vk: GPU time 14.2 ms a frame (the longest submission 8.4 ms); 11.4 textures made a frame: colour targets 2.1 (4.5 MB), depth 0.5 (2.3 MB), ASTC or BC 8.7 (16.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 2.40 MB unswizzled on the CPU in 0.32 ms [hos host] vk: the render thread (600 presents, 21.82 ms a frame): its jobs 16.78 ms a frame (submits 5.61, presents 5.21, others 5.95), idle 5.04; the game waited for room at 44 presents, 0.74 ms each (0.05 a frame; 16.73 jobs queued at a present); 7 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 12.55 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 289 captures at a CPU signal; the submits' waits read in 0.027 ms a frame; readbacks: 1705 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.59 ms after its hand-over (953.07 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 7.56 + 0.43 (the window's copy is in the idle), after it 1.75 + 1.21; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 31.62 (0.17 ms) [hos host] vk: frame ahead (600 presents): 6.10 parts a frame recorded in the signal's job, 65.7% of them took nothing that came with the signal; of the others, 2.06 a frame took pages of the signal's capture into a mirror (66.8 pages a frame) and 0.06 pages read as guest memory had them (132.5 pages a frame); a frame's first part took nothing in 115 of 289 frames, its first three in 115 [hos host] vk: frame ahead, what came with the signal: 290 signals with a capture, each of 561.2 pages in 14.3 pools; between a frame's last submit and that signal 0.34 texture and sampler registrations, 2.56 releases and 0.00 other CPU signals a signal (15 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.54+0.16, 0.47+0.12, 0.17+0.06, 0.17+0.06, 0.19+0.27, 0.07+0.19, 0.02+0.01, 0.06+0.13, 0.03+0.12, 0.02+0.07, 0.00+0.01, 0.01+0.00 [hos host] vk: pipeline cache: play file 8 saved (189 KB): 5 pipelines, 5 from the driver's cache, 0 compiled; 1786 more from the start's cache [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1603 dirty pages from their submits' captures, 44917 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 719 dispatches, 35564 chunks; 0 syncs as copies (no table room) [hos host] vk: read live, by pool: [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 120000 so far (x1 0x1178b9a000 x2 0x2 x3 0x1178b9a000 x4 0x654040) [hos host] vk: render targets drawn over from another start, let go: 9527 so far (45576 MB), 8517 of them (40855 MB) between the passes [hos host] vk: replay reads of guest memory as submitted: 3080000 (1970731 pages from the captures) [hos host] vk: program picks: 3925414 kept (1200685 by the last picks' textures, a uniform buffer bound since), 25976 new, 1848609 with other textures (of 5800000) [hos host] vk: 6 textures retired (1082 left, 2534 MB): budget 5, drawn over 1, dead targets 0 [hos host] vk: VRAM compaction 3 begins: the textures' blocks hold 3168 MB, 1874 MB of it free space [hos host] vk: replay reads of guest memory as submitted: 3100000 (1981642 pages from the captures) [hos host] vk: late pages' fix: stamps for 0x100d7f0000 (pool 0x100d400000; 64 blocks, 4096 KB) [hos host] vk: late pages' fix: stamps for 0x10387f0000 (pool 0x1038400000; 128 blocks, 8192 KB) [hos host] vk: late pages' fix: stamps for 0x103d3f0000 (pool 0x103d000000; 192 blocks, 12288 KB) [hos host] vk: late pages' fix: stamps for 0x103d7f0000 (pool 0x103d000000; 256 blocks, 16384 KB) [hos host] vk: mirror of pool 0x7fdfc00000 (12 MB) given back: idle, 6008 submissions since its sync [hos host] vk: late pages' fix: stamps for 0x117cff0000 (pool 0x1178400000; 320 blocks, 20480 KB) [hos host] vk: late pages' fix: stamps for 0x117d9f0000 (pool 0x1178400000; 384 blocks, 24576 KB) [hos host] vk: late pages' fix: stamps for 0x117dff0000 (pool 0x1178400000; 448 blocks, 28672 KB) [hos host] vk: write-back diff: 5 shadows made again (76 MB) [hos host] vk: texture #15466 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (4 so far) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 146, 1382 MB with memory of 1884 MB (501 MB spared); blocks given memory later 6149, given back 8008; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 401 binds (97.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 69 taken from the spares, 159 made new; 82 kept as spares, 0 destroyed; 13 spares now, 680 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 1856 dirty pages from their submits' captures, 44232 from earlier ones', 12349 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 709 dispatches, 35345 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 3392 runs, 191.2 MB compared; 8489 pages read into the shadows; 280 MB of shadows (34 MB of pages known, in 71 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 4674 in captures, 0 only in the collection, 12 only by the replay itself [hos host] vk: read live, by pool: 0x7fd6400000 (48 MB, captured, state 2): 12008; 0x7fa2000000 (8 MB, captured, state 2): 341; [hos host] vk: imports: 173 live (2200 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos host] vk: replay reads of guest memory as submitted: 3120000 (1992862 pages from the captures) [hos host] vk: mirror of pool 0x105d400000 (4 MB) given back: idle, 6003 submissions since its sync [hos ModuleSystemWorker4] vk: readbacks of GPU results: 908 of 3000 waited for their write-back, 0.11 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1723 kept their virtual core (nothing there to wait for), 2629 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos sead::MainThread] nvn: GPU waits in 5 s: 152 sync waits (2612 ms), 0 polls timed out (0 ms), 6 queue finishes (89 ms) [hos Presentation Thread] vi: 5445 frames presented (30.5 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 0, 17.8-22 105, 22-30 23, 30-40 4, over 40 21 [hos host] vk: replay reads of guest memory as submitted: 3140000 (2006290 pages from the captures) [hos host] vk: mirror of pool 0x7fd4c00000 (24 MB) given back: idle, 6042 submissions since its sync [hos rmbl::VibrationThread] hid: 16384 vibration values, strongest 0.00 since the last count [hos host] vk: replay reads of guest memory as submitted: 3160000 (2019122 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 2504931 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 6509 dirty pages from their submits' captures, 52867 from earlier ones', 157 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 770 dispatches, 36328 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7703 runs, 591.3 MB compared; 58 pages read into the shadows; 280 MB of shadows (34 MB of pages known, in 71 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14953 in captures, 0 only in the collection, 24 only by the replay itself [hos host] vk: read live, by pool: 0x7fa1800000 (4 MB, captured, state 2): 157; [hos host] vk: VRAM compaction 3 over: 332 textures moved (202 MB) in 29 passes, 0 left where they were; 53 blocks (1696 MB) given back to the driver, 2208 MB in all so far [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #85000: event 0x100d6c800c mode 0 value 59554 (now 59535) (lr 0x8066179c) [hos host] vk: replay reads of guest memory as submitted: 3180000 (2031996 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.01, scan 0.04 over 787 ranges, 4279 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.144; GPU wait for late pages 0.105 (32 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 24 kept (94 MB), 12 spare buffers, 5063 texture pages copied (19 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11881 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 4078197 kept (1270980 by the last picks' textures, a uniform buffer bound since), 25984 new, 1895818 with other textures (of 6000000) [hos host] vk: replay reads of guest memory as submitted: 3200000 (2045465 pages from the captures) [hos host] vk: texture #15522 0x10002199c00 120x67 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (5 so far) [hos host] vk: texture #15523 0x100021a9e00 60x33 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (6 so far) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.14, writes collected 0.05 [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #60000: event 0x100d6c802c mode 0 value 59999 (lr 0x80667688) [hos host] vk: sparse mirrors: 145, 1370 MB with memory of 1860 MB (489 MB spared); blocks given memory later 6180, given back 8008; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 407 binds (97.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 159 made new; 84 kept as spares, 0 destroyed; 14 spares now, 704 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 4487 dirty pages from their submits' captures, 45248 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 765 dispatches, 35626 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7357 runs, 586.5 MB compared; 47 pages read into the shadows; 280 MB of shadows (34 MB of pages known, in 72 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15085 in captures, 0 only in the collection, 26 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 174 live (2204 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos host] vk: replay reads of guest memory as submitted: 3220000 (2057354 pages from the captures) [hos host] vk: a frame: 261 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 14.62, present 0.67, draw 8.45, descriptors 3.75, texture lookups 0.29, uploads 0.05, pipelines 0.33, programs 1.49, commands 2.24, dispatch 1.02, memory sync 0.98, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 229 (1370 MB): 6.60 MB synced in 1104.2 copies, 18.60 MB written back [hos host] vk: 5520 presents, 47.1 fps; a frame: 2540 draws (0 skipped), 214 dispatches, 207 passes, 157 barriers, 6 submits, 0.7 uploads (0.2 MB), 84 ub updates (58 overlays), GPU-bound wait 0.0 ms (recording slots only); 12966 pipelines (3200 ms), 0 programs (10 failed), 1156 textures (2606 MB) [hos host] vk: host time in GPU resource waits: 11.45 ms a frame (fences 4.06, timelines 7.39, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 850.3 full barriers recorded, 349.0 left out (nothing recorded since the one before), 17.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.8; 0x8: 3.0; 0x9: 1.0; 0x10: 4.8; 0x11: 43.4; 0x12: 72.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 148.4, an op 1920.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 15.7 ms a frame (the longest submission 10.5 ms); 0.8 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.1 (0.0 MB), ASTC or BC 0.7 (0.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.07 MB unswizzled on the CPU in 0.01 ms [hos host] vk: mirror of pool 0x7f54000000 (4 MB) given back: idle, 6005 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 3240000 (2069045 pages from the captures) [hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 25 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 3146 dirty pages from their submits' captures, 44385 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 731 dispatches, 35325 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6928 runs, 578.1 MB compared; 42 pages read into the shadows; 280 MB of shadows (34 MB of pages known, in 71 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15008 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 2; [hos host] vk: replay reads of guest memory as submitted: 3260000 (2082490 pages from the captures) [hos nn.npln.Callback] vcores: 130930 takes, 27364 waits (mean 102 us, longest 68.4 ms, 1 over HOS_VCORE_LOG_MS), 12625 kicks, 12716 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: mirror of pool 0x7fb0000000 (8 MB) given back: idle, 6048 submissions since its sync [hos host] vk: mirror of pool 0x7fbc000000 (12 MB) given back: idle, 6048 submissions since its sync [hos host] vk: mirror of pool 0x7fdf000000 (12 MB) given back: idle, 6046 submissions since its sync [hos host] vk: mirror of pool 0x7fb2000000 (36 MB) given back: idle, 6046 submissions since its sync [hos host] vk: mirror of pool 0x7fde800000 (8 MB) given back: idle, 6046 submissions since its sync [hos host] vk: mirror of pool 0x7fd2400000 (4 MB) given back: idle, 6046 submissions since its sync [hos host] vk: mirror of pool 0x7fd1800000 (12 MB) given back: idle, 6046 submissions since its sync [hos host] vk: mirror of pool 0x7fbcc00000 (12 MB) given back: idle, 6046 submissions since its sync [hos host] vk: mirror of pool 0x7fb4400000 (24 MB) given back: idle, 6046 submissions since its sync [hos host] vk: mirror of pool 0x7fb0800000 (24 MB) given back: idle, 6046 submissions since its sync [hos host] vk: mirror of pool 0x7fd3400000 (4 MB) given back: idle, 6044 submissions since its sync [hos host] vk: mirror of pool 0x7fc1400000 (28 MB) given back: idle, 6044 submissions since its sync [hos host] vk: mirror of pool 0x7fbf800000 (28 MB) given back: idle, 6044 submissions since its sync [hos host] vk: mirror of pool 0x7fbd800000 (32 MB) given back: idle, 6044 submissions since its sync [hos host] vk: 99 mirrors given back (998 MB in all); 1173 MB of mirrors now, 152 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: replay reads of guest memory as submitted: 3280000 (2095551 pages from the captures) [hos ModuleSystemWorker4] longjmp to 0x81681690 (value 1, 0 slots above dropped) [hos host] vk: reaper: 1904 images (3681 MB) destroyed in 10 s, 55 of them where they were handed over (its cap or VRAM pressure); the queue at most 792 (1023 MB) [hos host] vk: pool 0x115c37d000+0x700000 imported in 3.4 ms as 0x115c000000..0x115d000000 (209 imports, 647 ms in all) [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.04 over 694 ranges, 4311 MB), 250 pages written, 250 copied (1.0 MB) [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 5102 texture pages copied (19 MB) [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: sparse mirrors: 132, 1187 MB with memory of 1636 MB (448 MB spared); blocks given memory later 6210, given back 8008; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 413 binds (100.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 99 kept as spares, 0 destroyed; 29 spares now, 952 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 727, 3233.2 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 2830 dirty pages from their submits' captures, 45989 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 698 dispatches, 37322 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5686 runs, 570.7 MB compared; 2065 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 88 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14571 in captures, 115 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: texture #15671 0x10001e00400 4096x4096 x1 NVN 0x32 (32 MB, rendered or copied into) finalized by the game [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10751 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 152 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 1 textures retired (1300 left, 2771 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: ui: the letters of zh-Hant: DFPT_C8, C:\WINDOWS\Fonts\msjh.ttc [hos host] vk: a frame: 100 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 6.56, present 0.13, draw 3.68, descriptors 1.72, texture lookups 0.33, uploads 0.11, pipelines 0.14, programs 0.55, commands 0.97, dispatch 0.47, memory sync 0.53, queue submit 0.00, slot wait 0.21 [hos host] vk: a frame: mirrors 231 (1187 MB): 3.93 MB synced in 785.3 copies, 7.06 MB written back [hos host] vk: 5640 presents, 54.5 fps; a frame: 976 draws (0 skipped), 62 dispatches, 159 passes, 107 barriers, 6 submits, 1.3 uploads (0.4 MB), 324 ub updates (147 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1300 textures (2771 MB) [hos host] vk: host time in GPU resource waits: 7.94 ms a frame (fences 5.13, timelines 2.81, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 425.2 full barriers recorded, 188.4 left out (nothing recorded since the one before), 51.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 3.2; 0x8: 1.1; 0x9: 0.4; 0x10: 2.5; 0x11: 15.8; 0x12: 70.2; 0x13: 1.4; 0x21: 2.8; 0x28: 5.9; 0x92: 3.7; 0x97: 0.4; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 16.9 from 11.0 batches at the game's barriers (18.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 102.1, an op 649.3, the end 3.8, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 0.7; 5: 9.5; 48: 2.9; [hos host] vk: GPU time 9.0 ms a frame (the longest submission 4.8 ms); 1.7 textures made a frame: colour targets 0.8 (0.7 MB), depth 0.1 (0.4 MB), ASTC or BC 0.8 (1.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.16 MB unswizzled on the CPU in 0.03 ms [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 135 dirty pages from their submits' captures, 47256 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42999 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 495 runs, 7.0 MB compared; 141 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7948 in captures, 31 only in the collection, 70 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.05 over 691 ranges, 4189 MB), 189 pages written, 189 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.015; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (88 MB), 12 spare buffers, 5614 texture pages copied (21 MB) [hos host] vk: program picks: 4224141 kept (1335362 by the last picks' textures, a uniform buffer bound since), 26010 new, 1949848 with other textures (of 6200000) [hos host] vk: 19 textures retired (1281 left, 2769 MB): budget 0, drawn over 19, dead targets 0 [hos host] vk: mirror of pool 0x7f8a800000 (4 MB) given back: idle, 6041 submissions since its sync [hos host] vk: mirror of pool 0x7f90c00000 (4 MB) given back: idle, 6041 submissions since its sync [hos host] vk: mirror of pool 0x7f9c000000 (4 MB) given back: idle, 6041 submissions since its sync [hos host] vk: mirror of pool 0x7f93800000 (8 MB) given back: idle, 6041 submissions since its sync [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: 1 textures retired (1280 left, 2769 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 2559140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 128, 1175 MB with memory of 1616 MB (440 MB spared); blocks given memory later 6210, given back 8024; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 415 binds (101.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 103 kept as spares, 0 destroyed; 33 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 174 dirty pages from their submits' captures, 47068 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42979 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 501 runs, 5.9 MB compared; 143 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7792 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos Presentation Thread] vi: 5719 frames presented (54.7 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 16, 15.5-17.8 134, 17.8-22 100, 22-30 20, 30-40 2, over 40 2 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11716 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: mirror of pool 0x7fa3400000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f90000000 (4 MB) given back: idle, 6002 submissions since its sync [hos host] vk: mirror of pool 0x7f8ec00000 (4 MB) given back: idle, 6064 submissions since its sync [hos host] vk: mirror of pool 0x7f93000000 (8 MB) given back: idle, 6063 submissions since its sync [hos host] vk: mirror of pool 0x7f8e000000 (4 MB) given back: idle, 6063 submissions since its sync [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 692 ranges, 4179 MB), 178 pages written, 178 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (90 MB), 10 spare buffers, 5680 texture pages copied (22 MB) [hos ModuleSystemWorker3] barrier: 80369 looks: 31217 found the masks moved already, 47019 waited until they moved, 2133 for the whole 1 ms (a wait 396 us on average); 64261 wakes [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.02, present 0.10, draw 0.74, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.10, memory sync 0.40, queue submit 0.00, slot wait 0.15 [hos host] vk: a frame: mirrors 231 (1164 MB): 2.60 MB synced in 603.3 copies, 0.08 MB written back [hos host] vk: 5760 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.67 ms a frame (fences 5.36, timelines 0.31, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.4 full barriers recorded, 95.5 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.8 ms a frame (the longest submission 21.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos sead::MainThread] nvn: 800000 recordings made, 71671 live copies (18.3 MB), 5794 texture copies; 2 command buffers' segments (4.3 MB) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 78 dirty pages from their submits' captures, 47163 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42982 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7790 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: mirror of pool 0x7fdc000000 (4 MB) given back: idle, 6016 submissions since its sync [hos host] vk: mirror of pool 0x7fdac00000 (4 MB) given back: idle, 6016 submissions since its sync [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: recorder: 4153686 packets run in 10.0 s (busy 1.21 s, spun idle 0.11 s, slept 10004 times, 3274 of them under 200 us), 0 calls made directly (no room); 13778 batches over 511 frames (27.0 a frame, 301 packets a batch), 7006 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 120 times, 9.00 ms (0.2 times, 0.02 ms a frame); every thread 821 times, 95.16 ms in all, the longest 1.29 ms; vmaDestroyBuffer 150/39.34 ms (max 1.29), vkWaitForFences 442/36.52 ms (max 0.48), the queue's lock 60/7.00 ms (max 0.40), vkDestroyImageView 69/4.45 ms (max 0.30), vmaFreeMemory 18/3.74 ms (max 0.89), vkQueueBindSparse 59/1.56 ms (max 0.35) [hos host] vk: recorder: of its busy 1.21 s, 0.34 s in parts' ends and submits (the longest 1.33 ms), 0.09 s of that blocked on the queue's lock (137 times, the longest 1.29 ms), so 0.25 s in the ends, submits and holds; 0.88 s running the other packets; 8129 packets a frame [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 7920 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos BaseProcExecutor1] vcores: 96582 takes, 17647 waits (mean 31 us, longest 11.4 ms, 0 over HOS_VCORE_LOG_MS), 8974 kicks, 8810 yields, 0 starved grants, 0 escapes in 5 s [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 693 ranges, 4161 MB), 191 pages written, 191 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (98 MB), 11 spare buffers, 5680 texture pages copied (22 MB) [hos host] vk: sparse mirrors: 121, 1163 MB with memory of 1584 MB (420 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 419 binds (101.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 110 kept as spares, 0 destroyed; 40 spares now, 1004 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 108 dirty pages from their submits' captures, 47800 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43590 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7906 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.08, present 0.10, draw 0.78, descriptors 0.40, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.17 [hos host] vk: a frame: mirrors 231 (1163 MB): 2.60 MB synced in 603.4 copies, 0.08 MB written back [hos host] vk: 5880 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.15 ms a frame (fences 5.15, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 8394062 lookups, 74097 missed (0.88 %) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos ModuleSystemWorker2] vk: readbacks of GPU results: 835 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1217 kept their virtual core (nothing there to wait for), 3135 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 7920 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 70, 490.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 62 dirty pages from their submits' captures, 46515 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 630 dispatches, 42358 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 486 runs, 5.7 MB compared; 139 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7679 in captures, 0 only in the collection, 69 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 696 ranges, 4127 MB), 182 pages written, 182 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 5680 texture pages copied (22 MB) [hos host] vk: presenter thread: the copy's fence 4.71 ms, the present 0.63 ms a present [hos host] vk: mirror of pool 0x7f90400000 (4 MB) given back: idle, 6011 submissions since its sync [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: memory: working set 15229 MB (peak 15236), private 21161 MB, staging 80 MB in 5 arenas, 37 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 37 allocations 672 MB; used 3416 of 15579 MB; heap 1 (device): 160 blocks 5155 MB, 20538 allocations 4626 MB; used 7149 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 93 images 283 MB; 53 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 47 blocks 1504 MB, 1178 images 1485 MB; 221 free ranges, the largest 13 MB [hos host] vk: render target images: 5676 reused from the pool, 3142 made new, 3046 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1025 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2854 idle [hos host] vk: VRAM: 7149 MB used of 11474 MB allowed (528 MB of it free in the blocks); textures 2769 MB (budget 7158 MB), 1280 textures [hos host] vk: budget: VRAM 5155 of 11474 MB (7149 MB by the driver, 1993 MB unaccounted): textures 2769, mirrors 1162, blocks free 528, shadows 296, other buffers 207, other images 111, staging 80; RAM 3346 MB (the process's private memory 21161 MB): imports 2228, capture spares 538, retired imports 416, captures 90, other buffers 44, capture table 30, others 0 [hos host] vk: textures: decoded ASTC 1503 MB, rendered 878 MB (90 drawn in one frame, kept: 22286 KB), other 345 MB; used in the last 8 presents 257 MB [hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 0 listed; 1028 MB of blocks hashed in 170 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 853 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 436 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 436 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 436 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 436 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 436 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 1 presents ago [hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 14 (10200 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 5 (8074 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 287 waited for it, 4.02 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 120, 1162 MB with memory of 1580 MB (417 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 420 binds (101.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 111 kept as spares, 0 destroyed; 41 spares now, 1008 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 84 dirty pages from their submits' captures, 47156 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42982 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 501 runs, 5.9 MB compared; 143 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7798 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos host] vk: status: 600 frames, the slowest 115.8 ms; VRAM 7.0 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5944 queued, 5944 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14147 ms and 315.2 MB read in all); 8040 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11149 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1994 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8566 (5944 ready, 0 loading), 15674 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 12586 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.01, present 0.10, draw 0.72, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.19, dispatch 0.11, memory sync 0.39, queue submit 0.00, slot wait 0.17 [hos host] vk: a frame: mirrors 231 (1162 MB): 2.60 MB synced in 603.3 copies, 0.08 MB written back [hos host] vk: 6000 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 4.42 ms a frame (fences 4.42, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.3 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 6.3 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 17.92 ms a frame): its jobs 11.39 ms a frame (submits 1.55, presents 5.73, others 4.11), idle 6.53; the game waited for room at 14 presents, 4.26 ms each (0.10 a frame; 6.47 jobs queued at a present); 6 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 3.08 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 362 captures at a CPU signal; the submits' waits read in 0.019 ms a frame; readbacks: 820 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 10.71 ms after its hand-over (128.64 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 7.82 + 0.18 (the window's copy is in the idle), after it 0.79 + 0.40; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 6.32 parts a frame recorded in the signal's job, 75.2% of them took nothing that came with the signal; of the others, 1.56 a frame took pages of the signal's capture into a mirror (53.3 pages a frame) and 0.01 pages read as guest memory had them (34.6 pages a frame); a frame's first part took nothing in 180 of 362 frames, its first three in 180 [hos host] vk: frame ahead, what came with the signal: 362 signals with a capture, each of 65.2 pages in 6.7 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.19 releases and 0.00 other CPU signals a signal (1 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.32+0.18, 0.09+0.03, 0.04+0.03, 0.16+0.03, 0.01+0.00, 0.01+0.00, 0.04+0.02, 0.07+0.05, 0.02+0.04, 0.02+0.02, 0.00+0.01, 0.01+0.00 [hos Presentation Thread] vi: real-time step on: ratio 0.980, game's frame rate 60 [hos host] vk: mirror of pool 0x7fa1800000 (4 MB) given back: idle, 6049 submissions since its sync [hos host] vk: 112 mirrors given back (1024 MB in all); 1159 MB of mirrors now, 640 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14656 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: render targets drawn over from another start, let go: 9600 so far (45692 MB), 8590 of them (40971 MB) between the passes [hos Presentation Thread] vi: 6019 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 6, 15.5-17.8 288, 17.8-22 6, 22-30 0, 30-40 0, over 40 0 [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.05 over 690 ranges, 4065 MB), 192 pages written, 192 copied (0.8 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (92 MB), 11 spare buffers, 7728 texture pages copied (30 MB) [hos sead::MainThread] nvn: CommandBufferWaitEvent #90000: event 0x100d6c8008 mode 0 value 62610 (now 62607) (lr 0x806661f8) [hos host] vk: readback pages refreshed from guest memory: 2571140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 88 dirty pages from their submits' captures, 47205 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 646 dispatches, 43003 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7795 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] nvn: GPU waits in 5 s: 131 sync waits (791 ms), 0 polls timed out (0 ms), 6 queue finishes (21 ms) [hos host] vk: mirror of pool 0x7f90800000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 8492 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.11, present 0.11, draw 0.75, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.46, queue submit 0.00, slot wait 0.19 [hos host] vk: a frame: mirrors 231 (1157 MB): 2.60 MB synced in 603.4 copies, 0.08 MB written back [hos host] vk: 6120 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 4.91 ms a frame (fences 4.90, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 6.1 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 689 ranges, 4051 MB), 196 pages written, 196 copied (0.8 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 4246 texture pages copied (16 MB) [hos host] vk: queues ordered by their events: 36000 submissions stopped at a wait, 35992 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 90568, not met 0; captures at a CPU event's wait 4646, 0 of them at a wait met on arrival [hos BaseProcExecutor1] vcores: 94374 takes, 16547 waits (mean 29 us, longest 5.6 ms, 0 over HOS_VCORE_LOG_MS), 8502 kicks, 8302 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: sparse mirrors: 118, 1157 MB with memory of 1572 MB (415 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 422 binds (101.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 113 kept as spares, 0 destroyed; 43 spares now, 1016 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 97 dirty pages from their submits' captures, 47759 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 641 dispatches, 43569 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7907 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: mirror of pool 0x7f8f000000 (8 MB) given back: idle, 6035 submissions since its sync [hos host] vk: mirror of pool 0x7f8e800000 (4 MB) given back: idle, 6035 submissions since its sync [hos host] vk: mirror of pool 0x1065800000 (8 MB) given back: idle, 6034 submissions since its sync [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12240 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: mirror of pool 0x7fa6800000 (4 MB) given back: idle, 6012 submissions since its sync [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 98 dirty pages from their submits' captures, 47860 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 655 dispatches, 43616 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7917 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 689 ranges, 4051 MB), 186 pages written, 186 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (106 MB), 10 spare buffers, 5086 texture pages copied (19 MB) [hos host] vk: mirror of pool 0x7fa2000000 (8 MB) given back: idle, 6058 submissions since its sync [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.05, present 0.10, draw 0.74, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 231 (1142 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back [hos host] vk: 6240 presents, 59.9 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 4.85 ms a frame (fences 4.85, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 202.9 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 6.0 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: mirror of pool 0x7fb8800000 (52 MB) given back: idle, 6005 submissions since its sync [hos host] vk: mirror of pool 0x7f81400000 (4 MB) given back: idle, 6064 submissions since its sync [hos host] vk: mirror of pool 0x7fd3c00000 (16 MB) given back: idle, 6064 submissions since its sync [hos host] vk: mirror of pool 0x7f8a000000 (4 MB) given back: idle, 6064 submissions since its sync [hos host] vk: mirror of pool 0x7f9ac00000 (20 MB) given back: idle, 6061 submissions since its sync [hos host] vk: mirror of pool 0x7f99000000 (16 MB) given back: idle, 6061 submissions since its sync [hos host] vk: mirror of pool 0x7f91000000 (32 MB) given back: idle, 6061 submissions since its sync [hos host] vk: mirror of pool 0x1142400000 (4 MB) given back: idle, 6061 submissions since its sync [hos host] vk: mirror of pool 0x7f86000000 (4 MB) given back: idle, 6061 submissions since its sync [hos host] vk: mirror of pool 0x7f84400000 (4 MB) given back: idle, 6061 submissions since its sync [hos host] vk: mirror of pool 0x7f97400000 (4 MB) given back: idle, 6060 submissions since its sync [hos host] vk: mirror of pool 0x7f61000000 (4 MB) given back: idle, 6060 submissions since its sync [hos host] vk: mirror of pool 0x7fff800000 (8 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7fa9000000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f83c00000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f84000000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7fd6400000 (48 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f84800000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f9a400000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f85800000 (8 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f9ec00000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f98800000 (8 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f86c00000 (20 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f88000000 (12 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f88c00000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f9f000000 (8 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f9fc00000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f8ac00000 (12 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f8b800000 (24 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f8f800000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7fa1400000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f94000000 (16 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f95000000 (4 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f9c400000 (32 MB) given back: idle, 6059 submissions since its sync [hos host] vk: mirror of pool 0x7f95400000 (4 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7f96800000 (4 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7f96c00000 (8 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7fab000000 (68 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7faf400000 (4 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7f89000000 (16 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7f86400000 (8 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7fb5c00000 (44 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7f85400000 (4 MB) given back: idle, 6058 submissions since its sync [hos host] vk: mirror of pool 0x7f62800000 (8 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x1042800000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x1042c00000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x7f98400000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x7f9f800000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x7f97800000 (12 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x7fa0400000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x7fa0800000 (8 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x104cc00000 (16 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x7fa1c00000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x7f95800000 (16 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x104e000000 (24 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x1066000000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x1165c00000 (8 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x1042400000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x1167000000 (8 MB) given back: idle, 6057 submissions since its sync [hos host] vk: mirror of pool 0x1042000000 (4 MB) given back: idle, 6057 submissions since its sync [hos host] vk: 178 mirrors given back (1525 MB in all); 658 MB of mirrors now, 640 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 128 dirty pages from their submits' captures, 47728 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 641 dispatches, 43574 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7911 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 25649 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vi: 6311 frames presented (58.3 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 12, 15.5-17.8 271, 17.8-22 6, 22-30 2, 30-40 0, over 40 1 [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.05 over 686 ranges, 4007 MB), 215 pages written, 215 copied (0.8 MB) [hos host] vk: a capture taken (ms): all 0.032; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.029, bytes copied 0.000; 25 kept (116 MB), 11 spare buffers, 2640 texture pages copied (10 MB) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 161 dirty pages from their submits' captures, 47081 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42962 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7698 in captures, 64 only in the collection, 70 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.32, present 0.10, draw 0.77, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.11, memory sync 0.51, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.2 copies, 0.08 MB written back [hos host] vk: 6360 presents, 56.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 7.18 ms a frame (fences 5.29, timelines 1.89, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.4 full barriers recorded, 95.5 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.5 ms a frame (the longest submission 5.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5280 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: recorder: 936684 packets run in 10.0 s (busy 0.47 s, spun idle 0.09 s, slept 8532 times, 3209 of them under 200 us), 0 calls made directly (no room); 9958 batches over 591 frames (16.8 a frame, 94 packets a batch), 4969 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 16 times, 0.50 ms (0.0 times, 0.00 ms a frame); every thread 578 times, 61.55 ms in all, the longest 0.44 ms; vkWaitForFences 553/60.85 ms (max 0.44), vmaFreeMemory 8/0.39 ms (max 0.17), vmaDestroyBuffer 9/0.20 ms (max 0.04), vkQueueBindSparse 8/0.11 ms (max 0.02) [hos host] vk: recorder: of its busy 0.47 s, 0.19 s in parts' ends and submits (the longest 0.85 ms), 0.00 s of that blocked on the queue's lock (5 times, the longest 0.81 ms), so 0.19 s in the ends, submits and holds; 0.28 s running the other packets; 1585 packets a frame [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 685 ranges, 3983 MB), 180 pages written, 180 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (118 MB), 9 spare buffers, 2640 texture pages copied (10 MB) [hos host] vk: readback pages refreshed from guest memory: 2583140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 105 dirty pages from their submits' captures, 47801 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43591 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 500 runs, 5.8 MB compared; 143 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7911 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos BaseProcExecutor2] vcores: 91943 takes, 16685 waits (mean 25 us, longest 4.6 ms, 0 over HOS_VCORE_LOG_MS), 8629 kicks, 8444 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.05, present 0.10, draw 0.74, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back [hos host] vk: 6480 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.12 ms a frame (fences 5.11, timelines 0.01, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.1 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.8 ms a frame (the longest submission 12.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos ModuleSystemWorker4] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 2844646 lookups, 6547 missed (0.23 %) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10560 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 683 ranges, 3969 MB), 190 pages written, 190 copied (0.7 MB) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 88 dirty pages from their submits' captures, 47155 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42987 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 501 runs, 5.9 MB compared; 143 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7806 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (110 MB), 10 spare buffers, 2640 texture pages copied (10 MB) [hos host] vk: presenter thread: the copy's fence 4.97 ms, the present 0.56 ms a present [hos host] vk: program picks: 4317354 kept (1344410 by the last picks' textures, a uniform buffer bound since), 26010 new, 2056635 with other textures (of 6400000) [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 73, 511.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 48471 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 657 dispatches, 44205 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 511 runs, 6.0 MB compared; 146 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8025 in captures, 0 only in the collection, 73 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 15 retired still imported (416 MB), 20 retired given back (404 MB) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 682 ranges, 3967 MB), 192 pages written, 192 copied (0.8 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9742 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (132 MB), 10 spare buffers, 3500 texture pages copied (13 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: memory: working set 15225 MB (peak 15259), private 21030 MB, staging 80 MB in 5 arenas, 37 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 36 allocations 686 MB; used 3416 of 15579 MB; heap 1 (device): 158 blocks 5027 MB, 12476 allocations 4123 MB; used 7022 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 93 images 283 MB; 53 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 47 blocks 1504 MB, 1178 images 1485 MB; 221 free ranges, the largest 13 MB [hos host] vk: render target images: 5676 reused from the pool, 3142 made new, 3046 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1025 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2854 idle [hos host] vk: VRAM: 7022 MB used of 11474 MB allowed (904 MB of it free in the blocks); textures 2769 MB (budget 7158 MB), 1280 textures [hos host] vk: budget: VRAM 5027 of 11474 MB (7022 MB by the driver, 1994 MB unaccounted): textures 2769, blocks free 904, mirrors 658, shadows 296, other buffers 207, other images 111, staging 80; RAM 3381 MB (the process's private memory 21030 MB): imports 2228, capture spares 574, retired imports 416, captures 132, capture table 31, others 0 [hos host] vk: textures: decoded ASTC 1503 MB, rendered 878 MB (90 drawn in one frame, kept: 22286 KB), other 345 MB; used in the last 8 presents 257 MB [hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 2 listed; 1028 MB of blocks hashed in 170 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1453 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1036 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1036 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1036 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1036 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1036 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 1 presents ago [hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 14 (10200 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 5 (8074 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 138 waited for it, 1.98 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 185.9 ms; VRAM 6.9 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5944 queued, 5944 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14147 ms and 315.2 MB read in all); 8040 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11149 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1994 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8566 (5944 ready, 0 loading), 15674 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 12586 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.06, present 0.09, draw 0.82, descriptors 0.44, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.22, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.21 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back [hos host] vk: 6600 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.14 ms a frame (fences 5.14, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 202.8 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 16.91 ms a frame): its jobs 8.52 ms a frame (submits 1.75, presents 6.23, others 0.54), idle 8.38; the game waited for room at 7 presents, 1.03 ms each (0.01 a frame; 5.79 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 215 captures at a CPU signal; the submits' waits read in 0.010 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.96 ms after its hand-over (176.08 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 10.33 + 0.17 (the window's copy is in the idle), after it 0.11 + 0.19; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 92.1% of them took nothing that came with the signal; of the others, 0.16 a frame took pages of the signal's capture into a mirror (0.9 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 192 of 215 frames, its first three in 192 [hos host] vk: frame ahead, what came with the signal: 215 signals with a capture, each of 11.3 pages in 2.1 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.04+0.17, 0.07+0.02, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: render targets drawn over from another start, let go: 9600 so far (45692 MB), 8590 of them (40971 MB) between the passes [hos Presentation Thread] vi: 6611 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 4, 15.5-17.8 291, 17.8-22 5, 22-30 0, 30-40 0, over 40 0 [hos host] vk: write-backs left out (600 submits): 73, 511.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 116 dirty pages from their submits' captures, 48458 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 657 dispatches, 44218 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 511 runs, 6.0 MB compared; 146 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8028 in captures, 0 only in the collection, 73 only by the replay itself [hos host] vk: read live, by pool: [hos rmbl::VibrationThread] hid: 20480 vibration values, strongest 0.00 since the last count [hos sead::MainThread] nvn: QueueSubmitCommands #40000: queue 000010100D6BF020 (its #6665), 1 handles (lr 0x80666be0) [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 681 ranges, 3961 MB), 184 pages written, 184 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 3500 texture pages copied (13 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5280 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 70, 490.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 72 dirty pages from their submits' captures, 46553 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 637 dispatches, 42385 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 490 runs, 5.8 MB compared; 140 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7684 in captures, 0 only in the collection, 70 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 14 retired still imported (388 MB), 21 retired given back (432 MB) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.09, present 0.10, draw 0.75, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.11, memory sync 0.41, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.4 copies, 0.08 MB written back [hos host] vk: 6720 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.16 ms a frame (fences 5.16, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.7 ms a frame (the longest submission 11.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos BaseProcExecutor0] vcores: 94018 takes, 17090 waits (mean 26 us, longest 3.3 ms, 0 over HOS_VCORE_LOG_MS), 8877 kicks, 8648 yields, 0 starved grants, 0 escapes in 5 s [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.05 over 680 ranges, 3943 MB), 188 pages written, 188 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 24 kept (96 MB), 12 spare buffers, 3500 texture pages copied (13 MB) [hos host] vk: readback pages refreshed from guest memory: 2595140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10560 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 92 dirty pages from their submits' captures, 47766 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 641 dispatches, 43569 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7914 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.02, present 0.09, draw 0.75, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.21, dispatch 0.11, memory sync 0.38, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back [hos host] vk: 6840 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.13 ms a frame (fences 5.13, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.1 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.6 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #65000: event 0x100d6c8000 mode 0 value 64999 (lr 0x80664f14) [hos sead::MainThread] barrier: 102787 looks: 37251 found the masks moved already, 63403 waited until they moved, 2133 for the whole 1 ms (a wait 313 us on average); 99077 wakes [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 91 dirty pages from their submits' captures, 47815 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43594 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7914 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 14 retired still imported (388 MB), 21 retired given back (432 MB) [hos sead::MainThread] nvn: CommandBufferWaitEvent #95000: event 0x100d6c8000 mode 0 value 65110 (now 65104) (lr 0x80664f90) [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.05 over 680 ranges, 3943 MB), 178 pages written, 178 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.008; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (88 MB), 11 spare buffers, 3981 texture pages copied (15 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3172 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos Presentation Thread] vi: 6911 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 4, 15.5-17.8 292, 17.8-22 4, 22-30 0, 30-40 0, over 40 0 [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 47808 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43592 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7772 in captures, 133 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.17, present 0.10, draw 0.76, descriptors 0.39, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.21, dispatch 0.11, memory sync 0.41, queue submit 0.00, slot wait 0.21 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.5 copies, 0.08 MB written back [hos host] vk: 6960 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.19 ms a frame (fences 5.19, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.0 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.3 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 680 ranges, 3941 MB), 192 pages written, 192 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (114 MB), 9 spare buffers, 3981 texture pages copied (15 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10560 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: recorder: 950517 packets run in 10.0 s (busy 0.46 s, spun idle 0.10 s, slept 8741 times, 3379 of them under 200 us), 0 calls made directly (no room); 10016 batches over 600 frames (16.7 a frame, 95 packets a batch), 5004 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 8 times, 0.41 ms (0.0 times, 0.00 ms a frame); every thread 565 times, 56.03 ms in all, the longest 0.29 ms; vkWaitForFences 550/55.38 ms (max 0.29), vkDestroyBuffer 8/0.41 ms (max 0.29), vmaDestroyBuffer 7/0.23 ms (max 0.04) [hos host] vk: recorder: of its busy 0.46 s, 0.19 s in parts' ends and submits (the longest 0.64 ms), 0.00 s of that blocked on the queue's lock (4 times, the longest 0.58 ms), so 0.19 s in the ends, submits and holds; 0.27 s running the other packets; 1584 packets a frame [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 108 dirty pages from their submits' captures, 47798 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43589 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7910 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 2 retired still imported (48 MB), 33 retired given back (772 MB) [hos BaseProcExecutor1] vcores: 93853 takes, 17366 waits (mean 26 us, longest 4.1 ms, 0 over HOS_VCORE_LOG_MS), 8999 kicks, 8759 yields, 0 starved grants, 0 escapes in 5 s [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.05 over 680 ranges, 3941 MB), 184 pages written, 184 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (124 MB), 11 spare buffers, 3981 texture pages copied (15 MB) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 78 dirty pages from their submits' captures, 47164 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 639 dispatches, 42978 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7799 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5280 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.09, present 0.09, draw 0.72, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.19, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.3 copies, 0.08 MB written back [hos host] vk: 7080 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.23 ms a frame (fences 5.23, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.4 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos SceneLowPriority] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 3179825 lookups, 7044 missed (0.22 %) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: readback pages refreshed from guest memory: 2607140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 71, 497.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 72 dirty pages from their submits' captures, 47219 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 646 dispatches, 43000 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 497 runs, 5.8 MB compared; 142 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7790 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 0 retired still imported (0 MB), 35 retired given back (820 MB) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.01, scan 0.05 over 634 ranges, 3175 MB), 185 pages written, 185 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (118 MB), 10 spare buffers, 2128 texture pages copied (8 MB) [hos host] vk: presenter thread: the copy's fence 5.16 ms, the present 0.57 ms a present [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 8512 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos host] vk: memory: working set 15220 MB (peak 15259), private 21025 MB, staging 32 MB in 2 arenas, 37 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 37 allocations 692 MB; used 3000 of 15579 MB; heap 1 (device): 158 blocks 5027 MB, 12473 allocations 4075 MB; used 7022 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 93 images 283 MB; 53 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 47 blocks 1504 MB, 1178 images 1485 MB; 221 free ranges, the largest 13 MB [hos host] vk: render target images: 5676 reused from the pool, 3142 made new, 3046 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1025 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2854 idle [hos host] vk: VRAM: 7022 MB used of 11474 MB allowed (952 MB of it free in the blocks); textures 2769 MB (budget 7158 MB), 1280 textures [hos host] vk: budget: VRAM 5027 of 11474 MB (7022 MB by the driver, 1994 MB unaccounted): textures 2769, blocks free 952, mirrors 658, shadows 296, other buffers 207, other images 111, staging 32; RAM 2961 MB (the process's private memory 21025 MB): imports 2228, capture spares 586, captures 116, capture table 31, others 0 [hos host] vk: textures: decoded ASTC 1503 MB, rendered 878 MB (90 drawn in one frame, kept: 22286 KB), other 345 MB; used in the last 8 presents 257 MB [hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 6 listed; 1028 MB of blocks hashed in 170 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 2053 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1636 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1636 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1636 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1636 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1636 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 1 presents ago [hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 14 (10200 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 5 (8074 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 153 waited for it, 0.19 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 19.7 ms; VRAM 6.9 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5944 queued, 5944 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14147 ms and 315.2 MB read in all); 8040 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11149 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 1994 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8566 (5944 ready, 0 loading), 15674 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 12586 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.01, present 0.09, draw 0.74, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.19, dispatch 0.11, memory sync 0.40, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.4 copies, 0.08 MB written back [hos host] vk: 7200 presents, 59.9 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.19 ms a frame (fences 5.19, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.1 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.4 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: queues ordered by their events: 40617 submissions stopped at a wait, 40613 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 96927, not met 0; captures at a CPU event's wait 5027, 0 of them at a wait met on arrival [hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 8.14 ms a frame (submits 1.70, presents 5.89, others 0.55), idle 8.53; the game waited for room at 0 presents, 0.00 ms each (0.00 a frame; 5.15 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 222 captures at a CPU signal; the submits' waits read in 0.009 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 7.63 ms after its hand-over (16.17 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 10.56 + 0.17 (the window's copy is in the idle), after it 0.11 + 0.19; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 96.8% of them took nothing that came with the signal; of the others, 0.06 a frame took pages of the signal's capture into a mirror (0.2 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 210 of 222 frames, its first three in 210 [hos host] vk: frame ahead, what came with the signal: 222 signals with a capture, each of 3.3 pages in 1.8 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.04+0.16, 0.07+0.02, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos ModuleSystemWorker1] nvn: 850000 recordings made, 71504 live copies (18.1 MB), 5794 texture copies; 5 command buffers' segments (4.3 MB) [hos host] vk: render targets drawn over from another start, let go: 9600 so far (45692 MB), 8590 of them (40971 MB) between the passes [hos Presentation Thread] vi: 7212 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 3, 15.5-17.8 294, 17.8-22 4, 22-30 0, 30-40 0, over 40 0 [hos host] vk: write-backs left out (600 submits): 72, 504.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 83 dirty pages from their submits' captures, 47825 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 648 dispatches, 43588 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 504 runs, 5.9 MB compared; 144 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7899 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 618 ranges, 2911 MB), 180 pages written, 180 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.008; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (130 MB), 11 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3232 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 73, 511.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 80 dirty pages from their submits' captures, 48441 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 650 dispatches, 44161 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 507 runs, 5.9 MB compared; 145 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8007 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 0 retired still imported (0 MB), 35 retired given back (820 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.02, present 0.10, draw 0.73, descriptors 0.37, texture lookups 0.03, uploads 0.00, pipelines 0.02, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.39, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.3 copies, 0.08 MB written back [hos host] vk: 7320 presents, 60.1 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.23 ms a frame (fences 5.23, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 202.9 full barriers recorded, 95.3 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.6 ms a frame (the longest submission 3.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos nn.npln.Worker] vcores: 93350 takes, 17048 waits (mean 26 us, longest 2.7 ms, 0 over HOS_VCORE_LOG_MS), 8716 kicks, 8550 yields, 0 starved grants, 0 escapes in 5 s [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 594 ranges, 2807 MB), 177 pages written, 177 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (114 MB), 10 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: write-backs left out (600 submits): 73, 511.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 93 dirty pages from their submits' captures, 48481 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 657 dispatches, 44204 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 511 runs, 6.0 MB compared; 146 pages read into the shadows; 296 MB of shadows (40 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 8025 in captures, 0 only in the collection, 73 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos ModuleSystemWorker1] vk: readbacks of GPU results: 0 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 77 kept their virtual core (nothing there to wait for), 2995 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 53, 658 MB with memory of 840 MB (181 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 427 binds (108.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 70 taken from the spares, 161 made new; 178 kept as spares, 0 destroyed; 108 spares now, 1748 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 70, 490.0 MB; 26 bindings followed, 37 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far [hos host] vk: mirror syncs (600 submits): 59 dirty pages from their submits' captures, 46517 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 630 dispatches, 42353 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 494 runs, 5.8 MB compared; 141 pages read into the shadows; 296 MB of shadows (41 MB of pages known, in 87 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 7681 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 176 live (2228 MB), 0 retired still imported (0 MB), 35 retired given back (820 MB) [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.01, present 0.09, draw 0.75, descriptors 0.38, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.06, commands 0.20, dispatch 0.11, memory sync 0.38, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 231 (658 MB): 2.60 MB synced in 603.2 copies, 0.08 MB written back [hos host] vk: 7440 presents, 60.0 fps; a frame: 219 draws (0 skipped), 2 dispatches, 129 passes, 77 barriers, 6 submits, 0.0 uploads (0.0 MB), 459 ub updates (198 overlays), GPU-bound wait 0.0 ms (recording slots only); 12983 pipelines (3202 ms), 0 programs (10 failed), 1280 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 5.06 ms a frame (fences 5.06, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 203.2 full barriers recorded, 95.4 left out (nothing recorded since the one before), 71.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x11: 1.0; 0x12: 69.0; 0x13: 1.0; 0x21: 2.0; 0x92: 4.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 74.0, an op 43.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 5.7 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 592 ranges, 2572 MB), 170 pages written, 170 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.004; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (102 MB), 10 spare buffers, 0 texture pages copied (0 MB) [hos host] vk: reaper: 73 images (117 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 15 (58 MB) [hos host] vk: pool 0x103cc00000..0x103d800000 took in 2 pools [hos host] vk: pool 0x7fa308ce00+0xe38 imported in 3.7 ms as 0x7fa3000000..0x7fa3400000 (212 imports, 653 ms in all) [hos host] vk: slow upload 36.7 ms: #15784 0x10f0851400 768x1536x1 x1 levels 11 NVN 0x7f ASTC (BC3), 515 of 515 KB [hos host] vk: slow upload 15.8 ms: #15793 0x10f0b09a00 1536x1536x1 x1 levels 11 NVN 0x8d ASTC (BC3), 1026 of 1026 KB [hos host] vk: a frame's uploads took 59 ms: ASTC on the GPU 10 (17.0 MB, 57 ms); on the CPU 17 (7.7 MB, 1 ms); [768x1536 x1 NVN 0x7f ASTC on the GPU 0.5 MB 37 ms] [1536x1536 x1 NVN 0x8d ASTC on the GPU 1.0 MB 16 ms] [hos sead::MainThread] nvn: GPU waits in 5 s: 11 sync waits (322 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: mirror of pool 0x103d000000 (8 MB) given back: retired, 63 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x103cc00000 (4 MB) given back: retired, 63 submissions since its sync, GPU-written [hos host] vk: the CPU wrote 0x103cddc000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cddd000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cff8000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cff9000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cffa000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cdde000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cffb000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cddf000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cffc000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cde0000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cffd000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cde1000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cffe000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cfff000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cde2000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103d000000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cde3000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103d001000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103cde4000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: the CPU wrote 0x103d002000 +0x1000, in a binding whose write-back was left out: written back from now on (RX_VK_WB_MAX_KB=0: every one) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 4411136 kept (1353274 by the last picks' textures, a uniform buffer bound since), 26032 new, 2162831 with other textures (of 6600000) [hos host] vk: late pages found at the replay (600 takes): 19 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06 [hos sead::MainThread] nvn: EventSignal (CPU) #15000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: readback pages refreshed from guest memory: 2619140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 157, 656.5 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 100 so far [hos host] vk: mirror syncs (600 submits): 146 dirty pages from their submits' captures, 60189 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 52304 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 885 runs, 84.2 MB compared; 1969 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18081 in captures, 47 only in the collection, 67 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vi: 7509 frames presented (59.3 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 4, 15.5-17.8 291, 17.8-22 1, 22-30 0, 30-40 0, over 40 1 [hos host] vk: queues ordered by their events: 42000 submissions stopped at a wait, 41994 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 98813, not met 0; captures at a CPU event's wait 5150, 0 of them at a wait met on arrival [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 171 pages written, 171 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (114 MB), 11 spare buffers, 15 texture pages copied (0 MB) [hos ModuleSystemWorker5] tex: the game's texture heap: 115 compaction rounds asked for in 7200 frames without one running [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.93, present 0.12, draw 1.45, descriptors 1.06, texture lookups 0.70, uploads 0.49, pipelines 0.03, programs 0.09, commands 0.19, dispatch 0.16, memory sync 0.40, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 237 (696 MB): 3.89 MB synced in 770.0 copies, 1.35 MB written back [hos host] vk: 7560 presents, 58.2 fps; a frame: 175 draws (0 skipped), 7 dispatches, 130 passes, 82 barriers, 6 submits, 0.6 uploads (0.2 MB), 455 ub updates (196 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1347 textures (2917 MB) [hos host] vk: host time in GPU resource waits: 4.15 ms a frame (fences 3.28, timelines 0.87, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 220.2 full barriers recorded, 101.0 left out (nothing recorded since the one before), 68.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.4; 0x10: 4.2; 0x11: 1.0; 0x12: 68.8; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 79.3, an op 50.3, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.8 ms a frame (the longest submission 7.2 ms); 0.7 textures made a frame: colour targets 0.4 (0.4 MB), depth 0.0 (0.1 MB), ASTC or BC 0.2 (1.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.07 MB unswizzled on the CPU in 0.01 ms [hos host] vk: late pages' fix: stamps for 0x103cdf0000 (pool 0x103cc00000; 384 blocks, 24576 KB) [hos host] vk: late pages' fix: stamps for 0x103d1f0000 (pool 0x103cc00000; 448 blocks, 28672 KB) [hos host] vk: late pages' fix: stamps for 0x103d5f0000 (pool 0x103cc00000; 512 blocks, 32768 KB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 4 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 210, 730.6 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 259 so far [hos host] vk: mirror syncs (600 submits): 114 dirty pages from their submits' captures, 64604 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 982 dispatches, 55728 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 1050 runs, 104.5 MB compared; 140 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23415 in captures, 0 only in the collection, 70 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos host] vk: 11 textures retired (1338 left, 2877 MB): budget 0, drawn over 11, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: recorder: 945269 packets run in 10.0 s (busy 0.47 s, spun idle 0.10 s, slept 8771 times, 3455 of them under 200 us), 0 calls made directly (no room); 10199 batches over 597 frames (17.1 a frame, 93 packets a batch), 5018 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 16 times, 1.41 ms (0.0 times, 0.00 ms a frame); every thread 550 times, 50.82 ms in all, the longest 0.52 ms; vkWaitForFences 515/48.09 ms (max 0.37), vkDestroyImageView 9/1.01 ms (max 0.38), vkQueueBindSparse 7/0.80 ms (max 0.52), vmaDestroyBuffer 13/0.51 ms (max 0.13), vkEndCommandBuffer 2/0.35 ms (max 0.24), the queue's lock 1/0.03 ms (max 0.03) [hos host] vk: recorder: of its busy 0.47 s, 0.19 s in parts' ends and submits (the longest 0.63 ms), 0.00 s of that blocked on the queue's lock (4 times, the longest 0.59 ms), so 0.19 s in the ends, submits and holds; 0.28 s running the other packets; 1583 packets a frame [hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 173 pages written, 173 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.001 (1 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (98 MB), 11 spare buffers, 19 texture pages copied (0 MB) [hos ModuleSystemWorker2] vcores: 93575 takes, 17031 waits (mean 26 us, longest 5.9 ms, 0 over HOS_VCORE_LOG_MS), 8727 kicks, 8566 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: write-backs left out (600 submits): 216, 706.0 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 423 so far [hos host] vk: mirror syncs (600 submits): 118 dirty pages from their submits' captures, 67468 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57543 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 1080 runs, 152.9 MB compared; 144 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 24840 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 2 textures retired (1361 left, 2878 MB): budget 0, drawn over 1, dead targets 1 [hos host] vk: a frame: 0 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.06, present 0.11, draw 0.75, descriptors 0.40, texture lookups 0.05, uploads 0.01, pipelines 0.02, programs 0.06, commands 0.19, dispatch 0.13, memory sync 0.37, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 237 (696 MB): 3.67 MB synced in 797.8 copies, 2.20 MB written back [hos host] vk: 7680 presents, 60.0 fps; a frame: 172 draws (0 skipped), 8 dispatches, 131 passes, 83 barriers, 6 submits, 0.2 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.91 ms a frame (fences 3.86, timelines 0.05, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 221.8 full barriers recorded, 101.6 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.1; 0x10: 4.1; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.2, an op 51.1, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.1 ms a frame (the longest submission 15.1 ms); 0.2 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.01 ms [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 6713644 lookups, 7689 missed (0.11 %) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #100000: event 0x100d6c8000 mode 0 value 67608 (now 67602) (lr 0x80664f90) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 213, 651.2 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 585 so far [hos host] vk: mirror syncs (600 submits): 97 dirty pages from their submits' captures, 67739 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56616 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 1065 runs, 195.9 MB compared; 142 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 25254 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 180 pages written, 180 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (100 MB), 9 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: presenter thread: the copy's fence 4.24 ms, the present 0.51 ms a present [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 213, 606.5 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 746 so far [hos host] vk: mirror syncs (600 submits): 101 dirty pages from their submits' captures, 69054 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56724 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 1065 runs, 240.5 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 25966 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: memory: working set 15087 MB (peak 15259), private 21148 MB, staging 48 MB in 3 arenas, 38 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 37 allocations 694 MB; used 3020 of 15579 MB; heap 1 (device): 162 blocks 5155 MB, 13226 allocations 4246 MB; used 7150 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 16 blocks 512 MB, 93 images 259 MB; 52 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1259 images 1618 MB; 228 free ranges, the largest 19 MB [hos host] vk: render target images: 5676 reused from the pool, 3152 made new, 3056 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1035 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2864 idle [hos host] vk: VRAM: 7150 MB used of 11474 MB allowed (909 MB of it free in the blocks); textures 2878 MB (budget 7158 MB), 1361 textures [hos host] vk: budget: VRAM 5155 of 11474 MB (7150 MB by the driver, 1994 MB unaccounted): textures 2878, blocks free 909, mirrors 696, shadows 300, other buffers 211, other images 111, staging 48; RAM 2977 MB (the process's private memory 21148 MB): imports 2236, capture spares 576, captures 122, capture table 31, retired imports 12, others 0 [hos host] vk: textures: decoded ASTC 1626 MB, rendered 856 MB (97 drawn in one frame, kept: 19496 KB), other 355 MB; used in the last 8 presents 311 MB [hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 7 listed; 1046 MB of blocks hashed in 175 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 2653 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 2236 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 2236 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 2236 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 2236 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 2236 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 349 presents ago [hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 13 (8585 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 4 (6459 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 210 waited for it, 0.59 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 100.8 ms; VRAM 7.0 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5946 queued, 5946 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14150 ms and 315.2 MB read in all); 8042 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11151 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2008 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8568 (5946 ready, 0 loading), 15676 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 12590 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.10, present 0.12, draw 0.80, descriptors 0.42, texture lookups 0.04, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 237 (696 MB): 3.80 MB synced in 795.9 copies, 3.27 MB written back [hos host] vk: 7800 presents, 60.0 fps; a frame: 202 draws (0 skipped), 8 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.89 ms a frame (fences 3.89, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 222.2 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 51.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.0 ms a frame (the longest submission 5.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 16.77 ms a frame): its jobs 9.32 ms a frame (submits 1.76, presents 6.91, others 0.65), idle 7.45; the game waited for room at 5 presents, 0.96 ms each (0.01 a frame; 5.66 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 250 captures at a CPU signal; the submits' waits read in 0.009 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 8.86 ms after its hand-over (108.66 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 9.46 + 0.17 (the window's copy is in the idle), after it 0.12 + 0.36; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 94.0% of them took nothing that came with the signal; of the others, 0.12 a frame took pages of the signal's capture into a mirror (0.7 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 222 of 250 frames, its first three in 222 [hos host] vk: frame ahead, what came with the signal: 250 signals with a capture, each of 4.1 pages in 2.2 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.06+0.34, 0.06+0.02, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: render targets drawn over from another start, let go: 9616 so far (45730 MB), 8605 of them (41009 MB) between the passes [hos Presentation Thread] vi: 7810 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 1, 15.5-17.8 299, 17.8-22 1, 22-30 0, 30-40 0, over 40 0 [hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 181 pages written, 181 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 10 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: readback pages refreshed from guest memory: 2631140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 216, 569.3 MB; 28 bindings followed, 41 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 910 so far [hos host] vk: mirror syncs (600 submits): 136 dirty pages from their submits' captures, 70729 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57531 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 1080 runs, 289.7 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27019 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos rmbl::VibrationThread] hid: 24576 vibration values, strongest 0.00 since the last count [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 190 pages written, 190 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (104 MB), 10 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.06, present 0.11, draw 0.82, descriptors 0.43, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 797.9 copies, 4.29 MB written back [hos host] vk: 7920 presents, 60.1 fps; a frame: 201 draws (0 skipped), 8 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.98 ms a frame (fences 3.98, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 221.4 full barriers recorded, 101.3 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 50.5, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.1 ms a frame (the longest submission 13.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos BaseProcExecutor1] vcores: 95156 takes, 17469 waits (mean 28 us, longest 4.7 ms, 0 over HOS_VCORE_LOG_MS), 8980 kicks, 8803 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: write-backs left out (600 submits): 216, 523.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 116 dirty pages from their submits' captures, 71710 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1016 dispatches, 57898 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 975 runs, 327.5 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27610 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 3300000 (2104954 pages from the captures) [hos ModuleSystemWorker5] barrier: 126566 looks: 44646 found the masks moved already, 79787 waited until they moved, 2133 for the whole 1 ms (a wait 264 us on average); 134098 wakes [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 222, 518.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 144 dirty pages from their submits' captures, 73128 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1028 dispatches, 59112 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 733 runs, 341.7 MB compared; 147 pages read into the shadows; 300 MB of shadows (43 MB of pages known, in 91 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 28382 in captures, 0 only in the collection, 73 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 190 pages written, 190 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 10 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.04, present 0.10, draw 0.82, descriptors 0.42, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.22, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 796.0 copies, 4.68 MB written back [hos host] vk: 8040 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.97 ms a frame (fences 3.97, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 218.8 full barriers recorded, 101.3 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.1 ms a frame (the longest submission 5.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 121 dirty pages from their submits' captures, 71309 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57606 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 727 runs, 341.7 MB compared; 145 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27634 in captures, 0 only in the collection, 73 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 188 pages written, 188 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (112 MB), 11 spare buffers, 19 texture pages copied (0 MB) [hos Presentation Thread] vi: 8110 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 300, 17.8-22 0, 22-30 0, 30-40 0, over 40 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 142 dirty pages from their submits' captures, 71278 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57613 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27672 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.03, present 0.11, draw 0.79, descriptors 0.41, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.20, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 237 (696 MB): 3.88 MB synced in 796.2 copies, 4.68 MB written back [hos host] vk: 8160 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.95 ms a frame (fences 3.95, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 218.8 full barriers recorded, 101.3 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.3 ms a frame (the longest submission 6.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos sead::MainThread] nvn: GPU waits in 5 s: 6 sync waits (16 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 188 pages written, 188 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (100 MB), 9 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: recorder: 989574 packets run in 10.0 s (busy 0.49 s, spun idle 0.09 s, slept 8543 times, 3247 of them under 200 us), 0 calls made directly (no room); 10602 batches over 600 frames (17.7 a frame, 93 packets a batch), 5001 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 0 times, 0.00 ms (0.0 times, 0.00 ms a frame); every thread 559 times, 57.15 ms in all, the longest 0.38 ms; vkWaitForFences 550/56.63 ms (max 0.38), vmaDestroyImage 1/0.19 ms (max 0.19), vkDestroyImageView 2/0.17 ms (max 0.14), vmaDestroyBuffer 6/0.16 ms (max 0.04) [hos host] vk: recorder: of its busy 0.49 s, 0.20 s in parts' ends and submits (the longest 0.35 ms), 0.00 s of that blocked on the queue's lock (1 times, the longest 0.31 ms), so 0.20 s in the ends, submits and holds; 0.30 s running the other packets; 1649 packets a frame [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: readback pages refreshed from guest memory: 2643140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 108 dirty pages from their submits' captures, 70605 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 57022 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 710 runs, 332.3 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27244 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vcores: 95239 takes, 17177 waits (mean 28 us, longest 3.8 ms, 0 over HOS_VCORE_LOG_MS), 8939 kicks, 8753 yields, 0 starved grants, 0 escapes in 5 s [hos sead::MainThread] nvn: 900000 recordings made, 71656 live copies (18.4 MB), 5837 texture copies; 2 command buffers' segments (4.3 MB) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.11, present 0.11, draw 0.82, descriptors 0.44, texture lookups 0.04, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 237 (696 MB): 3.88 MB synced in 797.8 copies, 4.68 MB written back [hos host] vk: 8280 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 4.06 ms a frame (fences 4.05, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 219.4 full barriers recorded, 101.5 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.4 ms a frame (the longest submission 13.8 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 184 pages written, 184 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 93 dirty pages from their submits' captures, 70338 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56794 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 710 runs, 332.3 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27276 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos SceneLowPriority] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 2989158 lookups, 7159 missed (0.24 %) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: presenter thread: the copy's fence 3.95 ms, the present 0.49 ms a present [hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 97 dirty pages from their submits' captures, 70318 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56797 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 710 runs, 332.3 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27216 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 186 pages written, 186 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (110 MB), 11 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: memory: working set 15087 MB (peak 15259), private 21148 MB, staging 48 MB in 3 arenas, 38 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 36 allocations 680 MB; used 3020 of 15579 MB; heap 1 (device): 162 blocks 5155 MB, 13226 allocations 4246 MB; used 7150 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 16 blocks 512 MB, 93 images 259 MB; 52 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1259 images 1618 MB; 228 free ranges, the largest 19 MB [hos host] vk: render target images: 5676 reused from the pool, 3152 made new, 3056 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1035 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2864 idle [hos host] vk: VRAM: 7150 MB used of 11474 MB allowed (909 MB of it free in the blocks); textures 2878 MB (budget 7158 MB), 1361 textures [hos host] vk: budget: VRAM 5155 of 11474 MB (7150 MB by the driver, 1994 MB unaccounted): textures 2878, blocks free 909, mirrors 696, shadows 300, other buffers 211, other images 111, staging 48; RAM 2977 MB (the process's private memory 21148 MB): imports 2236, capture spares 586, captures 112, capture table 31, retired imports 12, others 0 [hos host] vk: textures: decoded ASTC 1626 MB, rendered 856 MB (97 drawn in one frame, kept: 19496 KB), other 355 MB; used in the last 8 presents 311 MB [hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 7 listed; 1046 MB of blocks hashed in 175 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 3253 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 2836 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 2836 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 2836 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 2836 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 2836 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 949 presents ago [hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 13 (8585 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 4 (6459 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 151 waited for it, 0.07 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 17.9 ms; VRAM 7.0 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5946 queued, 5946 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14150 ms and 315.2 MB read in all); 8042 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11151 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2008 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8568 (5946 ready, 0 loading), 15676 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 12590 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.07, present 0.11, draw 0.81, descriptors 0.42, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.37, queue submit 0.00, slot wait 0.21 [hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 796.1 copies, 4.68 MB written back [hos host] vk: 8400 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.90 ms a frame (fences 3.90, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 219.1 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.3 ms a frame (the longest submission 6.3 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 9.69 ms a frame (submits 1.72, presents 7.42, others 0.54), idle 6.98; the game waited for room at 0 presents, 0.00 ms each (0.00 a frame; 5.04 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 201 captures at a CPU signal; the submits' waits read in 0.010 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 9.22 ms after its hand-over (16.26 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 9.01 + 0.15 (the window's copy is in the idle), after it 0.07 + 0.19; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 97.3% of them took nothing that came with the signal; of the others, 0.05 a frame took pages of the signal's capture into a mirror (0.3 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 190 of 201 frames, its first three in 190 [hos host] vk: frame ahead, what came with the signal: 201 signals with a capture, each of 3.9 pages in 2.1 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.03+0.18, 0.04+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: render targets drawn over from another start, let go: 9616 so far (45730 MB), 8605 of them (41009 MB) between the passes [hos Presentation Thread] vi: 8411 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 301, 17.8-22 0, 22-30 0, 30-40 0, over 40 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 219, 511.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 143 dirty pages from their submits' captures, 72283 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1022 dispatches, 58405 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 723 runs, 337.0 MB compared; 145 pages read into the shadows; 300 MB of shadows (43 MB of pages known, in 91 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 28050 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos host] vk: commit: 2539 MB left under Windows' limit as it is now, and the page files can grow by 79851 MB more: counted as room [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 184 pages written, 184 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: program picks: 4521066 kept (1356870 by the last picks' textures, a uniform buffer bound since), 26034 new, 2252899 with other textures (of 6800000) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 210, 490.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 125 dirty pages from their submits' captures, 69723 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 988 dispatches, 56306 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 707 runs, 332.3 MB compared; 141 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 26817 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] nvn: CommandBufferSignalEvent #70000: event 0x100d6c8000 mode 0 value 69999 (lr 0x80664f14) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.05, present 0.12, draw 0.77, descriptors 0.40, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.20, dispatch 0.13, memory sync 0.36, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 237 (696 MB): 3.88 MB synced in 797.7 copies, 4.68 MB written back [hos host] vk: 8520 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.95 ms a frame (fences 3.95, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 219.3 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.5 ms a frame (the longest submission 13.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos BaseProcExecutor1] vcores: 96102 takes, 17314 waits (mean 27 us, longest 3.8 ms, 0 over HOS_VCORE_LOG_MS), 8907 kicks, 8707 yields, 0 starved grants, 0 escapes in 5 s [hos sead::MainThread] nvn: CommandBufferWaitEvent #105000: event 0x1006e38004 mode 0 value 1 (now 0) (lr 0x802eb444) [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 187 pages written, 187 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (120 MB), 9 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: readback pages refreshed from guest memory: 2655140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 71190 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1000 dispatches, 57504 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27635 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.03, present 0.10, draw 0.83, descriptors 0.43, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.22, dispatch 0.13, memory sync 0.35, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 796.1 copies, 4.68 MB written back [hos host] vk: 8640 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.73 ms a frame (fences 3.73, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 219.1 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.5 ms a frame (the longest submission 6.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 115 dirty pages from their submits' captures, 71313 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57607 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27673 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 188 pages written, 188 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 10 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos Presentation Thread] vi: 8711 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 300, 17.8-22 0, 22-30 0, 30-40 0, over 40 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 121 dirty pages from their submits' captures, 70397 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1002 dispatches, 56882 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 710 runs, 332.3 MB compared; 142 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27166 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.02, present 0.11, draw 0.78, descriptors 0.41, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.12, memory sync 0.36, queue submit 0.00, slot wait 0.21 [hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 796.0 copies, 4.68 MB written back [hos host] vk: 8760 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.75 ms a frame (fences 3.75, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 219.1 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.6 ms a frame (the longest submission 6.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 186 pages written, 186 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 9 spare buffers, 19 texture pages copied (0 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 1 sync waits (5 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 144 dirty pages from their submits' captures, 71573 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57827 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27686 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: recorder: 990859 packets run in 10.0 s (busy 0.49 s, spun idle 0.09 s, slept 8600 times, 3294 of them under 200 us), 0 calls made directly (no room); 10670 batches over 600 frames (17.8 a frame, 93 packets a batch), 5035 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 0 times, 0.00 ms (0.0 times, 0.00 ms a frame); every thread 522 times, 53.87 ms in all, the longest 0.49 ms; vkWaitForFences 519/53.74 ms (max 0.49), vmaDestroyBuffer 3/0.12 ms (max 0.07) [hos host] vk: recorder: of its busy 0.49 s, 0.19 s in parts' ends and submits (the longest 0.58 ms), 0.00 s of that blocked on the queue's lock (1 times, the longest 0.55 ms), so 0.19 s in the ends, submits and holds; 0.29 s running the other packets; 1651 packets a frame [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos tera::FileResourceMgr] vcores: 95158 takes, 17161 waits (mean 26 us, longest 2.6 ms, 0 over HOS_VCORE_LOG_MS), 8788 kicks, 8595 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 186 pages written, 186 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (114 MB), 11 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 97 dirty pages from their submits' captures, 71185 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1000 dispatches, 57504 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 720 runs, 337.0 MB compared; 144 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27672 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.01, present 0.11, draw 0.80, descriptors 0.42, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.21, dispatch 0.13, memory sync 0.35, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 237 (696 MB): 3.88 MB synced in 797.7 copies, 4.68 MB written back [hos host] vk: 8880 presents, 60.0 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.83 ms a frame (fences 3.74, timelines 0.09, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 219.1 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.6 ms a frame (the longest submission 9.9 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: queues ordered by their events: 48000 submissions stopped at a wait, 47992 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 107040, not met 0; captures at a CPU event's wait 5664, 0 of them at a wait met on arrival [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos ModuleSystemWorker2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 2857641 lookups, 6724 missed (0.24 %) [hos sead::MainThread] vk: readbacks of GPU results: 0 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 0 kept their virtual core (nothing there to wait for), 3072 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: readback pages refreshed from guest memory: 2667140 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 216, 504.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 101 dirty pages from their submits' captures, 71315 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 57582 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 713 runs, 332.3 MB compared; 143 pages read into the shadows; 300 MB of shadows (44 MB of pages known, in 91 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27582 in captures, 0 only in the collection, 71 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 590 ranges, 2847 MB), 186 pages written, 186 copied (0.7 MB) [hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: presenter thread: the copy's fence 3.77 ms, the present 0.48 ms a present [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05 [hos host] vk: memory: working set 15087 MB (peak 15259), private 21148 MB, staging 48 MB in 3 arenas, 38 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 752 MB, 36 allocations 680 MB; used 3020 of 15579 MB; heap 1 (device): 162 blocks 5155 MB, 13226 allocations 4246 MB; used 7150 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 16 blocks 512 MB, 93 images 259 MB; 52 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1259 images 1618 MB; 228 free ranges, the largest 19 MB [hos host] vk: render target images: 5676 reused from the pool, 3152 made new, 3056 destroyed from it, 985 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1035 with the pool empty (0 of them with targets still in the dead queue), 283 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 495 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2864 idle [hos host] vk: VRAM: 7150 MB used of 11474 MB allowed (909 MB of it free in the blocks); textures 2878 MB (budget 7158 MB), 1361 textures [hos host] vk: budget: VRAM 5155 of 11474 MB (7150 MB by the driver, 1994 MB unaccounted): textures 2878, blocks free 909, mirrors 696, shadows 300, other buffers 211, other images 111, staging 48; RAM 2959 MB (the process's private memory 21148 MB): imports 2236, capture spares 572, captures 108, capture table 31, retired imports 12, others 0 [hos host] vk: textures: decoded ASTC 1626 MB, rendered 856 MB (96 drawn in one frame, kept: 19240 KB), other 355 MB; used in the last 8 presents 311 MB [hos host] vk: decoded ASTC used again: 18 textures (74 MB) copied from ones the game finalized, not decoded; 18 of those retired, 7 listed; 1046 MB of blocks hashed in 175 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 3853 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 3436 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 3436 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 3436 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 3436 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 3436 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 1549 presents ago [hos host] vk: 32 MB: 0x10001e00400 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 13 (8585 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 328x33x1 NVN 0x25: 1 (42 KB); 452x106x1 NVN 0x1: 1 (46 KB); 640x360x1 NVN 0x25: 1 (900 KB); 640x360x1 NVN 0x34: 1 (900 KB); 1466x282x1 NVN 0x25: 4 (6459 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 179 waited for it, 0.13 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 22.2 ms; VRAM 7.0 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5946 queued, 5946 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14150 ms and 315.2 MB read in all); 8042 programs (1614 of them compute variants without their code, 0 graphics programs no note names), 11151 graphics pipelines (1 not in a cache), 705 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2008 pipelines made at create taken by draws; 2622 create calls with nothing kept for their stages, 0 before the index; units 8568 (5946 ready, 0 loading), 15676 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 12590 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 703 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 2 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 2.03, present 0.11, draw 0.77, descriptors 0.40, texture lookups 0.03, uploads 0.00, pipelines 0.03, programs 0.07, commands 0.20, dispatch 0.13, memory sync 0.35, queue submit 0.00, slot wait 0.20 [hos host] vk: a frame: mirrors 237 (696 MB): 3.87 MB synced in 795.9 copies, 4.68 MB written back [hos host] vk: 9000 presents, 59.9 fps; a frame: 199 draws (0 skipped), 5 dispatches, 132 passes, 83 barriers, 6 submits, 0.0 uploads (0.0 MB), 463 ub updates (200 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1361 textures (2878 MB) [hos host] vk: host time in GPU resource waits: 3.52 ms a frame (fences 3.52, timelines 0.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 219.2 full barriers recorded, 101.4 left out (nothing recorded since the one before), 70.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x10: 4.0; 0x11: 1.0; 0x12: 70.0; 0x13: 1.0; 0x21: 2.0; 0x92: 3.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 0.0 from 0.0 batches at the game's barriers (0.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 80.0, an op 48.0, the end 2.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: none [hos host] vk: GPU time 7.7 ms a frame (the longest submission 7.6 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 10.01 ms a frame (submits 1.63, presents 7.78, others 0.60), idle 6.66; the game waited for room at 2 presents, 0.30 ms each (0.00 a frame; 5.15 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 228 captures at a CPU signal; the submits' waits read in 0.009 ms a frame; readbacks: 0 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 9.55 ms after its hand-over (22.07 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 8.67 + 0.16 (the window's copy is in the idle), after it 0.08 + 0.19; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 2.00 parts a frame recorded in the signal's job, 95.8% of them took nothing that came with the signal; of the others, 0.08 a frame took pages of the signal's capture into a mirror (0.6 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 209 of 228 frames, its first three in 209 [hos host] vk: frame ahead, what came with the signal: 228 signals with a capture, each of 4.2 pages in 2.1 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.03+0.18, 0.05+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos Presentation Thread] vi: real-time step on: ratio 0.996, game's frame rate 60 [hos host] vk: render targets drawn over from another start, let go: 9617 so far (45730 MB), 8605 of them (41009 MB) between the passes [hos host] vk: 1 textures retired (1361 left, 2878 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: sparse mirrors: 57, 696 MB with memory of 892 MB (195 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 434 binds (109.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 76 taken from the spares, 161 made new; 180 kept as spares, 0 destroyed; 104 spares now, 1696 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 213, 497.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 128 dirty pages from their submits' captures, 70308 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 56802 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 717 runs, 337.0 MB compared; 143 pages read into the shadows; 300 MB of shadows (45 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 27312 in captures, 0 only in the collection, 72 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2236 MB), 2 retired still imported (12 MB), 35 retired given back (820 MB) [hos host] vk: pool 0x1010eb8000+0xf00000 imported in 18.8 ms as 0x1010800000..0x1012000000 (213 imports, 672 ms in all) [hos host] vk: slow upload 18.8 ms: #15853 0x100032a1000 2560x1440x1 x1 levels 1 NVN 0xa, 15360 of 15360 KB [hos host] vk: texture #14767 0x10004b01a00 2560x1440 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564) [hos host] vk: texture #14768 0x10004fb1e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564) [hos host] vk: texture #14769 0x100050a2000 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564) [hos host] vk: texture #14770 0x10004ec1c00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564) [hos host] vk: texture #14801 0x10004921600 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 9010 (drawn at 5564) [hos host] vk: reaper: 28 images (78 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 5 (39 MB) [hos Presentation Thread] vi: 9011 frames presented (58.8 fps over the last 5.1 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 1, 15.5-17.8 297, 17.8-22 0, 22-30 1, 30-40 0, over 40 1 [hos host] vk: mirror of pool 0x1010800000 (20 MB) given back: retired, 28968 submissions since its sync, GPU-written [hos host] vk: replay reads of guest memory as submitted: 3320000 (2115936 pages from the captures) [hos host] vk: 1 textures retired (1413 left, 2963 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: replay reads of guest memory as submitted: 3340000 (2128850 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 594 ranges, 2913 MB), 203 pages written, 203 copied (0.8 MB) [hos host] vk: a capture taken (ms): all 0.009; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (98 MB), 12 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: write-backs left out (600 submits): 852, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3152 dirty pages from their submits' captures, 46163 from earlier ones', 112 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 721 dispatches, 37048 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6777 runs, 608.8 MB compared; 436 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15678 in captures, 1 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x1163400000 (8 MB, captured, state 2): 90; 0x1159c00000 (36 MB, captured, state 2): 20; 0x7ffe000000 (24 MB, captured, state 2): 1; 0x1051800000 (4 MB, captured, state 2): 1; [hos host] vk: program picks: 4648785 kept (1390759 by the last picks' textures, a uniform buffer bound since), 26034 new, 2325180 with other textures (of 7000000) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 3360000 (2141601 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 51 sync waits (305 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: texture #15688 0x10001f54a00 80x45 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15683 0x10001f5b600 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15685 0x100020b6800 320x180 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15766 0x10005661000 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15765 0x10005570e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15682 0x10001e00400 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15684 0x10001ef0600 320x180 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15686 0x10001f40800 160x90 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15664 0x10000000000 128x128 x6 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15768 0x10005570e00 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: texture #15661 0x10000000000 256x256 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9070 (drawn at 9009) [hos host] vk: replay reads of guest memory as submitted: 3380000 (2155760 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 11 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 117, 1180 MB with memory of 1592 MB (411 MB spared); blocks given memory later 6210, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 496 binds (146.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 44 spares now, 996 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3110 dirty pages from their submits' captures, 46514 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 727 dispatches, 37153 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6808 runs, 605.5 MB compared; 35 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15591 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 177 live (2240 MB), 3 retired still imported (32 MB), 35 retired given back (820 MB) [hos host] vk: 1 textures retired (1434 left, 2989 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: replay reads of guest memory as submitted: 3400000 (2169447 pages from the captures) [hos sead::MainThread] nvn: CommandBufferWaitEvent #110000: event 0x100d6c802c mode 0 value 73182 (now 73163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos sead::MainThread] nvn: 950000 recordings made, 71999 live copies (24.9 MB), 5844 texture copies; 2 command buffers' segments (4.1 MB) [hos rmbl::VibrationThread] hid: 28672 vibration values, strongest 0.00 since the last count [hos host] vk: replay reads of guest memory as submitted: 3420000 (2183145 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 849, 3772.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3366 dirty pages from their submits' captures, 46459 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 753 dispatches, 37257 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6649 runs, 588.1 MB compared; 38 pages read into the shadows; 300 MB of shadows (46 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15648 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos NavMeshCAStepDepWorker1] vcores: 109247 takes, 17514 waits (mean 51 us, longest 8.0 ms, 0 over HOS_VCORE_LOG_MS), 8525 kicks, 8536 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 1 textures retired (1465 left, 3069 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: replay reads of guest memory as submitted: 3440000 (2196415 pages from the captures) [hos host] vk: a frame: 274 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.72, present 0.19, draw 7.60, descriptors 3.66, texture lookups 0.78, uploads 0.40, pipelines 0.28, programs 1.16, commands 1.95, dispatch 0.94, memory sync 0.36, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 299 (1183 MB): 9.79 MB synced in 1068.0 copies, 16.86 MB written back [hos host] vk: 9120 presents, 52.8 fps; a frame: 2080 draws (0 skipped), 155 dispatches, 203 passes, 151 barriers, 6 submits, 0.8 uploads (0.9 MB), 112 ub updates (69 overlays), GPU-bound wait 0.0 ms (recording slots only); 12985 pipelines (3203 ms), 0 programs (10 failed), 1465 textures (3069 MB) [hos host] vk: host time in GPU resource waits: 9.40 ms a frame (fences 4.37, timelines 5.04, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 756.8 full barriers recorded, 326.5 left out (nothing recorded since the one before), 21.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 6.7; 0x8: 2.8; 0x9: 0.9; 0x10: 5.0; 0x11: 38.5; 0x12: 72.2; 0x13: 1.9; 0x21: 3.9; 0x28: 14.8; 0x92: 3.0; 0x97: 0.9; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 42.6 from 27.8 batches at the game's barriers (45.3 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 142.0, an op 1513.1, the end 6.6, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.9; 5: 24.1; 48: 7.4; [hos host] vk: GPU time 13.7 ms a frame (the longest submission 7.1 ms); 1.2 textures made a frame: colour targets 0.4 (0.6 MB), depth 0.2 (0.4 MB), ASTC or BC 0.6 (1.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.59 MB unswizzled on the CPU in 0.09 ms [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 600 ranges, 2963 MB), 237 pages written, 237 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 19 texture pages copied (0 MB) [hos host] vk: 8 textures retired (1461 left, 3072 MB): budget 0, drawn over 8, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 3460000 (2209996 pages from the captures) [hos host] vk: program picks: 4801422 kept (1459419 by the last picks' textures, a uniform buffer bound since), 26049 new, 2372528 with other textures (of 7200000) [hos host] vk: 1 textures retired (1460 left, 3072 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 2722646 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 118, 1183 MB with memory of 1596 MB (412 MB spared); blocks given memory later 6212, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 498 binds (147.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 138 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 43 spares now, 992 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 826, 3665.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3380 dirty pages from their submits' captures, 45518 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 741 dispatches, 36338 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6800 runs, 605.4 MB compared; 40 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15307 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 178 live (2244 MB), 3 retired still imported (32 MB), 35 retired given back (820 MB) [hos host] vk: replay reads of guest memory as submitted: 3480000 (2223551 pages from the captures) [hos host] vk: (its written part: 1 of 1 KB readable, 654 bytes not zero; levels L0) [hos host] vk: texture 0x10e52b7600 1224x1528x1 x3 levels 1 NVN 0x50 (5544 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 1 KB of it [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1016 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1064 kept their virtual core (nothing there to wait for), 2776 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 3500000 (2237223 pages from the captures) [hos sead::MainThread] barrier: 152194 looks: 53890 found the masks moved already, 96156 waited until they moved, 2148 for the whole 1 ms (a wait 236 us on average); 167746 wakes [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3342 dirty pages from their submits' captures, 47306 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 769 dispatches, 37511 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6785 runs, 605.1 MB compared; 36 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15719 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 3520000 (2250854 pages from the captures) [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #75000: event 0x100d6c8014 mode 0 value 75000 (lr 0x80660bcc) [hos host] vk: replay reads of guest memory as submitted: 3540000 (2264417 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 604 ranges, 2965 MB), 241 pages written, 241 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 148 texture pages copied (0 MB) [hos host] vk: sparse mirrors: 118, 1183 MB with memory of 1596 MB (412 MB spared); blocks given memory later 6213, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 499 binds (147.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 138 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 43 spares now, 992 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3203 dirty pages from their submits' captures, 46519 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 730 dispatches, 37232 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6768 runs, 604.9 MB compared; 39 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15659 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 178 live (2244 MB), 3 retired still imported (32 MB), 35 retired given back (820 MB) [hos host] vk: replay reads of guest memory as submitted: 3560000 (2277882 pages from the captures) [hos host] vk: program picks: 4955167 kept (1530953 by the last picks' textures, a uniform buffer bound since), 26050 new, 2418782 with other textures (of 7400000) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 129 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 3580000 (2291609 pages from the captures) [hos host] vk: texture 0x10d221bc00 768x768x1 x1 levels 10 NVN 0x7b -> vk 37 (ASTC decoded) (1500 textures, 3114 MB) [hos host] vk: a frame: 277 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.69, present 0.17, draw 7.38, descriptors 3.21, texture lookups 0.18, uploads 0.04, pipelines 0.30, programs 1.24, commands 2.04, dispatch 0.90, memory sync 0.41, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 301 (1203 MB): 6.06 MB synced in 1099.8 copies, 17.79 MB written back [hos host] vk: 9240 presents, 53.5 fps; a frame: 2288 draws (0 skipped), 166 dispatches, 211 passes, 150 barriers, 6 submits, 0.5 uploads (0.1 MB), 81 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12988 pipelines (3203 ms), 0 programs (10 failed), 1501 textures (3114 MB) [hos host] vk: host time in GPU resource waits: 9.84 ms a frame (fences 4.52, timelines 5.33, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 816.5 full barriers recorded, 344.7 left out (nothing recorded since the one before), 17.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.2; 0x8: 3.0; 0x9: 1.0; 0x10: 3.2; 0x11: 41.3; 0x12: 70.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 140.9, an op 1635.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 14.9 ms a frame (the longest submission 6.1 ms); 0.5 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.4 (0.5 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 3600000 (2305892 pages from the captures) [hos host] vk: 2 of 2 small textures the GPU drew written into their memory in 1.2 ms, 1 frames after their copies were recorded, the first #16055 0x116cf4d000 260x260 levels 9 NVN 0xd (26 so far, 5765 KB, 7 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 6006 dirty pages from their submits' captures, 51991 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 831 dispatches, 39056 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6653 runs, 617.7 MB compared; 54 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16092 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 140000 so far (x1 0x1179846000 x2 0 x3 0x1179846000 x4 0x654040) [hos host] vk: replay reads of guest memory as submitted: 3620000 (2319919 pages from the captures) [hos host] vk: 2 textures retired (1537 left, 3176 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: late pages' fix (600 recordings with one): of 1113 captures with late pages 1012 fixed (73.7 pages, 74.1 fixes a capture), waited as before: 0 no pool, 101 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 448 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 1 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 39, of an earlier recording 408, not recorded yet 0, of this part 0; pages fixed from stamps 73651, from snapshots 939; the shadows: 1288561 pages synced from captures, 1259005 made known by it, 27346 left unknown because the replay or the collection wrote them since [hos Presentation Thread] vi: 9278 frames presented (53.3 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 101, 17.8-22 153, 22-30 9, 30-40 4, over 40 0 [hos host] vk: replay reads of guest memory as submitted: 3640000 (2334660 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 121, 1209 MB with memory of 1620 MB (411 MB spared); blocks given memory later 6391, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 508 binds (147.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 141 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 40 spares now, 968 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2639 dirty pages from their submits' captures, 49944 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 803 dispatches, 39053 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6245 runs, 612.7 MB compared; 68 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16280 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: 0x7fd6400000 (48 MB, captured, state 2): 1; [hos host] vk: imports: 179 live (2256 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB) [hos host] vk: replay reads of guest memory as submitted: 3660000 (2349463 pages from the captures) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #115000: event 0x100d6c802c mode 0 value 76982 (now 76963) (lr 0x8028c554) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 606 ranges, 3192 MB), 265 pages written, 265 copied (1.0 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (100 MB), 12 spare buffers, 254 texture pages copied (0 MB) [hos host] vk: replay reads of guest memory as submitted: 3680000 (2364162 pages from the captures) [hos host] vk: program picks: 5106856 kept (1600648 by the last picks' textures, a uniform buffer bound since), 26064 new, 2467079 with other textures (of 7600000) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1092 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1380 kept their virtual core (nothing there to wait for), 2972 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: readback pages refreshed from guest memory: 2789369 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3321 dirty pages from their submits' captures, 44782 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 721 dispatches, 35120 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6692 runs, 588.7 MB compared; 36 pages read into the shadows; 300 MB of shadows (46 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15191 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 3700000 (2379282 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 260 sync waits (1458 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 3720000 (2393868 pages from the captures) [hos host] vk: 3 textures retired (1534 left, 3164 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: texture slots registered again at another size: 400000 (of 509712 registrations; slot 16221 640x360 -> 256x256) [hos host] vk: replay reads of guest memory as submitted: 3740000 (2408580 pages from the captures) [hos host] vk: recorder: 4643494 packets run in 10.0 s (busy 1.23 s, spun idle 0.12 s, slept 10974 times, 4093 of them under 200 us), 0 calls made directly (no room); 15604 batches over 547 frames (28.5 a frame, 298 packets a batch), 7753 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 86 times, 22.90 ms (0.2 times, 0.04 ms a frame); every thread 644 times, 64.01 ms in all, the longest 10.86 ms; vkWaitForFences 474/33.76 ms (max 0.51), the queue's lock 35/24.53 ms (max 10.86), vkQueueBindSparse 75/2.54 ms (max 0.23), vkDestroyImageView 28/2.32 ms (max 0.28), vmaDestroyBuffer 26/0.37 ms (max 0.04), vkEndCommandBuffer 2/0.34 ms (max 0.19) [hos host] vk: recorder: of its busy 1.23 s, 0.27 s in parts' ends and submits (the longest 1.15 ms), 0.01 s of that blocked on the queue's lock (25 times, the longest 1.11 ms), so 0.26 s in the ends, submits and holds; 0.96 s running the other packets; 8489 packets a frame [hos host] vk: sparse mirrors: 121, 1210 MB with memory of 1620 MB (409 MB spared); blocks given memory later 6411, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 509 binds (147.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 141 taken from the spares, 161 made new; 181 kept as spares, 0 destroyed; 40 spares now, 968 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3391 dirty pages from their submits' captures, 45525 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 730 dispatches, 36188 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6743 runs, 588.9 MB compared; 35 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14716 in captures, 0 only in the collection, 26 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 179 live (2256 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB) [hos host] vk: mirror of pool 0x7fa3000000 (4 MB) given back: idle, 6036 submissions since its sync [hos host] vk: mirror of pool 0x7fa3400000 (4 MB) given back: idle, 6036 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 3760000 (2423160 pages from the captures) [hos host] vk: a frame: 310 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.75, present 0.20, draw 7.75, descriptors 3.52, texture lookups 0.30, uploads 0.14, pipelines 0.32, programs 1.24, commands 2.07, dispatch 0.89, memory sync 0.39, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 303 (1211 MB): 5.83 MB synced in 1087.1 copies, 17.72 MB written back [hos host] vk: 9360 presents, 50.1 fps; a frame: 2282 draws (0 skipped), 163 dispatches, 206 passes, 148 barriers, 6 submits, 0.3 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12989 pipelines (3204 ms), 0 programs (10 failed), 1534 textures (3164 MB) [hos host] vk: host time in GPU resource waits: 10.88 ms a frame (fences 4.13, timelines 6.75, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 803.1 full barriers recorded, 339.4 left out (nothing recorded since the one before), 16.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 40.4; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 138.6, an op 1580.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 15.9 ms a frame (the longest submission 11.8 ms); 0.8 textures made a frame: colour targets 0.3 (1.7 MB), depth 0.2 (1.1 MB), ASTC or BC 0.2 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.06 MB unswizzled on the CPU in 0.01 ms [hos host] vk: replay reads of guest memory as submitted: 3780000 (2437742 pages from the captures) [hos nn.npln.Callback] vcores: 132370 takes, 23179 waits (mean 77 us, longest 13.2 ms, 0 over HOS_VCORE_LOG_MS), 10669 kicks, 10863 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2965 dirty pages from their submits' captures, 45280 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 724 dispatches, 36038 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6575 runs, 571.7 MB compared; 60 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14889 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 1; [hos host] vk: replay reads of guest memory as submitted: 3800000 (2452435 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 614 ranges, 3204 MB), 245 pages written, 245 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (96 MB), 12 spare buffers, 254 texture pages copied (0 MB) [hos host] vk: program picks: 5260085 kept (1672308 by the last picks' textures, a uniform buffer bound since), 26066 new, 2513848 with other textures (of 7800000) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 106 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 3820000 (2466846 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 3840000 (2480075 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 122, 1213 MB with memory of 1624 MB (410 MB spared); blocks given memory later 6426, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 516 binds (148.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 144 taken from the spares, 161 made new; 183 kept as spares, 0 destroyed; 39 spares now, 964 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 851, 3776.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 4073 dirty pages from their submits' captures, 47471 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 759 dispatches, 37784 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7303 runs, 628.3 MB compared; 37 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15443 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 180 live (2260 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB) [hos host] vk: replay reads of guest memory as submitted: 3860000 (2493453 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 3880000 (2506184 pages from the captures) [hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 41712489 lookups, 97484 missed (0.23 %) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3068 dirty pages from their submits' captures, 47063 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 740 dispatches, 37373 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7167 runs, 611.6 MB compared; 38 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15518 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 3900000 (2519678 pages from the captures) [hos sead::MainThread] nvn: CommandBufferSignalEvent #80000: event 0x100d6c8020 mode 0 value 80000 (lr 0x80664f14) [hos ModuleSystemWorker1] nvn: 1000000 recordings made, 73669 live copies (23.7 MB), 5862 texture copies; 5 command buffers' segments (4.3 MB) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 611 ranges, 3212 MB), 250 pages written, 250 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (120 MB), 12 spare buffers, 492 texture pages copied (1 MB) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1108 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1420 kept their virtual core (nothing there to wait for), 3188 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 3920000 (2532367 pages from the captures) [hos host] vk: program picks: 5411383 kept (1739339 by the last picks' textures, a uniform buffer bound since), 26070 new, 2562546 with other textures (of 8000000) [hos host] vk: a frame: 288 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.09, present 0.18, draw 7.93, descriptors 3.44, texture lookups 0.09, uploads 0.01, pipelines 0.32, programs 1.33, commands 2.22, dispatch 0.95, memory sync 0.45, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 305 (1214 MB): 5.83 MB synced in 1099.3 copies, 17.96 MB written back [hos host] vk: 9480 presents, 55.4 fps; a frame: 2247 draws (0 skipped), 171 dispatches, 201 passes, 138 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12989 pipelines (3204 ms), 0 programs (10 failed), 1550 textures (3178 MB) [hos host] vk: host time in GPU resource waits: 8.89 ms a frame (fences 4.72, timelines 4.17, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 782.6 full barriers recorded, 323.0 left out (nothing recorded since the one before), 17.2 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 30.9; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 128.9, an op 1591.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 14.1 ms a frame (the longest submission 5.1 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: readback pages refreshed from guest memory: 2856733 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 122, 1214 MB with memory of 1624 MB (409 MB spared); blocks given memory later 6442, given back 8035; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 518 binds (148.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 144 taken from the spares, 161 made new; 183 kept as spares, 0 destroyed; 39 spares now, 964 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 4165 dirty pages from their submits' captures, 47755 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 773 dispatches, 38061 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7170 runs, 611.5 MB compared; 38 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15885 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 180 live (2260 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB) [hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.7 ms, 1 frames after their copies were recorded, the first #16156 0x116d1f9000 260x260 levels 9 NVN 0xd (27 so far, 5975 KB, 8 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: replay reads of guest memory as submitted: 3940000 (2546318 pages from the captures) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #120000: event 0x100d6c802c mode 0 value 80782 (now 80763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 3960000 (2560246 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 3980000 (2574962 pages from the captures) [hos host] vk: write-backs left out (600 submits): 874, 3883.5 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3551 dirty pages from their submits' captures, 48737 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 757 dispatches, 38112 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7371 runs, 629.6 MB compared; 37 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15384 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 4000000 (2588790 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 4020000 (2602822 pages from the captures) [hos Presentation Thread] vi: 9546 frames presented (53.5 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 70, 17.8-22 193, 22-30 5, 30-40 0, over 40 0 [hos host] vk: replay reads of guest memory as submitted: 4040000 (2617337 pages from the captures) [hos host] vk: sparse mirrors: 124, 1225 MB with memory of 1632 MB (406 MB spared); blocks given memory later 6566, given back 8041; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 528 binds (148.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 146 taken from the spares, 161 made new; 183 kept as spares, 0 destroyed; 37 spares now, 956 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3976 dirty pages from their submits' captures, 46467 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 738 dispatches, 36830 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7147 runs, 596.0 MB compared; 40 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15253 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2264 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 591 ranges, 3223 MB), 267 pages written, 267 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (98 MB), 10 spare buffers, 886 texture pages copied (3 MB) [hos host] vk: replay reads of guest memory as submitted: 4060000 (2632214 pages from the captures) [hos host] vk: presenter thread: the copy's fence 4.33 ms, the present 0.78 ms a present [hos host] vk: program picks: 5563234 kept (1807667 by the last picks' textures, a uniform buffer bound since), 26080 new, 2610685 with other textures (of 8200000) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 685 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 4080000 (2646804 pages from the captures) [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3308 dirty pages from their submits' captures, 46109 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 731 dispatches, 36625 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7044 runs, 579.0 MB compared; 35 pages read into the shadows; 300 MB of shadows (47 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15292 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 4100000 (2661154 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] nvn: GPU waits in 5 s: 271 sync waits (1270 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 4120000 (2675815 pages from the captures) [hos host] vk: memory: working set 15103 MB (peak 15259), private 21420 MB, staging 80 MB in 5 arenas, 40 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 12 blocks 736 MB, 36 allocations 660 MB; used 3040 of 15579 MB; heap 1 (device): 169 blocks 5378 MB, 21892 allocations 5110 MB; used 7374 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 15 blocks 480 MB, 104 images 296 MB; 60 free ranges, the largest 17 MB [hos host] vk: image blocks (textures, compacted): 61 blocks 1952 MB, 1451 images 1917 MB; 251 free ranges, the largest 30 MB [hos host] vk: render target images: 5719 reused from the pool, 3222 made new, 3116 destroyed from it, 989 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1096 with the pool empty (0 of them with targets still in the dead queue), 286 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 501 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2924 idle [hos host] vk: VRAM: 7374 MB used of 11474 MB allowed (268 MB of it free in the blocks); textures 3181 MB (budget 7158 MB), 1563 textures [hos host] vk: budget: VRAM 5378 of 11474 MB (7374 MB by the driver, 1995 MB unaccounted): textures 3181, mirrors 1225, shadows 300, blocks free 268, other buffers 211, other images 111, staging 80; RAM 2975 MB (the process's private memory 21420 MB): imports 2264, capture spares 552, captures 98, capture table 31, retired imports 20, other buffers 10, others 0 [hos host] vk: textures: decoded ASTC 1837 MB, rendered 873 MB (106 drawn in one frame, kept: 47145 KB), other 426 MB; used in the last 8 presents 1965 MB [hos host] vk: decoded ASTC used again: 25 textures (86 MB) copied from ones the game finalized, not decoded; 25 of those retired, 2 listed; 1089 MB of blocks hashed in 184 ms [hos host] vk: 288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 4453 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 2149 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 4453 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 15 (6972 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 1 (175 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); 1466x282x1 NVN 0x25: 4 (6459 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 593 waited for it, 5.19 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 114.6 ms; VRAM 7.2 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 25 program records loaded (10.3 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 5964 queued, 5964 loaded (median 1.35 ms, 99th percentile 16.57 ms, worst 34.27 ms, 14177 ms and 315.6 MB read in all); 8062 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11165 graphics pipelines (1 not in a cache), 711 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2056 pipelines made at create taken by draws; 2627 create calls with nothing kept for their stages, 0 before the index; units 8591 (5964 ready, 0 loading), 15729 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 12618 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 305 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.94, present 0.20, draw 7.74, descriptors 3.32, texture lookups 0.10, uploads 0.01, pipelines 0.31, programs 1.32, commands 2.19, dispatch 0.92, memory sync 0.46, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 307 (1225 MB): 5.93 MB synced in 1105.6 copies, 18.06 MB written back [hos host] vk: 9600 presents, 54.1 fps; a frame: 2253 draws (0 skipped), 175 dispatches, 200 passes, 134 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12999 pipelines (3209 ms), 0 programs (10 failed), 1563 textures (3181 MB) [hos host] vk: host time in GPU resource waits: 9.92 ms a frame (fences 4.91, timelines 5.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 780.5 full barriers recorded, 317.5 left out (nothing recorded since the one before), 17.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.9 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1587.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 14.7 ms a frame (the longest submission 11.1 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 18.82 ms a frame): its jobs 15.94 ms a frame (submits 1.09, presents 3.85, others 11.00), idle 2.87; the game waited for room at 5 presents, 0.44 ms each (0.00 a frame; 2.51 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 594 captures at a CPU signal; the submits' waits read in 0.043 ms a frame; readbacks: 2945 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.01 ms after its hand-over (99.66 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.36 + 0.15 (the window's copy is in the idle), after it 2.04 + 0.67; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 10.69 parts a frame recorded in the signal's job, 71.6% of them took nothing that came with the signal; of the others, 3.03 a frame took pages of the signal's capture into a mirror (103.5 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 6 of 594 frames, its first three in 6 [hos host] vk: frame ahead, what came with the signal: 594 signals with a capture, each of 136.2 pages in 12.6 pools; between a frame's last submit and that signal 0.02 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (4 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.06+0.19, 0.16+0.05, 0.10+0.10, 0.27+0.07, 0.03+0.02, 0.09+0.04, 0.14+0.05, 0.10+0.06, 0.06+0.05, 0.02+0.03, 0.01+0.01, 0.00+0.00 [hos host] vk: mirror of pool 0x7fa1800000 (4 MB) given back: idle, 6034 submissions since its sync [hos host] vk: render targets drawn over from another start, let go: 9721 so far (46153 MB), 8706 of them (41432 MB) between the passes [hos host] vk: replay reads of guest memory as submitted: 4140000 (2690453 pages from the captures) [hos host] vk: mirror of pool 0x7f81800000 (12 MB) given back: idle, 6056 submissions since its sync [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 122, 1216 MB with memory of 1616 MB (399 MB spared); blocks given memory later 6574, given back 8041; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 532 binds (149.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 146 taken from the spares, 161 made new; 185 kept as spares, 0 destroyed; 39 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3124 dirty pages from their submits' captures, 45270 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 715 dispatches, 35446 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7252 runs, 597.5 MB compared; 32 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 14916 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2264 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB) [hos host] vk: replay reads of guest memory as submitted: 4160000 (2704458 pages from the captures) [hos sead::MainThread] vk: readbacks of GPU results: 1104 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1289 kept their virtual core (nothing there to wait for), 3063 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 4180000 (2718671 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 591 ranges, 3223 MB), 245 pages written, 245 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 24 kept (96 MB), 12 spare buffers, 886 texture pages copied (3 MB) [hos host] vk: readback pages refreshed from guest memory: 2923824 pages so far (0 held: may be written ahead in their submission) [hos NavMeshCAStepDepWorker1] vcores: 139238 takes, 25434 waits (mean 84 us, longest 19.4 ms, 0 over HOS_VCORE_LOG_MS), 11792 kicks, 12005 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3018 dirty pages from their submits' captures, 45946 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 707 dispatches, 36389 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7235 runs, 597.2 MB compared; 34 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15336 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: program picks: 5714423 kept (1874161 by the last picks' textures, a uniform buffer bound since), 26081 new, 2659495 with other textures (of 8400000) [hos host] vk: replay reads of guest memory as submitted: 4200000 (2732867 pages from the captures) [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 160000 so far (x1 0x117da08010 x2 0 x3 0x117da08000 x4 0x9b040) [hos host] vk: replay reads of guest memory as submitted: 4220000 (2747574 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 4240000 (2761945 pages from the captures) [hos host] vk: sparse mirrors: 123, 1224 MB with memory of 1620 MB (396 MB spared); blocks given memory later 6644, given back 8041; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 536 binds (149.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 147 taken from the spares, 161 made new; 185 kept as spares, 0 destroyed; 38 spares now, 968 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3561 dirty pages from their submits' captures, 47909 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 753 dispatches, 37736 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7382 runs, 615.3 MB compared; 34 pages read into the shadows; 300 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15700 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2264 MB), 1 retired still imported (20 MB), 37 retired given back (832 MB) [hos host] vk: replay reads of guest memory as submitted: 4260000 (2776112 pages from the captures) [hos host] vk: queues ordered by their events: 54000 submissions stopped at a wait, 53993 went on once it was reached, 4 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 1 (stopped 50.00 ms at most); waits met 124852, not met 0; captures at a CPU event's wait 6392, 0 of them at a wait met on arrival [hos host] vk: mirror of pool 0x7f8a400000 (4 MB) given back: idle, 6025 submissions since its sync [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #125000: event 0x100d6c802c mode 0 value 84582 (now 84563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: mirror of pool 0x7fa9000000 (4 MB) given back: idle, 6028 submissions since its sync [hos host] vk: replay reads of guest memory as submitted: 4280000 (2790517 pages from the captures) [hos host] vk: upload over a drawn image: #1212 0x10001fe0800 640x360x1 x1 NVN 0x14 at present 9706 (texture frame 9707), finalized, 842 KB dirty in 1 ranges [hos host] vk: reaper: 79 images (265 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 13 (80 MB) [hos host] vk: pool 0x100fcfca00+0xf00000 imported in 3.2 ms as 0x100f800000..0x1012000000 (219 imports, 680 ms in all) [hos host] vk: pool 0x100f800000..0x1012000000 took in 2 pools [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: texture #1928 0x1001efcd000 192x192x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4630 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #1925 0x1001efd9200 128x128x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4630 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14058 0x10004470e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4563 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14052 0x1001e071200 2560x1440x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 4563 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #506 0x1147c16000 12288x12288x1 x1 NVN 0x32 the GPU drew goes with no copy of its texels (live, unused for 4563 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 70 textures retired (1524 left, 2868 MB): budget 68, drawn over 2, dead targets 0; GPU-drawn with no copy kept: 1 live, 4 finalized, 316 MB [hos host] vk: mirror of pool 0x1010800000 (24 MB) given back: retired, 111 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x100f800000 (8 MB) given back: retired, 79 submissions since its sync, GPU-written [hos host] vk: mirror of pool 0x7fa2000000 (8 MB) given back: idle, 6053 submissions since its sync [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9712 (texture frame 9713), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9714 (texture frame 9715), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9716 (texture frame 9717), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9718 (texture frame 9719), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3006 dirty pages from their submits' captures, 50999 from earlier ones', 224 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 39758 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7679 runs, 650.7 MB compared; 150 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16523 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: 0x7f9ac00000 (20 MB, captured, state 2): 224; [hos sead::MainThread] nvn: CommandBufferSignalEvent #85000: event 0x100d6c8028 mode 0 value 84999 (lr 0x80667410) [hos host] vk: a frame: 307 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.86, present 0.30, draw 7.52, descriptors 3.27, texture lookups 0.27, uploads 0.14, pipelines 0.30, programs 1.32, commands 2.04, dispatch 0.98, memory sync 0.43, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 311 (1227 MB): 6.30 MB synced in 1101.7 copies, 18.10 MB written back [hos host] vk: 9720 presents, 53.7 fps; a frame: 2138 draws (0 skipped), 176 dispatches, 192 passes, 134 barriers, 6 submits, 0.8 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12999 pipelines (3209 ms), 0 programs (10 failed), 1552 textures (2914 MB) [hos host] vk: host time in GPU resource waits: 8.97 ms a frame (fences 4.28, timelines 4.69, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 752.6 full barriers recorded, 309.7 left out (nothing recorded since the one before), 17.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1443.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 14.4 ms a frame (the longest submission 6.8 ms); 0.8 textures made a frame: colour targets 0.3 (1.1 MB), depth 0.1 (0.6 MB), ASTC or BC 0.4 (0.5 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.54 MB unswizzled on the CPU in 0.08 ms [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9722 (texture frame 9723), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4300000 (2803053 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 598 ranges, 3301 MB), 250 pages written, 250 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (96 MB), 11 spare buffers, 886 texture pages copied (3 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9726 (texture frame 9727), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9728 (texture frame 9729), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9730 (texture frame 9731), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9732 (texture frame 9733), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9734 (texture frame 9735), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9736 (texture frame 9737), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9738 (texture frame 9739), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9740 (texture frame 9741), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4320000 (2813616 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 5863466 kept (1937369 by the last picks' textures, a uniform buffer bound since), 26133 new, 2710400 with other textures (of 8600000) [hos host] vk: mirror of pool 0x7f81400000 (4 MB) given back: idle, 6043 submissions since its sync [hos host] vk: sparse mirrors: 122, 1232 MB with memory of 1628 MB (395 MB spared); blocks given memory later 6644, given back 8043; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 546 binds (152.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 151 taken from the spares, 162 made new; 190 kept as spares, 0 destroyed; 39 spares now, 996 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 974, 4327.6 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1491 dirty pages from their submits' captures, 53486 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 865 dispatches, 41809 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8244 runs, 686.2 MB compared; 92 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17476 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: replay reads of guest memory as submitted: 4340000 (2823321 pages from the captures) [hos host] vk: texture #15867 0x10004ce1e00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705) [hos host] vk: texture #15871 0x10004fb1e00 160x90 x1 NVN 0x3 the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705) [hos host] vk: texture #15883 0x100023a0e00 1280x720 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705) [hos host] vk: texture #15890 0x100021aea00 120x67 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705) [hos host] vk: texture #15889 0x100020d0c00 120x67 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705) [hos host] vk: texture #16132 0x10009cf3200 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 9766 (drawn at 9705) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9776 (texture frame 9777), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9778 (texture frame 9779), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9780 (texture frame 9781), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: slow upload 16.3 ms: #16359 0x11142cb200 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB [hos host] vk: replay reads of guest memory as submitted: 4360000 (2834863 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9782 (texture frame 9783), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9784 (texture frame 9785), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9786 (texture frame 9787), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1100 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1111 kept their virtual core (nothing there to wait for), 2729 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9788 (texture frame 9789), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9790 (texture frame 9791), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9792 (texture frame 9793), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9794 (texture frame 9795), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3704 dirty pages from their submits' captures, 54448 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 884 dispatches, 41290 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8182 runs, 700.3 MB compared; 73 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17023 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9796 (texture frame 9797), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 4380000 (2846523 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 608 ranges, 3413 MB), 241 pages written, 241 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 24 kept (98 MB), 11 spare buffers, 1182 texture pages copied (4 MB) [hos Presentation Thread] vi: 9815 frames presented (53.7 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 2, 15.5-17.8 110, 17.8-22 141, 22-30 12, 30-40 3, over 40 1 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9818 (texture frame 9819), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4400000 (2856617 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9820 (texture frame 9821), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9822 (texture frame 9823), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9824 (texture frame 9825), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: (its written part: 3 of 3 KB readable, 1728 bytes not zero; levels L0) [hos host] vk: texture 0x111a589c00 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 3 KB of it [hos host] vk: (its written part: 3 of 3 KB readable, 1728 bytes not zero; levels L0) [hos host] vk: texture 0x111a589c00 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 3 KB of it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9826 (texture frame 9827), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9828 (texture frame 9829), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: 2 textures retired (1690 left, 3187 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 2991319 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 123, 1232 MB with memory of 1632 MB (399 MB spared); blocks given memory later 6647, given back 8044; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 551 binds (153.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 152 taken from the spares, 162 made new; 191 kept as spares, 0 destroyed; 39 spares now, 996 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2133 dirty pages from their submits' captures, 54838 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 887 dispatches, 43207 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8127 runs, 713.7 MB compared; 89 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17927 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: replay reads of guest memory as submitted: 4420000 (2868569 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 211 textures retired (1478 left, 2862 MB): budget 211, drawn over 0, dead targets 0 [hos host] vk: a frame: 177 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.24, present 0.30, draw 7.52, descriptors 3.80, texture lookups 0.81, uploads 0.83, pipelines 0.27, programs 1.23, commands 1.74, dispatch 0.88, memory sync 0.55, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 314 (1232 MB): 5.72 MB synced in 1075.4 copies, 17.95 MB written back [hos host] vk: 9840 presents, 51.0 fps; a frame: 1752 draws (0 skipped), 166 dispatches, 169 passes, 131 barriers, 6 submits, 3.3 uploads (3.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 12999 pipelines (3209 ms), 0 programs (10 failed), 1478 textures (2862 MB) [hos host] vk: host time in GPU resource waits: 8.90 ms a frame (fences 1.64, timelines 7.25, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 670.6 full barriers recorded, 286.0 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.1; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.5, an op 1064.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 14.2 ms a frame (the longest submission 9.5 ms); 1.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 1.2 (2.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 3.11 MB unswizzled on the CPU in 0.45 ms [hos host] vk: late pages' fix (600 recordings with one): of 673 captures with late pages 673 fixed (134.0 pages, 134.9 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 524 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 1 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 29, of an earlier recording 494, not recorded yet 0, of this part 0; pages fixed from stamps 89153, from snapshots 1027; the shadows: 264414 pages synced from captures, 262355 made known by it, 2059 left unknown because the replay or the collection wrote them since [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9850 (texture frame 9851), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker2] nvn: 1050000 recordings made, 74295 live copies (22.8 MB), 5872 texture copies; 6 command buffers' segments (4.3 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 259 sync waits (1546 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: program picks: 6005421 kept (1996383 by the last picks' textures, a uniform buffer bound since), 26166 new, 2768412 with other textures (of 8800000) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9854 (texture frame 9855), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4440000 (2880769 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9856 (texture frame 9857), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9858 (texture frame 9859), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9860 (texture frame 9861), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9865 (texture frame 9866), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9867 (texture frame 9868), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9869 (texture frame 9870), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9871 (texture frame 9872), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4460000 (2892704 pages from the captures) [hos host] vk: (its written part: 3 of 3 KB readable, 1728 bytes not zero; levels L0) [hos host] vk: texture 0x111a589c00 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 3 KB of it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9873 (texture frame 9874), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0) [hos host] vk: texture 0x111b189e00 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it [hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0) [hos host] vk: texture 0x111bd8a000 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9875 (texture frame 9876), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1997 dirty pages from their submits' captures, 55258 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 912 dispatches, 43503 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8190 runs, 714.9 MB compared; 92 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17975 in captures, 44 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: recorder: 6128515 packets run in 10.0 s (busy 1.64 s, spun idle 0.12 s, slept 11677 times, 4214 of them under 200 us), 0 calls made directly (no room); 17871 batches over 529 frames (33.8 a frame, 343 packets a batch), 9004 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 61 times, 5.51 ms (0.1 times, 0.01 ms a frame); every thread 637 times, 43.11 ms in all, the longest 0.66 ms; vkWaitForFences 465/29.57 ms (max 0.62), the queue's lock 43/5.29 ms (max 0.57), vkDestroyImageView 29/2.22 ms (max 0.35), vkQueueBindSparse 42/2.18 ms (max 0.33), vmaFreeMemory 11/2.16 ms (max 0.66), vmaDestroyBuffer 40/1.34 ms (max 0.48) [hos host] vk: recorder: of its busy 1.64 s, 0.36 s in parts' ends and submits (the longest 1.17 ms), 0.07 s of that blocked on the queue's lock (111 times, the longest 1.13 ms), so 0.30 s in the ends, submits and holds; 1.27 s running the other packets; 11585 packets a frame [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 250 textures retired (1237 left, 2512 MB): budget 250, drawn over 0, dead targets 0 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9882 (texture frame 9883), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9884 (texture frame 9885), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9886 (texture frame 9887), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4480000 (2905378 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9888 (texture frame 9889), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 46 textures retired (1191 left, 2441 MB): budget 46, drawn over 0, dead targets 0 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9890 (texture frame 9891), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9892 (texture frame 9893), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #130000: event 0x100d6c802c mode 0 value 88382 (now 88363) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.07, writes collected 0.05 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9897 (texture frame 9898), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 619 ranges, 3448 MB), 306 pages written, 306 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.072; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.052, bytes copied 0.000; 25 kept (112 MB), 11 spare buffers, 14339 texture pages copied (56 MB) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9899 (texture frame 9900), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9901 (texture frame 9902), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos BaseProcExecutor2] vcores: 132444 takes, 21912 waits (mean 74 us, longest 9.1 ms, 0 over HOS_VCORE_LOG_MS), 10361 kicks, 10542 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9903 (texture frame 9904), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9905 (texture frame 9906), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9907 (texture frame 9908), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4500000 (2915517 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 43619 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 123, 1232 MB with memory of 1632 MB (399 MB spared); blocks given memory later 6647, given back 8044; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 551 binds (153.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 152 taken from the spares, 162 made new; 191 kept as spares, 0 destroyed; 39 spares now, 996 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1715 dirty pages from their submits' captures, 55158 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 881 dispatches, 43356 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8264 runs, 716.2 MB compared; 82 pages read into the shadows; 292 MB of shadows (46 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18616 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9916 (texture frame 9917), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: VRAM compaction 4 begins: the textures' blocks hold 2208 MB, 706 MB of it free space [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9918 (texture frame 9919), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0) [hos host] vk: texture 0x11331b0400 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it [hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0) [hos host] vk: texture 0x11321b0200 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9920 (texture frame 9921), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9922 (texture frame 9923), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0) [hos host] vk: texture 0x111c18a000 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it [hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0) [hos host] vk: texture 0x111b189e00 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9924 (texture frame 9925), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4520000 (2925313 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9929 (texture frame 9930), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9931 (texture frame 9932), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9933 (texture frame 9934), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9935 (texture frame 9936), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9937 (texture frame 9938), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9939 (texture frame 9940), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] vk: readbacks of GPU results: 1073 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1419 kept their virtual core (nothing there to wait for), 3189 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 4540000 (2938009 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9948 (texture frame 9949), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9950 (texture frame 9951), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9952 (texture frame 9953), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: program picks: 6149747 kept (2056700 by the last picks' textures, a uniform buffer bound since), 26189 new, 2824063 with other textures (of 9000000) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9954 (texture frame 9955), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2049 dirty pages from their submits' captures, 56810 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 43788 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8259 runs, 715.7 MB compared; 95 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17528 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9956 (texture frame 9957), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4560000 (2950381 pages from the captures) [hos host] vk: a frame: 206 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.37, present 0.37, draw 7.94, descriptors 3.79, texture lookups 0.71, uploads 0.75, pipelines 0.30, programs 1.40, commands 1.91, dispatch 0.89, memory sync 0.74, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 315 (1233 MB): 5.64 MB synced in 1086.0 copies, 17.90 MB written back [hos host] vk: 9960 presents, 48.3 fps; a frame: 1917 draws (0 skipped), 162 dispatches, 179 passes, 131 barriers, 6 submits, 1.5 uploads (2.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13003 pipelines (3211 ms), 0 programs (10 failed), 1193 textures (2441 MB) [hos host] vk: host time in GPU resource waits: 8.82 ms a frame (fences 1.46, timelines 7.36, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 695.5 full barriers recorded, 295.4 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1116.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 15.0 ms a frame (the longest submission 12.6 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 2.80 MB unswizzled on the CPU in 0.38 ms [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9961 (texture frame 9962), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9963 (texture frame 9964), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9965 (texture frame 9966), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0) [hos host] vk: texture 0x11321b0200 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it [hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0) [hos host] vk: texture 0x1132db0400 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9967 (texture frame 9968), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos BaseProcExecutor4] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 13624206 lookups, 115902 missed (0.85 %) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9969 (texture frame 9970), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0) [hos host] vk: texture 0x111b189e00 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it [hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0) [hos host] vk: texture 0x111bd8a000 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9971 (texture frame 9972), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9978 (texture frame 9979), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4580000 (2960759 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9980 (texture frame 9981), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: CommandBufferSignalEvent #90000: event 0x100d6c8034 mode 0 value 89999 (lr 0x806616b0) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9982 (texture frame 9983), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] vk: capture: a submit 0.21 ms (lock 0.00, scan 0.04 over 620 ranges, 3444 MB), 334 pages written, 334 copied (1.3 MB) [hos host] vk: a capture taken (ms): all 0.068; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.050, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 14576 texture pages copied (56 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9984 (texture frame 9985), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: VRAM compaction 4 over: 368 textures moved (282 MB) in 34 passes, 0 left where they were; 20 blocks (640 MB) given back to the driver, 2848 MB in all so far [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9986 (texture frame 9987), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9993 (texture frame 9994), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9995 (texture frame 9996), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4600000 (2972360 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 125, 1230 MB with memory of 1640 MB (409 MB spared); blocks given memory later 6649, given back 8107; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 555 binds (153.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 154 taken from the spares, 162 made new; 191 kept as spares, 0 destroyed; 37 spares now, 988 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1582 dirty pages from their submits' captures, 55596 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 43459 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8160 runs, 714.9 MB compared; 112 pages read into the shadows; 292 MB of shadows (46 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18099 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos Presentation Thread] nvn: QueueSubmitCommands #60000: queue 000010100D640020 (its #9998), 1 handles (lr 0x805282f4) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.08, writes collected 0.05 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9997 (texture frame 9998), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: EventSignal (CPU) #20000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 9999 (texture frame 10000), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10001 (texture frame 10002), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10003 (texture frame 10004), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10008 (texture frame 10009), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10010 (texture frame 10011), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10012 (texture frame 10013), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10014 (texture frame 10015), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4620000 (2982233 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10016 (texture frame 10017), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10018 (texture frame 10019), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10025 (texture frame 10026), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10027 (texture frame 10028), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 1.0 ms, 1 frames after their copies were recorded, the first #16432 0x116d470000 260x260 levels 9 NVN 0xd (28 so far, 6184 KB, 9 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10029 (texture frame 10030), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10031 (texture frame 10032), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4640000 (2994205 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10033 (texture frame 10034), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: readback pages refreshed from guest memory: 3058429 pages so far (0 held: may be written ahead in their submission) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10035 (texture frame 10036), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1516 dirty pages from their submits' captures, 55840 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 883 dispatches, 43404 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8120 runs, 714.5 MB compared; 91 pages read into the shadows; 292 MB of shadows (46 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17382 in captures, 0 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10042 (texture frame 10043), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10044 (texture frame 10045), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: program picks: 6299013 kept (2122582 by the last picks' textures, a uniform buffer bound since), 26189 new, 2874797 with other textures (of 9200000) [hos host] vk: mirror of pool 0x7fa1c00000 (4 MB) given back: idle, 6062 submissions since its sync [hos host] vk: 192 mirrors given back (1612 MB in all); 1227 MB of mirrors now, 160 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10046 (texture frame 10047), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4660000 (3006962 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10048 (texture frame 10049), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10050 (texture frame 10051), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker1] nvn: DrawMeshTasksIndirect: 180000 so far (x1 0x117db40010 x2 0 x3 0x117db40000 x4 0x44b040) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10052 (texture frame 10053), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10057 (texture frame 10058), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos Presentation Thread] vi: 10060 frames presented (48.9 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 23, 17.8-22 196, 22-30 17, 30-40 2, over 40 7 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10059 (texture frame 10060), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10061 (texture frame 10062), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10063 (texture frame 10064), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10065 (texture frame 10066), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4680000 (3017318 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10067 (texture frame 10068), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0) [hos host] vk: texture 0x11331b0400 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it [hos host] vk: (its written part: 16384 of 16384 KB readable, 2164214 bytes not zero; levels L0) [hos host] vk: texture 0x11321b0200 4096x4096x1 x1 levels 1 NVN 0x4b (16384 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 16384 KB of it [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 623 ranges, 3444 MB), 264 pages written, 264 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.041; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.026, bytes copied 0.000; 24 kept (182 MB), 11 spare buffers, 8291 texture pages copied (32 MB) [hos host] vk: (its written part: 12288 of 12288 KB readable, 2449906 bytes not zero; levels L0) [hos host] vk: texture 0x111c18a000 2048x2048x1 x3 levels 1 NVN 0x50 (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 12288 KB of it [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 90536 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10074 (texture frame 10075), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: sparse mirrors: 125, 1226 MB with memory of 1640 MB (413 MB spared); blocks given memory later 6649, given back 8124; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 559 binds (155.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 155 taken from the spares, 162 made new; 192 kept as spares, 0 destroyed; 37 spares now, 988 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1720 dirty pages from their submits' captures, 54714 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 43164 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8121 runs, 714.5 MB compared; 103 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16926 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10076 (texture frame 10077), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10078 (texture frame 10079), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a frame: 232 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.61, present 0.29, draw 8.06, descriptors 3.90, texture lookups 0.65, uploads 0.74, pipelines 0.30, programs 1.31, commands 1.99, dispatch 0.89, memory sync 0.71, queue submit 0.00, slot wait 0.27 [hos host] vk: a frame: mirrors 317 (1226 MB): 5.58 MB synced in 1081.4 copies, 17.87 MB written back [hos host] vk: 10080 presents, 49.4 fps; a frame: 2111 draws (0 skipped), 161 dispatches, 182 passes, 131 barriers, 6 submits, 1.1 uploads (2.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13005 pipelines (3212 ms), 0 programs (10 failed), 1205 textures (2444 MB) [hos host] vk: host time in GPU resource waits: 8.53 ms a frame (fences 1.28, timelines 7.25, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 704.6 full barriers recorded, 298.4 left out (nothing recorded since the one before), 13.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.1, an op 1217.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 14.7 ms a frame (the longest submission 9.6 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 2.54 MB unswizzled on the CPU in 0.37 ms [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10080 (texture frame 10081), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10082 (texture frame 10083), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10084 (texture frame 10085), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4700000 (3026965 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10089 (texture frame 10090), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10091 (texture frame 10092), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10093 (texture frame 10094), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a 192 MB capture buffer made ahead in 74.1 ms, off the game's thread [hos host] vk: (its written part: 3 of 3 KB readable, 0 bytes not zero; levels L0) [hos host] vk: texture 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6 (39444 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it [hos BaseProcExecutor0] vcores: BaseProcExecutor0 (priority 16) waited 126.5 ms for a core: given by sead::MainThread (priority 16), which held it 126.6 ms and let it go 123.1 ms after it was asked; the cores it may use now: core 0: it; [hos PreActorModelCompose] vcores: PreActorModelCompose (priority 26) waited 59.9 ms for a core: given by sead::MainThread (priority 16), which held it 0.0 ms and was never asked; the cores it may use now: core 0: it; [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10095 (texture frame 10096), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a 128 MB capture buffer made ahead in 40.9 ms, off the game's thread [hos sead::MainThread] nvn: GPU waits in 5 s: 240 sync waits (1722 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #135000: event 0x100d6c802c mode 0 value 92182 (now 92163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.09 [hos host] vk: mirror of pool 0x7f9f000000 (8 MB) given back: idle, 6009 submissions since its sync [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10097 (texture frame 10098), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] vk: readbacks of GPU results: 1080 of 3000 waited for their write-back, 0.12 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1600 kept their virtual core (nothing there to wait for), 3008 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: mirror of pool 0x7f85800000 (8 MB) given back: idle, 6022 submissions since its sync [hos host] vk: mirror of pool 0x7f96800000 (4 MB) given back: idle, 6001 submissions since its sync [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10102 (texture frame 10103), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4720000 (3039150 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10104 (texture frame 10105), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: mirror of pool 0x7f84400000 (4 MB) given back: idle, 6051 submissions since its sync [hos host] vk: mirror of pool 0x7faf400000 (4 MB) given back: idle, 6049 submissions since its sync [hos host] vk: mirror of pool 0x7fd6400000 (48 MB) given back: idle, 6049 submissions since its sync [hos host] vk: mirror of pool 0x7f7ec00000 (4 MB) given back: idle, 6048 submissions since its sync [hos host] vk: mirror of pool 0x7fa0400000 (4 MB) given back: idle, 6048 submissions since its sync [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10106 (texture frame 10107), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10108 (texture frame 10109), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.5 ms, 1 frames after their copies were recorded, the first #16437 0x116d72b000 260x260 levels 9 NVN 0xd (29 so far, 6394 KB, 10 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: texture #14280 0x116c3ac000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 10110 (drawn at 5293) [hos host] vk: 208 mirrors given back (1687 MB in all); 1153 MB of mirrors now, 160 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10110 (texture frame 10111), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 7802 dirty pages from their submits' captures, 123514 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1084 dispatches, 49057 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8151 runs, 714.8 MB compared; 179 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16950 in captures, 125 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10117 (texture frame 10118), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4740000 (3051287 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10119 (texture frame 10120), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10121 (texture frame 10122), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10123 (texture frame 10124), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos BaseProcExecutor5] vcores: 122136 takes, 22285 waits (mean 89 us, longest 126.5 ms, 2 over HOS_VCORE_LOG_MS), 10814 kicks, 10908 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10125 (texture frame 10126), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10127 (texture frame 10128), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10132 (texture frame 10133), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos rmbl::VibrationThread] hid: 32768 vibration values, strongest 0.00 since the last count [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10134 (texture frame 10135), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker4] nvn: 1100000 recordings made, 66797 live copies (22.6 MB), 5874 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10136 (texture frame 10137), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4760000 (3061605 pages from the captures) [hos host] vk: program picks: 6447819 kept (2190306 by the last picks' textures, a uniform buffer bound since), 26189 new, 2925991 with other textures (of 9400000) [hos host] vk: (its written part: 3 of 3 KB readable, 1630 bytes not zero; levels L0) [hos host] vk: texture 0x11286a9400 3072x2096x1 x1 levels 1 NVN 0x4b (6336 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10138 (texture frame 10139), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10140 (texture frame 10141), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10142 (texture frame 10143), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10151 (texture frame 10152), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10153 (texture frame 10154), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10155 (texture frame 10156), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos Presentation Thread] vk: capture: a submit 0.85 ms (lock 0.08, scan 0.09 over 604 ranges, 3428 MB), 1043 pages written, 1043 copied (4.1 MB) [hos host] vk: a capture taken (ms): all 0.076; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.047, bytes copied 0.000; 25 kept (188 MB), 11 spare buffers, 3818 texture pages copied (14 MB) [hos host] vk: sparse mirrors: 107, 1065 MB with memory of 1484 MB (419 MB spared); blocks given memory later 7748, given back 10466; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 600 binds (184.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 155 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 55 spares now, 1144 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 10327 dirty pages from their submits' captures, 66398 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1002 dispatches, 44795 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8136 runs, 714.6 MB compared; 448 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16205 in captures, 118 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: replay reads of guest memory as submitted: 4780000 (3071388 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10157 (texture frame 10158), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10159 (texture frame 10160), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: presenter thread: the copy's fence 2.35 ms, the present 0.83 ms a present [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10164 (texture frame 10165), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10166 (texture frame 10167), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10168 (texture frame 10169), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10170 (texture frame 10171), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10172 (texture frame 10173), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10174 (texture frame 10175), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4800000 (3081823 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10176 (texture frame 10177), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10178 (texture frame 10179), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10180 (texture frame 10181), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10182 (texture frame 10183), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10184 (texture frame 10185), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10186 (texture frame 10187), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4820000 (3094061 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10193 (texture frame 10194), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10195 (texture frame 10196), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1403 dirty pages from their submits' captures, 52279 from earlier ones', 1172 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 853 dispatches, 41256 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8210 runs, 714.5 MB compared; 60 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16717 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 887; 0x7f9ac00000 (20 MB, captured, state 2): 285; [hos host] vk: reaper: 648 images (1388 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 221 (376 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.07, writes collected 0.18 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10197 (texture frame 10198), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: memory: working set 15561 MB (peak 15827), private 21495 MB, staging 80 MB in 5 arenas, 40 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 844 MB; used 3280 of 15579 MB; heap 1 (device): 159 blocks 4823 MB, 19030 allocations 4239 MB; used 6818 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 100 images 304 MB; 61 free ranges, the largest 24 MB [hos host] vk: image blocks (textures, compacted): 50 blocks 1600 MB, 1166 images 1540 MB; 267 free ranges, the largest 19 MB [hos host] vk: render target images: 5719 reused from the pool, 3265 made new, 3163 destroyed from it, 989 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1139 with the pool empty (0 of them with targets still in the dead queue), 286 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 501 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2971 idle [hos host] vk: VRAM: 6818 MB used of 11474 MB allowed (583 MB of it free in the blocks); textures 2460 MB (budget 7158 MB), 1264 textures [hos host] vk: budget: VRAM 4823 of 11474 MB (6818 MB by the driver, 1995 MB unaccounted): textures 2460, mirrors 1065, blocks free 583, shadows 292, other buffers 211, other images 111, staging 80, dead images 19; RAM 3208 MB (the process's private memory 21495 MB): imports 2280, capture spares 672, captures 152, retired imports 52, capture table 32, other buffers 20, others 0 [hos host] vk: textures: decoded ASTC 1391 MB, rendered 527 MB (94 drawn in one frame, kept: 42654 KB), other 508 MB; used in the last 8 presents 1692 MB [hos host] vk: decoded ASTC used again: 66 textures (184 MB) copied from ones the game finalized, not decoded; 66 of those retired, 0 listed; 1154 MB of blocks hashed in 198 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 5053 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x111b189e00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 22 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 18 (7500 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 4 (703 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); 1466x282x1 NVN 0x25: 4 (6459 KB); [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 22376 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos Presentation Thread] vk: presents after the CPU's signal (600): 590 waited for it, 6.84 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10199 (texture frame 10200), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: status: 600 frames, the slowest 146.0 ms; VRAM 6.7 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6151 queued, 6151 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15240 ms and 330.9 MB read in all); 8346 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11699 graphics pipelines (1 not in a cache), 711 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2147 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 8985 (6151 ready, 0 loading), 16724 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13162 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 759 known, 2 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 180 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 14.10, present 0.18, draw 7.86, descriptors 3.46, texture lookups 0.21, uploads 0.32, pipelines 0.32, programs 1.37, commands 2.12, dispatch 1.01, memory sync 1.51, queue submit 0.00, slot wait 0.27 [hos host] vk: a frame: mirrors 317 (1065 MB): 8.55 MB synced in 1078.2 copies, 17.87 MB written back [hos host] vk: 10200 presents, 42.7 fps; a frame: 2042 draws (0 skipped), 162 dispatches, 177 passes, 131 barriers, 6 submits, 1.3 uploads (1.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13553 pipelines (3509 ms), 0 programs (10 failed), 1264 textures (2460 MB) [hos host] vk: host time in GPU resource waits: 9.27 ms a frame (fences 2.34, timelines 6.93, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 691.3 full barriers recorded, 294.7 left out (nothing recorded since the one before), 13.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.4, an op 1217.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 13.8 ms a frame (the longest submission 15.3 ms); 0.8 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.8 (0.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.47 MB unswizzled on the CPU in 0.25 ms [hos host] vk: the render thread (600 presents, 20.52 ms a frame): its jobs 19.46 ms a frame (submits 1.69, presents 6.76, others 11.00), idle 1.07; the game waited for room at 36 presents, 0.60 ms each (0.04 a frame; 14.40 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 591 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 2985 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 18.66 ms after its hand-over (55.41 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.99 + 0.28 (the window's copy is in the idle), after it 2.03 + 0.63; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.62 parts a frame recorded in the signal's job, 70.5% of them took nothing that came with the signal; of the others, 2.54 a frame took pages of the signal's capture into a mirror (80.5 pages a frame) and 0.00 pages read as guest memory had them (2.4 pages a frame); a frame's first part took nothing in 0 of 589 frames, its first three in 0 [hos host] vk: frame ahead, what came with the signal: 590 signals with a capture, each of 367.2 pages in 12.8 pools; between a frame's last submit and that signal 0.54 texture and sampler registrations, 0.64 releases and 0.00 other CPU signals a signal (55 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.40+0.11, 0.17+0.12, 0.05+0.03, 0.10+0.03, 0.95+0.11, 0.12+0.18, 0.10+0.03, 0.10+0.02, 0.01+0.01, 0.01+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10201 (texture frame 10202), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10203 (texture frame 10204), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10208 (texture frame 10209), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: render targets drawn over from another start, let go: 9765 so far (46318 MB), 8750 of them (41596 MB) between the passes [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10210 (texture frame 10211), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4840000 (3103953 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10212 (texture frame 10213), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: late pages found at the replay (600 takes): 449 pages in 5 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10217 (texture frame 10218), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10219 (texture frame 10220), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10221 (texture frame 10222), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10226 (texture frame 10227), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10228 (texture frame 10229), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10230 (texture frame 10231), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4860000 (3112665 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 3125944 pages so far (0 held: may be written ahead in their submission) [hos host] vk: program picks: 6596448 kept (2259090 by the last picks' textures, a uniform buffer bound since), 26189 new, 2977362 with other textures (of 9600000) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10235 (texture frame 10236), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: sparse mirrors: 107, 1065 MB with memory of 1484 MB (419 MB spared); blocks given memory later 7748, given back 10466; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 600 binds (184.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 155 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 55 spares now, 1144 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 461 dirty pages from their submits' captures, 55761 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 43219 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8438 runs, 732.5 MB compared; 52 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17750 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10242 (texture frame 10243), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10244 (texture frame 10245), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: CommandBufferSignalEvent #95000: event 0x100d6c8040 mode 0 value 94999 (lr 0x806621e8) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10246 (texture frame 10247), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 606 ranges, 3438 MB), 244 pages written, 244 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (82 MB), 11 spare buffers, 4231 texture pages copied (16 MB) [hos host] vk: replay reads of guest memory as submitted: 4880000 (3121599 pages from the captures) [hos sead::MainThread] vk: readbacks of GPU results: 1042 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1476 kept their virtual core (nothing there to wait for), 2876 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10260 (texture frame 10261), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10264 (texture frame 10265), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10266 (texture frame 10267), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4900000 (3130480 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10278 (texture frame 10279), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 244 dirty pages from their submits' captures, 56958 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 893 dispatches, 44246 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8880 runs, 768.9 MB compared; 45 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17826 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10280 (texture frame 10281), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10282 (texture frame 10283), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10294 (texture frame 10295), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10296 (texture frame 10297), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: CommandBufferWaitEvent #140000: event 0x100d6c802c mode 0 value 95982 (now 95963) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 4920000 (3139328 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10298 (texture frame 10299), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10300 (texture frame 10301), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5568 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos Presentation Thread] vi: 10305 frames presented (48.9 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 19, 15.5-17.8 122, 17.8-22 58, 22-30 34, 30-40 6, over 40 6 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 4940000 (3149711 pages from the captures) [hos host] vk: a frame: 180 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.84, present 0.16, draw 6.30, descriptors 2.71, texture lookups 0.10, uploads 0.14, pipelines 0.26, programs 1.12, commands 1.71, dispatch 0.85, memory sync 0.42, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 318 (1065 MB): 5.29 MB synced in 1042.7 copies, 17.88 MB written back [hos host] vk: 10320 presents, 60.4 fps; a frame: 1878 draws (0 skipped), 164 dispatches, 158 passes, 131 barriers, 6 submits, 0.5 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13553 pipelines (3509 ms), 0 programs (10 failed), 1282 textures (2462 MB) [hos host] vk: host time in GPU resource waits: 1.82 ms a frame (fences 1.39, timelines 0.43, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 610.4 full barriers recorded, 273.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 1140.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.9 ms a frame (the longest submission 4.4 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.76 MB unswizzled on the CPU in 0.12 ms [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10320 (texture frame 10321), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: sparse mirrors: 108, 1065 MB with memory of 1488 MB (422 MB spared); blocks given memory later 7748, given back 10466; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 601 binds (185.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 156 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 54 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 261 dirty pages from their submits' captures, 55661 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 884 dispatches, 43008 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8527 runs, 733.6 MB compared; 47 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17427 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10322 (texture frame 10323), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10324 (texture frame 10325), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10326 (texture frame 10327), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4960000 (3161977 pages from the captures) [hos host] vk: program picks: 6745057 kept (2326487 by the last picks' textures, a uniform buffer bound since), 26189 new, 3028753 with other textures (of 9800000) [hos Presentation Thread] vk: capture: a submit 0.29 ms (lock 0.02, scan 0.05 over 596 ranges, 3440 MB), 658 pages written, 485 copied (1.9 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (282 MB), 12 spare buffers, 4627 texture pages copied (18 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10342 (texture frame 10343), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos OptimizeChangeBanc] nvn: 1150000 recordings made, 68335 live copies (24.1 MB), 5874 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10344 (texture frame 10345), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10346 (texture frame 10347), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10348 (texture frame 10349), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10350 (texture frame 10351), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 4980000 (3170716 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10352 (texture frame 10353), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: GPU waits in 5 s: 137 sync waits (724 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 16245 dirty pages from their submits' captures, 108939 from earlier ones', 40350 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1094 dispatches, 47348 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8529 runs, 733.2 MB compared; 260 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16632 in captures, 90 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: 0x7fab000000 (68 MB, captured, state 2): 16287; 0x7fb5c00000 (44 MB, captured, state 2): 10659; 0x7fb8800000 (52 MB, captured, state 2): 8665; 0x7fd3c00000 (16 MB, captured, state 2): 3252; 0x1145400000 (40 MB, captured, state 2): 533; 0x104e000000 (24 MB, captured, state 2): 226; 0x103cc00000 (12 MB, captured, state 2): 198; 0x1038400000 (4 MB, captured, state 2): 72; [hos host] vk: recorder: 5246293 packets run in 10.0 s (busy 1.53 s, spun idle 0.11 s, slept 10088 times, 3308 of them under 200 us), 0 calls made directly (no room); 15467 batches over 494 frames (31.3 a frame, 339 packets a batch), 7332 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 124 times, 8.49 ms (0.3 times, 0.02 ms a frame); every thread 662 times, 42.83 ms in all, the longest 2.42 ms; vkWaitForFences 422/25.22 ms (max 2.42), the queue's lock 46/5.37 ms (max 0.80), vkDestroyImageView 45/3.85 ms (max 0.84), vmaDestroyBuffer 56/3.21 ms (max 0.72), vkQueueBindSparse 72/2.33 ms (max 0.37), vmaFreeMemory 15/2.10 ms (max 0.45) [hos host] vk: recorder: of its busy 1.53 s, 0.39 s in parts' ends and submits (the longest 1.48 ms), 0.12 s of that blocked on the queue's lock (199 times, the longest 1.44 ms), so 0.27 s in the ends, submits and holds; 1.14 s running the other packets; 10620 packets a frame [hos host] vk: replay reads of guest memory as submitted: 5000000 (3179749 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10374 (texture frame 10375), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a submission's parts ended: 33785 where the game had ended its render pass, 6215 at 512 draws [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10376 (texture frame 10377), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10378 (texture frame 10379), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10380 (texture frame 10381), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10382 (texture frame 10383), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10384 (texture frame 10385), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5020000 (3188818 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.04, writes collected 0.19 [hos nn.npln.Worker] vcores: 150978 takes, 30278 waits (mean 67 us, longest 16.3 ms, 0 over HOS_VCORE_LOG_MS), 15559 kicks, 15572 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 108, 1135 MB with memory of 1488 MB (352 MB spared); blocks given memory later 8950, given back 10537; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 630 binds (188.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 156 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 54 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1778 dirty pages from their submits' captures, 52602 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 847 dispatches, 41823 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8251 runs, 714.2 MB compared; 71 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16766 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 181 live (2280 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10406 (texture frame 10407), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5040000 (3199876 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10408 (texture frame 10409), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10410 (texture frame 10411), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10412 (texture frame 10413), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1026 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 5 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1659 kept their virtual core (nothing there to wait for), 3205 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10414 (texture frame 10415), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10416 (texture frame 10417), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5060000 (3208884 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.22 ms (lock 0.03, scan 0.05 over 630 ranges, 3538 MB), 350 pages written, 350 copied (1.4 MB) [hos host] vk: a capture taken (ms): all 0.032; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (130 MB), 11 spare buffers, 4627 texture pages copied (18 MB) [hos host] vk: program picks: 6892419 kept (2392272 by the last picks' textures, a uniform buffer bound since), 26190 new, 3081390 with other textures (of 10000000) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15662 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10438 (texture frame 10439), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: readback pages refreshed from guest memory: 3193053 pages so far (0 held: may be written ahead in their submission) [hos host] vk: a frame: 196 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.21, present 0.16, draw 7.12, descriptors 3.02, texture lookups 0.07, uploads 0.12, pipelines 0.30, programs 1.28, commands 1.97, dispatch 0.94, memory sync 1.21, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 318 (1135 MB): 8.88 MB synced in 1056.6 copies, 17.87 MB written back [hos host] vk: 10440 presents, 52.6 fps; a frame: 1965 draws (0 skipped), 163 dispatches, 170 passes, 131 barriers, 6 submits, 0.2 uploads (0.7 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13553 pipelines (3509 ms), 0 programs (10 failed), 1282 textures (2462 MB) [hos host] vk: host time in GPU resource waits: 6.76 ms a frame (fences 1.93, timelines 4.83, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 660.5 full barriers recorded, 287.2 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1173.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 13.0 ms a frame (the longest submission 43.0 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.67 MB unswizzled on the CPU in 0.12 ms [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10440 (texture frame 10441), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1464 dirty pages from their submits' captures, 52649 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 845 dispatches, 41747 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8254 runs, 714.2 MB compared; 71 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16621 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10442 (texture frame 10443), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 200000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10444 (texture frame 10445), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10446 (texture frame 10447), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5080000 (3217827 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10448 (texture frame 10449), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: queues ordered by their events: 60000 submissions stopped at a wait, 59980 went on once it was reached, 18 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 2 (stopped 50.00 ms at most); waits met 143922, not met 0; captures at a CPU event's wait 7043, 0 of them at a wait met on arrival [hos host] vk: late pages found at the replay (600 takes): 15 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 5100000 (3226850 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10468 (texture frame 10469), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10470 (texture frame 10471), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10472 (texture frame 10473), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10474 (texture frame 10475), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10476 (texture frame 10477), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker5] nvn: 1200000 recordings made, 74180 live copies (23.6 MB), 5875 texture copies; 7 command buffers' segments (4.4 MB) [hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.5 ms, 1 frames after their copies were recorded, the first #16568 0x116dad4000 260x260 levels 9 NVN 0xd (30 so far, 6603 KB, 10 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10478 (texture frame 10479), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: sparse mirrors: 109, 1137 MB with memory of 1492 MB (354 MB spared); blocks given memory later 8950, given back 10537; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 631 binds (188.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 157 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 53 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1407 dirty pages from their submits' captures, 53531 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 861 dispatches, 41847 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8273 runs, 715.0 MB compared; 69 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16850 in captures, 1 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 182 live (2284 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos ModuleSystemWorker5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 21838561 lookups, 145583 missed (0.67 %) [hos host] vk: replay reads of guest memory as submitted: 5120000 (3236169 pages from the captures) [hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #145000: event 0x100d6c802c mode 0 value 99782 (now 99763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10500 (texture frame 10501), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10502 (texture frame 10503), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10504 (texture frame 10505), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5140000 (3245271 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10506 (texture frame 10507), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #100000: event 0x100d6c8000 mode 0 value 99999 (lr 0x80291440) [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 623 ranges, 3545 MB), 233 pages written, 233 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (118 MB), 10 spare buffers, 5326 texture pages copied (20 MB) [hos host] vk: replay reads of guest memory as submitted: 5160000 (3257834 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10522 (texture frame 10523), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1055 dirty pages from their submits' captures, 55587 from earlier ones', 8 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 898 dispatches, 43358 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8589 runs, 735.8 MB compared; 60 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17549 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7f9ac00000 (20 MB, captured, state 2): 8; [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10524 (texture frame 10525), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10526 (texture frame 10527), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: program picks: 7040380 kept (2458709 by the last picks' textures, a uniform buffer bound since), 26192 new, 3133427 with other textures (of 10200000) [hos host] vk: replay reads of guest memory as submitted: 5180000 (3268601 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10540 (texture frame 10541), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10542 (texture frame 10543), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10544 (texture frame 10545), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5200000 (3278373 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10558 (texture frame 10559), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a frame: 184 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.61, present 0.15, draw 6.86, descriptors 2.96, texture lookups 0.08, uploads 0.12, pipelines 0.29, programs 1.22, commands 1.85, dispatch 0.88, memory sync 0.46, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 320 (1142 MB): 5.44 MB synced in 1052.5 copies, 17.94 MB written back [hos host] vk: 10560 presents, 60.9 fps; a frame: 1953 draws (0 skipped), 167 dispatches, 169 passes, 131 barriers, 6 submits, 0.3 uploads (0.7 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13553 pipelines (3509 ms), 0 programs (10 failed), 1296 textures (2474 MB) [hos host] vk: host time in GPU resource waits: 3.14 ms a frame (fences 1.34, timelines 1.80, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 658.6 full barriers recorded, 285.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1170.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.5 ms a frame (the longest submission 4.7 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.66 MB unswizzled on the CPU in 0.10 ms [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10560 (texture frame 10561), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10562 (texture frame 10563), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: sparse mirrors: 110, 1142 MB with memory of 1500 MB (357 MB spared); blocks given memory later 8950, given back 10543; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 634 binds (189.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 52 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 160 dirty pages from their submits' captures, 57765 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 929 dispatches, 44582 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8872 runs, 755.9 MB compared; 52 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18037 in captures, 1 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: 0x7f8a800000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 182 live (2284 MB), 3 retired still imported (52 MB), 37 retired given back (832 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 8055 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1112 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1495 kept their virtual core (nothing there to wait for), 2857 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10574 (texture frame 10575), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10576 (texture frame 10577), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5220000 (3288336 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10578 (texture frame 10579), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10580 (texture frame 10581), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos BaseProcExecutor0] vcores: BaseProcExecutor0 (priority 16) waited 80.5 ms for a core: given by sead::MainThread (priority 16), which held it 80.6 ms and let it go 77.6 ms after it was asked; the cores it may use now: core 0: it; [hos Presentation Thread] vi: 10583 frames presented (54.7 fps over the last 5.1 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 44, 15.5-17.8 175, 17.8-22 33, 22-30 19, 30-40 3, over 40 4 [hos host] vk: late pages found at the replay (600 takes): 35 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 5240000 (3300109 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.14 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10598 (texture frame 10599), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10600 (texture frame 10601), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10602 (texture frame 10603), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10604 (texture frame 10605), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 4664 dirty pages from their submits' captures, 135697 from earlier ones', 6 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1104 dispatches, 49783 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8709 runs, 739.0 MB compared; 774 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16630 in captures, 534 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x7f9ac00000 (20 MB, captured, state 2): 6; [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10606 (texture frame 10607), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos Presentation Thread] vk: capture: a submit 0.44 ms (lock 0.02, scan 0.06 over 598 ranges, 3461 MB), 652 pages written, 652 copied (2.5 MB) [hos host] vk: a capture taken (ms): all 0.033; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 24 kept (90 MB), 12 spare buffers, 6126 texture pages copied (23 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10608 (texture frame 10609), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5260000 (3311953 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 5280000 (3324974 pages from the captures) [hos host] vk: program picks: 7189455 kept (2526672 by the last picks' textures, a uniform buffer bound since), 26201 new, 3184343 with other textures (of 10400000) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10628 (texture frame 10629), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10630 (texture frame 10631), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10632 (texture frame 10633), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10634 (texture frame 10635), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: GPU waits in 5 s: 208 sync waits (879 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10636 (texture frame 10637), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5300000 (3335137 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 3260610 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 110, 1070 MB with memory of 1500 MB (429 MB spared); blocks given memory later 9032, given back 11766; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 645 binds (191.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 210 kept as spares, 0 destroyed; 52 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1562 dirty pages from their submits' captures, 55625 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 915 dispatches, 43658 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8552 runs, 721.6 MB compared; 72 pages read into the shadows; 292 MB of shadows (47 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17829 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 182 live (2284 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10652 (texture frame 10653), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10654 (texture frame 10655), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10656 (texture frame 10657), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: 1250000 recordings made, 69744 live copies (21.4 MB), 5875 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: replay reads of guest memory as submitted: 5320000 (3344520 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10670 (texture frame 10671), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10672 (texture frame 10673), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10674 (texture frame 10675), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a frame: 183 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.83, present 0.16, draw 6.77, descriptors 2.93, texture lookups 0.10, uploads 0.15, pipelines 0.28, programs 1.21, commands 1.84, dispatch 0.92, memory sync 0.99, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 320 (1070 MB): 8.21 MB synced in 1073.5 copies, 18.04 MB written back [hos host] vk: 10680 presents, 51.8 fps; a frame: 1959 draws (0 skipped), 173 dispatches, 168 passes, 131 barriers, 6 submits, 0.4 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13554 pipelines (3509 ms), 0 programs (10 failed), 1322 textures (2516 MB) [hos host] vk: host time in GPU resource waits: 5.12 ms a frame (fences 2.24, timelines 2.88, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 659.6 full barriers recorded, 283.9 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.3, an op 1165.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.1 ms a frame (the longest submission 16.4 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.70 MB unswizzled on the CPU in 0.12 ms [hos AudioFrameworkThread] vcores: 148870 takes, 24741 waits (mean 75 us, longest 80.5 ms, 1 over HOS_VCORE_LOG_MS), 12169 kicks, 12197 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 5340000 (3355741 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10686 (texture frame 10687), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10688 (texture frame 10689), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 118 dirty pages from their submits' captures, 60247 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 948 dispatches, 44887 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9198 runs, 776.8 MB compared; 70 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18292 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10690 (texture frame 10691), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12723 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10692 (texture frame 10693), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10694 (texture frame 10695), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #150000: event 0x100d6c802c mode 0 value 103582 (now 103563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 595 ranges, 3471 MB), 254 pages written, 254 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 6126 texture pages copied (23 MB) [hos host] vk: replay reads of guest memory as submitted: 5360000 (3365347 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 345 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10708 (texture frame 10709), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10710 (texture frame 10711), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10712 (texture frame 10713), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5380000 (3376705 pages from the captures) [hos host] vk: program picks: 7341009 kept (2598217 by the last picks' textures, a uniform buffer bound since), 26203 new, 3232787 with other textures (of 10600000) [hos host] vk: sparse mirrors: 109, 1059 MB with memory of 1492 MB (432 MB spared); blocks given memory later 9032, given back 11888; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 648 binds (191.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 211 kept as spares, 0 destroyed; 53 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 852 dirty pages from their submits' captures, 56060 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 908 dispatches, 43573 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8884 runs, 742.3 MB compared; 62 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17774 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 182 live (2284 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10734 (texture frame 10735), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1067 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1606 kept their virtual core (nothing there to wait for), 3002 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10736 (texture frame 10737), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5400000 (3387557 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10750 (texture frame 10751), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10752 (texture frame 10753), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10754 (texture frame 10755), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5420000 (3396904 pages from the captures) [hos host] vk: presenter thread: the copy's fence 1.64 ms, the present 0.76 ms a present [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10766 (texture frame 10767), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10768 (texture frame 10769), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10770 (texture frame 10771), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #105000: event 0x100d6c8014 mode 0 value 105001 (lr 0x80660c94) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 454 dirty pages from their submits' captures, 57406 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 925 dispatches, 44504 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9079 runs, 760.1 MB compared; 46 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18312 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10772 (texture frame 10773), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5440000 (3406288 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10788 (texture frame 10789), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.6 ms, 1 frames after their copies were recorded, the first #16605 0x116dd8f000 260x260 levels 9 NVN 0xd (31 so far, 6813 KB, 11 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10790 (texture frame 10791), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10792 (texture frame 10793), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.05 over 596 ranges, 3471 MB), 293 pages written, 293 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (176 MB), 11 spare buffers, 7161 texture pages copied (27 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10794 (texture frame 10795), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5460000 (3417732 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: memory: working set 15275 MB (peak 15952), private 21269 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 896 MB, 36 allocations 822 MB; used 3232 of 15579 MB; heap 1 (device): 159 blocks 4823 MB, 19000 allocations 4256 MB; used 6818 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 102 images 304 MB; 61 free ranges, the largest 24 MB [hos host] vk: image blocks (textures, compacted): 50 blocks 1600 MB, 1220 images 1577 MB; 271 free ranges, the largest 4 MB [hos host] vk: render target images: 5719 reused from the pool, 3267 made new, 3163 destroyed from it, 989 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1141 with the pool empty (0 of them with targets still in the dead queue), 286 with none in it a present yet, 25 with none of a kind they can use, 463 with all smaller (the largest 51% short on average), 501 with all bigger (over 1.5 times), 850 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2971 idle [hos host] vk: VRAM: 6818 MB used of 11474 MB allowed (566 MB of it free in the blocks); textures 2518 MB (budget 7158 MB), 1328 textures [hos host] vk: budget: VRAM 4823 of 11474 MB (6818 MB by the driver, 1995 MB unaccounted): textures 2518, mirrors 1059, blocks free 566, shadows 292, other buffers 211, other images 111, staging 64; RAM 3180 MB (the process's private memory 21269 MB): imports 2284, capture spares 722, captures 110, retired imports 32, capture table 32, others 0 [hos host] vk: textures: decoded ASTC 1439 MB, rendered 527 MB (95 drawn in one frame, kept: 42890 KB), other 518 MB; used in the last 8 presents 1696 MB [hos host] vk: decoded ASTC used again: 79 textures (195 MB) copied from ones the game finalized, not decoded; 79 of those retired, 0 listed; 1165 MB of blocks hashed in 201 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 5653 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x111bd8a000 4096x4096x1 x1 levels 1 NVN 0x4b, last used 733 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 18 (7500 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 4 (703 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); 1466x282x1 NVN 0x25: 4 (6459 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 325 waited for it, 3.92 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 124.9 ms; VRAM 6.7 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 9 new graphics pipelines, pipelines from the driver's cache 11, from the start's cache 0, compiled 2; 9 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6153 queued, 6153 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15243 ms and 331.0 MB read in all); 8348 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11700 graphics pipelines (1 not in a cache), 712 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2158 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 8987 (6153 ready, 0 loading), 16669 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13164 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 760 known, 3 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 178 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.41, present 0.16, draw 6.80, descriptors 2.94, texture lookups 0.07, uploads 0.10, pipelines 0.29, programs 1.20, commands 1.85, dispatch 0.91, memory sync 0.48, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 320 (1059 MB): 5.52 MB synced in 1060.0 copies, 18.11 MB written back [hos host] vk: 10800 presents, 60.0 fps; a frame: 1991 draws (0 skipped), 176 dispatches, 167 passes, 131 barriers, 6 submits, 0.2 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13554 pipelines (3509 ms), 0 programs (10 failed), 1328 textures (2518 MB) [hos host] vk: host time in GPU resource waits: 2.13 ms a frame (fences 1.51, timelines 0.62, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 654.8 full barriers recorded, 282.3 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.3, an op 1168.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.4 ms a frame (the longest submission 18.6 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.63 MB unswizzled on the CPU in 0.10 ms [hos host] vk: queues ordered by their events: 62591 submissions stopped at a wait, 62559 went on once it was reached, 32 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 3 (stopped 50.30 ms at most); waits met 152524, not met 0; captures at a CPU event's wait 7218, 0 of them at a wait met on arrival [hos host] vk: the render thread (600 presents, 17.59 ms a frame): its jobs 16.01 ms a frame (submits 5.57, presents 4.90, others 5.54), idle 1.58; the game waited for room at 19 presents, 0.92 ms each (0.03 a frame; 7.80 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 327 captures at a CPU signal; the submits' waits read in 0.044 ms a frame; readbacks: 1717 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.19 ms after its hand-over (48.10 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.79 + 0.33 (the window's copy is in the idle), after it 1.30 + 0.38; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.01 parts a frame recorded in the signal's job, 73.6% of them took nothing that came with the signal; of the others, 2.11 a frame took pages of the signal's capture into a mirror (91.8 pages a frame) and 0.03 pages read as guest memory had them (123.4 pages a frame); a frame's first part took nothing in 22 of 327 frames, its first three in 21 [hos host] vk: frame ahead, what came with the signal: 328 signals with a capture, each of 509.0 pages in 12.2 pools; between a frame's last submit and that signal 0.09 texture and sampler registrations, 0.01 releases and 0.00 other CPU signals a signal (14 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.26+0.11, 0.30+0.13, 0.05+0.02, 0.08+0.02, 0.45+0.05, 0.05+0.03, 0.09+0.02, 0.02+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9554 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10806 (texture frame 10807), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10808 (texture frame 10809), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: render targets drawn over from another start, let go: 9765 so far (46318 MB), 8750 of them (41596 MB) between the passes [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10810 (texture frame 10811), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5480000 (3430038 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 109, 1059 MB with memory of 1492 MB (432 MB spared); blocks given memory later 9032, given back 11890; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 649 binds (191.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 211 kept as spares, 0 destroyed; 53 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 312 dirty pages from their submits' captures, 61374 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 939 dispatches, 44958 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9116 runs, 761.2 MB compared; 57 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18517 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 182 live (2284 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 7493383 kept (2672394 by the last picks' textures, a uniform buffer bound since), 26208 new, 3280408 with other textures (of 10800000) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10824 (texture frame 10825), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10826 (texture frame 10827), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10828 (texture frame 10829), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10830 (texture frame 10831), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5500000 (3440862 pages from the captures) [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 220000 so far (x1 0x117a4f2040 x2 0x6001d40 x3 0x117a4f2000 x4 0xf33080) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10842 (texture frame 10843), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10844 (texture frame 10845), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10846 (texture frame 10847), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5520000 (3449941 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 3327894 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 187 dirty pages from their submits' captures, 57793 from earlier ones', 12 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 948 dispatches, 44415 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9058 runs, 759.7 MB compared; 66 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18165 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7f83c00000 (4 MB, captured, state 2): 12; [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10860 (texture frame 10861), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10862 (texture frame 10863), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10864 (texture frame 10865), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5540000 (3461658 pages from the captures) [hos Presentation Thread] vi: 10876 frames presented (58.5 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 35, 15.5-17.8 231, 17.8-22 16, 22-30 6, 30-40 4, over 40 1 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10878 (texture frame 10879), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10880 (texture frame 10881), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5560000 (3472007 pages from the captures) [hos host] vk: reaper: 29 images (48 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 8 (19 MB) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 589 ranges, 3175 MB), 275 pages written, 275 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 7513 texture pages copied (29 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10896 (texture frame 10897), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #155000: event 0x100d6c802c mode 0 value 107382 (now 107363) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: sparse mirrors: 109, 1059 MB with memory of 1492 MB (432 MB spared); blocks given memory later 9034, given back 11899; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 652 binds (192.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 158 taken from the spares, 162 made new; 211 kept as spares, 0 destroyed; 53 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 143 dirty pages from their submits' captures, 57900 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 941 dispatches, 44420 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9046 runs, 759.5 MB compared; 56 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18235 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 182 live (2284 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10898 (texture frame 10899), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1155 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1658 kept their virtual core (nothing there to wait for), 2950 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10900 (texture frame 10901), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10902 (texture frame 10903), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10904 (texture frame 10905), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5580000 (3483571 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10906 (texture frame 10907), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11560 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 7645334 kept (2746963 by the last picks' textures, a uniform buffer bound since), 26208 new, 3328457 with other textures (of 11000000) [hos host] vk: a frame: 173 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.02, present 0.16, draw 6.33, descriptors 2.76, texture lookups 0.07, uploads 0.09, pipelines 0.27, programs 1.09, commands 1.73, dispatch 0.88, memory sync 0.49, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 320 (1059 MB): 5.42 MB synced in 1059.0 copies, 18.09 MB written back [hos host] vk: 10920 presents, 60.0 fps; a frame: 1968 draws (0 skipped), 175 dispatches, 162 passes, 131 barriers, 6 submits, 0.2 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13554 pipelines (3509 ms), 0 programs (10 failed), 1328 textures (2518 MB) [hos host] vk: host time in GPU resource waits: 1.42 ms a frame (fences 1.39, timelines 0.02, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 634.1 full barriers recorded, 277.1 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1118.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.0 ms a frame (the longest submission 4.3 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.62 MB unswizzled on the CPU in 0.09 ms [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 5600000 (3495708 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10922 (texture frame 10923), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10924 (texture frame 10925), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10926 (texture frame 10927), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10928 (texture frame 10929), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10935 (texture frame 10936), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10937 (texture frame 10938), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 192 dirty pages from their submits' captures, 56914 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 920 dispatches, 43682 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8878 runs, 740.6 MB compared; 51 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17668 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10939 (texture frame 10940), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5620000 (3508225 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10941 (texture frame 10942), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: GPU waits in 5 s: 45 sync waits (92 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10946 (texture frame 10947), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10948 (texture frame 10949), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: recorder: 5972872 packets run in 10.0 s (busy 1.72 s, spun idle 0.11 s, slept 10383 times, 3466 of them under 200 us), 0 calls made directly (no room); 17526 batches over 580 frames (30.2 a frame, 341 packets a batch), 8435 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 41 times, 3.25 ms (0.1 times, 0.01 ms a frame); every thread 602 times, 44.73 ms in all, the longest 0.76 ms; vkWaitForFences 502/34.97 ms (max 0.59), vmaDestroyBuffer 41/4.42 ms (max 0.76), vmaFreeMemory 14/2.60 ms (max 0.67), the queue's lock 19/1.64 ms (max 0.22), vkQueueBindSparse 23/0.60 ms (max 0.16), a part's submit 1/0.39 ms (max 0.39) [hos host] vk: recorder: of its busy 1.72 s, 0.43 s in parts' ends and submits (the longest 1.17 ms), 0.12 s of that blocked on the queue's lock (259 times, the longest 1.12 ms), so 0.31 s in the ends, submits and holds; 1.29 s running the other packets; 10298 packets a frame [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10950 (texture frame 10951), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10952 (texture frame 10953), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10957 (texture frame 10958), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5640000 (3520786 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10959 (texture frame 10960), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10961 (texture frame 10962), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10964 (texture frame 10965), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10966 (texture frame 10967), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10968 (texture frame 10969), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10975 (texture frame 10976), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10977 (texture frame 10978), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5660000 (3532521 pages from the captures) [hos host] vk: sparse mirrors: 111, 1060 MB with memory of 1500 MB (439 MB spared); blocks given memory later 9034, given back 11900; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 655 binds (193.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 160 taken from the spares, 162 made new; 211 kept as spares, 0 destroyed; 51 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 860 dirty pages from their submits' captures, 60135 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 927 dispatches, 44836 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9092 runs, 757.5 MB compared; 57 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 95 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18009 in captures, 3 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7f83c00000 (4 MB, captured, state 2): 2; 0x7fe2800000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 183 live (2288 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos BaseProcExecutor3] vcores: 152285 takes, 23508 waits (mean 68 us, longest 6.5 ms, 0 over HOS_VCORE_LOG_MS), 11165 kicks, 10966 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10984 (texture frame 10985), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10986 (texture frame 10987), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.04 over 587 ranges, 3167 MB), 299 pages written, 299 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (134 MB), 11 spare buffers, 7675 texture pages copied (29 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10988 (texture frame 10989), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10993 (texture frame 10994), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10995 (texture frame 10996), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 10997 (texture frame 10998), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5680000 (3544826 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11002 (texture frame 11003), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11004 (texture frame 11005), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11006 (texture frame 11007), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11011 (texture frame 11012), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11013 (texture frame 11014), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11015 (texture frame 11016), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5700000 (3556983 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 23394 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11017 (texture frame 11018), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11019 (texture frame 11020), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: program picks: 7795370 kept (2818288 by the last picks' textures, a uniform buffer bound since), 26212 new, 3378417 with other textures (of 11200000) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11021 (texture frame 11022), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 425 dirty pages from their submits' captures, 56320 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 916 dispatches, 43524 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8656 runs, 736.9 MB compared; 54 pages read into the shadows; 292 MB of shadows (48 MB of pages known, in 94 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17700 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 1; [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11023 (texture frame 11024), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11025 (texture frame 11026), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11027 (texture frame 11028), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 5720000 (3569823 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11032 (texture frame 11033), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #110000: event 0x100d6c801c mode 0 value 110000 (lr 0x802915c8) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11034 (texture frame 11035), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.6 ms, 1 frames after their copies were recorded, the first #16647 0x116e006000 260x260 levels 9 NVN 0xd (32 so far, 7022 KB, 11 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11036 (texture frame 11037), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11038 (texture frame 11039), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a frame: 159 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.23, present 0.16, draw 6.51, descriptors 2.91, texture lookups 0.13, uploads 0.20, pipelines 0.26, programs 1.11, commands 1.73, dispatch 0.91, memory sync 0.55, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 323 (1063 MB): 5.55 MB synced in 1060.1 copies, 17.98 MB written back [hos host] vk: 11040 presents, 59.8 fps; a frame: 1816 draws (0 skipped), 170 dispatches, 170 passes, 132 barriers, 6 submits, 0.6 uploads (1.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1364 textures (2555 MB) [hos host] vk: host time in GPU resource waits: 1.92 ms a frame (fences 1.24, timelines 0.67, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 629.9 full barriers recorded, 278.4 left out (nothing recorded since the one before), 23.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.9; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.9, an op 1049.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.1 ms a frame (the longest submission 4.8 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.25 MB unswizzled on the CPU in 0.18 ms [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11040 (texture frame 11041), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11045 (texture frame 11046), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11047 (texture frame 11048), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11049 (texture frame 11050), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5740000 (3583051 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11051 (texture frame 11052), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11053 (texture frame 11054), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11055 (texture frame 11056), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11057 (texture frame 11058), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1117 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1713 kept their virtual core (nothing there to wait for), 3151 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: readback pages refreshed from guest memory: 3395639 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 111, 1049 MB with memory of 1488 MB (438 MB spared); blocks given memory later 9034, given back 11900; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 657 binds (193.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 161 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 51 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 28 bindings followed, 39 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1284 dirty pages from their submits' captures, 54301 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 42410 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7723 runs, 708.1 MB compared; 74 pages read into the shadows; 292 MB of shadows (47 MB of pages known, in 92 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17607 in captures, 1 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 183 live (2288 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11062 (texture frame 11063), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 18964628 lookups, 140714 missed (0.74 %) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11064 (texture frame 11065), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11066 (texture frame 11067), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11068 (texture frame 11069), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5760000 (3595251 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11070 (texture frame 11071), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11072 (texture frame 11073), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11074 (texture frame 11075), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11076 (texture frame 11077), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: pool 0x1163f6e000+0x7e9000 imported in 2.2 ms as 0x1163c00000..0x1164800000 (222 imports, 685 ms in all) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 579 ranges, 3131 MB), 279 pages written, 279 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.037; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 25 kept (132 MB), 12 spare buffers, 8301 texture pages copied (32 MB) [hos host] vk: late pages' fix (600 recordings with one): of 1013 captures with late pages 1013 fixed (95.7 pages, 96.5 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 442 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 544 blocks (34816 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 162, of an earlier recording 280, not recorded yet 0, of this part 0; pages fixed from stamps 95932, from snapshots 1017; the shadows: 546305 pages synced from captures, 528079 made known by it, 18226 left unknown because the replay or the collection wrote them since [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11078 (texture frame 11079), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: 1 textures retired (1367 left, 2567 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11080 (texture frame 11081), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11082 (texture frame 11083), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11084 (texture frame 11085), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11086 (texture frame 11087), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11088 (texture frame 11089), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11091 (texture frame 11092), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5780000 (3606260 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11093 (texture frame 11094), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11095 (texture frame 11096), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #160000: event 0x100d6c802c mode 0 value 111182 (now 111163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06 [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11097 (texture frame 11098), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11099 (texture frame 11100), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1856 dirty pages from their submits' captures, 55141 from earlier ones', 1013 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 866 dispatches, 41469 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7471 runs, 705.7 MB compared; 2096 pages read into the shadows; 304 MB of shadows (54 MB of pages known, in 99 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18046 in captures, 40 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: 0x1163c00000 (12 MB, captured, state 2): 1013; [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11101 (texture frame 11102), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11103 (texture frame 11104), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11105 (texture frame 11106), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11110 (texture frame 11111), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11112 (texture frame 11113), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5800000 (3618601 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11114 (texture frame 11115), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11116 (texture frame 11117), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11118 (texture frame 11119), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: program picks: 7948036 kept (2891733 by the last picks' textures, a uniform buffer bound since), 26216 new, 3425747 with other textures (of 11400000) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11120 (texture frame 11121), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11122 (texture frame 11123), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos sead::MainThread] nvn: 1300000 recordings made, 70235 live copies (23.0 MB), 5878 texture copies; 4 command buffers' segments (4.4 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11124 (texture frame 11125), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11126 (texture frame 11127), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11128 (texture frame 11129), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11130 (texture frame 11131), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11132 (texture frame 11133), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11134 (texture frame 11135), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11136 (texture frame 11137), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5820000 (3628447 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11138 (texture frame 11139), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: sparse mirrors: 112, 1063 MB with memory of 1500 MB (436 MB spared); blocks given memory later 9115, given back 11938; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 660 binds (193.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 162 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 50 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 974, 4327.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1368 dirty pages from their submits' captures, 52931 from earlier ones', 1013 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 883 dispatches, 41064 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7454 runs, 698.0 MB compared; 79 pages read into the shadows; 304 MB of shadows (51 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18416 in captures, 40 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: 0x1163c00000 (12 MB, captured, state 2): 1013; [hos host] vk: imports: 184 live (2300 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11140 (texture frame 11141), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11142 (texture frame 11143), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11144 (texture frame 11145), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11146 (texture frame 11147), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11148 (texture frame 11149), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: a frame: 147 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.57, present 0.16, draw 7.51, descriptors 3.28, texture lookups 0.08, uploads 0.24, pipelines 0.31, programs 1.34, commands 2.01, dispatch 0.95, memory sync 0.58, queue submit 0.00, slot wait 0.29 [hos host] vk: a frame: mirrors 325 (1079 MB): 5.80 MB synced in 1053.5 copies, 17.89 MB written back [hos host] vk: 11160 presents, 57.9 fps; a frame: 1946 draws (0 skipped), 149 dispatches, 187 passes, 131 barriers, 6 submits, 0.4 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1367 textures (2567 MB) [hos host] vk: host time in GPU resource waits: 4.96 ms a frame (fences 1.21, timelines 3.75, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 683.4 full barriers recorded, 296.6 left out (nothing recorded since the one before), 21.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1048.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.1 ms a frame (the longest submission 5.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.1 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.47 MB unswizzled on the CPU in 0.24 ms [hos ModuleSystemWorker4] barrier: 180943 looks: 66255 found the masks moved already, 112380 waited until they moved, 2308 for the whole 1 ms (a wait 273 us on average); 188127 wakes [hos host] vk: replay reads of guest memory as submitted: 5840000 (3638638 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 569 ranges, 3120 MB), 261 pages written, 261 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (118 MB), 12 spare buffers, 2386 texture pages copied (9 MB) [hos Presentation Thread] vi: 11172 frames presented (59.2 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 33, 15.5-17.8 210, 17.8-22 51, 22-30 2, 30-40 0, over 40 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6542 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11175 (texture frame 11176), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1371 dirty pages from their submits' captures, 51150 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 852 dispatches, 39992 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7346 runs, 672.9 MB compared; 67 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16232 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11177 (texture frame 11178), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11179 (texture frame 11180), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: replay reads of guest memory as submitted: 5860000 (3650630 pages from the captures) [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11181 (texture frame 11182), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11183 (texture frame 11184), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: upload over a drawn image: #16215 0x10001e00400 1280x720x1 x1 NVN 0x3f at present 11185 (texture frame 11186), finalized, 0 KB dirty in 0 ranges, a GPU buffer wrote it [hos host] vk: late pages found at the replay (600 takes): 12 pages in 12 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: replay reads of guest memory as submitted: 5880000 (3662281 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 113, 1061 MB with memory of 1548 MB (486 MB spared); blocks given memory later 9115, given back 12228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 663 binds (194.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 163 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 49 spares now, 1080 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 370 dirty pages from their submits' captures, 54923 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 909 dispatches, 42084 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7710 runs, 708.6 MB compared; 64 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17364 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 184 live (2300 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos host] vk: program picks: 8100063 kept (2963733 by the last picks' textures, a uniform buffer bound since), 26228 new, 3473708 with other textures (of 11600000) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1111 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1511 kept their virtual core (nothing there to wait for), 2841 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 5900000 (3674176 pages from the captures) [hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 240000 so far (x1 0x117a4f2040 x2 0x6001d40 x3 0x117a4f2000 x4 0xf33080) [hos sead::MainThread] nvn: GPU waits in 5 s: 167 sync waits (528 ms), 0 polls timed out (0 ms), 3 queue finishes (38 ms) [hos host] vk: replay reads of guest memory as submitted: 5920000 (3683773 pages from the captures) [hos host] vk: queues ordered by their events: 66000 submissions stopped at a wait, 65966 went on once it was reached, 32 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 3 (stopped 50.30 ms at most); waits met 163918, not met 0; captures at a CPU event's wait 7435, 0 of them at a wait met on arrival [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 562 ranges, 3108 MB), 258 pages written, 258 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (130 MB), 11 spare buffers, 3653 texture pages copied (14 MB) [hos host] vk: readback pages refreshed from guest memory: 3462809 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 324 dirty pages from their submits' captures, 56080 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 926 dispatches, 43180 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7755 runs, 724.1 MB compared; 61 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17643 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 5940000 (3695021 pages from the captures) [hos ModuleSystemWorker5] vcores: 157095 takes, 33222 waits (mean 61 us, longest 11.3 ms, 0 over HOS_VCORE_LOG_MS), 15990 kicks, 15991 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a frame: 144 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.73, present 0.14, draw 6.31, descriptors 2.72, texture lookups 0.06, uploads 0.18, pipelines 0.26, programs 1.14, commands 1.71, dispatch 0.81, memory sync 0.47, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 325 (1060 MB): 5.38 MB synced in 1045.7 copies, 17.70 MB written back [hos host] vk: 11280 presents, 60.2 fps; a frame: 1925 draws (0 skipped), 150 dispatches, 186 passes, 131 barriers, 6 submits, 0.4 uploads (1.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1367 textures (2567 MB) [hos host] vk: host time in GPU resource waits: 1.98 ms a frame (fences 1.47, timelines 0.50, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 680.6 full barriers recorded, 296.1 left out (nothing recorded since the one before), 20.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.4, an op 1024.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.2 ms a frame (the longest submission 5.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.29 MB unswizzled on the CPU in 0.18 ms [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10428 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 1 of 1 small textures the GPU drew written into their memory in 0.6 ms, 1 frames after their copies were recorded, the first #16653 0x116e25b000 260x260 levels 9 NVN 0xd (33 so far, 7232 KB, 12 ms; 0 left, the CPU wrote their memory since; 0 dropped, drawn over or gone; 0 failed) [hos host] vk: replay reads of guest memory as submitted: 5960000 (3705722 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #165000: event 0x100d6c802c mode 0 value 114982 (now 114963) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #115000: event 0x100d6c801c mode 0 value 114997 (lr 0x802911b8) [hos host] vk: sparse mirrors: 113, 1060 MB with memory of 1548 MB (487 MB spared); blocks given memory later 9115, given back 12242; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 665 binds (194.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 163 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 49 spares now, 1080 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 480 dirty pages from their submits' captures, 54561 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 41610 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7658 runs, 707.5 MB compared; 66 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17243 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fb5c00000 (44 MB, captured, state 2): 1; [hos host] vk: imports: 184 live (2300 MB), 2 retired still imported (32 MB), 38 retired given back (852 MB) [hos host] vk: replay reads of guest memory as submitted: 5980000 (3718162 pages from the captures) [hos rmbl::VibrationThread] hid: 36864 vibration values, strongest 0.00 since the last count [hos host] vk: program picks: 8251640 kept (3036090 by the last picks' textures, a uniform buffer bound since), 26228 new, 3522131 with other textures (of 11800000) [hos host] vk: texture #15310 0x100089d2a00 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 11334 (drawn at 9705) [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 254 dirty pages from their submits' captures, 55429 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 877 dispatches, 42515 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7589 runs, 691.5 MB compared; 48 pages read into the shadows; 304 MB of shadows (51 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17467 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 6000000 (3726749 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 552 ranges, 3106 MB), 286 pages written, 286 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.015; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 4250 texture pages copied (16 MB) [hos host] vk: texture #14274 0x116c30d000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 11362 (drawn at 5293) [hos host] vk: presenter thread: the copy's fence 1.34 ms, the present 0.76 ms a present [hos host] vk: replay reads of guest memory as submitted: 6020000 (3735336 pages from the captures) [hos host] vk: sparse mirrors: 113, 1059 MB with memory of 1548 MB (488 MB spared); blocks given memory later 9115, given back 12263; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 668 binds (194.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 163 taken from the spares, 162 made new; 212 kept as spares, 0 destroyed; 49 spares now, 1080 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 191 dirty pages from their submits' captures, 54686 from earlier ones', 4 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 915 dispatches, 41612 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7761 runs, 710.0 MB compared; 56 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17013 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 4; [hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1129 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1730 kept their virtual core (nothing there to wait for), 3390 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 7930 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 6040000 (3744362 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: memory: working set 15182 MB (peak 15952), private 21247 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 16 blocks 928 MB, 37 allocations 836 MB; used 3248 of 15579 MB; heap 1 (device): 161 blocks 4887 MB, 19045 allocations 4466 MB; used 6882 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 149 images 465 MB; 101 free ranges, the largest 21 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1267 images 1617 MB; 278 free ranges, the largest 29 MB [hos host] vk: render target images: 5725 reused from the pool, 3332 made new, 3181 destroyed from it, 989 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1194 with the pool empty (0 of them with targets still in the dead queue), 287 with none in it a present yet, 25 with none of a kind they can use, 466 with all smaller (the largest 51% short on average), 504 with all bigger (over 1.5 times), 855 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 2989 idle [hos host] vk: VRAM: 6882 MB used of 11474 MB allowed (420 MB of it free in the blocks); textures 2719 MB (budget 7158 MB), 1422 textures [hos host] vk: budget: VRAM 4887 of 11474 MB (6882 MB by the driver, 1995 MB unaccounted): textures 2719, mirrors 1056, blocks free 420, shadows 304, other buffers 211, other images 111, staging 64; RAM 3168 MB (the process's private memory 21247 MB): imports 2300, capture spares 732, captures 104, capture table 32, others 0 [hos host] vk: textures: decoded ASTC 1475 MB, rendered 669 MB (130 drawn in one frame, kept: 152242 KB), other 522 MB; used in the last 8 presents 1717 MB [hos host] vk: decoded ASTC used again: 91 textures (212 MB) copied from ones the game finalized, not decoded; 91 of those retired, 18 listed; 1176 MB of blocks hashed in 203 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 6253 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 421 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 20 (7697 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 137x134x1 NVN 0x1: 1 (17 KB); 204x106x1 NVN 0x1: 1 (21 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 5 (879 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); 1466x282x1 NVN 0x25: 4 (6459 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 253 waited for it, 2.35 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 23.4 ms; VRAM 6.7 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6153 queued, 6153 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15243 ms and 331.0 MB read in all); 8348 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11700 graphics pipelines (1 not in a cache), 712 compute pipelines (32 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2172 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 8987 (6153 ready, 0 loading), 16189 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13164 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 760 known, 3 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 152 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.21, present 0.15, draw 6.70, descriptors 2.98, texture lookups 0.15, uploads 0.22, pipelines 0.27, programs 1.15, commands 1.79, dispatch 0.83, memory sync 0.46, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 325 (1056 MB): 5.45 MB synced in 1060.5 copies, 17.75 MB written back [hos host] vk: 11400 presents, 60.0 fps; a frame: 2010 draws (0 skipped), 152 dispatches, 180 passes, 131 barriers, 6 submits, 0.6 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1422 textures (2719 MB) [hos host] vk: host time in GPU resource waits: 1.76 ms a frame (fences 1.56, timelines 0.20, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 683.1 full barriers recorded, 295.3 left out (nothing recorded since the one before), 15.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1078.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.8 ms a frame (the longest submission 5.0 ms); 0.7 textures made a frame: colour targets 0.2 (0.9 MB), depth 0.4 (1.5 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.51 MB unswizzled on the CPU in 0.21 ms [hos host] vk: the render thread (600 presents, 16.78 ms a frame): its jobs 15.43 ms a frame (submits 6.19, presents 4.88, others 4.35), idle 1.36; the game waited for room at 30 presents, 0.10 ms each (0.01 a frame; 10.78 jobs queued at a present); 3 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 11.09 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 261 captures at a CPU signal; the submits' waits read in 0.041 ms a frame; readbacks: 1411 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.73 ms after its hand-over (21.02 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.43 + 0.27 (the window's copy is in the idle), after it 1.11 + 0.33; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.48 parts a frame recorded in the signal's job, 79.2% of them took nothing that came with the signal; of the others, 1.75 a frame took pages of the signal's capture into a mirror (33.0 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 37 of 261 frames, its first three in 36 [hos host] vk: frame ahead, what came with the signal: 261 signals with a capture, each of 77.0 pages in 9.9 pools; between a frame's last submit and that signal 0.01 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.21+0.06, 0.34+0.13, 0.03+0.02, 0.03+0.02, 0.20+0.02, 0.18+0.03, 0.06+0.02, 0.05+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: render targets drawn over from another start, let go: 9788 so far (46474 MB), 8773 of them (41752 MB) between the passes [hos host] vk: program picks: 8403939 kept (3109804 by the last picks' textures, a uniform buffer bound since), 26228 new, 3569832 with other textures (of 12000000) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 157 dirty pages from their submits' captures, 54358 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 871 dispatches, 41925 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7664 runs, 694.0 MB compared; 52 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17257 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 6060000 (3753504 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: texture #15694 0x1001e7cf000 1120x1333x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 5845 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14767 0x10004b01a00 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1735 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14761 0x10005552400 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 1735 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14769 0x100050a2000 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1708 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14768 0x10004fb1e00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1696 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #14770 0x10004ec1c00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1645 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 341 textures retired (1046 left, 2104 MB): budget 337, drawn over 4, dead targets 0; GPU-drawn with no copy kept: 0 live, 6 finalized, 26 MB [hos host] vk: replay reads of guest memory as submitted: 6080000 (3764635 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 558 ranges, 3108 MB), 279 pages written, 279 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (138 MB), 11 spare buffers, 4250 texture pages copied (16 MB) [hos host] vk: texture #16733 0x10008612200 957x957 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 11449 (drawn at 11388) [hos host] vk: readback pages refreshed from guest memory: 3530428 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 107, 1046 MB with memory of 1516 MB (469 MB spared); blocks given memory later 9121, given back 12304; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 678 binds (196.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 163 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 55 spares now, 1112 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 285 dirty pages from their submits' captures, 54312 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 878 dispatches, 41932 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7877 runs, 711.5 MB compared; 58 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17670 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 6100000 (3774630 pages from the captures) [hos Presentation Thread] vi: 11472 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 9, 15.5-17.8 280, 17.8-22 11, 22-30 0, 30-40 0, over 40 0 [hos host] vk: texture #16647 0x116e006000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 11481 (drawn at 11034) [hos host] vk: replay reads of guest memory as submitted: 6120000 (3786066 pages from the captures) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 385 dirty pages from their submits' captures, 53582 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 870 dispatches, 41418 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7561 runs, 677.3 MB compared; 70 pages read into the shadows; 304 MB of shadows (51 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17298 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12254 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #170000: event 0x100d6c802c mode 0 value 118782 (now 118763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 6140000 (3795198 pages from the captures) [hos host] vk: program picks: 8556862 kept (3182304 by the last picks' textures, a uniform buffer bound since), 26228 new, 3616909 with other textures (of 12200000) [hos host] vk: 3 textures retired (1045 left, 2092 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: a frame: 161 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.91, present 0.21, draw 6.37, descriptors 2.80, texture lookups 0.18, uploads 0.23, pipelines 0.27, programs 1.11, commands 1.71, dispatch 0.81, memory sync 0.45, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 325 (1046 MB): 5.39 MB synced in 1058.4 copies, 17.80 MB written back [hos host] vk: 11520 presents, 60.0 fps; a frame: 1957 draws (0 skipped), 153 dispatches, 178 passes, 131 barriers, 6 submits, 0.4 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13555 pipelines (3509 ms), 0 programs (10 failed), 1045 textures (2092 MB) [hos host] vk: host time in GPU resource waits: 1.38 ms a frame (fences 1.22, timelines 0.17, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 685.3 full barriers recorded, 296.1 left out (nothing recorded since the one before), 13.3 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 1068.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.1 ms a frame (the longest submission 5.1 ms); 0.5 textures made a frame: colour targets 0.2 (1.6 MB), depth 0.2 (1.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.53 MB unswizzled on the CPU in 0.22 ms [hos host] vk: texture #14272 0x116c2d8000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 11522 (drawn at 5293) [hos host] vk: reaper: 431 images (863 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 282 (412 MB) [hos host] vk: sparse mirrors: 109, 1056 MB with memory of 1532 MB (475 MB spared); blocks given memory later 9121, given back 12304; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 680 binds (196.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 165 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 53 spares now, 1096 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 337 dirty pages from their submits' captures, 53514 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 867 dispatches, 41464 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7792 runs, 711.0 MB compared; 64 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17180 in captures, 2 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 6160000 (3806084 pages from the captures) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1112 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1302 kept their virtual core (nothing there to wait for), 2794 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 554 ranges, 3060 MB), 286 pages written, 286 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (128 MB), 11 spare buffers, 4488 texture pages copied (17 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 52 sync waits (73 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: recorder: 6121759 packets run in 10.0 s (busy 1.79 s, spun idle 0.11 s, slept 10352 times, 3143 of them under 200 us), 0 calls made directly (no room); 18332 batches over 596 frames (30.8 a frame, 334 packets a batch), 9002 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 59 times, 6.79 ms (0.1 times, 0.01 ms a frame); every thread 631 times, 45.11 ms in all, the longest 1.28 ms; vkWaitForFences 522/31.67 ms (max 0.45), vmaFreeMemory 18/3.46 ms (max 0.60), vmaDestroyBuffer 14/3.16 ms (max 1.28), the queue's lock 21/2.53 ms (max 0.39), vkQueueBindSparse 29/2.05 ms (max 0.52), vkDestroyImageView 18/1.75 ms (max 0.31) [hos host] vk: recorder: of its busy 1.79 s, 0.52 s in parts' ends and submits (the longest 2.66 ms), 0.20 s of that blocked on the queue's lock (349 times, the longest 2.61 ms), so 0.32 s in the ends, submits and holds; 1.28 s running the other packets; 10271 packets a frame [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 2 textures retired (1060 left, 2103 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6180000 (3818659 pages from the captures) [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #120000: event 0x100d6c8030 mode 0 value 119999 (lr 0x806612a8) [hos host] vk: replay reads of guest memory as submitted: 6200000 (3831028 pages from the captures) [hos ModuleSystemWorker3] nvn: 1350000 recordings made, 70785 live copies (23.2 MB), 5950 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 623 dirty pages from their submits' captures, 55304 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 937 dispatches, 42082 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8286 runs, 703.0 MB compared; 65 pages read into the shadows; 304 MB of shadows (51 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17249 in captures, 2 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos aal::LowPrioWorkerThread] vcores: 152095 takes, 22663 waits (mean 68 us, longest 7.1 ms, 0 over HOS_VCORE_LOG_MS), 10699 kicks, 10553 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 6220000 (3842972 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: texture #16057 0x116cf82000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 11598 (drawn at 9242) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 4548 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 6240000 (3854371 pages from the captures) [hos host] vk: program picks: 8709400 kept (3255707 by the last picks' textures, a uniform buffer bound since), 26230 new, 3664369 with other textures (of 12400000) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 112, 1065 MB with memory of 1552 MB (487 MB spared); blocks given memory later 9122, given back 12350; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 688 binds (198.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 168 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 50 spares now, 1076 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 207 dirty pages from their submits' captures, 56984 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 920 dispatches, 43988 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9104 runs, 758.4 MB compared; 62 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17803 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 17 textures retired (1061 left, 2094 MB): budget 17, drawn over 0, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 35 textures retired (1099 left, 2193 MB): budget 35, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6260000 (3867177 pages from the captures) [hos ModuleSystemWorker4] nvn: DrawMeshTasksIndirect: 260000 so far (x1 0x117a4f2040 x2 0x6001d40 x3 0x117a4f2000 x4 0xf33080) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.04 over 564 ranges, 3150 MB), 263 pages written, 263 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (122 MB), 11 spare buffers, 5309 texture pages copied (20 MB) [hos host] vk: texture 0x10001e00400 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 25 textures retired (1260 left, 2497 MB): budget 25, drawn over 0, dead targets 0 [hos host] vk: a frame: 179 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.93, present 0.15, draw 7.26, descriptors 3.81, texture lookups 1.20, uploads 1.24, pipelines 0.26, programs 1.04, commands 1.68, dispatch 0.86, memory sync 0.48, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 344 (1112 MB): 6.00 MB synced in 1059.3 copies, 18.01 MB written back [hos host] vk: 11640 presents, 56.5 fps; a frame: 1859 draws (0 skipped), 169 dispatches, 163 passes, 132 barriers, 6 submits, 2.8 uploads (2.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1272 textures (2510 MB) [hos host] vk: host time in GPU resource waits: 4.36 ms a frame (fences 2.13, timelines 2.22, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 657.8 full barriers recorded, 280.5 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.6; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.9 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.7, an op 1090.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.0 ms a frame (the longest submission 5.0 ms); 2.8 textures made a frame: colour targets 0.4 (1.1 MB), depth 0.1 (0.6 MB), ASTC or BC 2.3 (3.5 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.92 MB unswizzled on the CPU in 0.27 ms [hos host] vk: replay reads of guest memory as submitted: 6280000 (3879327 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 3597538 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1724 dirty pages from their submits' captures, 50481 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 847 dispatches, 40189 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8064 runs, 683.5 MB compared; 72 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16595 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 2; [hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 14961078 lookups, 126778 missed (0.85 %) [hos host] vk: replay reads of guest memory as submitted: 6300000 (3890708 pages from the captures) [hos host] vk: texture #17190 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (7 so far) [hos host] vk: sparse mirrors: 126, 1112 MB with memory of 1636 MB (523 MB spared); blocks given memory later 9122, given back 12350; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 702 binds (201.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 182 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 36 spares now, 992 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 365 dirty pages from their submits' captures, 55198 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 875 dispatches, 42641 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8558 runs, 720.3 MB compared; 58 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17566 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 184 live (2300 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: texture #16214 0x10004bf1600 160x90 x1 NVN 0x3 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635) [hos host] vk: texture #16190 0x10004651000 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635) [hos host] vk: texture #16194 0x100053ea200 640x360 x1 NVN 0xa the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635) [hos host] vk: texture #16193 0x10005372000 640x360 x1 NVN 0x5 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635) [hos host] vk: texture #16195 0x100054da400 640x360 x1 NVN 0x11 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635) [hos host] vk: texture #16197 0x10005191c00 1600x1600 x3 NVN 0x32 the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635) [hos host] vk: texture #16227 0x10004380c00 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 11696 (drawn at 11635) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #175000: event 0x100d6c802c mode 0 value 122582 (now 122563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 6320000 (3902172 pages from the captures) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1099 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 19 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1464 kept their virtual core (nothing there to wait for), 2888 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 5 textures retired (1334 left, 2623 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6340000 (3915189 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 16 textures retired (1328 left, 2642 MB): budget 16, drawn over 0, dead targets 0 [hos host] vk: program picks: 8859341 kept (3318222 by the last picks' textures, a uniform buffer bound since), 26233 new, 3714425 with other textures (of 12600000) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.04 over 572 ranges, 3284 MB), 257 pages written, 257 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (138 MB), 8 spare buffers, 6127 texture pages copied (23 MB) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 49 textures retired (1355 left, 2651 MB): budget 49, drawn over 0, dead targets 0 [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1624 dirty pages from their submits' captures, 48847 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 827 dispatches, 38231 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7201 runs, 627.7 MB compared; 77 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15701 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos ModuleSystemWorker4] tex: the game's texture heap: 189 compaction rounds asked for in 10800 frames without one running [hos host] vk: replay reads of guest memory as submitted: 6360000 (3928180 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13321 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 6 textures retired (1397 left, 2696 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: texture #16204 0x10004741200 2560x1440 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11742 (drawn at 11635) [hos host] vk: texture #16205 0x10004bf1600 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11742 (drawn at 11635) [hos host] vk: texture #16206 0x10004b01400 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11742 (drawn at 11635) [hos host] vk: texture 0x100023a0e00 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since) [hos host] vk: texture #16228 0x10001e00400 640x360 x1 NVN 0x14 the GPU drew, finalized by the game, sampled again at frame 11742 (drawn at 11635) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 25 textures retired (1378 left, 2697 MB): budget 25, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6380000 (3941778 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 4 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: a frame: 265 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.64, present 0.17, draw 6.87, descriptors 3.24, texture lookups 0.49, uploads 0.28, pipelines 0.27, programs 1.11, commands 1.76, dispatch 0.85, memory sync 0.46, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 348 (1119 MB): 5.53 MB synced in 1066.4 copies, 17.88 MB written back [hos host] vk: 11760 presents, 59.0 fps; a frame: 1887 draws (0 skipped), 164 dispatches, 187 passes, 134 barriers, 6 submits, 2.1 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1378 textures (2697 MB) [hos host] vk: host time in GPU resource waits: 4.83 ms a frame (fences 1.80, timelines 3.03, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 729.4 full barriers recorded, 304.2 left out (nothing recorded since the one before), 16.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 26.5; 0x12: 69.3; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 124.8, an op 1174.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.2 ms a frame (the longest submission 4.2 ms); 2.3 textures made a frame: colour targets 0.2 (0.7 MB), depth 0.1 (0.4 MB), ASTC or BC 2.0 (3.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.29 MB unswizzled on the CPU in 0.05 ms [hos Presentation Thread] vi: 11763 frames presented (58.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 50, 15.5-17.8 197, 17.8-22 28, 22-30 10, 30-40 6, over 40 0 [hos host] vk: sparse mirrors: 130, 1119 MB with memory of 1656 MB (536 MB spared); blocks given memory later 9126, given back 12351; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 708 binds (202.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 186 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 32 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1836 dirty pages from their submits' captures, 52988 from earlier ones', 4 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 873 dispatches, 41954 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7344 runs, 702.9 MB compared; 74 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16993 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: 0x7f86400000 (8 MB, captured, state 2): 4; [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 4 textures retired (1374 left, 2696 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6400000 (3954130 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 6 textures retired (1368 left, 2690 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 20 textures retired (1352 left, 2663 MB): budget 20, drawn over 0, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 17 textures retired (1340 left, 2647 MB): budget 17, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6420000 (3965802 pages from the captures) [hos host] vk: texture #17008 0x10004bf1c00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741) [hos host] vk: texture #17026 0x100021aea00 120x67 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741) [hos host] vk: texture #17031 0x100020e5a00 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741) [hos host] vk: texture #16990 0x10005552400 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741) [hos host] vk: texture #17003 0x100050a2000 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741) [hos host] vk: texture #17002 0x10004fb1e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 11802 (drawn at 11741) [hos host] vk: texture #17033 0x10004921600 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 11803 (drawn at 11741) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 9 textures retired (1337 left, 2640 MB): budget 8, drawn over 1, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 182 dirty pages from their submits' captures, 74789 from earlier ones', 41178 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 987 dispatches, 45762 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7659 runs, 736.7 MB compared; 62 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18460 in captures, 2 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: 0x7fab000000 (68 MB, captured, state 2): 16936; 0x7fb5c00000 (44 MB, captured, state 2): 10657; 0x7fb8800000 (52 MB, captured, state 2): 8785; 0x7fd3c00000 (16 MB, captured, state 2): 3830; 0x7fa9000000 (4 MB, captured, state 2): 644; 0x7faf400000 (4 MB, captured, state 2): 155; 0x113b400000 (24 MB, captured, state 2): 29; 0x7f94000000 (16 MB, captured, state 2): 26; [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 38 textures retired (1317 left, 2607 MB): budget 38, drawn over 0, dead targets 0 [hos sead::MainThread] vk: capture: a submit 0.28 ms (lock 0.02, scan 0.05 over 567 ranges, 3340 MB), 704 pages written, 533 copied (2.1 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (318 MB), 11 spare buffers, 6348 texture pages copied (24 MB) [hos OptimizeChangeBanc] nvn: 1400000 recordings made, 69330 live copies (23.3 MB), 5950 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: program picks: 9006207 kept (3367113 by the last picks' textures, a uniform buffer bound since), 26274 new, 3767518 with other textures (of 12800000) [hos sead::MainThread] nvn: CommandBufferSignalEvent #125000: event 0x100d6c803c mode 0 value 124999 (lr 0x80661bfc) [hos host] vk: replay reads of guest memory as submitted: 6440000 (3977550 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 118 sync waits (649 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 9 textures retired (1308 left, 2590 MB): budget 9, drawn over 0, dead targets 0 [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 7 textures retired (1303 left, 2586 MB): budget 7, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6460000 (3990729 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 12 textures retired (1296 left, 2586 MB): budget 12, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 3665028 pages so far (0 held: may be written ahead in their submission) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10276 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 130, 1195 MB with memory of 1656 MB (460 MB spared); blocks given memory later 10338, given back 12351; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 743 binds (211.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 186 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 32 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 7944 dirty pages from their submits' captures, 87919 from earlier ones', 3774 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1051 dispatches, 46111 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7136 runs, 699.4 MB compared; 359 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16523 in captures, 236 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 3774; [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos sead::MainThread] vk: readbacks of GPU results: 1019 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1452 kept their virtual core (nothing there to wait for), 3156 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos BaseProcExecutor2] vcores: 151035 takes, 28215 waits (mean 68 us, longest 17.1 ms, 0 over HOS_VCORE_LOG_MS), 14212 kicks, 14221 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: late pages found at the replay (600 takes): 371 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 2 textures retired (1294 left, 2586 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6480000 (4001677 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 10 textures retired (1289 left, 2569 MB): budget 10, drawn over 0, dead targets 0 [hos host] vk: a frame: 266 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.63, present 0.18, draw 6.79, descriptors 2.90, texture lookups 0.16, uploads 0.03, pipelines 0.29, programs 1.22, commands 1.86, dispatch 0.88, memory sync 1.12, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 348 (1187 MB): 8.67 MB synced in 1065.1 copies, 17.50 MB written back [hos host] vk: 11880 presents, 55.2 fps; a frame: 1809 draws (0 skipped), 138 dispatches, 167 passes, 131 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1289 textures (2569 MB) [hos host] vk: host time in GPU resource waits: 3.72 ms a frame (fences 1.69, timelines 2.03, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 630.0 full barriers recorded, 286.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 995.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.2 ms a frame (the longest submission 5.7 ms); 0.4 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.4 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 5 textures retired (1285 left, 2562 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 871 dirty pages from their submits' captures, 55953 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 933 dispatches, 43815 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7446 runs, 733.7 MB compared; 78 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17502 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 6500000 (4013890 pages from the captures) [hos sead::MainThread] nvn: CommandBufferWaitEvent #180000: event 0x100d6c802c mode 0 value 126382 (now 126363) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.18 [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 21 textures retired (1272 left, 2524 MB): budget 21, drawn over 0, dead targets 0 [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos Presentation Thread] vk: capture: a submit 0.21 ms (lock 0.03, scan 0.06 over 613 ranges, 3388 MB), 324 pages written, 324 copied (1.3 MB) [hos host] vk: a capture taken (ms): all 0.029; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (114 MB), 11 spare buffers, 8463 texture pages copied (33 MB) [hos host] vk: replay reads of guest memory as submitted: 6520000 (4026453 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: program picks: 9151396 kept (3411432 by the last picks' textures, a uniform buffer bound since), 26280 new, 3822323 with other textures (of 13000000) [hos host] vk: sparse mirrors: 130, 1189 MB with memory of 1656 MB (466 MB spared); blocks given memory later 10360, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 749 binds (212.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 186 taken from the spares, 162 made new; 218 kept as spares, 0 destroyed; 32 spares now, 972 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 396 dirty pages from their submits' captures, 57256 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 987 dispatches, 44313 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7548 runs, 734.7 MB compared; 68 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18161 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7f86000000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: replay reads of guest memory as submitted: 6540000 (4038464 pages from the captures) [hos ModuleSystemWorker2] nvn: 1450000 recordings made, 75348 live copies (23.7 MB), 5950 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: 1 textures retired (1277 left, 2512 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: presenter thread: the copy's fence 1.44 ms, the present 0.75 ms a present [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17717 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 6560000 (4050721 pages from the captures) [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 434 dirty pages from their submits' captures, 58607 from earlier ones', 910 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 45389 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7719 runs, 751.5 MB compared; 72 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18459 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 910; [hos host] vk: texture #533 0x1159c17000 4096x4096 x1 NVN 0x32 the GPU drew kept, not evicted: its texels cannot be written back [hos host] vk: replay reads of guest memory as submitted: 6580000 (4062584 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.07 [hos host] vk: memory: working set 15171 MB (peak 15952), private 21363 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 17 blocks 784 MB, 36 allocations 692 MB; used 3108 of 15579 MB; heap 1 (device): 166 blocks 5047 MB, 20797 allocations 4379 MB; used 7042 of 8467 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 111 images 372 MB; 66 free ranges, the largest 28 MB [hos host] vk: image blocks (textures, compacted): 57 blocks 1824 MB, 1176 images 1506 MB; 413 free ranges, the largest 32 MB [hos host] vk: render target images: 5753 reused from the pool, 3435 made new, 3322 destroyed from it, 993 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1294 with the pool empty (0 of them with targets still in the dead queue), 287 with none in it a present yet, 25 with none of a kind they can use, 467 with all smaller (the largest 51% short on average), 506 with all bigger (over 1.5 times), 855 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3130 idle [hos host] vk: VRAM: 7042 MB used of 8467 MB allowed (667 MB of it free in the blocks); textures 2514 MB (budget 2140 MB), 1293 textures [hos host] vk: budget: VRAM 5047 of 8467 MB (7042 MB by the driver, 1995 MB unaccounted): textures 2514, mirrors 1174, blocks free 667, shadows 304, other buffers 211, other images 111, staging 64; RAM 3034 MB (the process's private memory 21363 MB): imports 2304, capture spares 564, captures 134, capture table 32, others 0 [hos host] vk: textures: decoded ASTC 1544 MB, rendered 589 MB (88 drawn in one frame, kept: 45085 KB), other 345 MB; used in the last 8 presents 1899 MB [hos host] vk: decoded ASTC used again: 185 textures (477 MB) copied from ones the game finalized, not decoded; 185 of those retired, 1 listed; 1360 MB of blocks hashed in 231 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 6853 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 28 (8783 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 102x30x1 NVN 0x25: 1 (11 KB); 137x134x1 NVN 0x1: 1 (17 KB); 202x38x1 NVN 0x25: 1 (29 KB); 204x106x1 NVN 0x1: 1 (21 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 8 (1407 KB); 266x102x1 NVN 0x1: 1 (26 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 261 waited for it, 3.48 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 48.6 ms; VRAM 6.9 of 8.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6168 queued, 6168 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15261 ms and 331.4 MB read in all); 8363 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11712 graphics pipelines (1 not in a cache), 713 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2212 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9002 (6168 ready, 0 loading), 16426 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13188 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 763 known, 6 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 221 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.57, present 0.16, draw 6.00, descriptors 2.58, texture lookups 0.19, uploads 0.09, pipelines 0.26, programs 1.07, commands 1.63, dispatch 0.79, memory sync 0.57, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 348 (1174 MB): 5.41 MB synced in 1055.5 copies, 17.49 MB written back [hos host] vk: 12000 presents, 60.0 fps; a frame: 1691 draws (0 skipped), 138 dispatches, 168 passes, 131 barriers, 6 submits, 0.8 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1293 textures (2514 MB) [hos host] vk: host time in GPU resource waits: 2.16 ms a frame (fences 1.47, timelines 0.70, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 634.9 full barriers recorded, 287.5 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 923.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.8 ms a frame (the longest submission 4.7 ms); 0.8 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.7 (1.7 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 17.22 ms a frame): its jobs 15.76 ms a frame (submits 6.17, presents 5.01, others 4.57), idle 1.46; the game waited for room at 21 presents, 1.17 ms each (0.04 a frame; 8.24 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 267 captures at a CPU signal; the submits' waits read in 0.040 ms a frame; readbacks: 1488 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.96 ms after its hand-over (37.83 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.60 + 0.27 (the window's copy is in the idle), after it 1.26 + 0.81; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.67 parts a frame recorded in the signal's job, 73.2% of them took nothing that came with the signal; of the others, 2.30 a frame took pages of the signal's capture into a mirror (61.0 pages a frame) and 0.03 pages read as guest memory had them (168.4 pages a frame); a frame's first part took nothing in 31 of 267 frames, its first three in 31 [hos host] vk: frame ahead, what came with the signal: 267 signals with a capture, each of 560.8 pages in 11.1 pools; between a frame's last submit and that signal 0.05 texture and sampler registrations, 0.14 releases and 0.00 other CPU signals a signal (7 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.36+0.09, 0.41+0.13, 0.05+0.04, 0.03+0.02, 0.10+0.05, 0.16+0.27, 0.09+0.18, 0.05+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 616 ranges, 3342 MB), 268 pages written, 268 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (128 MB), 10 spare buffers, 4776 texture pages copied (18 MB) [hos Presentation Thread] vi: real-time step on: ratio 0.997, game's frame rate 60 [hos host] vk: render targets drawn over from another start, let go: 9953 so far (47123 MB), 8938 of them (42401 MB) between the passes [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1086 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1406 kept their virtual core (nothing there to wait for), 2946 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 6600000 (4073949 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 126, 1160 MB with memory of 1588 MB (427 MB spared); blocks given memory later 10383, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 757 binds (212.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 187 taken from the spares, 162 made new; 222 kept as spares, 0 destroyed; 35 spares now, 1036 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 322 dirty pages from their submits' captures, 56737 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 958 dispatches, 44051 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7550 runs, 733.8 MB compared; 83 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18036 in captures, 1 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: 224 mirrors given back (1757 MB in all); 1158 MB of mirrors now, 172 MB of them idle and kept for what the GPU wrote there alone [hos sead::MainThread] nvn: DrawMeshTasks: 280000 so far (x1 0 x2 0x20 x3 0x10201414 x4 0x6001ba0) [hos host] vk: commit: 2481 MB left under Windows' limit as it is now, and the page files can grow by 79851 MB more: counted as room [hos host] vk: replay reads of guest memory as submitted: 6620000 (4085301 pages from the captures) [hos host] vk: program picks: 9295234 kept (3454419 by the last picks' textures, a uniform buffer bound since), 26280 new, 3878485 with other textures (of 13200000) [hos Presentation Thread] vi: 12053 frames presented (57.9 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 35, 15.5-17.8 210, 17.8-22 31, 22-30 10, 30-40 3, over 40 1 [hos host] vk: readback pages refreshed from guest memory: 3732143 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 1099, 4882.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 99 dirty pages from their submits' captures, 60368 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 970 dispatches, 46189 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7908 runs, 768.4 MB compared; 77 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18965 in captures, 1 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: queues ordered by their events: 72000 submissions stopped at a wait, 71966 went on once it was reached, 32 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 3 (stopped 50.30 ms at most); waits met 184199, not met 0; captures at a CPU event's wait 7756, 0 of them at a wait met on arrival [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9456 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 6640000 (4096845 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] nvn: CommandBufferSignalEvent #130000: event 0x100d6c8048 mode 0 value 129999 (lr 0x80666598) [hos host] vk: 2 textures retired (1293 left, 2500 MB): budget 0, drawn over 0, dead targets 2 [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #185000: event 0x100d6c802c mode 0 value 130182 (now 130163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: 1 textures retired (1292 left, 2486 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6660000 (4107675 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 615 ranges, 3342 MB), 278 pages written, 278 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 25 kept (140 MB), 11 spare buffers, 7124 texture pages copied (27 MB) [hos host] vk: sparse mirrors: 123, 1178 MB with memory of 1596 MB (417 MB spared); blocks given memory later 10423, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 763 binds (213.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 188 taken from the spares, 162 made new; 227 kept as spares, 0 destroyed; 39 spares now, 1032 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 221 dirty pages from their submits' captures, 57591 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 928 dispatches, 44835 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7694 runs, 750.4 MB compared; 97 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18535 in captures, 1 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: a frame: 256 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 8.77, present 0.18, draw 5.42, descriptors 2.27, texture lookups 0.08, uploads 0.00, pipelines 0.24, programs 1.00, commands 1.49, dispatch 0.77, memory sync 0.47, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 350 (1178 MB): 5.52 MB synced in 1044.4 copies, 17.45 MB written back [hos host] vk: 12120 presents, 60.0 fps; a frame: 1617 draws (0 skipped), 136 dispatches, 166 passes, 131 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1292 textures (2486 MB) [hos host] vk: host time in GPU resource waits: 1.69 ms a frame (fences 1.47, timelines 0.22, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 622.7 full barriers recorded, 285.6 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 831.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.5 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 6680000 (4119607 pages from the captures) [hos host] vk: recorder: 5707589 packets run in 10.0 s (busy 1.61 s, spun idle 0.11 s, slept 10472 times, 2900 of them under 200 us), 0 calls made directly (no room); 17238 batches over 580 frames (29.7 a frame, 331 packets a batch), 8506 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 110 times, 8.90 ms (0.2 times, 0.02 ms a frame); every thread 704 times, 43.13 ms in all, the longest 0.44 ms; vkWaitForFences 495/28.15 ms (max 0.41), vkQueueBindSparse 83/4.77 ms (max 0.36), vkDestroyImageView 49/3.43 ms (max 0.29), vmaFreeMemory 16/2.32 ms (max 0.39), vkEndCommandBuffer 9/1.68 ms (max 0.44), the queue's lock 17/1.55 ms (max 0.22) [hos host] vk: recorder: of its busy 1.61 s, 0.40 s in parts' ends and submits (the longest 1.14 ms), 0.09 s of that blocked on the queue's lock (164 times, the longest 1.10 ms), so 0.30 s in the ends, submits and holds; 1.21 s running the other packets; 9841 packets a frame [hos sead::MainThread] nvn: GPU waits in 5 s: 67 sync waits (175 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 405 dirty pages from their submits' captures, 64795 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 958 dispatches, 45046 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7280 runs, 729.1 MB compared; 93 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18653 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 6700000 (4131296 pages from the captures) [hos BaseProcExecutor0] vcores: 158125 takes, 28237 waits (mean 70 us, longest 21.3 ms, 0 over HOS_VCORE_LOG_MS), 13413 kicks, 13178 yields, 0 starved grants, 0 escapes in 5 s [hos PlayReporter] vcores: PlayReporter (priority 29) waited 71.5 ms for a core: given by PreActorModelCompose (priority 26), which held it 0.0 ms and was never asked; the cores it may use now: core 0: it; [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1081 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1473 kept their virtual core (nothing there to wait for), 2879 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: program picks: 9440102 kept (3495015 by the last picks' textures, a uniform buffer bound since), 26280 new, 3933617 with other textures (of 13400000) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21010 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 6720000 (4142673 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 119, 1181 MB with memory of 1580 MB (398 MB spared); blocks given memory later 10517, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 769 binds (214.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 188 taken from the spares, 162 made new; 231 kept as spares, 0 destroyed; 43 spares now, 1048 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 274 dirty pages from their submits' captures, 59406 from earlier ones', 5 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 957 dispatches, 44754 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7425 runs, 746.3 MB compared; 99 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18121 in captures, 1 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 3; 0x7fd3c00000 (16 MB, captured, state 2): 1; 0x1042000000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 622 ranges, 3347 MB), 304 pages written, 304 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 25 kept (122 MB), 10 spare buffers, 7360 texture pages copied (28 MB) [hos host] vk: replay reads of guest memory as submitted: 6740000 (4154135 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 6760000 (4166426 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 295 dirty pages from their submits' captures, 57883 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 950 dispatches, 45054 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7429 runs, 746.3 MB compared; 85 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18581 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a frame: 249 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.07, present 0.15, draw 5.69, descriptors 2.45, texture lookups 0.08, uploads 0.18, pipelines 0.24, programs 1.02, commands 1.54, dispatch 0.76, memory sync 0.50, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 350 (1171 MB): 5.61 MB synced in 1048.1 copies, 17.36 MB written back [hos host] vk: 12240 presents, 60.0 fps; a frame: 1602 draws (0 skipped), 130 dispatches, 169 passes, 132 barriers, 6 submits, 0.5 uploads (1.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1308 textures (2503 MB) [hos host] vk: host time in GPU resource waits: 1.76 ms a frame (fences 1.59, timelines 0.17, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 628.4 full barriers recorded, 288.8 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.5, an op 823.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.6 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.15 MB unswizzled on the CPU in 0.17 ms [hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 39708874 lookups, 142783 missed (0.36 %) [hos host] vk: replay reads of guest memory as submitted: 6780000 (4179928 pages from the captures) [hos host] vk: 1 textures retired (1319 left, 2506 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: reaper: 515 images (1015 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 48 (86 MB) [hos host] vk: replay reads of guest memory as submitted: 6800000 (4193927 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 3799676 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 119, 1177 MB with memory of 1576 MB (398 MB spared); blocks given memory later 10517, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 775 binds (214.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 191 taken from the spares, 162 made new; 234 kept as spares, 0 destroyed; 43 spares now, 1052 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 538 dirty pages from their submits' captures, 56964 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 943 dispatches, 44224 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7359 runs, 730.2 MB compared; 85 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18097 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: 0x7f9c000000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 24304 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 9584186 kept (3539728 by the last picks' textures, a uniform buffer bound since), 26288 new, 3989525 with other textures (of 13600000) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 622 ranges, 3369 MB), 275 pages written, 275 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.021, bytes copied 0.000; 25 kept (126 MB), 11 spare buffers, 8785 texture pages copied (34 MB) [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #190000: event 0x100d6c802c mode 0 value 133982 (now 133963) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 6820000 (4205116 pages from the captures) [hos host] vk: write-backs left out (600 submits): 1100, 4884.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 59240 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 974 dispatches, 45919 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7893 runs, 768.2 MB compared; 91 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 98 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19136 in captures, 0 only in the collection, 44 only by the replay itself [hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 1; [hos ModuleSystemWorker5] vk: readbacks of GPU results: 1094 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1641 kept their virtual core (nothing there to wait for), 3223 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 21 textures retired (1305 left, 2498 MB): budget 21, drawn over 0, dead targets 0 [hos host] vk: 156 textures retired (1166 left, 2270 MB): budget 156, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6840000 (4216898 pages from the captures) [hos host] vk: texture #17032 0x10001fe0800 640x360x1 x1 NVN 0x14 the GPU drew goes with no copy of its texels (finalized, unused for 609 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #17004 0x10004ec1c00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 609 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #17001 0x10004b01a00 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 609 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #16995 0x10005552400 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 609 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 39 textures retired (1185 left, 2291 MB): budget 39, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 19 MB [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #135000: event 0x100d6c8010 mode 0 value 135001 (lr 0x80660bcc) [hos Presentation Thread] vi: 12352 frames presented (59.8 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 22, 15.5-17.8 254, 17.8-22 21, 22-30 2, 30-40 0, over 40 0 [hos host] vk: a frame: 187 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 8.76, present 0.15, draw 5.45, descriptors 2.42, texture lookups 0.20, uploads 0.30, pipelines 0.23, programs 0.94, commands 1.45, dispatch 0.76, memory sync 0.53, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 358 (1192 MB): 5.50 MB synced in 1047.5 copies, 17.43 MB written back [hos host] vk: 12360 presents, 59.3 fps; a frame: 1547 draws (0 skipped), 134 dispatches, 163 passes, 131 barriers, 6 submits, 1.3 uploads (1.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1205 textures (2306 MB) [hos host] vk: host time in GPU resource waits: 3.03 ms a frame (fences 1.73, timelines 1.31, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 611.1 full barriers recorded, 281.9 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 731.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.1 ms a frame (the longest submission 4.9 ms); 0.9 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.9 (1.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.75 MB unswizzled on the CPU in 0.26 ms [hos host] vk: replay reads of guest memory as submitted: 6860000 (4230340 pages from the captures) [hos host] vk: sparse mirrors: 125, 1194 MB with memory of 1608 MB (413 MB spared); blocks given memory later 10517, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 781 binds (216.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 197 taken from the spares, 162 made new; 234 kept as spares, 0 destroyed; 37 spares now, 1020 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1055 dirty pages from their submits' captures, 56413 from earlier ones', 17 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 940 dispatches, 44141 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7371 runs, 732.3 MB compared; 78 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18433 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 17; [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: 1 textures retired (1208 left, 2307 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: (its written part: 3 of 3 KB readable, 2016 bytes not zero; levels L0) [hos host] vk: texture 0x10faa26e00 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0): what the CPU wrote, 3 KB of it [hos host] vk: replay reads of guest memory as submitted: 6880000 (4241792 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.04 over 620 ranges, 3411 MB), 253 pages written, 253 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (118 MB), 12 spare buffers, 2716 texture pages copied (10 MB) [hos host] vk: 3 textures retired (1214 left, 2326 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos ModuleSystemWorker2] nvn: 1500000 recordings made, 73611 live copies (23.4 MB), 5950 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: replay reads of guest memory as submitted: 6900000 (4254905 pages from the captures) [hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1491 dirty pages from their submits' captures, 51816 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 875 dispatches, 40991 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7640 runs, 692.7 MB compared; 68 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17105 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: program picks: 9732966 kept (3602077 by the last picks' textures, a uniform buffer bound since), 26288 new, 4040745 with other textures (of 13800000) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9881 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 6920000 (4268132 pages from the captures) [hos host] vk: 20 textures retired (1240 left, 2412 MB): budget 20, drawn over 0, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 3 textures retired (1237 left, 2412 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 6940000 (4280516 pages from the captures) [hos ModuleSystemWorker4] nvn: DrawMeshTasksIndirect: 300000 so far (x1 0x1179e9c000 x2 0 x3 0x1179e9c000 x4 0x654040) [hos host] vk: 6 textures retired (1237 left, 2412 MB): budget 6, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: GPU waits in 5 s: 110 sync waits (431 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 128, 1207 MB with memory of 1628 MB (420 MB spared); blocks given memory later 10517, given back 12472; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 786 binds (217.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 201 taken from the spares, 162 made new; 235 kept as spares, 0 destroyed; 34 spares now, 1000 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1220 dirty pages from their submits' captures, 53354 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 895 dispatches, 41556 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8363 runs, 703.9 MB compared; 52 pages read into the shadows; 304 MB of shadows (52 MB of pages known, in 97 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17129 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 6960000 (4291812 pages from the captures) [hos aal::LowPrioWorkerThread] vcores: 151761 takes, 22324 waits (mean 66 us, longest 71.5 ms, 1 over HOS_VCORE_LOG_MS), 10711 kicks, 10619 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 6980000 (4303613 pages from the captures) [hos host] vk: 3 textures retired (1241 left, 2409 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: late pages' fix: stamps for 0x113c9f0000 (pool 0x113b400000; 576 blocks, 36864 KB) [hos Presentation Thread] vk: capture: a submit 0.33 ms (lock 0.02, scan 0.05 over 589 ranges, 3355 MB), 591 pages written, 591 copied (2.3 MB) [hos host] vk: a capture taken (ms): all 0.048; GPU wait for late pages 0.019 (1 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (752 MB), 3 spare buffers, 3622 texture pages copied (14 MB) [hos host] vk: a frame: 171 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.31, present 0.16, draw 6.96, descriptors 3.12, texture lookups 0.21, uploads 0.31, pipelines 0.28, programs 1.21, commands 1.83, dispatch 0.89, memory sync 0.82, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 363 (1209 MB): 7.22 MB synced in 1054.7 copies, 17.97 MB written back [hos host] vk: 12480 presents, 55.2 fps; a frame: 1979 draws (0 skipped), 166 dispatches, 174 passes, 131 barriers, 6 submits, 1.1 uploads (2.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1241 textures (2409 MB) [hos host] vk: host time in GPU resource waits: 4.98 ms a frame (fences 1.49, timelines 3.48, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 682.3 full barriers recorded, 291.3 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 1104.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.1 ms a frame (the longest submission 11.6 ms); 0.7 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.6 (1.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.86 MB unswizzled on the CPU in 0.27 ms [hos host] vk: replay reads of guest memory as submitted: 7000000 (4314559 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 3866900 pages so far (0 held: may be written ahead in their submission) [hos sead::MainThread] vk: readbacks of GPU results: 1059 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 19 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1680 kept their virtual core (nothing there to wait for), 3184 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 21296 dirty pages from their submits' captures, 88983 from earlier ones', 4 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1068 dispatches, 47535 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9168 runs, 760.4 MB compared; 254 pages read into the shadows; 316 MB of shadows (51 MB of pages known, in 101 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18908 in captures, 118 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7f81400000 (4 MB, captured, state 2): 4; [hos sead::MainThread] nvn: CommandBufferWaitEvent #195000: event 0x100d6c802c mode 0 value 137782 (now 137763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.16 [hos host] vk: replay reads of guest memory as submitted: 7020000 (4326198 pages from the captures) [hos host] vk: program picks: 9883916 kept (3671438 by the last picks' textures, a uniform buffer bound since), 26288 new, 4089795 with other textures (of 14000000) [hos host] vk: 8 textures retired (1233 left, 2408 MB): budget 8, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: EventSignal (CPU) #25000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos rmbl::VibrationThread] hid: 40960 vibration values, strongest 0.00 since the last count [hos host] vk: 8 textures retired (1225 left, 2390 MB): budget 8, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 7040000 (4335886 pages from the captures) [hos host] vk: 1 textures retired (1224 left, 2390 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: sparse mirrors: 128, 1139 MB with memory of 1628 MB (488 MB spared); blocks given memory later 10636, given back 13668; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 806 binds (222.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 201 taken from the spares, 162 made new; 235 kept as spares, 0 destroyed; 34 spares now, 1000 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1999 dirty pages from their submits' captures, 50765 from earlier ones', 7 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 873 dispatches, 39956 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8487 runs, 687.8 MB compared; 78 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18029 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: 0x7f9ac00000 (20 MB, captured, state 2): 6; 0x7f81400000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 185 live (2304 MB), 0 retired still imported (0 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 7060000 (4347328 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21374 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 9 textures retired (1217 left, 2384 MB): budget 9, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 7080000 (4358466 pages from the captures) [hos host] vk: 15 textures retired (1202 left, 2383 MB): budget 15, drawn over 0, dead targets 0 [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1364 dirty pages from their submits' captures, 52022 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 877 dispatches, 40811 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8716 runs, 707.0 MB compared; 63 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19120 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: late pages found at the replay (600 takes): 38 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: presenter thread: the copy's fence 1.54 ms, the present 0.73 ms a present [hos Presentation Thread] vk: capture: a submit 0.24 ms (lock 0.03, scan 0.04 over 594 ranges, 3363 MB), 309 pages written, 309 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.042; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 24 kept (104 MB), 11 spare buffers, 6879 texture pages copied (26 MB) [hos host] vk: 3 textures retired (1199 left, 2382 MB): budget 3, drawn over 0, dead targets 0 [hos ModuleSystemWorker2] nvn: 1550000 recordings made, 69412 live copies (22.9 MB), 5950 texture copies; 5 command buffers' segments (4.3 MB) [hos host] vk: replay reads of guest memory as submitted: 7100000 (4368349 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 7120000 (4380519 pages from the captures) [hos host] vk: pool 0x7fab000000..0x7faf400000 (68 MB) trimmed to its 2 live part(s), 12 MB: the rest outside every game pool for 1856 submissions (2 trims, 92 MB in all) [hos host] vk: 365 pools mirrored in VRAM (1137 MB) [hos host] vk: program picks: 10034788 kept (3741539 by the last picks' textures, a uniform buffer bound since), 26288 new, 4138923 with other textures (of 14200000) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: memory: working set 15334 MB (peak 15952), private 21569 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 11 blocks 864 MB, 35 allocations 836 MB; used 3200 of 15579 MB; heap 1 (device): 166 blocks 5047 MB, 20136 allocations 4235 MB; used 7042 of 8466 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 107 images 339 MB; 65 free ranges, the largest 28 MB [hos host] vk: image blocks (textures, compacted): 57 blocks 1824 MB, 1091 images 1418 MB; 410 free ranges, the largest 32 MB [hos host] vk: render target images: 5753 reused from the pool, 3435 made new, 3326 destroyed from it, 993 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1294 with the pool empty (0 of them with targets still in the dead queue), 287 with none in it a present yet, 25 with none of a kind they can use, 467 with all smaller (the largest 51% short on average), 506 with all bigger (over 1.5 times), 855 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3134 idle [hos host] vk: VRAM: 7042 MB used of 8466 MB allowed (811 MB of it free in the blocks); textures 2394 MB (budget 2107 MB), 1204 textures [hos host] vk: budget: VRAM 5047 of 8466 MB (7042 MB by the driver, 1995 MB unaccounted): textures 2394, mirrors 1136, blocks free 811, shadows 316, other buffers 213, other images 111, staging 64; RAM 3184 MB (the process's private memory 21569 MB): imports 2248, capture spares 750, captures 86, retired imports 68, capture table 32, others 0 [hos host] vk: textures: decoded ASTC 1417 MB, rendered 540 MB (88 drawn in one frame, kept: 45085 KB), other 405 MB; used in the last 8 presents 1729 MB [hos host] vk: decoded ASTC used again: 227 textures (552 MB) copied from ones the game finalized, not decoded; 227 of those retired, 6 listed; 1412 MB of blocks hashed in 239 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 7453 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 28 (8783 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 102x30x1 NVN 0x25: 1 (11 KB); 137x134x1 NVN 0x1: 1 (17 KB); 202x38x1 NVN 0x25: 1 (29 KB); 204x106x1 NVN 0x1: 1 (21 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 8 (1407 KB); 266x102x1 NVN 0x1: 1 (26 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); [hos host] vk: sparse mirrors: 128, 1136 MB with memory of 1568 MB (431 MB spared); blocks given memory later 10636, given back 13668; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 810 binds (224.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 203 taken from the spares, 162 made new; 237 kept as spares, 0 destroyed; 34 spares now, 1060 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 881 dirty pages from their submits' captures, 53011 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 854 dispatches, 41092 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8773 runs, 707.0 MB compared; 68 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19032 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos Presentation Thread] vk: presents after the CPU's signal (600): 267 waited for it, 3.40 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 85.5 ms; VRAM 6.9 of 8.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6168 queued, 6168 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15261 ms and 331.4 MB read in all); 8363 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11712 graphics pipelines (1 not in a cache), 713 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2220 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9002 (6168 ready, 0 loading), 16669 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13188 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 763 known, 6 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 184 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.04, present 0.18, draw 7.37, descriptors 3.17, texture lookups 0.14, uploads 0.13, pipelines 0.30, programs 1.27, commands 2.02, dispatch 1.01, memory sync 0.78, queue submit 0.00, slot wait 0.28 [hos host] vk: a frame: mirrors 365 (1136 MB): 5.41 MB synced in 1049.3 copies, 18.12 MB written back [hos host] vk: 12600 presents, 57.4 fps; a frame: 1967 draws (0 skipped), 176 dispatches, 168 passes, 131 barriers, 6 submits, 0.5 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1204 textures (2394 MB) [hos host] vk: host time in GPU resource waits: 4.08 ms a frame (fences 1.49, timelines 2.59, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 670.2 full barriers recorded, 285.0 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1100.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.8 ms a frame (the longest submission 8.2 ms); 0.4 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.4 (0.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.59 MB unswizzled on the CPU in 0.10 ms [hos host] vk: the render thread (600 presents, 17.14 ms a frame): its jobs 15.24 ms a frame (submits 5.71, presents 5.06, others 4.47), idle 1.90; the game waited for room at 33 presents, 0.42 ms each (0.02 a frame; 10.37 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 268 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 1481 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.50 ms after its hand-over (34.54 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.04 + 0.28 (the window's copy is in the idle), after it 1.19 + 0.44; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.31 parts a frame recorded in the signal's job, 74.5% of them took nothing that came with the signal; of the others, 2.12 a frame took pages of the signal's capture into a mirror (114.0 pages a frame) and 0.01 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 14 of 267 frames, its first three in 14 [hos host] vk: frame ahead, what came with the signal: 268 signals with a capture, each of 516.6 pages in 11.9 pools; between a frame's last submit and that signal 0.07 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (5 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.16+0.07, 0.49+0.14, 0.03+0.02, 0.04+0.02, 0.16+0.08, 0.20+0.07, 0.07+0.02, 0.03+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: replay reads of guest memory as submitted: 7140000 (4391213 pages from the captures) [hos host] vk: render targets drawn over from another start, let go: 9956 so far (47138 MB), 8939 of them (42402 MB) between the passes [hos host] vk: 3 textures retired (1210 left, 2396 MB): budget 3, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: CommandBufferSignalEvent #140000: event 0x100d6c8018 mode 0 value 140000 (lr 0x80660c94) [hos host] vk: replay reads of guest memory as submitted: 7160000 (4401260 pages from the captures) [hos Presentation Thread] vi: 12637 frames presented (56.8 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 48, 15.5-17.8 183, 17.8-22 32, 22-30 16, 30-40 4, over 40 2 [hos host] vk: 1 textures retired (1209 left, 2380 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 273 dirty pages from their submits' captures, 60488 from earlier ones', 10770 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 934 dispatches, 44325 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9328 runs, 760.6 MB compared; 69 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20555 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb5c00000 (44 MB, captured, state 2): 10770; [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1108 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1516 kept their virtual core (nothing there to wait for), 2836 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 7180000 (4413672 pages from the captures) [hos host] vk: 240 mirrors given back (1810 MB in all); 1108 MB of mirrors now, 172 MB of them idle and kept for what the GPU wrote there alone [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 597 ranges, 3321 MB), 282 pages written, 282 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (74 MB), 9 spare buffers, 6880 texture pages copied (26 MB) [hos host] vk: 4 textures retired (1205 left, 2379 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5118 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 7200000 (4423823 pages from the captures) [hos host] vk: 1 textures retired (1204 left, 2379 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: readback pages refreshed from guest memory: 3934517 pages so far (0 held: may be written ahead in their submission) [hos host] vk: replay reads of guest memory as submitted: 7220000 (4434612 pages from the captures) [hos host] vk: sparse mirrors: 116, 1079 MB with memory of 1480 MB (400 MB spared); blocks given memory later 10636, given back 13668; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 817 binds (224.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 203 taken from the spares, 162 made new; 249 kept as spares, 0 destroyed; 46 spares now, 1148 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 654 dirty pages from their submits' captures, 57345 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 928 dispatches, 43261 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9149 runs, 743.6 MB compared; 239 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20050 in captures, 129 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos host] vk: program picks: 10185445 kept (3810041 by the last picks' textures, a uniform buffer bound since), 26290 new, 4188264 with other textures (of 14400000) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #200000: event 0x100d6c802c mode 0 value 141582 (now 141563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 7240000 (4444192 pages from the captures) [hos host] vk: texture #17002 0x10004fb1e00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 600 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 1 textures retired (1203 left, 2378 MB): budget 1, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: recorder: 5668541 packets run in 10.0 s (busy 1.64 s, spun idle 0.11 s, slept 10394 times, 3169 of them under 200 us), 0 calls made directly (no room); 17297 batches over 583 frames (29.7 a frame, 328 packets a batch), 8506 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 73 times, 5.97 ms (0.1 times, 0.01 ms a frame); every thread 803 times, 57.59 ms in all, the longest 1.42 ms; vkWaitForFences 508/35.80 ms (max 0.50), vmaDestroyBuffer 164/11.77 ms (max 1.42), vmaFreeMemory 16/2.99 ms (max 0.75), vkQueueBindSparse 52/2.70 ms (max 0.40), the queue's lock 21/2.05 ms (max 0.25), vkDestroyImageView 24/1.62 ms (max 0.31) [hos host] vk: recorder: of its busy 1.64 s, 0.41 s in parts' ends and submits (the longest 1.74 ms), 0.10 s of that blocked on the queue's lock (198 times, the longest 1.68 ms), so 0.31 s in the ends, submits and holds; 1.23 s running the other packets; 9723 packets a frame [hos host] vk: a frame: 180 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.02, present 0.20, draw 6.25, descriptors 2.65, texture lookups 0.11, uploads 0.01, pipelines 0.27, programs 1.13, commands 1.71, dispatch 0.89, memory sync 0.44, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 365 (1079 MB): 5.86 MB synced in 1049.4 copies, 18.12 MB written back [hos host] vk: 12720 presents, 60.1 fps; a frame: 1914 draws (0 skipped), 177 dispatches, 164 passes, 131 barriers, 6 submits, 0.2 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1203 textures (2378 MB) [hos host] vk: host time in GPU resource waits: 1.88 ms a frame (fences 1.44, timelines 0.44, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 642.5 full barriers recorded, 279.0 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 1063.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.8 ms a frame (the longest submission 5.2 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 7260000 (4454760 pages from the captures) [hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 162 dirty pages from their submits' captures, 59005 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 944 dispatches, 45270 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9490 runs, 778.0 MB compared; 55 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21156 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: texture #17003 0x100050a2000 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 606 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 1 textures retired (1202 left, 2377 MB): budget 1, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: replay reads of guest memory as submitted: 7280000 (4466694 pages from the captures) [hos BaseProcExecutor2] vcores: 151691 takes, 26919 waits (mean 71 us, longest 40.2 ms, 0 over HOS_VCORE_LOG_MS), 13432 kicks, 13332 yields, 0 starved grants, 0 escapes in 5 s [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 590 ranges, 3308 MB), 270 pages written, 270 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (74 MB), 11 spare buffers, 8474 texture pages copied (33 MB) [hos host] vk: replay reads of guest memory as submitted: 7300000 (4476711 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9824 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 113, 1067 MB with memory of 1464 MB (396 MB spared); blocks given memory later 10636, given back 13762; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 822 binds (226.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 203 taken from the spares, 162 made new; 252 kept as spares, 0 destroyed; 49 spares now, 1164 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 124 dirty pages from their submits' captures, 57618 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 940 dispatches, 44339 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9304 runs, 760.3 MB compared; 58 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20621 in captures, 1 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 7320000 (4488442 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] nvn: GPU waits in 5 s: 124 sync waits (408 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: 6 textures retired (1196 left, 2357 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 7340000 (4497832 pages from the captures) [hos host] vk: program picks: 10335912 kept (3879975 by the last picks' textures, a uniform buffer bound since), 26290 new, 4237797 with other textures (of 14600000) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1121 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1696 kept their virtual core (nothing there to wait for), 3168 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 239 dirty pages from their submits' captures, 59099 from earlier ones', 2478 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 912 dispatches, 43329 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8994 runs, 740.2 MB compared; 57 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19815 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fbc000000 (12 MB, captured, state 2): 2478; [hos host] vk: replay reads of guest memory as submitted: 7360000 (4509281 pages from the captures) [hos host] vk: 1 textures retired (1203 left, 2361 MB): budget 1, drawn over 0, dead targets 0 [hos BaseProcExecutor1] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 19190882 lookups, 141020 missed (0.73 %) [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 320000 so far (x1 0x117c952020 x2 0 x3 0x117c952000 x4 0x168040) [hos host] vk: replay reads of guest memory as submitted: 7380000 (4518259 pages from the captures) [hos host] vk: 5 textures retired (1198 left, 2346 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: a frame: 159 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.83, present 0.15, draw 6.25, descriptors 2.64, texture lookups 0.08, uploads 0.00, pipelines 0.26, programs 1.13, commands 1.73, dispatch 0.90, memory sync 0.43, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 366 (1067 MB): 5.54 MB synced in 1051.7 copies, 18.03 MB written back [hos host] vk: 12840 presents, 59.9 fps; a frame: 1889 draws (0 skipped), 171 dispatches, 165 passes, 131 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1198 textures (2346 MB) [hos host] vk: host time in GPU resource waits: 1.46 ms a frame (fences 1.34, timelines 0.11, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 644.9 full barriers recorded, 281.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 1024.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.0 ms a frame (the longest submission 4.2 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 115, 1069 MB with memory of 1476 MB (406 MB spared); blocks given memory later 10637, given back 13762; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 827 binds (227.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 253 kept as spares, 0 destroyed; 47 spares now, 1152 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 252 dirty pages from their submits' captures, 54746 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 42145 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8393 runs, 715.2 MB compared; 52 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19455 in captures, 1 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos host] vk: 55 textures retired (1154 left, 2237 MB): budget 55, drawn over 0, dead targets 0 [hos sead::MainThread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.04 over 590 ranges, 3308 MB), 303 pages written, 303 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (124 MB), 12 spare buffers, 6950 texture pages copied (27 MB) [hos host] vk: replay reads of guest memory as submitted: 7400000 (4530623 pages from the captures) [hos host] vk: 21 textures retired (1135 left, 2193 MB): budget 21, drawn over 0, dead targets 0 [hos host] vk: reaper: 447 images (671 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 136 (191 MB) [hos host] vk: 22 textures retired (1113 left, 2132 MB): budget 22, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 7420000 (4542497 pages from the captures) [hos host] vk: queues ordered by their events: 78000 submissions stopped at a wait, 77951 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 204385, not met 0; captures at a CPU event's wait 8093, 0 of them at a wait met on arrival [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14323 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos ModuleSystemWorker2] nvn: CommandBufferSignalEvent #145000: event 0x100d6c8024 mode 0 value 145000 (lr 0x80665068) [hos host] vk: readback pages refreshed from guest memory: 4002129 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 923 dirty pages from their submits' captures, 51522 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 850 dispatches, 40169 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8067 runs, 679.3 MB compared; 85 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18273 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: program picks: 10487821 kept (3952924 by the last picks' textures, a uniform buffer bound since), 26290 new, 4285888 with other textures (of 14800000) [hos host] vk: replay reads of guest memory as submitted: 7440000 (4553997 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #205000: event 0x100d6c802c mode 0 value 145382 (now 145363) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: 24 textures retired (1093 left, 2102 MB): budget 24, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 7460000 (4562497 pages from the captures) [hos host] vk: sparse mirrors: 115, 1069 MB with memory of 1476 MB (406 MB spared); blocks given memory later 10637, given back 13762; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 827 binds (227.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 253 kept as spares, 0 destroyed; 47 spares now, 1152 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 730 dirty pages from their submits' captures, 54150 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 874 dispatches, 40206 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8121 runs, 680.4 MB compared; 65 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18672 in captures, 1 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos Presentation Thread] vi: 12938 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 25, 15.5-17.8 254, 17.8-22 21, 22-30 1, 30-40 0, over 40 0 [hos host] vk: replay reads of guest memory as submitted: 7480000 (4573281 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.01, scan 0.05 over 576 ranges, 3303 MB), 290 pages written, 290 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.029; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (112 MB), 10 spare buffers, 10983 texture pages copied (42 MB) [hos host] vk: a frame: 163 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.73, present 0.15, draw 6.84, descriptors 2.86, texture lookups 0.11, uploads 0.02, pipelines 0.29, programs 1.24, commands 1.90, dispatch 0.87, memory sync 0.60, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 368 (1054 MB): 5.54 MB synced in 1055.8 copies, 17.90 MB written back [hos host] vk: 12960 presents, 60.0 fps; a frame: 2105 draws (0 skipped), 163 dispatches, 178 passes, 131 barriers, 6 submits, 0.3 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1096 textures (2103 MB) [hos host] vk: host time in GPU resource waits: 2.64 ms a frame (fences 1.37, timelines 1.27, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 700.1 full barriers recorded, 297.3 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.9 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1099.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.9 ms a frame (the longest submission 3.8 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 7500000 (4583982 pages from the captures) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 303 dirty pages from their submits' captures, 55149 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 950 dispatches, 41871 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8359 runs, 698.7 MB compared; 74 pages read into the shadows; 316 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19438 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fb5c00000 (44 MB, captured, state 2): 1; [hos sead::MainThread] vk: readbacks of GPU results: 1118 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 10 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1482 kept their virtual core (nothing there to wait for), 2614 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: (its written part: 3 of 3 KB readable, 1560 bytes not zero; levels L0) [hos host] vk: texture 0x10d787b400 2480x2480x1 x1 levels 1 NVN 0x4b (6045 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it [hos host] vk: (its written part: 3 of 3 KB readable, 1560 bytes not zero; levels L0) [hos host] vk: texture 0x10d787b400 2480x2480x1 x1 levels 1 NVN 0x4b (6045 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it [hos host] vk: program picks: 10641908 kept (4026683 by the last picks' textures, a uniform buffer bound since), 26290 new, 4331801 with other textures (of 15000000) [hos host] vk: replay reads of guest memory as submitted: 7520000 (4595958 pages from the captures) [hos host] vk: VRAM compaction 5 begins: the textures' blocks hold 1696 MB, 552 MB of it free space [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21770 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06 [hos host] vk: sparse mirrors: 113, 1053 MB with memory of 1452 MB (398 MB spared); blocks given memory later 10638, given back 13785; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 833 binds (228.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 255 kept as spares, 0 destroyed; 49 spares now, 1176 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 345 dirty pages from their submits' captures, 54029 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 891 dispatches, 41518 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8488 runs, 700.9 MB compared; 67 pages read into the shadows; 316 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19270 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 7540000 (4605855 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture slots registered again at another size: 600000 (of 759767 registrations; slot 7164 1280x720 -> 2560x1440) [hos ModuleSystemWorker3] nvn: 1600000 recordings made, 70455 live copies (23.0 MB), 5950 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: replay reads of guest memory as submitted: 7560000 (4614180 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.05 over 574 ranges, 3309 MB), 282 pages written, 282 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.032; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.024, bytes copied 0.000; 24 kept (126 MB), 12 spare buffers, 9513 texture pages copied (37 MB) [hos BaseProcExecutor0] vcores: 154595 takes, 24209 waits (mean 67 us, longest 11.5 ms, 0 over HOS_VCORE_LOG_MS), 11597 kicks, 11466 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 759 dirty pages from their submits' captures, 52291 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 40800 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8563 runs, 701.9 MB compared; 64 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18666 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: VRAM compaction 5 over: 386 textures moved (408 MB) in 33 passes, 0 left where they were; 15 blocks (480 MB) given back to the driver, 3328 MB in all so far [hos host] vk: replay reads of guest memory as submitted: 7580000 (4625021 pages from the captures) [hos host] vk: program picks: 10798108 kept (4104657 by the last picks' textures, a uniform buffer bound since), 26290 new, 4375601 with other textures (of 15200000) [hos host] vk: replay reads of guest memory as submitted: 7600000 (4635916 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 4069375 pages so far (0 held: may be written ahead in their submission) [hos host] vk: a frame: 142 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.14, present 0.37, draw 6.41, descriptors 2.78, texture lookups 0.19, uploads 0.02, pipelines 0.27, programs 1.11, commands 1.75, dispatch 0.89, memory sync 0.52, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 368 (1053 MB): 5.42 MB synced in 1059.9 copies, 17.98 MB written back [hos host] vk: 13080 presents, 60.0 fps; a frame: 2044 draws (0 skipped), 168 dispatches, 178 passes, 131 barriers, 6 submits, 0.2 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1106 textures (2106 MB) [hos host] vk: host time in GPU resource waits: 1.83 ms a frame (fences 1.32, timelines 0.50, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 701.1 full barriers recorded, 297.0 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1042.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.9 ms a frame (the longest submission 4.8 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: texture #16653 0x116e25b000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 13081 (drawn at 11285) [hos host] vk: sparse mirrors: 113, 1053 MB with memory of 1452 MB (398 MB spared); blocks given memory later 10638, given back 13786; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 834 binds (229.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 255 kept as spares, 0 destroyed; 49 spares now, 1176 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 161 dirty pages from their submits' captures, 55374 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 42447 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8687 runs, 718.1 MB compared; 72 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19743 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 83 sync waits (209 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos sead::MainThread] nvn: CommandBufferWaitEvent #210000: event 0x100d6c802c mode 0 value 149182 (now 149163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: a submission's parts ended: 51879 where the game had ended its render pass, 8121 at 512 draws [hos host] vk: replay reads of guest memory as submitted: 7620000 (4644358 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17923 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 577 dirty pages from their submits' captures, 54482 from earlier ones', 32 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 875 dispatches, 41865 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8421 runs, 700.7 MB compared; 54 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19328 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: 0x7fe2800000 (4 MB, captured, state 2): 32; [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1118 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1451 kept their virtual core (nothing there to wait for), 3157 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 7640000 (4655028 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.04 over 573 ranges, 3319 MB), 285 pages written, 285 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 6150 texture pages copied (24 MB) [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #150000: event 0x100d6c802c mode 0 value 149999 (lr 0x80667688) [hos host] vk: replay reads of guest memory as submitted: 7660000 (4664788 pages from the captures) [hos host] vk: sparse mirrors: 113, 1052 MB with memory of 1452 MB (399 MB spared); blocks given memory later 10638, given back 13799; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 835 binds (229.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 206 taken from the spares, 162 made new; 255 kept as spares, 0 destroyed; 49 spares now, 1176 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 388 dirty pages from their submits' captures, 54538 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 881 dispatches, 41999 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8435 runs, 701.6 MB compared; 54 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19361 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos host] vk: presenter thread: the copy's fence 1.35 ms, the present 0.74 ms a present [hos host] vk: program picks: 10954590 kept (4186845 by the last picks' textures, a uniform buffer bound since), 26292 new, 4419117 with other textures (of 15400000) [hos ModuleSystemWorker2] barrier: 209606 looks: 78534 found the masks moved already, 128670 waited until they moved, 2402 for the whole 1 ms (a wait 298 us on average); 208977 wakes [hos host] vk: replay reads of guest memory as submitted: 7680000 (4675370 pages from the captures) [hos host] vk: 256 mirrors given back (1863 MB in all); 1070 MB of mirrors now, 172 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: replay reads of guest memory as submitted: 7700000 (4686225 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 517 dirty pages from their submits' captures, 57933 from earlier ones', 24 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 902 dispatches, 42184 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8429 runs, 701.5 MB compared; 63 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19236 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fbc000000 (12 MB, captured, state 2): 19; 0x1043000000 (4 MB, captured, state 2): 5; [hos host] vk: memory: working set 15224 MB (peak 15952), private 20919 MB, staging 64 MB in 4 arenas, 40 made so far (199 ms) [hos host] vk: memory heaps: heap 0 (system): 12 blocks 864 MB, 35 allocations 840 MB; used 3200 of 15579 MB; heap 1 (device): 148 blocks 4471 MB, 19027 allocations 3928 MB; used 6466 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 17 blocks 544 MB, 112 images 355 MB; 72 free ranges, the largest 17 MB [hos host] vk: image blocks (textures, compacted): 39 blocks 1248 MB, 1030 images 1162 MB; 274 free ranges, the largest 19 MB [hos host] vk: render target images: 5753 reused from the pool, 3478 made new, 3364 destroyed from it, 993 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1325 with the pool empty (0 of them with targets still in the dead queue), 288 with none in it a present yet, 25 with none of a kind they can use, 470 with all smaller (the largest 51% short on average), 509 with all bigger (over 1.5 times), 860 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3172 idle [hos host] vk: VRAM: 6466 MB used of 11474 MB allowed (542 MB of it free in the blocks); textures 2153 MB (budget 7158 MB), 1148 textures [hos host] vk: budget: VRAM 4471 of 11474 MB (6466 MB by the driver, 1995 MB unaccounted): textures 2153, mirrors 1070, blocks free 542, shadows 316, other buffers 213, other images 111, staging 64; RAM 3189 MB (the process's private memory 20919 MB): imports 2248, capture spares 754, captures 86, retired imports 68, capture table 33, others 0 [hos host] vk: textures: decoded ASTC 1218 MB, rendered 563 MB (93 drawn in one frame, kept: 70981 KB), other 332 MB; used in the last 8 presents 1737 MB [hos host] vk: decoded ASTC used again: 333 textures (771 MB) copied from ones the game finalized, not decoded; 333 of those retired, 8 listed; 1456 MB of blocks hashed in 249 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 8053 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 29 (8959 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 73x70x1 NVN 0x25: 1 (19 KB); 102x30x1 NVN 0x25: 1 (11 KB); 137x134x1 NVN 0x1: 1 (17 KB); 202x38x1 NVN 0x25: 1 (29 KB); 204x106x1 NVN 0x1: 1 (21 KB); 208x105x1 NVN 0x1: 1 (21 KB); 260x260x1 NVN 0xd: 9 (1583 KB); 266x102x1 NVN 0x1: 1 (26 KB); 298x41x1 NVN 0x25: 1 (47 KB); 328x33x1 NVN 0x25: 1 (42 KB); 349x42x1 NVN 0x25: 1 (57 KB); 362x105x1 NVN 0x1: 1 (37 KB); 429x106x1 NVN 0x1: 1 (44 KB); 452x106x1 NVN 0x1: 1 (46 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 203 waited for it, 1.83 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 27.0 ms; VRAM 6.3 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6168 queued, 6168 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15261 ms and 331.4 MB read in all); 8363 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11712 graphics pipelines (1 not in a cache), 713 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2226 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9002 (6168 ready, 0 loading), 16176 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13188 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 763 known, 6 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 143 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.26, present 0.16, draw 6.55, descriptors 2.94, texture lookups 0.26, uploads 0.20, pipelines 0.27, programs 1.08, commands 1.76, dispatch 0.85, memory sync 0.50, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 371 (1070 MB): 5.78 MB synced in 1073.2 copies, 17.99 MB written back [hos host] vk: 13200 presents, 60.1 fps; a frame: 2094 draws (0 skipped), 168 dispatches, 173 passes, 131 barriers, 6 submits, 1.1 uploads (0.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1148 textures (2153 MB) [hos host] vk: host time in GPU resource waits: 2.56 ms a frame (fences 1.56, timelines 1.00, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 679.7 full barriers recorded, 290.4 left out (nothing recorded since the one before), 13.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 1131.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.5 ms a frame (the longest submission 4.6 ms); 1.2 textures made a frame: colour targets 0.1 (0.4 MB), depth 0.3 (1.1 MB), ASTC or BC 0.8 (1.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.86 MB unswizzled on the CPU in 0.14 ms [hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 15.01 ms a frame (submits 6.99, presents 4.63, others 3.39), idle 1.66; the game waited for room at 11 presents, 0.10 ms each (0.00 a frame; 5.74 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 209 captures at a CPU signal; the submits' waits read in 0.041 ms a frame; readbacks: 1193 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.24 ms after its hand-over (25.73 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.64 + 0.28 (the window's copy is in the idle), after it 1.00 + 0.31; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.78 parts a frame recorded in the signal's job, 80.8% of them took nothing that came with the signal; of the others, 1.68 a frame took pages of the signal's capture into a mirror (26.2 pages a frame) and 0.01 pages read as guest memory had them (12.0 pages a frame); a frame's first part took nothing in 40 of 209 frames, its first three in 40 [hos host] vk: frame ahead, what came with the signal: 209 signals with a capture, each of 66.9 pages in 9.0 pools; between a frame's last submit and that signal 0.01 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.18+0.06, 0.24+0.13, 0.03+0.02, 0.03+0.01, 0.09+0.02, 0.30+0.02, 0.05+0.03, 0.06+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: render targets drawn over from another start, let go: 9992 so far (47293 MB), 8975 of them (42557 MB) between the passes [hos host] vk: replay reads of guest memory as submitted: 7720000 (4697297 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17572 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] vk: capture: a submit 0.20 ms (lock 0.01, scan 0.05 over 565 ranges, 3326 MB), 341 pages written, 341 copied (1.3 MB) [hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 8826 texture pages copied (34 MB) [hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 340000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 7740000 (4706447 pages from the captures) [hos Presentation Thread] vi: 13239 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 26, 15.5-17.8 254, 17.8-22 19, 22-30 2, 30-40 0, over 40 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 115, 1065 MB with memory of 1476 MB (410 MB spared); blocks given memory later 10638, given back 13853; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 844 binds (230.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 210 taken from the spares, 162 made new; 257 kept as spares, 0 destroyed; 47 spares now, 1152 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 231 dirty pages from their submits' captures, 57408 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 945 dispatches, 44300 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9496 runs, 760.4 MB compared; 87 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20778 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 186 live (2248 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos host] vk: 2 textures retired (1157 left, 2161 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 7760000 (4717419 pages from the captures) [hos host] vk: 6 textures retired (1151 left, 2136 MB): budget 0, drawn over 6, dead targets 0 [hos host] vk: program picks: 11104857 kept (4257044 by the last picks' textures, a uniform buffer bound since), 26302 new, 4468840 with other textures (of 15600000) [hos host] vk: readback pages refreshed from guest memory: 4137034 pages so far (0 held: may be written ahead in their submission) [hos host] vk: replay reads of guest memory as submitted: 7780000 (4728206 pages from the captures) [hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 481 dirty pages from their submits' captures, 58992 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 977 dispatches, 45238 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9686 runs, 764.4 MB compared; 88 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21183 in captures, 2 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1107 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 10 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1334 kept their virtual core (nothing there to wait for), 2762 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos sead::MainThread] nvn: CommandBufferWaitEvent #215000: event 0x100d6c802c mode 0 value 152982 (now 152963) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.07 [hos host] vk: replay reads of guest memory as submitted: 7800000 (4738134 pages from the captures) [hos host] vk: recorder: 6141123 packets run in 10.0 s (busy 1.68 s, spun idle 0.11 s, slept 10520 times, 3228 of them under 200 us), 0 calls made directly (no room); 18321 batches over 600 frames (30.5 a frame, 335 packets a batch), 9010 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 86 times, 5.67 ms (0.1 times, 0.01 ms a frame); every thread 669 times, 38.33 ms in all, the longest 0.67 ms; vkWaitForFences 509/27.36 ms (max 0.48), vmaDestroyBuffer 33/3.13 ms (max 0.67), vkDestroyImageView 61/2.56 ms (max 0.38), vmaFreeMemory 21/2.23 ms (max 0.49), vkQueueBindSparse 31/1.58 ms (max 0.30), the queue's lock 6/1.08 ms (max 0.64) [hos host] vk: recorder: of its busy 1.68 s, 0.43 s in parts' ends and submits (the longest 1.56 ms), 0.11 s of that blocked on the queue's lock (243 times, the longest 1.46 ms), so 0.32 s in the ends, submits and holds; 1.25 s running the other packets; 10235 packets a frame [hos host] vk: a frame: 168 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.28, present 0.18, draw 5.67, descriptors 2.45, texture lookups 0.09, uploads 0.09, pipelines 0.24, programs 0.99, commands 1.55, dispatch 0.89, memory sync 0.52, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 374 (1065 MB): 5.46 MB synced in 1063.7 copies, 18.16 MB written back [hos host] vk: 13320 presents, 60.0 fps; a frame: 1699 draws (0 skipped), 179 dispatches, 155 passes, 131 barriers, 6 submits, 0.3 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13567 pipelines (3513 ms), 0 programs (10 failed), 1153 textures (2137 MB) [hos host] vk: host time in GPU resource waits: 1.39 ms a frame (fences 1.21, timelines 0.18, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 613.6 full barriers recorded, 272.1 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.7 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 937.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.9 ms a frame (the longest submission 4.4 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.56 MB unswizzled on the CPU in 0.08 ms [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 549 ranges, 3336 MB), 271 pages written, 271 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (86 MB), 10 spare buffers, 3822 texture pages copied (14 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5813 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 117, 1065 MB with memory of 1484 MB (418 MB spared); blocks given memory later 10640, given back 13858; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 851 binds (230.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 212 taken from the spares, 162 made new; 257 kept as spares, 0 destroyed; 45 spares now, 1144 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 58496 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 928 dispatches, 44688 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9585 runs, 762.8 MB compared; 60 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21106 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2256 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 7820000 (4750100 pages from the captures) [hos Presentation Thread] nvn: QueueSubmitCommands #80000: queue 000010100D5EB020 (its #13331), 8 handles (lr 0x805282f4) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos BaseProcExecutor0] vcores: 153669 takes, 22243 waits (mean 71 us, longest 12.5 ms, 0 over HOS_VCORE_LOG_MS), 10624 kicks, 10535 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 7840000 (4760843 pages from the captures) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 737 dirty pages from their submits' captures, 57067 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 954 dispatches, 44283 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9284 runs, 759.3 MB compared; 60 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20990 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x105dc00000 (4 MB, captured, state 2): 2; [hos host] vk: replay reads of guest memory as submitted: 7860000 (4773197 pages from the captures) [hos host] vk: program picks: 11254306 kept (4325002 by the last picks' textures, a uniform buffer bound since), 26312 new, 4519381 with other textures (of 15800000) [hos host] vk: replay reads of guest memory as submitted: 7880000 (4785337 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 57 sync waits (145 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06 [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #155000: event 0x100d6c8038 mode 0 value 154999 (lr 0x8066189c) [hos host] vk: sparse mirrors: 119, 1080 MB with memory of 1536 MB (455 MB spared); blocks given memory later 10642, given back 13858; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 854 binds (231.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 214 taken from the spares, 162 made new; 257 kept as spares, 0 destroyed; 43 spares now, 1092 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 989 dirty pages from their submits' captures, 56617 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 948 dispatches, 44155 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8673 runs, 748.1 MB compared; 78 pages read into the shadows; 316 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20758 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 189 live (2260 MB), 1 retired still imported (68 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 7900000 (4796852 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.18 ms (lock 0.00, scan 0.05 over 545 ranges, 3414 MB), 297 pages written, 297 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.039; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.028, bytes copied 0.000; 24 kept (90 MB), 11 spare buffers, 11044 texture pages copied (43 MB) [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 15368935 lookups, 134707 missed (0.88 %) [hos host] vk: texture 0x10001e00400 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since) [hos host] vk: replay reads of guest memory as submitted: 7920000 (4809590 pages from the captures) [hos host] vk: pool 0x1011f97c00+0xf00000 imported in 4.9 ms as 0x100f800000..0x1014000000 (227 imports, 693 ms in all) [hos host] vk: pool 0x100f800000..0x1014000000 took in 2 pools [hos host] vk: texture 0x10001e00400 2560x1440x1 x1 levels 1 NVN 0x34 (15360 KB of storage) re-uploaded (its memory was written; flags 0x8): a GPU write into its storage, 15360 KB of it, over what the GPU wrote (finalized since) [hos host] vk: 1 textures retired (1366 left, 2599 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: a frame: 195 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.69, present 0.18, draw 6.81, descriptors 3.61, texture lookups 1.23, uploads 0.94, pipelines 0.24, programs 0.97, commands 1.54, dispatch 0.92, memory sync 0.60, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 389 (1107 MB): 6.34 MB synced in 1053.6 copies, 17.94 MB written back [hos host] vk: 13440 presents, 56.1 fps; a frame: 1688 draws (0 skipped), 166 dispatches, 156 passes, 132 barriers, 6 submits, 2.3 uploads (2.7 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1409 textures (2642 MB) [hos host] vk: host time in GPU resource waits: 4.18 ms a frame (fences 1.76, timelines 2.43, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 618.5 full barriers recorded, 273.0 left out (nothing recorded since the one before), 14.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 26.3; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.4, an op 955.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.5 ms a frame (the longest submission 5.6 ms); 3.2 textures made a frame: colour targets 0.9 (2.7 MB), depth 0.2 (1.2 MB), ASTC or BC 2.1 (3.7 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.81 MB unswizzled on the CPU in 0.30 ms [hos host] [debug] vk: a staging arena of 22 MB took 5.8 ms to make [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 27955 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] vk: readbacks of GPU results: 1109 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 23 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1589 kept their virtual core (nothing there to wait for), 3275 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: reaper: 214 images (574 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 23 (74 MB) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1828 dirty pages from their submits' captures, 52671 from earlier ones', 76 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 904 dispatches, 41559 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8237 runs, 697.7 MB compared; 71 pages read into the shadows; 284 MB of shadows (51 MB of pages known, in 101 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19990 in captures, 1 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x7f98800000 (8 MB, captured, state 2): 55; 0x7f9ec00000 (4 MB, captured, state 2): 14; 0x7f8e000000 (4 MB, captured, state 2): 4; 0x7fa9000000 (4 MB, captured, state 2): 3; [hos host] vk: texture #15756 0x1001edcae00 122x98x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 5900 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15754 0x1001edcf000 58x34x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 5900 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15733 0x10000032200 800x62x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 4441 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 280 textures retired (1218 left, 2513 MB): budget 280, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 3 finalized, 0 MB [hos host] vk: replay reads of guest memory as submitted: 7940000 (4821890 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 45 textures retired (1193 left, 2473 MB): budget 45, drawn over 0, dead targets 0 [hos ModuleSystemWorker5] nvn: 1650000 recordings made, 70958 live copies (23.7 MB), 5974 texture copies; 4 command buffers' segments (4.2 MB) [hos host] vk: replay reads of guest memory as submitted: 7960000 (4832405 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 4204161 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 132, 1113 MB with memory of 1604 MB (490 MB spared); blocks given memory later 10646, given back 13865; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 874 binds (235.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 228 taken from the spares, 163 made new; 259 kept as spares, 0 destroyed; 31 spares now, 1096 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1161 dirty pages from their submits' captures, 50591 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 841 dispatches, 39635 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8031 runs, 665.4 MB compared; 54 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18152 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: 41 textures retired (1181 left, 2430 MB): budget 0, drawn over 0, dead targets 41 [hos host] vk: program picks: 11405083 kept (4392933 by the last picks' textures, a uniform buffer bound since), 26312 new, 4568604 with other textures (of 16000000) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #220000: event 0x100d6c802c mode 0 value 156782 (now 156763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 7980000 (4844384 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 547 ranges, 3442 MB), 276 pages written, 276 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (86 MB), 10 spare buffers, 2513 texture pages copied (9 MB) [hos host] vk: write-backs left out (600 submits): 1001, 4442.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 208 dirty pages from their submits' captures, 55305 from earlier ones', 103 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 921 dispatches, 42577 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8734 runs, 720.0 MB compared; 53 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19795 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fe1000000 (4 MB, captured, state 2): 45; 0x1043400000 (4 MB, captured, state 2): 35; 0x7fe2800000 (4 MB, captured, state 2): 11; 0x105dc00000 (4 MB, captured, state 2): 4; 0x7f9ac00000 (20 MB, captured, state 2): 3; 0x7fa3400000 (4 MB, captured, state 2): 2; 0x1044400000 (4 MB, captured, state 2): 2; 0x7fbc000000 (12 MB, captured, state 2): 1; [hos Presentation Thread] vi: 13526 frames presented (57.3 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 49, 15.5-17.8 199, 17.8-22 20, 22-30 10, 30-40 6, over 40 3 [hos host] vk: replay reads of guest memory as submitted: 8000000 (4857068 pages from the captures) [hos host] vk: texture #18378 0x100020e5a00 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 13549 (drawn at 13488) [hos host] vk: 2 textures retired (1205 left, 2437 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8020000 (4868864 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 4098 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a frame: 207 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.20, present 0.25, draw 6.47, descriptors 3.16, texture lookups 0.79, uploads 0.55, pipelines 0.25, programs 0.97, commands 1.61, dispatch 0.92, memory sync 0.46, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 391 (1110 MB): 5.46 MB synced in 1056.9 copies, 18.01 MB written back [hos host] vk: 13560 presents, 57.5 fps; a frame: 1681 draws (0 skipped), 172 dispatches, 165 passes, 133 barriers, 6 submits, 1.2 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1205 textures (2437 MB) [hos host] vk: host time in GPU resource waits: 3.75 ms a frame (fences 1.76, timelines 1.99, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 655.2 full barriers recorded, 282.5 left out (nothing recorded since the one before), 15.2 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 26.2; 0x12: 69.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.8 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.6, an op 961.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.3 ms a frame (the longest submission 4.1 ms); 1.4 textures made a frame: colour targets 0.2 (0.5 MB), depth 0.1 (0.4 MB), ASTC or BC 1.1 (1.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.44 MB unswizzled on the CPU in 0.07 ms [hos host] vk: sparse mirrors: 131, 1110 MB with memory of 1600 MB (490 MB spared); blocks given memory later 10647, given back 13865; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 876 binds (235.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 228 taken from the spares, 163 made new; 260 kept as spares, 0 destroyed; 32 spares now, 1100 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 154 dirty pages from their submits' captures, 57517 from earlier ones', 78 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 976 dispatches, 43965 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9363 runs, 758.8 MB compared; 60 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21542 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 78; [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 8040000 (4880458 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 8060000 (4891665 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 542 ranges, 3372 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (92 MB), 11 spare buffers, 4245 texture pages copied (16 MB) [hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 125 dirty pages from their submits' captures, 57528 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 909 dispatches, 43912 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9323 runs, 743.8 MB compared; 54 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20829 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: program picks: 11555015 kept (4458513 by the last picks' textures, a uniform buffer bound since), 26315 new, 4618669 with other textures (of 16200000) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1120 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1337 kept their virtual core (nothing there to wait for), 2759 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 8080000 (4902253 pages from the captures) [hos ModuleSystemWorker1] nvn: DrawMeshTasksIndirect: 360000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040) [hos PreActorModelCompose] vcores: 148170 takes, 22496 waits (mean 68 us, longest 7.4 ms, 0 over HOS_VCORE_LOG_MS), 10579 kicks, 10536 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 1 textures retired (1210 left, 2436 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 132, 1113 MB with memory of 1604 MB (490 MB spared); blocks given memory later 10651, given back 13865; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 880 binds (235.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 230 taken from the spares, 163 made new; 261 kept as spares, 0 destroyed; 31 spares now, 1096 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: 76 textures retired (1203 left, 2449 MB): budget 76, drawn over 0, dead targets 0 [hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 308 dirty pages from their submits' captures, 57390 from earlier ones', 23 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 967 dispatches, 43720 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9305 runs, 760.0 MB compared; 53 pages read into the shadows; 284 MB of shadows (51 MB of pages known, in 102 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21313 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7fa6800000 (4 MB, captured, state 2): 6; 0x7fa2000000 (8 MB, captured, state 2): 6; 0x7f9ec00000 (4 MB, captured, state 2): 6; 0x7fb5c00000 (44 MB, captured, state 2): 5; [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 8100000 (4913313 pages from the captures) [hos host] vk: 20 textures retired (1217 left, 2474 MB): budget 20, drawn over 0, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10246 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #160000: event 0x100d6c8044 mode 0 value 159999 (lr 0x802742f0) [hos host] vk: late pages' fix (600 recordings with one): of 1230 captures with late pages 1229 fixed (74.3 pages, 74.9 fixes a capture), waited as before: 0 no pool, 1 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 470 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 576 blocks (36864 KB); of the reads on the CPU, 1 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 171, of an earlier recording 298, not recorded yet 0, of this part 0; pages fixed from stamps 90276, from snapshots 1076; the shadows: 1228694 pages synced from captures, 1197451 made known by it, 31243 left unknown because the replay or the collection wrote them since [hos sead::MainThread] nvn: GPU waits in 5 s: 91 sync waits (564 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 8120000 (4924719 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #15753 0x1001ede6200 122x98x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4671 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15750 0x1001edea400 58x34x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4671 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 3 textures retired (1270 left, 2583 MB): budget 3, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 2 finalized, 0 MB [hos OptimizeChangeBanc] nvn: 1700000 recordings made, 69995 live copies (23.2 MB), 6003 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: a frame: 167 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.77, present 0.25, draw 6.35, descriptors 2.97, texture lookups 0.60, uploads 0.41, pipelines 0.25, programs 1.03, commands 1.64, dispatch 0.94, memory sync 0.93, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 394 (1128 MB): 7.65 MB synced in 1070.1 copies, 18.09 MB written back [hos host] vk: 13680 presents, 54.3 fps; a frame: 1706 draws (0 skipped), 175 dispatches, 160 passes, 131 barriers, 6 submits, 1.2 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1270 textures (2583 MB) [hos host] vk: host time in GPU resource waits: 3.46 ms a frame (fences 1.77, timelines 1.69, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 645.4 full barriers recorded, 277.8 left out (nothing recorded since the one before), 13.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.2, an op 919.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.8 ms a frame (the longest submission 5.5 ms); 2.0 textures made a frame: colour targets 0.3 (2.1 MB), depth 0.4 (2.0 MB), ASTC or BC 1.2 (1.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.60 MB unswizzled on the CPU in 0.11 ms [hos host] vk: readback pages refreshed from guest memory: 4271702 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 949, 4216.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 31530 dirty pages from their submits' captures, 89124 from earlier ones', 20 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1085 dispatches, 45848 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8127 runs, 682.7 MB compared; 108 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18515 in captures, 2 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: 0x7fe2800000 (4 MB, captured, state 2): 16; 0x7f98800000 (8 MB, captured, state 2): 2; 0x105dc00000 (4 MB, captured, state 2): 2; [hos host] vk: queues ordered by their events: 84000 submissions stopped at a wait, 83952 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 224729, not met 0; captures at a CPU event's wait 8398, 0 of them at a wait met on arrival [hos host] vk: texture #15760 0x1001edbea00 146x105x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4681 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15757 0x1001edc4c00 82x41x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4681 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 7 textures retired (1287 left, 2615 MB): budget 7, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 2 finalized, 0 MB [hos host] vk: replay reads of guest memory as submitted: 8140000 (4934601 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.40 ms (lock 0.04, scan 0.06 over 572 ranges, 3534 MB), 772 pages written, 772 copied (3.0 MB) [hos host] vk: a capture taken (ms): all 0.039; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.020, bytes copied 0.000; 25 kept (110 MB), 10 spare buffers, 7928 texture pages copied (30 MB) [hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #225000: event 0x100d6c802c mode 0 value 160582 (now 160563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.04, writes collected 0.16 [hos host] vk: texture #15739 0x1001ee02800 42x41x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4691 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 1 textures retired (1293 left, 2643 MB): budget 1, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: texture #17983 0x10008942600 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13708 (drawn at 13647) [hos host] vk: texture #18670 0x1000a0b2c00 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 13708 (drawn at 13647) [hos host] vk: texture #18528 0x10008612200 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13709 (drawn at 13647) [hos host] vk: texture #15592 0x1001ef3d200 640x360x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 8148 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15743 0x1001edfbc00 74x32x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15742 0x1001edfe600 106x105x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15747 0x1001edf2800 42x41x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15749 0x1001edee600 106x105x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #15746 0x1001edf5a00 138x96x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 4701 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 8 textures retired (1293 left, 2668 MB): budget 6, drawn over 2, dead targets 0; GPU-drawn with no copy kept: 0 live, 6 finalized, 0 MB [hos host] vk: program picks: 11706106 kept (4531198 by the last picks' textures, a uniform buffer bound since), 26319 new, 4667574 with other textures (of 16400000) [hos rmbl::VibrationThread] hid: 45056 vibration values, strongest 0.00 since the last count [hos host] vk: replay reads of guest memory as submitted: 8160000 (4946228 pages from the captures) [hos host] vk: texture #18778 0x10008612200 839x839 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13719 (drawn at 13657) [hos host] vk: texture #15337 0x1001f05fc00 637x133x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 4573 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 15 textures retired (1283 left, 2660 MB): budget 10, drawn over 5, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: texture #18779 0x10008612200 853x853 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13720 (drawn at 13658) [hos host] vk: texture #18781 0x10008612200 872x872 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13721 (drawn at 13659) [hos host] vk: texture #18782 0x10008612200 890x890 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 13722 (drawn at 13660) [hos host] vk: sparse mirrors: 133, 1132 MB with memory of 1612 MB (479 MB spared); blocks given memory later 11018, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 907 binds (242.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 231 taken from the spares, 163 made new; 261 kept as spares, 0 destroyed; 30 spares now, 1088 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1311 dirty pages from their submits' captures, 52085 from earlier ones', 22 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 875 dispatches, 40518 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8033 runs, 680.4 MB compared; 100 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18756 in captures, 2 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: 0x1043400000 (4 MB, captured, state 2): 14; 0x1044400000 (4 MB, captured, state 2): 8; [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: 51 textures retired (1241 left, 2619 MB): budget 43, drawn over 8, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8180000 (4957463 pages from the captures) [hos host] vk: 39 textures retired (1246 left, 2633 MB): budget 34, drawn over 5, dead targets 0 [hos host] vk: 8 new ASTC textures left for a later frame (this frame's uploads took their budget; 12000 so far) [hos host] vk: 23 textures retired (1229 left, 2605 MB): budget 14, drawn over 9, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8200000 (4969343 pages from the captures) [hos host] vk: 9 textures retired (1228 left, 2604 MB): budget 3, drawn over 6, dead targets 0 [hos host] vk: texture #16156 0x116d1f9000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 13762 (drawn at 9483) [hos host] vk: presenter thread: the copy's fence 1.41 ms, the present 0.71 ms a present [hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1346 dirty pages from their submits' captures, 52793 from earlier ones', 822 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 869 dispatches, 40816 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8070 runs, 680.5 MB compared; 108 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18388 in captures, 1 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x7fa5c00000 (4 MB, captured, state 2): 796; 0x7fa3800000 (8 MB, captured, state 2): 26; [hos sead::MainThread] vk: readbacks of GPU results: 1059 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 15 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1512 kept their virtual core (nothing there to wait for), 3096 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 6 textures retired (1229 left, 2601 MB): budget 0, drawn over 6, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8220000 (4980198 pages from the captures) [hos host] vk: 12 textures retired (1226 left, 2610 MB): budget 8, drawn over 4, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10077 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 585 ranges, 3555 MB), 247 pages written, 247 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (86 MB), 12 spare buffers, 8040 texture pages copied (31 MB) [hos host] vk: 7 textures retired (1227 left, 2609 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 321 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: memory: working set 15409 MB (peak 15952), private 21735 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 16 blocks 912 MB, 37 allocations 816 MB; used 3332 of 15579 MB; heap 1 (device): 170 blocks 5175 MB, 20210 allocations 4452 MB; used 7170 of 8421 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 19 blocks 608 MB, 165 images 531 MB; 111 free ranges, the largest 17 MB [hos host] vk: image blocks (textures, compacted): 59 blocks 1888 MB, 1065 images 1473 MB; 390 free ranges, the largest 21 MB [hos host] vk: render target images: 5853 reused from the pool, 3676 made new, 3509 destroyed from it, 1019 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 144 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1471 with the pool empty (0 of them with targets still in the dead queue), 296 with none in it a present yet, 34 with none of a kind they can use, 485 with all smaller (the largest 50% short on average), 528 with all bigger (over 1.5 times), 861 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3317 idle [hos host] vk: VRAM: 7170 MB used of 8421 MB allowed (723 MB of it free in the blocks); textures 2640 MB (budget 2200 MB), 1236 textures [hos host] vk: budget: VRAM 5175 of 8421 MB (7170 MB by the driver, 1995 MB unaccounted): textures 2640, mirrors 1138, blocks free 723, shadows 284, other buffers 213, other images 111, staging 64; RAM 3249 MB (the process's private memory 21735 MB): imports 2260, capture spares 722, retired imports 140, captures 92, capture table 33, other buffers 2, others 0 [hos host] vk: textures: decoded ASTC 1413 MB, rendered 765 MB (111 drawn in one frame, kept: 198416 KB), other 401 MB; used in the last 8 presents 1752 MB [hos host] vk: decoded ASTC used again: 359 textures (814 MB) copied from ones the game finalized, not decoded; 359 of those retired, 4 listed; 1607 MB of blocks hashed in 281 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 8653 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34, last used 2 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 52 (11349 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB); [hos host] vk: 10 textures retired (1226 left, 2621 MB): budget 8, drawn over 2, dead targets 0 [hos Presentation Thread] vk: presents after the CPU's signal (600): 237 waited for it, 4.24 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 53.1 ms; VRAM 7.0 of 8.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6171 queued, 6171 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15263 ms and 331.5 MB read in all); 8366 programs (1620 of them compute variants without their code, 0 graphics programs no note names), 11714 graphics pipelines (1 not in a cache), 714 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2241 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9005 (6171 ready, 0 loading), 16283 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13192 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 709 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 149 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.26, present 0.18, draw 6.99, descriptors 3.12, texture lookups 0.53, uploads 0.30, pipelines 0.30, programs 1.17, commands 1.87, dispatch 0.89, memory sync 0.66, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 395 (1138 MB): 5.68 MB synced in 1069.0 copies, 17.92 MB written back [hos host] vk: 13800 presents, 59.6 fps; a frame: 2048 draws (0 skipped), 165 dispatches, 177 passes, 132 barriers, 6 submits, 0.5 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1227 textures (2624 MB) [hos host] vk: host time in GPU resource waits: 4.24 ms a frame (fences 1.74, timelines 2.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 705.4 full barriers recorded, 296.6 left out (nothing recorded since the one before), 12.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.7, an op 1122.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.4 ms a frame (the longest submission 4.4 ms); 1.2 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.7 (2.5 MB), ASTC or BC 0.5 (0.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.18 MB unswizzled on the CPU in 0.03 ms [hos host] vk: the render thread (600 presents, 17.41 ms a frame): its jobs 15.39 ms a frame (submits 6.31, presents 4.74, others 4.34), idle 2.02; the game waited for room at 44 presents, 0.55 ms each (0.04 a frame; 13.31 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 244 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 1443 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.60 ms after its hand-over (50.32 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.24 + 0.34 (the window's copy is in the idle), after it 1.21 + 0.83; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.73 parts a frame recorded in the signal's job, 72.5% of them took nothing that came with the signal; of the others, 2.36 a frame took pages of the signal's capture into a mirror (164.2 pages a frame) and 0.04 pages read as guest memory had them (3.7 pages a frame); a frame's first part took nothing in 19 of 244 frames, its first three in 19 [hos host] vk: frame ahead, what came with the signal: 244 signals with a capture, each of 680.2 pages in 12.0 pools; between a frame's last submit and that signal 0.10 texture and sampler registrations, 0.09 releases and 0.00 other CPU signals a signal (15 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.22+0.10, 0.30+0.13, 0.06+0.02, 0.03+0.02, 0.14+0.12, 0.23+0.27, 0.14+0.15, 0.07+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: sparse mirrors: 134, 1138 MB with memory of 1620 MB (481 MB spared); blocks given memory later 11037, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 910 binds (243.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 232 taken from the spares, 163 made new; 261 kept as spares, 0 destroyed; 29 spares now, 1080 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 539 dirty pages from their submits' captures, 53984 from earlier ones', 867 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 887 dispatches, 41613 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8385 runs, 700.2 MB compared; 59 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19490 in captures, 2 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fa6400000 (4 MB, captured, state 2): 864; 0x7f81400000 (4 MB, captured, state 2): 3; [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: program picks: 11858053 kept (4605563 by the last picks' textures, a uniform buffer bound since), 26319 new, 4715627 with other textures (of 16600000) [hos host] vk: replay reads of guest memory as submitted: 8240000 (4988378 pages from the captures) [hos host] vk: render targets drawn over from another start, let go: 10207 so far (48225 MB), 9148 of them (43361 MB) between the passes [hos host] vk: 8 textures retired (1227 left, 2622 MB): budget 0, drawn over 8, dead targets 0 [hos Presentation Thread] vi: 13813 frames presented (57.3 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 50, 15.5-17.8 195, 17.8-22 24, 22-30 12, 30-40 4, over 40 2 [hos host] vk: 10 textures retired (1223 left, 2601 MB): budget 0, drawn over 10, dead targets 0 [hos host] vk: 10 textures retired (1219 left, 2586 MB): budget 0, drawn over 10, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8260000 (4995757 pages from the captures) [hos host] vk: 8 textures retired (1213 left, 2552 MB): budget 0, drawn over 8, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 172 dirty pages from their submits' captures, 54173 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 845 dispatches, 41442 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8406 runs, 700.7 MB compared; 54 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19533 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 5 textures retired (1217 left, 2560 MB): budget 0, drawn over 5, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8280000 (5005118 pages from the captures) [hos host] vk: 7 textures retired (1216 left, 2559 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: 9 textures retired (1216 left, 2556 MB): budget 0, drawn over 9, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8300000 (5015846 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 4338812 pages so far (0 held: may be written ahead in their submission) [hos host] vk: 3 textures retired (1221 left, 2576 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: sparse mirrors: 129, 1130 MB with memory of 1596 MB (466 MB spared); blocks given memory later 11094, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 917 binds (245.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 232 taken from the spares, 163 made new; 266 kept as spares, 0 destroyed; 34 spares now, 1104 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 425 dirty pages from their submits' captures, 55953 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 911 dispatches, 42436 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8365 runs, 700.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19808 in captures, 1 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 590 ranges, 3562 MB), 275 pages written, 275 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (106 MB), 12 spare buffers, 3159 texture pages copied (12 MB) [hos host] vk: recorder: 5597598 packets run in 10.0 s (busy 1.67 s, spun idle 0.11 s, slept 10270 times, 3164 of them under 200 us), 0 calls made directly (no room); 17345 batches over 573 frames (30.3 a frame, 323 packets a batch), 8640 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 161 times, 15.00 ms (0.3 times, 0.03 ms a frame); every thread 836 times, 73.03 ms in all, the longest 9.46 ms; vkWaitForFences 487/40.17 ms (max 9.46), vmaDestroyBuffer 101/11.51 ms (max 1.04), vkDestroyImageView 124/10.30 ms (max 0.34), vkQueueBindSparse 66/4.57 ms (max 0.44), vmaFreeMemory 12/1.74 ms (max 0.39), vkEndCommandBuffer 12/1.63 ms (max 0.42) [hos host] vk: recorder: of its busy 1.67 s, 0.42 s in parts' ends and submits (the longest 1.36 ms), 0.10 s of that blocked on the queue's lock (185 times, the longest 1.23 ms), so 0.32 s in the ends, submits and holds; 1.25 s running the other packets; 9769 packets a frame [hos host] vk: 273 mirrors given back (1973 MB in all); 1112 MB of mirrors now, 140 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: 5 textures retired (1223 left, 2585 MB): budget 0, drawn over 5, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9736 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 12012421 kept (4683370 by the last picks' textures, a uniform buffer bound since), 26321 new, 4761257 with other textures (of 16800000) [hos sead::MainThread] nvn: CommandBufferWaitEvent #230000: event 0x100d6c802c mode 0 value 164382 (now 164363) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: 7 textures retired (1224 left, 2587 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8320000 (5024694 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 6 textures retired (1225 left, 2589 MB): budget 0, drawn over 6, dead targets 0 [hos sead::MainThread] nvn: 1750000 recordings made, 75461 live copies (25.3 MB), 6098 texture copies; 2 command buffers' segments (4.1 MB) [hos ModuleSystemWorker2] vcores: 155732 takes, 30602 waits (mean 69 us, longest 20.0 ms, 0 over HOS_VCORE_LOG_MS), 15200 kicks, 15142 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 8 textures retired (1225 left, 2590 MB): budget 0, drawn over 8, dead targets 0 [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 418 dirty pages from their submits' captures, 55117 from earlier ones', 761 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 904 dispatches, 42503 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8422 runs, 701.8 MB compared; 69 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19538 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 761; [hos host] vk: a frame: 166 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.51, present 0.17, draw 6.71, descriptors 2.90, texture lookups 0.21, uploads 0.00, pipelines 0.29, programs 1.15, commands 1.84, dispatch 0.87, memory sync 0.47, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 395 (1088 MB): 5.53 MB synced in 1069.2 copies, 17.98 MB written back [hos host] vk: 13920 presents, 60.0 fps; a frame: 2129 draws (0 skipped), 168 dispatches, 178 passes, 131 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1226 textures (2593 MB) [hos host] vk: host time in GPU resource waits: 1.70 ms a frame (fences 1.45, timelines 0.25, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 699.5 full barriers recorded, 295.2 left out (nothing recorded since the one before), 13.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1111.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.6 ms a frame (the longest submission 3.4 ms); 0.7 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.7 (2.2 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 8340000 (5034316 pages from the captures) [hos host] vk: 9 textures retired (1229 left, 2595 MB): budget 0, drawn over 9, dead targets 0 [hos sead::MainThread] nvn: CommandBufferSignalEvent #165000: event 0x100d6c800c mode 0 value 165001 (lr 0x80660adc) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1127 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1549 kept their virtual core (nothing there to wait for), 3059 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 10 textures retired (1225 left, 2577 MB): budget 0, drawn over 10, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8360000 (5045015 pages from the captures) [hos host] vk: 9 textures retired (1221 left, 2560 MB): budget 0, drawn over 9, dead targets 0 [hos host] vk: sparse mirrors: 117, 1076 MB with memory of 1484 MB (407 MB spared); blocks given memory later 11094, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 923 binds (246.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 232 taken from the spares, 163 made new; 278 kept as spares, 0 destroyed; 46 spares now, 1216 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: 7 textures retired (1222 left, 2564 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 216 dirty pages from their submits' captures, 55003 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 870 dispatches, 42107 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8455 runs, 717.3 MB compared; 60 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19471 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: replay reads of guest memory as submitted: 8380000 (5055799 pages from the captures) [hos host] vk: 4 textures retired (1220 left, 2555 MB): budget 0, drawn over 4, dead targets 0 [hos sead::MainThread] nvn: GPU waits in 5 s: 103 sync waits (310 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 592 ranges, 3563 MB), 268 pages written, 268 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (128 MB), 11 spare buffers, 3778 texture pages copied (14 MB) [hos host] vk: 6 textures retired (1222 left, 2560 MB): budget 0, drawn over 6, dead targets 0 [hos host] vk: program picks: 12165641 kept (4759416 by the last picks' textures, a uniform buffer bound since), 26321 new, 4808037 with other textures (of 17000000) [hos host] vk: replay reads of guest memory as submitted: 8400000 (5066247 pages from the captures) [hos host] vk: 30 textures retired (1225 left, 2582 MB): budget 22, drawn over 8, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 945 dirty pages from their submits' captures, 54051 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 870 dispatches, 41941 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7910 runs, 694.1 MB compared; 58 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19289 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 14 textures retired (1223 left, 2578 MB): budget 8, drawn over 6, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8420000 (5079184 pages from the captures) [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 29086233 lookups, 139017 missed (0.48 %) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6955 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 5 textures retired (1230 left, 2596 MB): budget 0, drawn over 5, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 11 textures retired (1227 left, 2587 MB): budget 0, drawn over 11, dead targets 0 [hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 380000 so far (x1 0x117daa3070 x2 0 x3 0x117daa3040 x4 0x9b040) [hos host] vk: replay reads of guest memory as submitted: 8440000 (5091500 pages from the captures) [hos host] vk: 7 textures retired (1232 left, 2578 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: reaper: 791 images (1605 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 246 (279 MB) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 117, 1078 MB with memory of 1484 MB (405 MB spared); blocks given memory later 11097, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 927 binds (248.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 233 taken from the spares, 163 made new; 279 kept as spares, 0 destroyed; 46 spares now, 1216 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 593 dirty pages from their submits' captures, 55819 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 42960 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7776 runs, 707.6 MB compared; 66 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19823 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: 7 textures retired (1231 left, 2571 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: a frame: 187 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.76, present 0.17, draw 6.81, descriptors 2.89, texture lookups 0.24, uploads 0.03, pipelines 0.29, programs 1.21, commands 1.87, dispatch 0.84, memory sync 0.52, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 396 (1078 MB): 5.53 MB synced in 1073.4 copies, 17.81 MB written back [hos host] vk: 14040 presents, 59.9 fps; a frame: 2147 draws (0 skipped), 157 dispatches, 179 passes, 131 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1232 textures (2575 MB) [hos host] vk: host time in GPU resource waits: 2.05 ms a frame (fences 1.28, timelines 0.76, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 693.9 full barriers recorded, 296.3 left out (nothing recorded since the one before), 13.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1200.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.4 ms a frame (the longest submission 4.3 ms); 1.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.7 (2.0 MB), ASTC or BC 0.3 (0.5 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms [hos host] vk: replay reads of guest memory as submitted: 8460000 (5104000 pages from the captures) [hos host] vk: 9 textures retired (1235 left, 2577 MB): budget 0, drawn over 9, dead targets 0 [hos host] vk: texture #19093 0x10008612200 799x799 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14051 (drawn at 13967) [hos host] vk: texture #19102 0x10008612200 794x794 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14052 (drawn at 13966) [hos host] vk: texture #19152 0x10008612200 900x900 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14056 (drawn at 13994) [hos host] vk: 56 textures retired (1191 left, 2538 MB): budget 53, drawn over 3, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8480000 (5114089 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 595 ranges, 3565 MB), 264 pages written, 264 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 25 kept (110 MB), 12 spare buffers, 6911 texture pages copied (26 MB) [hos host] vk: texture #18190 0x100050a2000 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #18230 0x10001fe0800 640x360x1 x1 NVN 0x14 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #18188 0x10004b01a00 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #18183 0x10005552400 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #18191 0x10004ec1c00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #18189 0x10004fb1e00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 582 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 51 textures retired (1152 left, 2526 MB): budget 41, drawn over 10, dead targets 0; GPU-drawn with no copy kept: 0 live, 6 finalized, 21 MB [hos host] vk: program picks: 12319093 kept (4828623 by the last picks' textures, a uniform buffer bound since), 26321 new, 4854585 with other textures (of 17200000) [hos host] vk: readback pages refreshed from guest memory: 4406327 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1263 dirty pages from their submits' captures, 53227 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 854 dispatches, 41673 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7580 runs, 689.8 MB compared; 59 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19525 in captures, 2 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 87 textures retired (1098 left, 2448 MB): budget 79, drawn over 8, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8500000 (5127226 pages from the captures) [hos host] vk: 6 textures retired (1107 left, 2513 MB): budget 0, drawn over 6, dead targets 0 [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1120 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1382 kept their virtual core (nothing there to wait for), 2970 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #19151 0x10008612200 904x904 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14095 (drawn at 14033) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #235000: event 0x100d6c802c mode 0 value 168182 (now 168163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06 [hos host] vk: 32 textures retired (1092 left, 2505 MB): budget 23, drawn over 9, dead targets 0 [hos host] vk: 33 textures retired (1083 left, 2503 MB): budget 23, drawn over 10, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8520000 (5139525 pages from the captures) [hos Presentation Thread] vi: 14112 frames presented (59.7 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 30, 15.5-17.8 241, 17.8-22 25, 22-30 3, 30-40 0, over 40 0 [hos host] vk: sparse mirrors: 121, 1087 MB with memory of 1500 MB (412 MB spared); blocks given memory later 11098, given back 13987; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 932 binds (248.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 237 taken from the spares, 163 made new; 279 kept as spares, 0 destroyed; 42 spares now, 1200 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2125 dirty pages from their submits' captures, 54911 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 934 dispatches, 43397 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7575 runs, 704.9 MB compared; 203 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19513 in captures, 127 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: 26 textures retired (1034 left, 2175 MB): budget 21, drawn over 5, dead targets 0 [hos host] vk: VRAM compaction 6 begins: the textures' blocks hold 1664 MB, 549 MB of it free space [hos host] vk: replay reads of guest memory as submitted: 8540000 (5150345 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 29470 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 3 textures retired (1059 left, 2258 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #19310 0x10008612200 1421x1421 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14154 (drawn at 14093) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 595 ranges, 3609 MB), 290 pages written, 290 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.039; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.029, bytes copied 0.000; 25 kept (206 MB), 9 spare buffers, 9831 texture pages copied (38 MB) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 468 dirty pages from their submits' captures, 56934 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 920 dispatches, 44107 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8136 runs, 741.4 MB compared; 64 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20718 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fa9000000 (4 MB, captured, state 2): 2; [hos host] vk: a frame: 194 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.88, present 0.32, draw 6.91, descriptors 3.07, texture lookups 0.64, uploads 0.20, pipelines 0.29, programs 1.17, commands 1.86, dispatch 0.83, memory sync 0.56, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 401 (1089 MB): 5.55 MB synced in 1065.8 copies, 17.66 MB written back [hos host] vk: 14160 presents, 59.8 fps; a frame: 1960 draws (0 skipped), 147 dispatches, 177 passes, 131 barriers, 6 submits, 0.9 uploads (1.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1097 textures (2360 MB) [hos host] vk: host time in GPU resource waits: 3.46 ms a frame (fences 1.40, timelines 2.06, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 668.2 full barriers recorded, 292.9 left out (nothing recorded since the one before), 14.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.8 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1060.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.3 ms a frame (the longest submission 4.2 ms); 1.9 textures made a frame: colour targets 0.3 (1.2 MB), depth 1.0 (5.7 MB), ASTC or BC 0.7 (1.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.99 MB unswizzled on the CPU in 0.16 ms [hos host] vk: replay reads of guest memory as submitted: 8560000 (5161528 pages from the captures) [hos host] vk: program picks: 12468951 kept (4891719 by the last picks' textures, a uniform buffer bound since), 26321 new, 4904727 with other textures (of 17400000) [hos host] vk: VRAM compaction 6 over: 300 textures moved (270 MB) in 27 passes, 0 left where they were; 15 blocks (480 MB) given back to the driver, 3808 MB in all so far [hos host] vk: replay reads of guest memory as submitted: 8580000 (5172744 pages from the captures) [hos sead::MainThread] nvn: CommandBufferSignalEvent #170000: event 0x100d6c8014 mode 0 value 170000 (lr 0x802915c8) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: sparse mirrors: 121, 1092 MB with memory of 1512 MB (420 MB spared); blocks given memory later 11102, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 937 binds (249.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 239 taken from the spares, 163 made new; 281 kept as spares, 0 destroyed; 42 spares now, 1188 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 647 dirty pages from their submits' captures, 58914 from earlier ones', 129 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 975 dispatches, 45765 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8293 runs, 741.9 MB compared; 64 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21013 in captures, 2 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7fa3800000 (8 MB, captured, state 2): 129; [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: 1 textures retired (1148 left, 2524 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8600000 (5183875 pages from the captures) [hos BaseProcExecutor1] vcores: 154795 takes, 24921 waits (mean 69 us, longest 22.4 ms, 0 over HOS_VCORE_LOG_MS), 11644 kicks, 11676 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 3 textures retired (1158 left, 2525 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: pool 0x7f96c00000 (8 MB): mirror made at its first binding (401 so) [hos host] vk: 8 textures retired (1195 left, 2545 MB): budget 0, drawn over 8, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8620000 (5196577 pages from the captures) [hos host] vk: 9 textures retired (1202 left, 2515 MB): budget 0, drawn over 9, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1100 dirty pages from their submits' captures, 56327 from earlier ones', 1771 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 953 dispatches, 43867 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7933 runs, 736.4 MB compared; 85 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19954 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7f9c400000 (32 MB, captured, state 2): 1756; 0x7f9ec00000 (4 MB, captured, state 2): 14; 0x7fd3c00000 (16 MB, captured, state 2): 1; [hos host] vk: 10 textures retired (1199 left, 2491 MB): budget 0, drawn over 10, dead targets 0 [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1072 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 25 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1470 kept their virtual core (nothing there to wait for), 3138 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 605 ranges, 3601 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 7656 texture pages copied (29 MB) [hos host] vk: 1 textures retired (1120 left, 2216 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 23158 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 8640000 (5208228 pages from the captures) [hos host] vk: 2 textures retired (1119 left, 2211 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] nvn: GPU waits in 5 s: 146 sync waits (422 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: 5 textures retired (1115 left, 2183 MB): budget 0, drawn over 5, dead targets 0 [hos host] vk: a frame: 171 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.99, present 0.25, draw 6.33, descriptors 2.89, texture lookups 0.49, uploads 0.27, pipelines 0.25, programs 1.11, commands 1.62, dispatch 0.79, memory sync 0.53, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 407 (1127 MB): 5.77 MB synced in 1066.7 copies, 17.54 MB written back [hos host] vk: 14280 presents, 60.1 fps; a frame: 1750 draws (0 skipped), 142 dispatches, 171 passes, 132 barriers, 6 submits, 0.8 uploads (1.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13569 pipelines (3513 ms), 0 programs (10 failed), 1115 textures (2183 MB) [hos host] vk: host time in GPU resource waits: 2.87 ms a frame (fences 1.41, timelines 1.47, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 636.0 full barriers recorded, 286.6 left out (nothing recorded since the one before), 15.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.5, an op 1011.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.9 ms a frame (the longest submission 4.6 ms); 1.4 textures made a frame: colour targets 0.1 (0.9 MB), depth 0.6 (1.7 MB), ASTC or BC 0.7 (1.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.03 MB unswizzled on the CPU in 0.15 ms [hos host] vk: readback pages refreshed from guest memory: 4473439 pages so far (0 held: may be written ahead in their submission) [hos host] vk: replay reads of guest memory as submitted: 8660000 (5219767 pages from the captures) [hos host] vk: sparse mirrors: 126, 1127 MB with memory of 1584 MB (456 MB spared); blocks given memory later 11218, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 946 binds (251.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 244 taken from the spares, 163 made new; 281 kept as spares, 0 destroyed; 37 spares now, 1116 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 618 dirty pages from their submits' captures, 59978 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 975 dispatches, 46357 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7806 runs, 748.7 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20867 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 2; [hos host] vk: imports: 188 live (2260 MB), 3 retired still imported (140 MB), 40 retired given back (884 MB) [hos host] vk: program picks: 12609757 kept (4946761 by the last picks' textures, a uniform buffer bound since), 26324 new, 4963918 with other textures (of 17600000) [hos host] vk: 3 textures retired (1112 left, 2171 MB): budget 0, drawn over 3, dead targets 0 [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #240000: event 0x100d6c802c mode 0 value 171982 (now 171963) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.07 [hos host] vk: replay reads of guest memory as submitted: 8680000 (5231520 pages from the captures) [hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 121 dirty pages from their submits' captures, 61295 from earlier ones', 78 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 964 dispatches, 46593 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7891 runs, 750.7 MB compared; 68 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21229 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 77; 0x7fd3c00000 (16 MB, captured, state 2): 1; [hos host] vk: replay reads of guest memory as submitted: 8700000 (5243354 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 612 ranges, 3614 MB), 272 pages written, 272 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 24 kept (102 MB), 12 spare buffers, 7701 texture pages copied (30 MB) [hos ModuleSystemWorker2] nvn: 1800000 recordings made, 77844 live copies (25.1 MB), 6219 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: pool 0x7fbf733900+0x20d6e0 imported in 7.4 ms as 0x7fbd800000..0x7fc1400000 (228 imports, 700 ms in all) [hos host] vk: pool 0x7fbd800000..0x7fc1400000 took in 2 pools [hos host] vk: 288 mirrors given back (2071 MB in all); 1176 MB of mirrors now, 140 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: replay reads of guest memory as submitted: 8720000 (5255523 pages from the captures) [hos host] vk: presenter thread: the copy's fence 1.37 ms, the present 0.73 ms a present [hos host] vk: sparse mirrors: 126, 1180 MB with memory of 1640 MB (459 MB spared); blocks given memory later 11222, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 960 binds (254.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 250 taken from the spares, 164 made new; 288 kept as spares, 0 destroyed; 38 spares now, 1120 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1126 dirty pages from their submits' captures, 56337 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 904 dispatches, 43766 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7133 runs, 682.2 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19659 in captures, 1 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 187 live (2260 MB), 4 retired still imported (132 MB), 41 retired given back (952 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 7445 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 8740000 (5267640 pages from the captures) [hos host] vk: program picks: 12756202 kept (5010234 by the last picks' textures, a uniform buffer bound since), 26325 new, 5017472 with other textures (of 17800000) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: memory: working set 15190 MB (peak 15952), private 21222 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 864 MB, 37 allocations 778 MB; used 3276 of 15579 MB; heap 1 (device): 160 blocks 4855 MB, 20726 allocations 4225 MB; used 6850 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 21 blocks 672 MB, 105 images 259 MB; 69 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 47 blocks 1504 MB, 1099 images 1484 MB; 268 free ranges, the largest 11 MB [hos host] vk: render target images: 5883 reused from the pool, 3929 made new, 3822 destroyed from it, 1146 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 148 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1574 with the pool empty (0 of them with targets still in the dead queue), 331 with none in it a present yet, 35 with none of a kind they can use, 548 with all smaller (the largest 49% short on average), 564 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3630 idle [hos host] vk: VRAM: 6850 MB used of 11474 MB allowed (629 MB of it free in the blocks); textures 2380 MB (budget 7158 MB), 1210 textures [hos host] vk: budget: VRAM 4855 of 11474 MB (6850 MB by the driver, 1995 MB unaccounted): textures 2380, mirrors 1172, blocks free 629, shadows 284, other buffers 213, other images 111, staging 64; RAM 3203 MB (the process's private memory 21222 MB): imports 2260, capture spares 646, retired imports 132, captures 116, capture table 33, other buffers 16, others 0 [hos host] vk: textures: decoded ASTC 1427 MB, rendered 514 MB (69 drawn in one frame, kept: 42597 KB), other 405 MB; used in the last 8 presents 1885 MB [hos host] vk: decoded ASTC used again: 359 textures (814 MB) copied from ones the game finalized, not decoded; 359 of those retired, 3 listed; 1684 MB of blocks hashed in 294 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 9253 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 52 (11349 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 277 waited for it, 2.77 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 37.0 ms; VRAM 6.7 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6174 queued, 6174 loaded (median 1.40 ms, 99th percentile 16.63 ms, worst 38.13 ms, 15269 ms and 331.5 MB read in all); 8369 programs (1621 of them compute variants without their code, 0 graphics programs no note names), 11716 graphics pipelines (1 not in a cache), 715 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2247 pipelines made at create taken by draws; 2834 create calls with nothing kept for their stages, 0 before the index; units 9008 (6174 ready, 0 loading), 16392 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13196 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 710 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 190 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.44, present 0.14, draw 8.18, descriptors 3.57, texture lookups 0.40, uploads 0.35, pipelines 0.30, programs 1.21, commands 2.56, dispatch 0.79, memory sync 0.47, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 414 (1172 MB): 6.55 MB synced in 1088.6 copies, 17.50 MB written back [hos host] vk: 14400 presents, 57.9 fps; a frame: 2027 draws (0 skipped), 138 dispatches, 182 passes, 131 barriers, 6 submits, 1.1 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13571 pipelines (3514 ms), 0 programs (10 failed), 1210 textures (2380 MB) [hos host] vk: host time in GPU resource waits: 4.30 ms a frame (fences 1.66, timelines 2.64, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 670.1 full barriers recorded, 294.8 left out (nothing recorded since the one before), 19.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 121.9, an op 1210.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.2 ms a frame (the longest submission 5.3 ms); 1.0 textures made a frame: colour targets 0.1 (0.6 MB), depth 0.1 (0.3 MB), ASTC or BC 0.8 (1.7 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.25 MB unswizzled on the CPU in 0.19 ms [hos host] vk: queues ordered by their events: 89326 submissions stopped at a wait, 89280 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 242510, not met 0; captures at a CPU event's wait 8747, 0 of them at a wait met on arrival [hos host] vk: the render thread (600 presents, 16.80 ms a frame): its jobs 15.56 ms a frame (submits 6.15, presents 4.64, others 4.77), idle 1.23; the game waited for room at 17 presents, 0.77 ms each (0.02 a frame; 7.03 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 280 captures at a CPU signal; the submits' waits read in 0.042 ms a frame; readbacks: 1536 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.83 ms after its hand-over (34.54 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.26 + 0.29 (the window's copy is in the idle), after it 1.42 + 0.56; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.38 parts a frame recorded in the signal's job, 76.5% of them took nothing that came with the signal; of the others, 1.96 a frame took pages of the signal's capture into a mirror (37.9 pages a frame) and 0.01 pages read as guest memory had them (9.0 pages a frame); a frame's first part took nothing in 18 of 280 frames, its first three in 18 [hos host] vk: frame ahead, what came with the signal: 280 signals with a capture, each of 65.5 pages in 10.0 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.05 releases and 0.00 other CPU signals a signal (1 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.26+0.06, 0.42+0.14, 0.11+0.04, 0.04+0.02, 0.19+0.17, 0.29+0.09, 0.06+0.02, 0.05+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: replay reads of guest memory as submitted: 8760000 (5279301 pages from the captures) [hos host] vk: write-backs left out (600 submits): 926, 4109.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1428 dirty pages from their submits' captures, 51502 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 805 dispatches, 40522 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7027 runs, 666.9 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17852 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: readbacks of GPU results: 1069 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 22 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1244 kept their virtual core (nothing there to wait for), 2596 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: render targets drawn over from another start, let go: 10396 so far (49226 MB), 9336 of them (44360 MB) between the passes [hos Presentation Thread] vi: 14410 frames presented (59.4 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 62, 15.5-17.8 198, 17.8-22 27, 22-30 9, 30-40 1, over 40 1 [hos host] vk: replay reads of guest memory as submitted: 8780000 (5291840 pages from the captures) [hos host] vk: pool 0x1017549a00+0xf00000 imported in 3.7 ms as 0x1017400000..0x101a000000 (229 imports, 704 ms in all) [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 400000 so far (x1 0x117a4f2040 x2 0x6001d40 x3 0x117a4f2000 x4 0xf33080) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 611 ranges, 3618 MB), 250 pages written, 250 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 24 kept (78 MB), 12 spare buffers, 8400 texture pages copied (32 MB) [hos host] vk: replay reads of guest memory as submitted: 8800000 (5304316 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 123, 1174 MB with memory of 1624 MB (449 MB spared); blocks given memory later 11236, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 966 binds (255.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 251 taken from the spares, 164 made new; 292 kept as spares, 0 destroyed; 41 spares now, 1136 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1280 dirty pages from their submits' captures, 54409 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 850 dispatches, 42221 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7054 runs, 671.7 MB compared; 56 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19149 in captures, 0 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 187 live (2276 MB), 5 retired still imported (160 MB), 41 retired given back (952 MB) [hos host] vk: texture storage read as submitted (2001 times; 111350 pages from the copies, 0 of them rewritten in guest memory since their submit; 3787 pages kept, 271 spans, 957 MB) [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #175000: event 0x100d6c8020 mode 0 value 175000 (lr 0x80664f14) [hos host] vk: replay reads of guest memory as submitted: 8820000 (5316626 pages from the captures) [hos host] vk: 87 textures retired (1242 left, 2440 MB): budget 87, drawn over 0, dead targets 0 [hos host] vk: program picks: 12906179 kept (5073449 by the last picks' textures, a uniform buffer bound since), 26332 new, 5067488 with other textures (of 18000000) [hos host] vk: recorder: 6355175 packets run in 10.0 s (busy 1.76 s, spun idle 0.12 s, slept 11223 times, 3543 of them under 200 us), 0 calls made directly (no room); 18316 batches over 593 frames (30.9 a frame, 347 packets a batch), 8835 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 191 times, 19.08 ms (0.3 times, 0.03 ms a frame); every thread 913 times, 62.23 ms in all, the longest 1.33 ms; vkWaitForFences 519/26.28 ms (max 0.45), vkDestroyImageView 193/15.25 ms (max 0.44), vmaDestroyBuffer 96/10.39 ms (max 1.33), vmaFreeMemory 18/4.22 ms (max 0.82), vkQueueBindSparse 50/3.41 ms (max 0.38), the queue's lock 26/2.00 ms (max 0.19) [hos host] vk: recorder: of its busy 1.76 s, 0.44 s in parts' ends and submits (the longest 1.34 ms), 0.13 s of that blocked on the queue's lock (255 times, the longest 1.30 ms), so 0.31 s in the ends, submits and holds; 1.32 s running the other packets; 10717 packets a frame [hos host] vk: replay reads of guest memory as submitted: 8840000 (5330624 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 4540955 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3423 dirty pages from their submits' captures, 63001 from earlier ones', 5 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 906 dispatches, 44031 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7348 runs, 701.3 MB compared; 74 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18879 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7f86000000 (4 MB, captured, state 2): 5; [hos host] vk: queues ordered by their events: 90000 submissions stopped at a wait, 89953 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 244685, not met 0; captures at a CPU event's wait 8811, 0 of them at a wait met on arrival [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #245000: event 0x100d6c802c mode 0 value 175782 (now 175763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: replay reads of guest memory as submitted: 8860000 (5343224 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11526 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos BaseProcExecutor1] vcores: 152064 takes, 25000 waits (mean 68 us, longest 7.4 ms, 0 over HOS_VCORE_LOG_MS), 11876 kicks, 11879 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: texture #19745 0x10004bf1c00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454) [hos host] vk: texture #19746 0x10004b01a00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454) [hos host] vk: texture #19729 0x1000589ac00 640x360 x1 NVN 0x11 the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454) [hos host] vk: texture #19744 0x10004ce1e00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454) [hos host] vk: texture #19743 0x10009cf3200 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 14515 (drawn at 14454) [hos host] vk: a frame: 209 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.90, present 0.16, draw 7.00, descriptors 3.10, texture lookups 0.40, uploads 0.18, pipelines 0.29, programs 1.21, commands 1.87, dispatch 0.80, memory sync 0.51, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 415 (1168 MB): 5.95 MB synced in 1092.7 copies, 17.59 MB written back [hos host] vk: 14520 presents, 58.8 fps; a frame: 2080 draws (0 skipped), 142 dispatches, 185 passes, 132 barriers, 6 submits, 1.4 uploads (0.8 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13571 pipelines (3514 ms), 0 programs (10 failed), 1273 textures (2487 MB) [hos host] vk: host time in GPU resource waits: 3.66 ms a frame (fences 1.34, timelines 2.32, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 695.9 full barriers recorded, 300.1 left out (nothing recorded since the one before), 17.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.2; 0x12: 68.9; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.6, an op 1211.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.2 ms a frame (the longest submission 3.5 ms); 2.3 textures made a frame: colour targets 0.7 (2.5 MB), depth 0.3 (1.8 MB), ASTC or BC 1.2 (1.7 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.38 MB unswizzled on the CPU in 0.06 ms [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 8880000 (5354701 pages from the captures) [hos host] vk: sparse mirrors: 121, 1168 MB with memory of 1612 MB (443 MB spared); blocks given memory later 11245, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 969 binds (255.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 251 taken from the spares, 164 made new; 294 kept as spares, 0 destroyed; 43 spares now, 1148 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 296 dirty pages from their submits' captures, 57809 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 914 dispatches, 44544 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7647 runs, 720.4 MB compared; 47 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19993 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 187 live (2276 MB), 5 retired still imported (160 MB), 41 retired given back (952 MB) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 615 ranges, 3642 MB), 293 pages written, 293 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.017, bytes copied 0.000; 25 kept (92 MB), 11 spare buffers, 5309 texture pages copied (20 MB) [hos host] vk: 2 textures retired (1278 left, 2494 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8900000 (5366261 pages from the captures) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1071 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 34 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1601 kept their virtual core (nothing there to wait for), 3263 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 598 dirty pages from their submits' captures, 56367 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 43457 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7540 runs, 703.9 MB compared; 63 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19363 in captures, 2 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 8920000 (5378568 pages from the captures) [hos host] vk: program picks: 13055698 kept (5140932 by the last picks' textures, a uniform buffer bound since), 26336 new, 5117965 with other textures (of 18200000) [hos sead::MainThread] nvn: GPU waits in 5 s: 155 sync waits (606 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: texture #19891 0x10008612200 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14571 (drawn at 14510) [hos host] vk: texture #19606 0x10009932a00 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 14571 (drawn at 14510) [hos host] vk: replay reads of guest memory as submitted: 8940000 (5389411 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos ModuleSystemWorker3] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 37 made): 16216970 lookups, 124086 missed (0.77 %) [hos host] vk: sparse mirrors: 132, 1194 MB with memory of 1668 MB (473 MB spared); blocks given memory later 11247, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 982 binds (257.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 262 taken from the spares, 164 made new; 294 kept as spares, 0 destroyed; 32 spares now, 1092 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1091 dirty pages from their submits' captures, 52761 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 849 dispatches, 41002 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8037 runs, 680.9 MB compared; 59 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18434 in captures, 1 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2280 MB), 5 retired still imported (160 MB), 41 retired given back (952 MB) [hos host] vk: replay reads of guest memory as submitted: 8960000 (5397798 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 611 ranges, 3661 MB), 254 pages written, 254 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (104 MB), 11 spare buffers, 6800 texture pages copied (26 MB) [hos host] vk: 36 textures retired (1371 left, 2669 MB): budget 0, drawn over 36, dead targets 0 [hos host] vk: reaper: 714 images (2416 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 86 (178 MB) [hos host] vk: 74 textures retired (1305 left, 2602 MB): budget 73, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 8980000 (5409110 pages from the captures) [hos host] vk: texture #19734 0x100089d2a00 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 14631 (drawn at 14454) [hos host] vk: texture #19739 0x10004fb1e00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #19741 0x10004ec1c00 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #19740 0x100050a2000 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #19765 0x10001fe0800 640x360x1 x1 NVN 0x14 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #19731 0x10005552400 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #19738 0x10004b01a00 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 186 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 146 textures retired (1203 left, 2418 MB): budget 127, drawn over 5, dead targets 14; GPU-drawn with no copy kept: 0 live, 6 finalized, 21 MB [hos host] vk: a frame: 159 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.72, present 0.18, draw 6.88, descriptors 3.03, texture lookups 0.49, uploads 0.19, pipelines 0.29, programs 1.16, commands 1.89, dispatch 0.84, memory sync 0.50, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 428 (1206 MB): 5.86 MB synced in 1079.8 copies, 17.81 MB written back [hos host] vk: 14640 presents, 59.1 fps; a frame: 2103 draws (0 skipped), 157 dispatches, 177 passes, 131 barriers, 6 submits, 1.3 uploads (0.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13571 pipelines (3514 ms), 0 programs (10 failed), 1203 textures (2417 MB) [hos host] vk: host time in GPU resource waits: 4.30 ms a frame (fences 1.32, timelines 2.98, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 690.1 full barriers recorded, 293.5 left out (nothing recorded since the one before), 12.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1067.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.6 ms a frame (the longest submission 3.8 ms); 2.1 textures made a frame: colour targets 0.1 (0.8 MB), depth 0.7 (2.3 MB), ASTC or BC 1.3 (2.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.59 MB unswizzled on the CPU in 0.11 ms [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1317 dirty pages from their submits' captures, 52333 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 857 dispatches, 40906 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8102 runs, 681.5 MB compared; 63 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18412 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 9000000 (5421917 pages from the captures) [hos host] vk: 21 textures retired (1185 left, 2368 MB): budget 19, drawn over 2, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 24059 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 13207653 kept (5218825 by the last picks' textures, a uniform buffer bound since), 26336 new, 5166010 with other textures (of 18400000) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 9020000 (5432813 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 4608064 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 134, 1206 MB with memory of 1684 MB (477 MB spared); blocks given memory later 11255, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 986 binds (257.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 264 taken from the spares, 164 made new; 294 kept as spares, 0 destroyed; 30 spares now, 1076 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1189 dirty pages from their submits' captures, 56191 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 915 dispatches, 43976 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8792 runs, 736.4 MB compared; 64 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19800 in captures, 4 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2280 MB), 5 retired still imported (160 MB), 41 retired given back (952 MB) [hos host] vk: replay reads of guest memory as submitted: 9040000 (5443107 pages from the captures) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #250000: event 0x100d6c802c mode 0 value 179582 (now 179563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.07 [hos host] vk: 3 textures retired (1194 left, 2368 MB): budget 3, drawn over 0, dead targets 0 [hos Presentation Thread] vi: 14706 frames presented (59.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 68, 15.5-17.8 178, 17.8-22 37, 22-30 13, 30-40 0, over 40 0 [hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.01, scan 0.04 over 613 ranges, 3726 MB), 291 pages written, 291 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.041; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.027, bytes copied 0.000; 24 kept (92 MB), 12 spare buffers, 8352 texture pages copied (32 MB) [hos host] vk: replay reads of guest memory as submitted: 9060000 (5454577 pages from the captures) [hos sead::MainThread] nvn: CommandBufferSignalEvent #180000: event 0x100d6c8028 mode 0 value 179999 (lr 0x80667410) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2425 dirty pages from their submits' captures, 59734 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 946 dispatches, 43820 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8821 runs, 737.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20077 in captures, 4 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1121 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 28 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1511 kept their virtual core (nothing there to wait for), 2841 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 9080000 (5464908 pages from the captures) [hos host] vk: pool 0x7f9ab25200+0x1c3b60 imported in 2.3 ms as 0x7f9a800000..0x7f9c000000 (231 imports, 707 ms in all) [hos ModuleSystemWorker4] nvn: 1850000 recordings made, 79245 live copies (25.0 MB), 6228 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: 1 textures retired (1200 left, 2365 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9100000 (5473232 pages from the captures) [hos host] vk: a frame: 170 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.27, present 0.17, draw 6.51, descriptors 2.71, texture lookups 0.13, uploads 0.03, pipelines 0.26, programs 1.15, commands 1.89, dispatch 0.92, memory sync 0.61, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 429 (1220 MB): 6.08 MB synced in 1063.7 copies, 17.99 MB written back [hos host] vk: 14760 presents, 60.7 fps; a frame: 1863 draws (0 skipped), 170 dispatches, 156 passes, 132 barriers, 6 submits, 0.2 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13573 pipelines (3515 ms), 0 programs (10 failed), 1202 textures (2365 MB) [hos host] vk: host time in GPU resource waits: 2.68 ms a frame (fences 1.45, timelines 1.23, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 613.0 full barriers recorded, 274.5 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.6, an op 1030.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.9 ms a frame (the longest submission 5.5 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17059 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 13357451 kept (5288733 by the last picks' textures, a uniform buffer bound since), 26336 new, 5216212 with other textures (of 18600000) [hos host] vk: sparse mirrors: 134, 1220 MB with memory of 1688 MB (467 MB spared); blocks given memory later 11429, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 995 binds (258.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 265 taken from the spares, 164 made new; 295 kept as spares, 0 destroyed; 30 spares now, 1072 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1074, 4771.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 325 dirty pages from their submits' captures, 68915 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 949 dispatches, 46384 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9290 runs, 774.5 MB compared; 64 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21027 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 188 live (2284 MB), 6 retired still imported (180 MB), 41 retired given back (952 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 9120000 (5483935 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 612 ranges, 3725 MB), 284 pages written, 284 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.015; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (94 MB), 12 spare buffers, 6868 texture pages copied (26 MB) [hos host] vk: replay reads of guest memory as submitted: 9140000 (5494062 pages from the captures) [hos AudioFrameworkThread] vcores: 154519 takes, 26002 waits (mean 72 us, longest 31.5 ms, 0 over HOS_VCORE_LOG_MS), 12383 kicks, 12325 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: write-backs left out (600 submits): 1051, 4664.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 57739 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 927 dispatches, 44502 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9057 runs, 756.1 MB compared; 59 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20359 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 3 textures retired (1204 left, 2371 MB): budget 3, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 420000 so far (x1 0x117db40010 x2 0 x3 0x117db40000 x4 0x44b040) [hos host] vk: replay reads of guest memory as submitted: 9160000 (5506589 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 9180000 (5516498 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 133, 1218 MB with memory of 1684 MB (465 MB spared); blocks given memory later 11429, given back 13989; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 996 binds (258.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 265 taken from the spares, 164 made new; 296 kept as spares, 0 destroyed; 31 spares now, 1076 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1308 dirty pages from their submits' captures, 69224 from earlier ones', 734 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 931 dispatches, 42830 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8461 runs, 701.3 MB compared; 139 pages read into the shadows; 284 MB of shadows (51 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19164 in captures, 94 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fa2000000 (8 MB, captured, state 2): 534; 0x7fa0800000 (8 MB, captured, state 2): 200; [hos host] vk: imports: 188 live (2284 MB), 6 retired still imported (180 MB), 41 retired given back (952 MB) [hos host] vk: 2 textures retired (1203 left, 2349 MB): budget 2, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: GPU waits in 5 s: 129 sync waits (420 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: program picks: 13506894 kept (5360236 by the last picks' textures, a uniform buffer bound since), 26336 new, 5266769 with other textures (of 18800000) [hos host] vk: replay reads of guest memory as submitted: 9200000 (5524998 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 40989 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 4 textures retired (1209 left, 2348 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: a frame: 157 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.42, present 0.22, draw 6.55, descriptors 2.80, texture lookups 0.12, uploads 0.03, pipelines 0.27, programs 1.17, commands 1.80, dispatch 1.02, memory sync 1.07, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 429 (1158 MB): 8.56 MB synced in 1073.4 copies, 17.99 MB written back [hos host] vk: 14880 presents, 53.7 fps; a frame: 1847 draws (0 skipped), 170 dispatches, 157 passes, 131 barriers, 6 submits, 0.2 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 13742 pipelines (3589 ms), 0 programs (10 failed), 1209 textures (2348 MB) [hos host] vk: host time in GPU resource waits: 2.26 ms a frame (fences 1.46, timelines 0.80, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 618.5 full barriers recorded, 275.3 left out (nothing recorded since the one before), 13.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.1, an op 1007.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.2 ms a frame (the longest submission 6.6 ms); 0.2 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.00 ms [hos host] vk: texture #16605 0x116dd8f000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 14881 (drawn at 10787) [hos host] vk: late pages found at the replay (600 takes): 57 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: readback pages refreshed from guest memory: 4675763 pages so far (0 held: may be written ahead in their submission) [hos ModuleSystemWorker5] vk: readbacks of GPU results: 1105 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1550 kept their virtual core (nothing there to wait for), 3058 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 12355 dirty pages from their submits' captures, 148738 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1043 dispatches, 50097 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8627 runs, 719.3 MB compared; 278 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18693 in captures, 141 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 9220000 (5536415 pages from the captures) [hos host] vk: 304 mirrors given back (2177 MB in all); 1149 MB of mirrors now, 140 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: (its written part: 3 of 3 KB readable, 0 bytes not zero; levels L0) [hos host] vk: texture 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6 (39444 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it [hos Presentation Thread] vk: capture: a submit 0.49 ms (lock 0.02, scan 0.08 over 582 ranges, 3590 MB), 922 pages written, 922 copied (3.6 MB) [hos host] vk: a capture taken (ms): all 0.098; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.071, bytes copied 0.000; 24 kept (188 MB), 12 spare buffers, 23211 texture pages copied (90 MB) [hos sead::MainThread] nvn: CommandBufferWaitEvent #255000: event 0x100d6c802c mode 0 value 183382 (now 183363) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.10, writes collected 0.14 [hos host] vk: (its written part: 2 of 2 KB readable, 2531 bytes not zero; levels L0) [hos host] vk: texture 0x10debdc600 1240x1240x1 x3 levels 1 NVN 0x4d (4563 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 2 KB of it [hos host] vk: 89 textures retired (1123 left, 2224 MB): budget 89, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9240000 (5549029 pages from the captures) [hos rmbl::VibrationThread] hid: 49152 vibration values, strongest 0.00 since the last count [hos host] vk: 38 textures retired (1099 left, 2206 MB): budget 38, drawn over 0, dead targets 0 [hos host] vk: sparse mirrors: 126, 1055 MB with memory of 1580 MB (524 MB spared); blocks given memory later 11507, given back 15533; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1034 binds (269.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 267 taken from the spares, 164 made new; 305 kept as spares, 0 destroyed; 38 spares now, 1180 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 15361 dirty pages from their submits' captures, 88940 from earlier ones', 6577 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1033 dispatches, 46454 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8142 runs, 711.7 MB compared; 114 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18963 in captures, 67 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: 0x7fd6400000 (48 MB, captured, state 2): 4081; 0x7fbc000000 (12 MB, captured, state 2): 2489; 0x7fa9000000 (4 MB, captured, state 2): 6; 0x7fff800000 (8 MB, captured, state 2): 1; [hos host] vk: imports: 188 live (2284 MB), 6 retired still imported (180 MB), 41 retired given back (952 MB) [hos host] vk: replay reads of guest memory as submitted: 9260000 (5561653 pages from the captures) [hos host] vk: 3 textures retired (1119 left, 2219 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: 39 textures retired (1117 left, 2213 MB): budget 39, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9280000 (5573557 pages from the captures) [hos host] vk: 143 textures retired (1100 left, 2297 MB): budget 143, drawn over 0, dead targets 0 [hos host] vk: presenter thread: the copy's fence 1.35 ms, the present 0.79 ms a present [hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1717 dirty pages from their submits' captures, 51266 from earlier ones', 55 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 944 dispatches, 40635 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7655 runs, 675.0 MB compared; 119 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17958 in captures, 8 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: 0x7f84000000 (4 MB, captured, state 2): 48; 0x7fa2000000 (8 MB, captured, state 2): 7; [hos host] vk: program picks: 13658400 kept (5433737 by the last picks' textures, a uniform buffer bound since), 26345 new, 5315254 with other textures (of 19000000) [hos host] vk: pool 0x7fb5c00000..0x7fb8800000 (44 MB) trimmed to its 2 live part(s), 16 MB: the rest outside every game pool for 1856 submissions (3 trims, 120 MB in all) [hos host] vk: pool 0x7fb6000000..0x7fb6400000: 344 KB the CPU wrote since retired pool 0x7fb5c00000's last sync kept from guest memory, not its mirror [hos host] vk: pool 0x7fb7c00000..0x7fb8800000: 536 KB the CPU wrote since retired pool 0x7fb5c00000's last sync kept from guest memory, not its mirror [hos host] vk: replay reads of guest memory as submitted: 9300000 (5584189 pages from the captures) [hos Presentation Thread] vi: 14978 frames presented (54.4 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 26, 15.5-17.8 164, 17.8-22 49, 22-30 22, 30-40 9, over 40 2 [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #185000: event 0x100d6c8034 mode 0 value 184999 (lr 0x806616b0) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15346 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] vk: capture: a submit 0.37 ms (lock 0.07, scan 0.06 over 591 ranges, 3606 MB), 565 pages written, 565 copied (2.2 MB) [hos host] vk: a capture taken (ms): all 0.052; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.030, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 6896 texture pages copied (26 MB) [hos sead::MainThread] nvn: 1900000 recordings made, 72864 live copies (24.3 MB), 6231 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.05, writes collected 0.19 [hos host] vk: replay reads of guest memory as submitted: 9320000 (5596500 pages from the captures) [hos host] vk: memory: working set 15472 MB (peak 15952), private 21787 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 912 MB, 35 allocations 832 MB; used 3412 of 15579 MB; heap 1 (device): 163 blocks 4951 MB, 19498 allocations 4009 MB; used 7074 of 8421 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 19 blocks 608 MB, 115 images 302 MB; 69 free ranges, the largest 28 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 966 images 1294 MB; 381 free ranges, the largest 26 MB [hos host] vk: render target images: 5911 reused from the pool, 4185 made new, 4068 destroyed from it, 1159 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 150 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1792 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 576 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3876 idle [hos host] vk: VRAM: 7074 MB used of 8421 MB allowed (941 MB of it free in the blocks); textures 2232 MB (budget 2116 MB), 1087 textures [hos host] vk: budget: VRAM 4951 of 8421 MB (7074 MB by the driver, 2123 MB unaccounted): textures 2232, mirrors 1103, blocks free 941, shadows 284, other buffers 213, other images 111, staging 64; RAM 3346 MB (the process's private memory 21787 MB): imports 2256, capture spares 738, retired imports 224, captures 94, capture table 34, others 0 [hos host] vk: textures: decoded ASTC 1279 MB, rendered 548 MB (70 drawn in one frame, kept: 42773 KB), other 370 MB; used in the last 8 presents 1570 MB [hos host] vk: decoded ASTC used again: 402 textures (932 MB) copied from ones the game finalized, not decoded; 402 of those retired, 2 listed; 1823 MB of blocks hashed in 316 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 9853 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10e42b7400 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 56 (12053 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB); [hos sead::MainThread] nvn: EventSignal (CPU) #30000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos Presentation Thread] vk: presents after the CPU's signal (600): 337 waited for it, 3.62 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 51.6 ms; VRAM 6.9 of 8.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6457 queued, 6457 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16325 ms and 347.6 MB read in all); 8761 programs (1626 of them compute variants without their code, 0 graphics programs no note names), 12277 graphics pipelines (1 not in a cache), 716 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2268 pipelines made at create taken by draws; 3153 create calls with nothing kept for their stages, 0 before the index; units 9610 (6457 ready, 0 loading), 17770 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13948 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 711 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: 9452 programs from the program cache [hos host] vk: a frame: 172 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.36, present 0.20, draw 6.86, descriptors 3.22, texture lookups 0.71, uploads 0.54, pipelines 0.26, programs 1.12, commands 1.75, dispatch 1.00, memory sync 1.55, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 437 (1103 MB): 8.24 MB synced in 1078.5 copies, 17.78 MB written back [hos host] vk: 15000 presents, 53.4 fps; a frame: 1851 draws (0 skipped), 157 dispatches, 154 passes, 132 barriers, 6 submits, 1.6 uploads (1.7 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14144 pipelines (3721 ms), 0 programs (10 failed), 1087 textures (2232 MB) [hos host] vk: host time in GPU resource waits: 4.98 ms a frame (fences 1.60, timelines 3.38, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 606.8 full barriers recorded, 273.5 left out (nothing recorded since the one before), 14.3 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.7; 0x12: 68.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.0, an op 986.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.6 ms a frame (the longest submission 5.7 ms); 2.2 textures made a frame: colour targets 0.6 (1.5 MB), depth 0.2 (1.1 MB), ASTC or BC 1.5 (2.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.81 MB unswizzled on the CPU in 0.17 ms [hos host] vk: the render thread (600 presents, 17.55 ms a frame): its jobs 16.26 ms a frame (submits 5.51, presents 4.89, others 5.86), idle 1.29; the game waited for room at 34 presents, 0.83 ms each (0.05 a frame; 14.74 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 339 captures at a CPU signal; the submits' waits read in 0.040 ms a frame; readbacks: 1875 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.42 ms after its hand-over (38.22 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.96 + 0.30 (the window's copy is in the idle), after it 1.42 + 0.76; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.55 parts a frame recorded in the signal's job, 71.1% of them took nothing that came with the signal; of the others, 2.45 a frame took pages of the signal's capture into a mirror (126.2 pages a frame) and 0.02 pages read as guest memory had them (14.4 pages a frame); a frame's first part took nothing in 9 of 339 frames, its first three in 8 [hos host] vk: frame ahead, what came with the signal: 339 signals with a capture, each of 403.4 pages in 13.3 pools; between a frame's last submit and that signal 0.68 texture and sampler registrations, 0.88 releases and 0.00 other CPU signals a signal (65 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.21+0.12, 0.40+0.13, 0.08+0.03, 0.06+0.02, 0.18+0.05, 0.33+0.35, 0.08+0.04, 0.06+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: pipeline cache: play file 9 saved (36 KB): 1 pipelines, 1 from the driver's cache, 0 compiled; 1189 more from the start's cache [hos host] vk: sparse mirrors: 127, 1103 MB with memory of 1552 MB (448 MB spared); blocks given memory later 12262, given back 15533; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1080 binds (275.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 273 taken from the spares, 164 made new; 310 kept as spares, 0 destroyed; 37 spares now, 1208 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 671 dirty pages from their submits' captures, 56345 from earlier ones', 5 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 966 dispatches, 43306 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8033 runs, 709.3 MB compared; 65 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19569 in captures, 5 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fa5c00000 (4 MB, captured, state 2): 2; 0x7f86000000 (4 MB, captured, state 2): 2; 0x7f84000000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 189 live (2256 MB), 7 retired still imported (224 MB), 41 retired given back (952 MB) [hos Presentation Thread] vi: real-time step on: ratio 0.995, game's frame rate 60 [hos host] vk: render targets drawn over from another start, let go: 10647 so far (50241 MB), 9573 of them (45308 MB) between the passes [hos host] vk: texture #20270 0x116f87b000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 15010 (drawn at 14830) [hos host] vk: replay reads of guest memory as submitted: 9340000 (5608736 pages from the captures) [hos host] vk: 2 textures retired (1086 left, 2219 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: texture #20481 0x10004bf1c00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 15041 (drawn at 14980) [hos host] vk: replay reads of guest memory as submitted: 9360000 (5618773 pages from the captures) [hos sead::MainThread] vk: readbacks of GPU results: 1053 of 3000 waited for their write-back, 0.09 ms each; 0 looked at the stopped queues first, 37 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1662 kept their virtual core (nothing there to wait for), 3202 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: recorder: 5640656 packets run in 10.0 s (busy 1.63 s, spun idle 0.11 s, slept 10428 times, 3313 of them under 200 us), 0 calls made directly (no room); 17330 batches over 571 frames (30.4 a frame, 325 packets a batch), 8508 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 157 times, 9.69 ms (0.3 times, 0.02 ms a frame); every thread 782 times, 47.42 ms in all, the longest 1.04 ms; vkWaitForFences 490/27.34 ms (max 0.59), vmaDestroyBuffer 59/4.94 ms (max 1.04), vkDestroyImageView 65/4.90 ms (max 0.31), vkQueueBindSparse 114/4.80 ms (max 0.67), vmaFreeMemory 20/2.63 ms (max 0.42), vkEndCommandBuffer 3/0.90 ms (max 0.36) [hos host] vk: recorder: of its busy 1.63 s, 0.40 s in parts' ends and submits (the longest 1.45 ms), 0.08 s of that blocked on the queue's lock (154 times, the longest 1.38 ms), so 0.32 s in the ends, submits and holds; 1.23 s running the other packets; 9879 packets a frame [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 340 dirty pages from their submits' captures, 58162 from earlier ones', 249 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 933 dispatches, 44865 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8394 runs, 759.5 MB compared; 68 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20042 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 249; [hos host] vk: replay reads of guest memory as submitted: 9380000 (5629559 pages from the captures) [hos host] vk: program picks: 13811946 kept (5508826 by the last picks' textures, a uniform buffer bound since), 26347 new, 5361706 with other textures (of 19200000) [hos ModuleSystemWorker1] vcores: 158529 takes, 37801 waits (mean 84 us, longest 35.8 ms, 0 over HOS_VCORE_LOG_MS), 18511 kicks, 18361 yields, 0 starved grants, 0 escapes in 5 s [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 591 ranges, 3606 MB), 273 pages written, 273 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 6443 texture pages copied (25 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 18067 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: readback pages refreshed from guest memory: 4743016 pages so far (0 held: may be written ahead in their submission) [hos host] vk: replay reads of guest memory as submitted: 9400000 (5642192 pages from the captures) [hos host] vk: sparse mirrors: 120, 1042 MB with memory of 1472 MB (429 MB spared); blocks given memory later 12262, given back 15535; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1085 binds (276.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 273 taken from the spares, 164 made new; 317 kept as spares, 0 destroyed; 44 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 142 dirty pages from their submits' captures, 60322 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 951 dispatches, 45059 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8194 runs, 741.8 MB compared; 64 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20707 in captures, 3 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 189 live (2256 MB), 7 retired still imported (224 MB), 41 retired given back (952 MB) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #260000: event 0x100d6c802c mode 0 value 187182 (now 187163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #20338 0x116fbd1000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 15115 (drawn at 14946) [hos host] vk: replay reads of guest memory as submitted: 9420000 (5650967 pages from the captures) [hos host] vk: a frame: 152 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.04, present 0.17, draw 5.55, descriptors 2.34, texture lookups 0.08, uploads 0.01, pipelines 0.23, programs 1.02, commands 1.51, dispatch 0.82, memory sync 0.48, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 437 (1037 MB): 5.44 MB synced in 1053.8 copies, 17.67 MB written back [hos host] vk: 15120 presents, 60.1 fps; a frame: 1837 draws (0 skipped), 150 dispatches, 150 passes, 131 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14144 pipelines (3721 ms), 0 programs (10 failed), 1095 textures (2224 MB) [hos host] vk: host time in GPU resource waits: 1.53 ms a frame (fences 1.40, timelines 0.13, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 569.3 full barriers recorded, 268.8 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.0, an op 1020.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.1 ms a frame (the longest submission 4.3 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 103 dirty pages from their submits' captures, 58666 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 950 dispatches, 45239 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8383 runs, 759.4 MB compared; 57 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20701 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 9440000 (5661198 pages from the captures) [hos sead::MainThread] barrier: 238370 looks: 90914 found the masks moved already, 144864 waited until they moved, 2592 for the whole 1 ms (a wait 320 us on average); 229561 wakes [hos host] vk: replay reads of guest memory as submitted: 9460000 (5671851 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 94 sync waits (343 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 38 made): 28202372 lookups, 147255 missed (0.52 %) [hos host] vk: program picks: 13964327 kept (5583623 by the last picks' textures, a uniform buffer bound since), 26354 new, 5409318 with other textures (of 19400000) [hos host] vk: texture #18847 0x116f0e1000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 15177 (drawn at 13683) [hos host] vk: sparse mirrors: 121, 1039 MB with memory of 1496 MB (456 MB spared); blocks given memory later 12262, given back 15621; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1087 binds (276.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 274 taken from the spares, 164 made new; 317 kept as spares, 0 destroyed; 43 spares now, 1264 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1099, 4882.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 208 dirty pages from their submits' captures, 59409 from earlier ones', 599 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 982 dispatches, 45780 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8607 runs, 777.3 MB compared; 86 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20637 in captures, 1 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: 0x7fbc000000 (12 MB, captured, state 2): 563; 0x7fd6400000 (48 MB, captured, state 2): 36; [hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 593 ranges, 3596 MB), 288 pages written, 288 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.024, bytes copied 0.000; 25 kept (112 MB), 10 spare buffers, 5350 texture pages copied (20 MB) [hos host] vk: replay reads of guest memory as submitted: 9480000 (5684149 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 19335 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #20474 0x10004b01a00 2560x1440 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980) [hos host] vk: texture #20475 0x10004fb1e00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980) [hos host] vk: texture #20476 0x100050a2000 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980) [hos host] vk: texture #20477 0x10004ec1c00 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980) [hos host] vk: texture #20511 0x10001fe0800 640x360 x1 NVN 0x14 the GPU drew, finalized by the game, sampled again at frame 15207 (drawn at 14980) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1120 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1189 kept their virtual core (nothing there to wait for), 2907 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 440000 so far (x1 0x1178b9a000 x2 0 x3 0x1178b9a000 x4 0x654040) [hos host] vk: reaper: 711 images (1493 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 142 (223 MB) [hos host] vk: replay reads of guest memory as submitted: 9500000 (5696363 pages from the captures) [hos host] vk: 53 textures retired (1126 left, 2334 MB): budget 53, drawn over 0, dead targets 0 [hos host] vk: write-backs left out (600 submits): 1026, 4553.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 817 dirty pages from their submits' captures, 56731 from earlier ones', 15 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 971 dispatches, 43865 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7957 runs, 723.0 MB compared; 90 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19501 in captures, 2 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 10; 0x1042000000 (4 MB, captured, state 2): 4; 0x7fa2000000 (8 MB, captured, state 2): 1; [hos host] vk: 60 textures retired (1115 left, 2297 MB): budget 60, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9520000 (5708354 pages from the captures) [hos host] vk: 3 textures retired (1114 left, 2293 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: a frame: 188 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.73, present 0.15, draw 6.20, descriptors 2.84, texture lookups 0.51, uploads 0.38, pipelines 0.24, programs 1.06, commands 1.58, dispatch 0.81, memory sync 0.48, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 438 (1039 MB): 5.47 MB synced in 1060.8 copies, 17.68 MB written back [hos host] vk: 15240 presents, 58.3 fps; a frame: 1896 draws (0 skipped), 151 dispatches, 156 passes, 132 barriers, 6 submits, 1.1 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14144 pipelines (3721 ms), 0 programs (10 failed), 1114 textures (2293 MB) [hos host] vk: host time in GPU resource waits: 3.04 ms a frame (fences 1.52, timelines 1.53, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 601.1 full barriers recorded, 274.2 left out (nothing recorded since the one before), 14.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.6; 0x12: 68.9; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.7 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.1, an op 1043.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.5 ms a frame (the longest submission 4.6 ms); 1.4 textures made a frame: colour targets 0.3 (1.0 MB), depth 0.1 (0.6 MB), ASTC or BC 1.0 (1.7 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.06 MB unswizzled on the CPU in 0.01 ms [hos host] vk: replay reads of guest memory as submitted: 9540000 (5720276 pages from the captures) [hos sead::MainThread] nvn: CommandBufferSignalEvent #190000: event 0x100d6c8040 mode 0 value 189999 (lr 0x806621e8) [hos host] vk: texture #18104 0x116e859000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 15246 (drawn at 13394) [hos host] vk: slow upload 22.5 ms: #20779 0x10f72be800 1536x1536x1 x1 levels 11 NVN 0x7b ASTC, 2030 of 2030 KB [hos host] vk: 27 textures retired (1147 left, 2379 MB): budget 27, drawn over 0, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 126, 1099 MB with memory of 1576 MB (476 MB spared); blocks given memory later 12270, given back 15631; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1096 binds (278.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 279 taken from the spares, 164 made new; 317 kept as spares, 0 destroyed; 38 spares now, 1184 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1870 dirty pages from their submits' captures, 50352 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 869 dispatches, 39815 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7450 runs, 658.7 MB compared; 55 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17980 in captures, 2 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB) [hos host] vk: replay reads of guest memory as submitted: 9560000 (5733596 pages from the captures) [hos host] vk: 5 textures retired (1173 left, 2405 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: texture #19901 0x100021c0600 960x540 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 15267 (drawn at 15206) [hos host] vk: texture #20521 0x10008942600 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 15267 (drawn at 15206) [hos host] vk: texture #20534 0x10008942600 640x360 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 15267 (drawn at 15206) [hos host] vk: texture #20526 0x10004651000 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 15267 (drawn at 15206) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 591 ranges, 3642 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (116 MB), 11 spare buffers, 6358 texture pages copied (24 MB) [hos host] vk: program picks: 14117171 kept (5657953 by the last picks' textures, a uniform buffer bound since), 26357 new, 5456471 with other textures (of 19600000) [hos host] vk: replay reads of guest memory as submitted: 9580000 (5748010 pages from the captures) [hos host] vk: 1 textures retired (1172 left, 2404 MB): budget 0, drawn over 1, dead targets 0 [hos Presentation Thread] vi: 15270 frames presented (58.3 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 27, 15.5-17.8 229, 17.8-22 26, 22-30 7, 30-40 1, over 40 2 [hos host] vk: replay reads of guest memory as submitted: 9600000 (5761879 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 4810773 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1980 dirty pages from their submits' captures, 44027 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 763 dispatches, 34822 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6371 runs, 566.7 MB compared; 55 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15530 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 2 textures retired (1171 left, 2402 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: queues ordered by their events: 96000 submissions stopped at a wait, 95951 went on once it was reached, 46 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 4 (stopped 50.30 ms at most); waits met 264735, not met 0; captures at a CPU event's wait 9195, 0 of them at a wait met on arrival [hos host] vk: 320 mirrors given back (2262 MB in all); 1093 MB of mirrors now, 140 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: replay reads of guest memory as submitted: 9620000 (5776128 pages from the captures) [hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #265000: event 0x100d6c802c mode 0 value 190982 (now 190963) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05 [hos host] vk: 1 textures retired (1172 left, 2401 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9640000 (5789140 pages from the captures) [hos host] vk: late pages' fix (600 recordings with one): of 1048 captures with late pages 1048 fixed (81.2 pages, 81.7 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 451 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 576 blocks (36864 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 153, of an earlier recording 298, not recorded yet 0, of this part 0; pages fixed from stamps 84022, from snapshots 1075; the shadows: 820081 pages synced from captures, 799886 made known by it, 20195 left unknown because the replay or the collection wrote them since [hos host] vk: replay reads of guest memory as submitted: 9660000 (5803710 pages from the captures) [hos host] vk: sparse mirrors: 122, 1091 MB with memory of 1556 MB (464 MB spared); blocks given memory later 12282, given back 15631; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1101 binds (278.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 279 taken from the spares, 164 made new; 321 kept as spares, 0 destroyed; 42 spares now, 1204 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2247 dirty pages from their submits' captures, 45284 from earlier ones', 6 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 810 dispatches, 36026 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6580 runs, 584.8 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15995 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: 0x7fd3c00000 (16 MB, captured, state 2): 6; [hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB) [hos host] vk: 1 textures retired (1180 left, 2402 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9680000 (5817441 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21888 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 2 textures retired (1183 left, 2405 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9700000 (5830796 pages from the captures) [hos host] vk: program picks: 14272498 kept (5737562 by the last picks' textures, a uniform buffer bound since), 26357 new, 5501144 with other textures (of 19800000) [hos host] vk: 21 textures retired (1162 left, 2373 MB): budget 21, drawn over 0, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1841 dirty pages from their submits' captures, 43820 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 780 dispatches, 34883 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6430 runs, 568.1 MB compared; 77 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15598 in captures, 0 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 590 ranges, 3640 MB), 249 pages written, 249 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (96 MB), 11 spare buffers, 6958 texture pages copied (27 MB) [hos host] vk: replay reads of guest memory as submitted: 9720000 (5844432 pages from the captures) [hos host] vk: 2 textures retired (1160 left, 2370 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: a frame: 306 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.30, present 0.17, draw 7.87, descriptors 3.42, texture lookups 0.41, uploads 0.28, pipelines 0.31, programs 1.31, commands 2.20, dispatch 0.86, memory sync 0.51, queue submit 0.00, slot wait 0.29 [hos host] vk: a frame: mirrors 444 (1096 MB): 6.14 MB synced in 1087.6 copies, 17.74 MB written back [hos host] vk: 15360 presents, 56.4 fps; a frame: 2476 draws (0 skipped), 155 dispatches, 189 passes, 135 barriers, 6 submits, 0.9 uploads (0.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14144 pipelines (3721 ms), 0 programs (10 failed), 1160 textures (2370 MB) [hos host] vk: host time in GPU resource waits: 7.08 ms a frame (fences 2.35, timelines 4.73, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 721.5 full barriers recorded, 307.5 left out (nothing recorded since the one before), 17.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.7, an op 1613.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.2 ms a frame (the longest submission 3.7 ms); 0.9 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.8 (1.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.27 MB unswizzled on the CPU in 0.04 ms [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1101 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 11 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1343 kept their virtual core (nothing there to wait for), 3009 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos nn.npln.Worker] vcores: 143257 takes, 24474 waits (mean 71 us, longest 17.1 ms, 0 over HOS_VCORE_LOG_MS), 11744 kicks, 11708 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 9740000 (5858675 pages from the captures) [hos host] vk: 15 textures retired (1145 left, 2347 MB): budget 15, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9760000 (5871928 pages from the captures) [hos host] vk: sparse mirrors: 122, 1097 MB with memory of 1564 MB (466 MB spared); blocks given memory later 12321, given back 15631; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1108 binds (279.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 280 taken from the spares, 164 made new; 322 kept as spares, 0 destroyed; 42 spares now, 1196 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 8944 dirty pages from their submits' captures, 61324 from earlier ones', 18757 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 921 dispatches, 38108 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6592 runs, 585.2 MB compared; 90 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15902 in captures, 78 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: 0x7fbd800000 (60 MB, captured, state 2): 12488; 0x7fd6400000 (48 MB, captured, state 2): 3302; 0x7fab000000 (8 MB, captured, state 2): 1636; 0x7f9c400000 (32 MB, captured, state 2): 594; 0x7faf400000 (4 MB, captured, state 2): 295; 0x7f8e000000 (4 MB, captured, state 2): 278; 0x7fa9000000 (4 MB, captured, state 2): 98; 0x7fdfc00000 (12 MB, captured, state 2): 28; [hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB) [hos OptimizeChangeBanc] nvn: 1950000 recordings made, 72629 live copies (25.3 MB), 6235 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: 27 textures retired (1118 left, 2293 MB): budget 27, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9780000 (5886425 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.12 [hos host] vk: replay reads of guest memory as submitted: 9800000 (5899561 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 826, 3665.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 7661 dirty pages from their submits' captures, 62643 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 938 dispatches, 38010 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6423 runs, 583.2 MB compared; 89 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15938 in captures, 1 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: texture #13893 0x10005280e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 214 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 10 textures retired (1132 left, 2286 MB): budget 10, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB [hos host] vk: replay reads of guest memory as submitted: 9820000 (5913333 pages from the captures) [hos host] vk: program picks: 14428169 kept (5817159 by the last picks' textures, a uniform buffer bound since), 26357 new, 5545473 with other textures (of 20000000) [hos host] vk: replay reads of guest memory as submitted: 9840000 (5927465 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 219 sync waits (1031 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos Presentation Thread] vk: capture: a submit 0.43 ms (lock 0.04, scan 0.08 over 578 ranges, 3642 MB), 797 pages written, 681 copied (2.7 MB) [hos host] vk: a capture taken (ms): all 0.053; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.027, bytes copied 0.000; 24 kept (218 MB), 12 spare buffers, 11599 texture pages copied (45 MB) [hos host] vk: replay reads of guest memory as submitted: 9860000 (5940847 pages from the captures) [hos host] vk: 26 textures retired (1122 left, 2291 MB): budget 26, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 4877348 pages so far (0 held: may be written ahead in their submission) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16531 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 122, 1112 MB with memory of 1564 MB (451 MB spared); blocks given memory later 13381, given back 16453; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1141 binds (286.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 280 taken from the spares, 164 made new; 322 kept as spares, 0 destroyed; 42 spares now, 1196 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 13005 dirty pages from their submits' captures, 73557 from earlier ones', 2567 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 915 dispatches, 38592 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6248 runs, 565.5 MB compared; 87 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15949 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: 0x7fb7c00000 (12 MB, captured, state 2): 2567; [hos host] vk: imports: 189 live (2256 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB) [hos host] vk: replay reads of guest memory as submitted: 9880000 (5954870 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 6 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 1 textures retired (1129 left, 2292 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 9900000 (5968038 pages from the captures) [hos OptimizeChangeBanc] nvn: 2000000 recordings made, 77008 live copies (29.1 MB), 6235 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: a frame: 340 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 15.67, present 0.18, draw 8.81, descriptors 3.71, texture lookups 0.29, uploads 0.17, pipelines 0.35, programs 1.52, commands 2.53, dispatch 1.03, memory sync 1.75, queue submit 0.00, slot wait 0.30 [hos host] vk: a frame: mirrors 445 (1129 MB): 10.25 MB synced in 1115.1 copies, 17.68 MB written back [hos host] vk: 15480 presents, 47.4 fps; a frame: 2568 draws (0 skipped), 151 dispatches, 195 passes, 134 barriers, 6 submits, 0.6 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14146 pipelines (3722 ms), 0 programs (10 failed), 1141 textures (2302 MB) [hos host] vk: host time in GPU resource waits: 6.82 ms a frame (fences 1.93, timelines 4.89, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 722.9 full barriers recorded, 308.2 left out (nothing recorded since the one before), 22.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1644.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.6 ms a frame (the longest submission 5.9 ms); 0.5 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.5 (0.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 9920000 (5981625 pages from the captures) [hos ParallelDecompThread] zstd: 4 stream(s) decoding on the host at once [hos BaseProcExecutor0] vcores: BaseProcExecutor0 (priority 16) waited 70.6 ms for a core: given by sead::MainThread (priority 16), which held it 70.7 ms and let it go 68.0 ms after it was asked; the cores it may use now: core 0: it; [hos host] vk: late pages' fix: stamps for 0x1051bf0000 (pool 0x1051800000; 640 blocks, 40960 KB) [hos host] vk: late pages' fix: stamps for 0x11373f0000 (pool 0x1137000000; 704 blocks, 45056 KB) [hos host] vk: late pages' fix: stamps for 0x11377f0000 (pool 0x1137000000; 768 blocks, 49152 KB) [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 14881 dirty pages from their submits' captures, 90635 from earlier ones', 84 from guest memory as it was at the replay; 12 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1006 dispatches, 39888 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6353 runs, 581.7 MB compared; 92 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15461 in captures, 5 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: 0x7fdfc00000 (12 MB, captured, state 2): 84; [hos host] vk: texture #16205 0x10004bf1600 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 253 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 19 textures retired (1140 left, 2291 MB): budget 19, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: (its written part: 3 of 3 KB readable, 0 bytes not zero; levels L0) [hos host] vk: texture 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6 (39444 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 3 KB of it [hos host] vk: (its written part: 0 of 0 KB readable, 288 bytes not zero; levels L0) [hos host] vk: texture 0x10dff20200 2048x2048x1 x3 levels 1 NVN 0x4d (12288 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 0 KB of it [hos host] vk: replay reads of guest memory as submitted: 9940000 (5995005 pages from the captures) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #270000: event 0x100d6c802c mode 0 value 194782 (now 194763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.01, capture taken 0.10, writes collected 0.25 [hos host] vk: program picks: 14583155 kept (5895985 by the last picks' textures, a uniform buffer bound since), 26358 new, 5590486 with other textures (of 20200000) [hos host] vk: 1 textures retired (1145 left, 2287 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: slow upload 15.5 ms: #20943 0x10dae87a00 768x768x1 x1 levels 10 NVN 0x7b ASTC, 528 of 528 KB [hos host] vk: replay reads of guest memory as submitted: 9960000 (6008064 pages from the captures) [hos Presentation Thread] vi: 15509 frames presented (47.6 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 5, 15.5-17.8 74, 17.8-22 104, 22-30 42, 30-40 10, over 40 4 [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #195000: event 0x100d6c800c mode 0 value 195002 (lr 0x80660adc) [hos host] vk: replay reads of guest memory as submitted: 9980000 (6021927 pages from the captures) [hos host] vk: sparse mirrors: 122, 1107 MB with memory of 1536 MB (428 MB spared); blocks given memory later 13728, given back 16473; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1186 binds (289.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 282 taken from the spares, 164 made new; 324 kept as spares, 0 destroyed; 42 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 8793 dirty pages from their submits' captures, 80196 from earlier ones', 2 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 934 dispatches, 38828 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6178 runs, 564.3 MB compared; 387 pages read into the shadows; 284 MB of shadows (51 MB of pages known, in 102 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15149 in captures, 465 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: 0x7f98400000 (4 MB, captured, state 2): 2; [hos host] vk: imports: 190 live (2260 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 949 of 3000 waited for their write-back, 0.15 ms each; 0 looked at the stopped queues first, 8 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1108 kept their virtual core (nothing there to wait for), 2988 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 10000000 (6034958 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.61 ms (lock 0.06, scan 0.06 over 590 ranges, 3610 MB), 931 pages written, 931 copied (3.6 MB) [hos host] vk: a capture taken (ms): all 0.109; GPU wait for late pages 0.014 (1 of 600 takes), pools lock 0.000, retires and stamps 0.066, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 22828 texture pages copied (89 MB) [hos host] vk: replay reads of guest memory as submitted: 10020000 (6047343 pages from the captures) [hos host] vk: texture #20946 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (8 so far) [hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2158 dirty pages from their submits' captures, 45066 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 788 dispatches, 36174 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6356 runs, 581.8 MB compared; 75 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15464 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 10040000 (6060549 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 34538 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 10060000 (6073721 pages from the captures) [hos host] vk: presenter thread: the copy's fence 1.77 ms, the present 0.75 ms a present [hos host] vk: recorder: 6159760 packets run in 10.0 s (busy 1.65 s, spun idle 0.12 s, slept 11017 times, 3419 of them under 200 us), 0 calls made directly (no room); 16947 batches over 521 frames (32.5 a frame, 363 packets a batch), 8771 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 144 times, 10.07 ms (0.3 times, 0.02 ms a frame); every thread 806 times, 66.68 ms in all, the longest 1.07 ms; vkWaitForFences 467/37.13 ms (max 1.07), the queue's lock 84/11.12 ms (max 0.75), vmaFreeMemory 21/5.22 ms (max 0.89), vkDestroyImageView 40/3.61 ms (max 0.96), vkQueueBindSparse 108/3.01 ms (max 0.50), vmaDestroyBuffer 57/2.79 ms (max 0.73) [hos host] vk: recorder: of its busy 1.65 s, 0.44 s in parts' ends and submits (the longest 1.26 ms), 0.12 s of that blocked on the queue's lock (210 times, the longest 1.22 ms), so 0.32 s in the ends, submits and holds; 1.22 s running the other packets; 11823 packets a frame [hos host] vk: program picks: 14736938 kept (5972729 by the last picks' textures, a uniform buffer bound since), 26361 new, 5636700 with other textures (of 20400000) [hos host] vk: replay reads of guest memory as submitted: 10080000 (6085698 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 376 pages in 5 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 112, 1070 MB with memory of 1496 MB (425 MB spared); blocks given memory later 13728, given back 16889; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1196 binds (290.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 282 taken from the spares, 164 made new; 334 kept as spares, 0 destroyed; 52 spares now, 1264 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 826, 3665.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1785 dirty pages from their submits' captures, 45841 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 785 dispatches, 36556 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6322 runs, 581.9 MB compared; 79 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16297 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 190 live (2260 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB) [hos host] vk: replay reads of guest memory as submitted: 10100000 (6097447 pages from the captures) [hos host] vk: late pages' fix: stamps for 0x10383f0000 (pool 0x1038000000; 832 blocks, 53248 KB) [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 460000 so far (x1 0x117da08020 x2 0 x3 0x117da08000 x4 0x9b040) [hos BaseProcExecutor1] vcores: 147891 takes, 42524 waits (mean 92 us, longest 70.6 ms, 1 over HOS_VCORE_LOG_MS), 20751 kicks, 20682 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 10120000 (6109020 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.12 [hos host] vk: memory: working set 15681 MB (peak 16021), private 21990 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 12 blocks 928 MB, 36 allocations 850 MB; used 3360 of 15579 MB; heap 1 (device): 163 blocks 4951 MB, 19364 allocations 4078 MB; used 7074 of 8536 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 18 blocks 576 MB, 125 images 322 MB; 71 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 53 blocks 1696 MB, 1037 images 1356 MB; 377 free ranges, the largest 32 MB [hos host] vk: render target images: 5911 reused from the pool, 4234 made new, 4107 destroyed from it, 1160 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1836 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 581 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3915 idle [hos host] vk: VRAM: 7074 MB used of 8536 MB allowed (873 MB of it free in the blocks); textures 2314 MB (budget 2134 MB), 1168 textures [hos host] vk: budget: VRAM 4951 of 8536 MB (7074 MB by the driver, 2123 MB unaccounted): textures 2314, mirrors 1074, blocks free 873, shadows 284, other buffers 229, other images 111, staging 64; RAM 3297 MB (the process's private memory 21990 MB): imports 2260, capture spares 748, retired imports 152, captures 88, capture table 35, other buffers 14, others 0 [hos host] vk: textures: decoded ASTC 1409 MB, rendered 535 MB (82 drawn in one frame, kept: 47264 KB), other 337 MB; used in the last 8 presents 2098 MB [hos host] vk: decoded ASTC used again: 404 textures (933 MB) copied from ones the game finalized, not decoded; 404 of those retired, 0 listed; 1912 MB of blocks hashed in 328 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 10453 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10e42b7400 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 61 (12932 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 407 waited for it, 5.40 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 126.5 ms; VRAM 6.9 of 8.3 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6459 queued, 6459 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16328 ms and 347.7 MB read in all); 8763 programs (1626 of them compute variants without their code, 0 graphics programs no note names), 12279 graphics pipelines (1 not in a cache), 716 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2282 pipelines made at create taken by draws; 3153 create calls with nothing kept for their stages, 0 before the index; units 9612 (6459 ready, 0 loading), 17669 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13952 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 711 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 273 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 15.55, present 0.16, draw 9.17, descriptors 3.85, texture lookups 0.29, uploads 0.18, pipelines 0.36, programs 1.62, commands 2.62, dispatch 0.95, memory sync 1.41, queue submit 0.00, slot wait 0.32 [hos host] vk: a frame: mirrors 447 (1074 MB): 7.81 MB synced in 1106.9 copies, 17.64 MB written back [hos host] vk: 15600 presents, 45.2 fps; a frame: 2706 draws (0 skipped), 149 dispatches, 218 passes, 135 barriers, 6 submits, 0.5 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14146 pipelines (3722 ms), 0 programs (10 failed), 1168 textures (2314 MB) [hos host] vk: host time in GPU resource waits: 9.93 ms a frame (fences 2.61, timelines 7.32, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 722.7 full barriers recorded, 309.8 left out (nothing recorded since the one before), 44.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.5, an op 1803.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 13.7 ms a frame (the longest submission 14.8 ms); 0.4 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms [hos host] vk: the render thread (600 presents, 18.96 ms a frame): its jobs 17.32 ms a frame (submits 4.45, presents 4.67, others 8.20), idle 1.64; the game waited for room at 29 presents, 0.37 ms each (0.02 a frame; 15.24 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 408 captures at a CPU signal; the submits' waits read in 0.041 ms a frame; readbacks: 2148 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 16.40 ms after its hand-over (45.09 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.73 + 0.21 (the window's copy is in the idle), after it 2.32 + 0.89; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 10.96 parts a frame recorded in the signal's job, 68.3% of them took nothing that came with the signal; of the others, 3.46 a frame took pages of the signal's capture into a mirror (166.6 pages a frame) and 0.01 pages read as guest memory had them (7.0 pages a frame); a frame's first part took nothing in 11 of 407 frames, its first three in 11 [hos host] vk: frame ahead, what came with the signal: 408 signals with a capture, each of 565.7 pages in 14.4 pools; between a frame's last submit and that signal 2.02 texture and sampler registrations, 0.09 releases and 0.00 other CPU signals a signal (116 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.71+0.15, 0.37+0.10, 0.25+0.07, 0.30+0.06, 0.05+0.02, 0.10+0.29, 0.22+0.05, 0.17+0.03, 0.05+0.06, 0.07+0.05, 0.02+0.01, 0.01+0.00 [hos host] vk: replay reads of guest memory as submitted: 10140000 (6122830 pages from the captures) [hos host] vk: render targets drawn over from another start, let go: 10689 so far (50405 MB), 9614 of them (45466 MB) between the passes [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 597 ranges, 3583 MB), 260 pages written, 260 copied (1.0 MB) [hos host] vk: readback pages refreshed from guest memory: 4944795 pages so far (0 held: may be written ahead in their submission) [hos host] vk: a capture taken (ms): all 0.037; GPU wait for late pages 0.002 (1 of 600 takes), pools lock 0.000, retires and stamps 0.019, bytes copied 0.000; 24 kept (144 MB), 11 spare buffers, 4432 texture pages copied (17 MB) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2578 dirty pages from their submits' captures, 43019 from earlier ones', 166 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 774 dispatches, 34707 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6291 runs, 566.0 MB compared; 72 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15594 in captures, 0 only in the collection, 22 only by the replay itself [hos host] vk: read live, by pool: 0x1038000000 (4 MB, captured, state 2): 166; [hos host] vk: replay reads of guest memory as submitted: 10160000 (6135351 pages from the captures) [hos ModuleSystemWorker2] nvn: 2050000 recordings made, 73816 live copies (25.4 MB), 6237 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: replay reads of guest memory as submitted: 10180000 (6148284 pages from the captures) [hos host] vk: a submission's parts ended: 69145 where the game had ended its render pass, 10855 at 512 draws [hos host] vk: commit: 1770 MB left under Windows' limit as it is now, and the page files can grow by 79851 MB more: counted as room [hos host] vk: program picks: 14889496 kept (6049570 by the last picks' textures, a uniform buffer bound since), 26361 new, 5684142 with other textures (of 20600000) [hos host] vk: replay reads of guest memory as submitted: 10200000 (6160443 pages from the captures) [hos host] vk: sparse mirrors: 113, 1074 MB with memory of 1500 MB (425 MB spared); blocks given memory later 13729, given back 16891; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1200 binds (291.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 283 taken from the spares, 164 made new; 334 kept as spares, 0 destroyed; 51 spares now, 1260 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3730 dirty pages from their submits' captures, 47063 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 803 dispatches, 36624 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6620 runs, 584.5 MB compared; 79 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15567 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 190 live (2260 MB), 5 retired still imported (152 MB), 43 retired given back (1024 MB) [hos host] vk: replay reads of guest memory as submitted: 10220000 (6174128 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10240000 (6186790 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17791 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] vk: readbacks of GPU results: 1079 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1147 kept their virtual core (nothing there to wait for), 2949 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2253 dirty pages from their submits' captures, 43528 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 792 dispatches, 34917 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6271 runs, 549.1 MB compared; 75 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 105 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15621 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 10260000 (6197786 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 235 sync waits (1375 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 39 made): 22622743 lookups, 148417 missed (0.66 %) [hos host] vk: late pages found at the replay (600 takes): 4 pages in 4 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 10280000 (6208637 pages from the captures) [hos sead::MainThread] nvn: CommandBufferWaitEvent #275000: event 0x100d6c802c mode 0 value 198582 (now 198563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06 [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.04 over 611 ranges, 3597 MB), 297 pages written, 297 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.029; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 24 kept (136 MB), 12 spare buffers, 5526 texture pages copied (21 MB) [hos host] vk: pool 0x7fb1c8f900+0x773a90 imported in 6.4 ms as 0x7fb0800000..0x7fb4400000 (233 imports, 714 ms in all) [hos host] vk: replay reads of guest memory as submitted: 10300000 (6222169 pages from the captures) [hos host] vk: 337 mirrors given back (2335 MB in all); 1113 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: sparse mirrors: 114, 1113 MB with memory of 1596 MB (482 MB spared); blocks given memory later 13731, given back 16891; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1207 binds (291.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 286 taken from the spares, 165 made new; 337 kept as spares, 0 destroyed; 51 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 799, 3550.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2459 dirty pages from their submits' captures, 43471 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 803 dispatches, 34964 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6441 runs, 566.7 MB compared; 83 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15019 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 189 live (2260 MB), 7 retired still imported (212 MB), 43 retired given back (1024 MB) [hos host] vk: replay reads of guest memory as submitted: 10320000 (6235239 pages from the captures) [hos host] vk: program picks: 15042135 kept (6127554 by the last picks' textures, a uniform buffer bound since), 26377 new, 5731487 with other textures (of 20800000) [hos host] vk: a frame: 261 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 13.59, present 0.16, draw 8.64, descriptors 3.55, texture lookups 0.11, uploads 0.03, pipelines 0.34, programs 1.64, commands 2.43, dispatch 0.90, memory sync 0.64, queue submit 0.00, slot wait 0.33 [hos host] vk: a frame: mirrors 451 (1113 MB): 6.35 MB synced in 1103.6 copies, 17.72 MB written back [hos host] vk: 15720 presents, 52.3 fps; a frame: 2671 draws (0 skipped), 153 dispatches, 220 passes, 134 barriers, 6 submits, 0.3 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14153 pipelines (3725 ms), 0 programs (10 failed), 1201 textures (2358 MB) [hos host] vk: host time in GPU resource waits: 7.28 ms a frame (fences 1.57, timelines 5.70, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 732.9 full barriers recorded, 311.5 left out (nothing recorded since the one before), 45.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1764.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 13.2 ms a frame (the longest submission 3.8 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.15 MB unswizzled on the CPU in 0.02 ms [hos host] vk: replay reads of guest memory as submitted: 10340000 (6247330 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10360000 (6260685 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2220 dirty pages from their submits' captures, 44012 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 801 dispatches, 35044 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6421 runs, 566.6 MB compared; 78 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 103 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15939 in captures, 1 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 10380000 (6274114 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10400000 (6288002 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10420000 (6301288 pages from the captures) [hos Presentation Thread] vi: 15771 frames presented (52.4 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 61, 17.8-22 189, 22-30 11, 30-40 0, over 40 1 [hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #200000: event 0x100d6c8010 mode 0 value 200000 (lr 0x80660adc) [hos host] vk: readback pages refreshed from guest memory: 5011880 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 115, 1116 MB with memory of 1600 MB (483 MB spared); blocks given memory later 13731, given back 16900; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1212 binds (292.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 288 taken from the spares, 165 made new; 338 kept as spares, 0 destroyed; 50 spares now, 1220 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2649 dirty pages from their submits' captures, 44301 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 807 dispatches, 35536 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6441 runs, 567.1 MB compared; 74 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15262 in captures, 0 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 189 live (2260 MB), 7 retired still imported (212 MB), 43 retired given back (1024 MB) [hos host] vk: replay reads of guest memory as submitted: 10440000 (6314784 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 615 ranges, 3609 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.033; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (116 MB), 6 spare buffers, 6286 texture pages copied (24 MB) [hos host] vk: replay reads of guest memory as submitted: 10460000 (6326851 pages from the captures) [hos host] vk: program picks: 15192250 kept (6202680 by the last picks' textures, a uniform buffer bound since), 26383 new, 5781366 with other textures (of 21000000) [hos host] vk: replay reads of guest memory as submitted: 10480000 (6338835 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16039 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: pool 0x7fb0800000..0x7fb4400000 (60 MB) trimmed to its 1 live part(s), 20 MB: the rest outside every game pool for 1856 submissions (4 trims, 160 MB in all) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2254 dirty pages from their submits' captures, 44641 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 775 dispatches, 35713 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6449 runs, 567.0 MB compared; 63 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16283 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 10500000 (6350926 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 10520000 (6364029 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10540000 (6377142 pages from the captures) [hos sead::MainThread] vk: readbacks of GPU results: 1095 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 8 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1257 kept their virtual core (nothing there to wait for), 3095 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: sparse mirrors: 115, 1124 MB with memory of 1568 MB (443 MB spared); blocks given memory later 13732, given back 16900; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1217 binds (292.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 290 taken from the spares, 165 made new; 340 kept as spares, 0 destroyed; 50 spares now, 1252 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 849, 3772.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1927 dirty pages from their submits' captures, 46259 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 816 dispatches, 36915 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6647 runs, 601.0 MB compared; 87 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15866 in captures, 0 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 189 live (2220 MB), 8 retired still imported (272 MB), 43 retired given back (1024 MB) [hos host] vk: replay reads of guest memory as submitted: 10560000 (6389602 pages from the captures) [hos host] vk: a frame: 276 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 13.53, present 0.15, draw 8.73, descriptors 3.58, texture lookups 0.10, uploads 0.02, pipelines 0.35, programs 1.61, commands 2.50, dispatch 0.92, memory sync 0.51, queue submit 0.00, slot wait 0.30 [hos host] vk: a frame: mirrors 455 (1124 MB): 5.92 MB synced in 1100.7 copies, 17.71 MB written back [hos host] vk: 15840 presents, 51.8 fps; a frame: 2651 draws (0 skipped), 152 dispatches, 216 passes, 134 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14153 pipelines (3725 ms), 0 programs (10 failed), 1247 textures (2385 MB) [hos host] vk: host time in GPU resource waits: 7.75 ms a frame (fences 1.94, timelines 5.81, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 730.7 full barriers recorded, 312.2 left out (nothing recorded since the one before), 40.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.2, an op 1725.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 13.7 ms a frame (the longest submission 3.1 ms); 0.4 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.05 MB unswizzled on the CPU in 0.01 ms [hos host] vk: texture #14173 0x10004ce1800 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 608 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 13 textures retired (1235 left, 2363 MB): budget 13, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: reaper: 319 images (563 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 59 (86 MB) [hos host] vk: replay reads of guest memory as submitted: 10580000 (6402191 pages from the captures) [hos BaseProcExecutor1] vcores: 136465 takes, 23648 waits (mean 81 us, longest 22.6 ms, 0 over HOS_VCORE_LOG_MS), 11121 kicks, 11283 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 10600000 (6416067 pages from the captures) [hos host] vk: program picks: 15342233 kept (6275489 by the last picks' textures, a uniform buffer bound since), 26386 new, 5831380 with other textures (of 21200000) [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 620 ranges, 3569 MB), 258 pages written, 258 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 7568 texture pages copied (29 MB) [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2991 dirty pages from their submits' captures, 45131 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 36051 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6004 runs, 562.4 MB compared; 68 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16172 in captures, 0 only in the collection, 22 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 10620000 (6429191 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10640000 (6443738 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10660000 (6457791 pages from the captures) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #280000: event 0x100d6c802c mode 0 value 202382 (now 202363) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 10680000 (6471437 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 117, 1129 MB with memory of 1576 MB (447 MB spared); blocks given memory later 13732, given back 16901; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1220 binds (293.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 292 taken from the spares, 165 made new; 340 kept as spares, 0 destroyed; 48 spares now, 1244 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 801, 3554.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3584 dirty pages from their submits' captures, 44288 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 783 dispatches, 35507 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5944 runs, 577.5 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15882 in captures, 0 only in the collection, 22 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 189 live (2220 MB), 8 retired still imported (272 MB), 43 retired given back (1024 MB) [hos ModuleSystemWorker3] tex: the game's texture heap: 261 compaction rounds asked for in 14400 frames without one running [hos host] vk: 19 textures retired (1245 left, 2408 MB): budget 19, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 10700000 (6484123 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10720000 (6497564 pages from the captures) [hos host] vk: texture #16197 0x10005191c00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 493 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 3 textures retired (1251 left, 2391 MB): budget 3, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB [hos host] vk: readback pages refreshed from guest memory: 5078691 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3508 dirty pages from their submits' captures, 44910 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 804 dispatches, 35904 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5825 runs, 560.9 MB compared; 68 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16430 in captures, 1 only in the collection, 24 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 10740000 (6511951 pages from the captures) [hos host] vk: program picks: 15492266 kept (6345596 by the last picks' textures, a uniform buffer bound since), 26386 new, 5881347 with other textures (of 21400000) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 25678 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] nvn: GPU waits in 5 s: 264 sync waits (1403 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 10760000 (6525057 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 603 ranges, 3532 MB), 291 pages written, 291 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.036; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 24 kept (92 MB), 12 spare buffers, 8939 texture pages copied (34 MB) [hos host] vk: a frame: 289 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.50, present 0.18, draw 8.25, descriptors 3.48, texture lookups 0.18, uploads 0.07, pipelines 0.33, programs 1.52, commands 2.28, dispatch 0.78, memory sync 0.47, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 460 (1137 MB): 5.87 MB synced in 1095.8 copies, 17.54 MB written back [hos host] vk: 15960 presents, 55.6 fps; a frame: 2578 draws (0 skipped), 138 dispatches, 213 passes, 134 barriers, 6 submits, 0.5 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14153 pipelines (3725 ms), 0 programs (10 failed), 1252 textures (2392 MB) [hos host] vk: host time in GPU resource waits: 8.05 ms a frame (fences 3.25, timelines 4.79, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 715.1 full barriers recorded, 311.2 left out (nothing recorded since the one before), 38.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1692.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 13.4 ms a frame (the longest submission 4.2 ms); 0.5 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.5 (0.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.19 MB unswizzled on the CPU in 0.02 ms [hos host] vk: replay reads of guest memory as submitted: 10780000 (6537532 pages from the captures) [hos host] vk: late pages' fix (600 recordings with one): of 924 captures with late pages 922 fixed (77.4 pages, 77.8 fixes a capture), waited as before: 0 no pool, 2 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 314 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 6 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 77, of an earlier recording 231, not recorded yet 0, of this part 0; pages fixed from stamps 70535, from snapshots 785; the shadows: 321762 pages synced from captures, 314443 made known by it, 7319 left unknown because the replay or the collection wrote them since [hos host] vk: late pages found at the replay (600 takes): 17 pages in 17 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 120, 1137 MB with memory of 1592 MB (454 MB spared); blocks given memory later 13732, given back 16901; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1223 binds (295.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 295 taken from the spares, 165 made new; 340 kept as spares, 0 destroyed; 45 spares now, 1228 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2883 dirty pages from their submits' captures, 44322 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 776 dispatches, 35436 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5958 runs, 578.3 MB compared; 66 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15899 in captures, 0 only in the collection, 26 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 191 live (2232 MB), 8 retired still imported (272 MB), 43 retired given back (1024 MB) [hos host] vk: replay reads of guest memory as submitted: 10800000 (6550719 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 10820000 (6563737 pages from the captures) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1099 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 21 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1275 kept their virtual core (nothing there to wait for), 3077 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 480000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040) [hos BaseProcCreateAndDeleteThread] nvn: 2100000 recordings made, 74329 live copies (25.6 MB), 6242 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: replay reads of guest memory as submitted: 10840000 (6577437 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos rmbl::VibrationThread] hid: 53248 vibration values, strongest 0.00 since the last count [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3117 dirty pages from their submits' captures, 45226 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 857 dispatches, 35915 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5768 runs, 560.6 MB compared; 74 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16275 in captures, 0 only in the collection, 17 only by the replay itself [hos host] vk: read live, by pool: 0x7f9a400000 (4 MB, captured, state 2): 2; [hos host] vk: replay reads of guest memory as submitted: 10860000 (6590486 pages from the captures) [hos host] vk: program picks: 15642906 kept (6414825 by the last picks' textures, a uniform buffer bound since), 26386 new, 5930707 with other textures (of 21600000) [hos host] vk: texture #16206 0x10004b01400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 544 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 1 textures retired (1255 left, 2392 MB): budget 1, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: replay reads of guest memory as submitted: 10880000 (6602954 pages from the captures) [hos host] vk: texture #16204 0x10004741200 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 537 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 34 textures retired (1237 left, 2391 MB): budget 34, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 3 MB [hos host] vk: replay reads of guest memory as submitted: 10900000 (6616535 pages from the captures) [hos host] vk: sparse mirrors: 120, 1138 MB with memory of 1588 MB (449 MB spared); blocks given memory later 13734, given back 16905; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1229 binds (296.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 297 taken from the spares, 165 made new; 342 kept as spares, 0 destroyed; 45 spares now, 1232 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2945 dirty pages from their submits' captures, 44036 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 791 dispatches, 35397 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5889 runs, 577.0 MB compared; 65 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15882 in captures, 0 only in the collection, 20 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 6 retired still imported (212 MB), 45 retired given back (1084 MB) [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #205000: event 0x100d6c8008 mode 0 value 204995 (lr 0x80660f34) [hos host] vk: queues ordered by their events: 102000 submissions stopped at a wait, 101937 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 283496, not met 0; captures at a CPU event's wait 9932, 0 of them at a wait met on arrival [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 595 ranges, 3524 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 24 kept (100 MB), 10 spare buffers, 6018 texture pages copied (23 MB) [hos host] vk: replay reads of guest memory as submitted: 10920000 (6629971 pages from the captures) [hos Presentation Thread] vi: 16046 frames presented (54.9 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 131, 17.8-22 139, 22-30 5, 30-40 0, over 40 0 [hos host] vk: 12 textures retired (1234 left, 2390 MB): budget 12, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 10940000 (6642903 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3205 dirty pages from their submits' captures, 44912 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 817 dispatches, 36189 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5878 runs, 576.6 MB compared; 65 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16390 in captures, 0 only in the collection, 18 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 10960000 (6656527 pages from the captures) [hos host] vk: a frame: 297 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.29, present 0.16, draw 7.92, descriptors 3.31, texture lookups 0.12, uploads 0.03, pipelines 0.32, programs 1.47, commands 2.19, dispatch 0.79, memory sync 0.49, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 462 (1138 MB): 5.75 MB synced in 1091.3 copies, 17.49 MB written back [hos host] vk: 16080 presents, 56.5 fps; a frame: 2535 draws (0 skipped), 136 dispatches, 214 passes, 134 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14153 pipelines (3725 ms), 0 programs (10 failed), 1258 textures (2410 MB) [hos host] vk: host time in GPU resource waits: 7.94 ms a frame (fences 2.93, timelines 5.01, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 713.6 full barriers recorded, 311.1 left out (nothing recorded since the one before), 38.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.2, an op 1604.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 13.0 ms a frame (the longest submission 4.1 ms); 0.4 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.4 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 10980000 (6669004 pages from the captures) [hos host] vk: 2 textures retired (1257 left, 2383 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11907 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 11000000 (6682688 pages from the captures) [hos host] vk: program picks: 15794073 kept (6484367 by the last picks' textures, a uniform buffer bound since), 26397 new, 5979529 with other textures (of 21800000) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #285000: event 0x100d6c802c mode 0 value 206182 (now 206163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: readback pages refreshed from guest memory: 5145875 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 121, 1140 MB with memory of 1592 MB (451 MB spared); blocks given memory later 13741, given back 16905; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1234 binds (297.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 298 taken from the spares, 165 made new; 342 kept as spares, 0 destroyed; 44 spares now, 1228 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 800, 3552.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3393 dirty pages from their submits' captures, 43672 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 822 dispatches, 35275 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5737 runs, 560.2 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15890 in captures, 0 only in the collection, 17 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 5 retired still imported (184 MB), 46 retired given back (1112 MB) [hos host] vk: replay reads of guest memory as submitted: 11020000 (6696672 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 29 pages in 29 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: recorder: 7366661 packets run in 10.0 s (busy 1.82 s, spun idle 0.13 s, slept 12599 times, 4172 of them under 200 us), 0 calls made directly (no room); 18766 batches over 540 frames (34.8 a frame, 393 packets a batch), 10026 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 65 times, 8.00 ms (0.1 times, 0.01 ms a frame); every thread 709 times, 70.58 ms in all, the longest 1.00 ms; vkWaitForFences 503/47.78 ms (max 0.79), the queue's lock 99/11.78 ms (max 0.65), vmaFreeMemory 14/4.40 ms (max 1.00), vkQueueBindSparse 39/2.58 ms (max 0.40), vmaDestroyBuffer 29/1.81 ms (max 0.98), vkDestroyImageView 13/1.69 ms (max 0.54) [hos host] vk: recorder: of its busy 1.82 s, 0.49 s in parts' ends and submits (the longest 1.64 ms), 0.16 s of that blocked on the queue's lock (251 times, the longest 1.60 ms), so 0.33 s in the ends, submits and holds; 1.34 s running the other packets; 13642 packets a frame [hos host] vk: replay reads of guest memory as submitted: 11040000 (6710275 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 598 ranges, 3532 MB), 270 pages written, 270 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (102 MB), 12 spare buffers, 7648 texture pages copied (29 MB) [hos host] vk: write-backs left out (600 submits): 775, 3441.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2740 dirty pages from their submits' captures, 42109 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 753 dispatches, 33779 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5632 runs, 547.5 MB compared; 62 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15429 in captures, 0 only in the collection, 17 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 11060000 (6723314 pages from the captures) [hos host] vk: texture #21160 0x10002199c00 120x67 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (9 so far) [hos host] vk: texture #21161 0x100021a9e00 60x33 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (10 so far) [hos NavMeshCAStepDepWorker1] vcores: 146298 takes, 26205 waits (mean 80 us, longest 15.2 ms, 0 over HOS_VCORE_LOG_MS), 12215 kicks, 12413 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 11080000 (6735981 pages from the captures) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1095 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 3 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1370 kept their virtual core (nothing there to wait for), 2726 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 11100000 (6749191 pages from the captures) [hos host] vk: sparse mirrors: 121, 1130 MB with memory of 1584 MB (453 MB spared); blocks given memory later 13741, given back 16905; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1236 binds (297.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 343 kept as spares, 0 destroyed; 44 spares now, 1236 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2886 dirty pages from their submits' captures, 42602 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 784 dispatches, 33940 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5636 runs, 547.6 MB compared; 65 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 15944 in captures, 0 only in the collection, 27 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 5 retired still imported (184 MB), 46 retired given back (1112 MB) [hos host] vk: presenter thread: the copy's fence 2.48 ms, the present 0.85 ms a present [hos host] vk: replay reads of guest memory as submitted: 11120000 (6762453 pages from the captures) [hos host] vk: program picks: 15947037 kept (6555303 by the last picks' textures, a uniform buffer bound since), 26401 new, 6026561 with other textures (of 22000000) [hos host] vk: replay reads of guest memory as submitted: 11140000 (6775001 pages from the captures) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2860 dirty pages from their submits' captures, 46195 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 820 dispatches, 37089 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6268 runs, 602.3 MB compared; 43 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16613 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 11160000 (6788503 pages from the captures) [hos host] vk: memory: working set 15368 MB (peak 16021), private 21609 MB, staging 64 MB in 4 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 854 MB; used 3368 of 15579 MB; heap 1 (device): 162 blocks 4919 MB, 20364 allocations 4210 MB; used 7042 of 8536 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 18 blocks 576 MB, 132 images 306 MB; 70 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1134 images 1449 MB; 377 free ranges, the largest 32 MB [hos host] vk: render target images: 5911 reused from the pool, 4242 made new, 4108 destroyed from it, 1160 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1844 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 581 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3916 idle [hos host] vk: VRAM: 7042 MB used of 8536 MB allowed (708 MB of it free in the blocks); textures 2391 MB (budget 2134 MB), 1272 textures [hos host] vk: budget: VRAM 4919 of 8536 MB (7042 MB by the driver, 2123 MB unaccounted): textures 2391, mirrors 1130, blocks free 708, shadows 284, other buffers 229, other images 111, staging 64; RAM 3309 MB (the process's private memory 21609 MB): imports 2236, capture spares 754, retired imports 184, captures 100, capture table 35, others 0 [hos host] vk: textures: decoded ASTC 1505 MB, rendered 516 MB (81 drawn in one frame, kept: 47260 KB), other 338 MB; used in the last 8 presents 2006 MB [hos host] vk: decoded ASTC used again: 427 textures (948 MB) copied from ones the game finalized, not decoded; 427 of those retired, 2 listed; 1941 MB of blocks hashed in 333 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 11053 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 68 (14164 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 596 waited for it, 5.03 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 33.5 ms; VRAM 6.9 of 8.3 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2312 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 307 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.69, present 0.17, draw 7.63, descriptors 3.23, texture lookups 0.09, uploads 0.01, pipelines 0.31, programs 1.36, commands 2.12, dispatch 0.78, memory sync 0.50, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 464 (1130 MB): 5.71 MB synced in 1091.5 copies, 17.66 MB written back [hos host] vk: 16200 presents, 58.7 fps; a frame: 2433 draws (0 skipped), 139 dispatches, 205 passes, 134 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1272 textures (2391 MB) [hos host] vk: host time in GPU resource waits: 7.94 ms a frame (fences 3.54, timelines 4.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 718.4 full barriers recorded, 313.0 left out (nothing recorded since the one before), 28.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1479.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.7 ms a frame (the longest submission 3.7 ms); 0.1 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 18.22 ms a frame): its jobs 17.19 ms a frame (submits 1.19, presents 4.23, others 11.76), idle 1.04; the game waited for room at 28 presents, 0.23 ms each (0.01 a frame; 10.52 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 597 captures at a CPU signal; the submits' waits read in 0.042 ms a frame; readbacks: 2995 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 16.38 ms after its hand-over (47.75 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.55 + 0.14 (the window's copy is in the idle), after it 2.72 + 0.62; after a frame's last recording that reads or writes the game's memory come 1.04 recordings (0.00 ms of GPU time), after its last that reads it 1.04 (0.00 ms) [hos host] vk: frame ahead (600 presents): 11.51 parts a frame recorded in the signal's job, 70.7% of them took nothing that came with the signal; of the others, 3.38 a frame took pages of the signal's capture into a mirror (88.8 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 1 of 597 frames, its first three in 1 [hos host] vk: frame ahead, what came with the signal: 597 signals with a capture, each of 174.1 pages in 13.9 pools; between a frame's last submit and that signal 0.08 texture and sampler registrations, 0.08 releases and 0.00 other CPU signals a signal (24 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.99+0.22, 0.11+0.03, 0.34+0.11, 0.47+0.07, 0.02+0.01, 0.12+0.03, 0.27+0.03, 0.21+0.02, 0.07+0.02, 0.06+0.05, 0.03+0.02, 0.02+0.01 [hos host] vk: render targets drawn over from another start, let go: 10689 so far (50405 MB), 9614 of them (45466 MB) between the passes [hos host] vk: replay reads of guest memory as submitted: 11180000 (6802494 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 602 ranges, 3504 MB), 264 pages written, 264 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (130 MB), 8 spare buffers, 3311 texture pages copied (12 MB) [hos host] vk: replay reads of guest memory as submitted: 11200000 (6815978 pages from the captures) [hos host] vk: 1 textures retired (1276 left, 2382 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 120, 1126 MB with memory of 1580 MB (453 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1238 binds (297.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 344 kept as spares, 0 destroyed; 45 spares now, 1240 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2768 dirty pages from their submits' captures, 47616 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 849 dispatches, 38127 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6488 runs, 623.1 MB compared; 66 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17552 in captures, 0 only in the collection, 26 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 5 retired still imported (184 MB), 46 retired given back (1112 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 288 sync waits (1269 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos BaseProcExecutor3] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 25648112 lookups, 124201 missed (0.48 %) [hos host] vk: replay reads of guest memory as submitted: 11220000 (6828705 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16792 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 11240000 (6841408 pages from the captures) [hos host] vk: program picks: 16099651 kept (6624340 by the last picks' textures, a uniform buffer bound since), 26401 new, 6073947 with other textures (of 22200000) [hos host] vk: readback pages refreshed from guest memory: 5212970 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2756 dirty pages from their submits' captures, 45745 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 794 dispatches, 36722 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6216 runs, 602.0 MB compared; 68 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17085 in captures, 0 only in the collection, 27 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 11260000 (6854080 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 23 pages in 23 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 11280000 (6866943 pages from the captures) [hos sead::MainThread] nvn: CommandBufferWaitEvent #290000: event 0x100d6c802c mode 0 value 209982 (now 209963) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #210000: event 0x100d6c801c mode 0 value 209997 (lr 0x802911b8) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 601 ranges, 3464 MB), 246 pages written, 246 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 24 kept (102 MB), 12 spare buffers, 5478 texture pages copied (21 MB) [hos host] vk: sparse mirrors: 117, 1119 MB with memory of 1568 MB (448 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1240 binds (297.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 347 kept as spares, 0 destroyed; 48 spares now, 1252 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2715 dirty pages from their submits' captures, 47119 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 37797 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6288 runs, 617.8 MB compared; 70 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17507 in captures, 0 only in the collection, 25 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 5 retired still imported (184 MB), 46 retired given back (1112 MB) [hos host] vk: replay reads of guest memory as submitted: 11300000 (6879774 pages from the captures) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1092 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1549 kept their virtual core (nothing there to wait for), 3059 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 11320000 (6892659 pages from the captures) [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.08, present 0.17, draw 7.16, descriptors 2.98, texture lookups 0.08, uploads 0.01, pipelines 0.31, programs 1.24, commands 2.06, dispatch 0.79, memory sync 0.43, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 464 (1119 MB): 5.58 MB synced in 1079.5 copies, 17.70 MB written back [hos host] vk: 16320 presents, 59.5 fps; a frame: 2312 draws (0 skipped), 139 dispatches, 198 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1276 textures (2382 MB) [hos host] vk: host time in GPU resource waits: 6.93 ms a frame (fences 2.36, timelines 4.56, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 705.4 full barriers recorded, 309.2 left out (nothing recorded since the one before), 23.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.2, an op 1377.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.7 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 11340000 (6905598 pages from the captures) [hos host] vk: write-backs left out (600 submits): 874, 3883.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2631 dirty pages from their submits' captures, 46906 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 812 dispatches, 37650 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6029 runs, 598.7 MB compared; 69 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17353 in captures, 0 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vi: 16339 frames presented (58.6 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 6, 15.5-17.8 230, 17.8-22 55, 22-30 2, 30-40 0, over 40 0 [hos host] vk: program picks: 16251970 kept (6693314 by the last picks' textures, a uniform buffer bound since), 26401 new, 6121628 with other textures (of 22400000) [hos host] vk: replay reads of guest memory as submitted: 11360000 (6918426 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 11380000 (6931345 pages from the captures) [hos ModuleSystemWorker3] nvn: 2150000 recordings made, 72002 live copies (24.6 MB), 6246 texture copies; 4 command buffers' segments (4.4 MB) [hos host] vk: sparse mirrors: 117, 1119 MB with memory of 1568 MB (448 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1240 binds (297.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 347 kept as spares, 0 destroyed; 48 spares now, 1252 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2381 dirty pages from their submits' captures, 47040 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 814 dispatches, 37488 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6189 runs, 616.3 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17235 in captures, 0 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 4 retired still imported (164 MB), 47 retired given back (1132 MB) [hos host] vk: replay reads of guest memory as submitted: 11400000 (6944183 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 600 ranges, 3436 MB), 236 pages written, 236 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 24 kept (106 MB), 12 spare buffers, 5479 texture pages copied (21 MB) [hos host] vk: replay reads of guest memory as submitted: 11420000 (6956828 pages from the captures) [hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 500000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 851, 3776.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2815 dirty pages from their submits' captures, 45553 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 789 dispatches, 36614 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6168 runs, 616.3 MB compared; 72 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16886 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 11440000 (6969720 pages from the captures) [hos host] vk: texture slots registered again at another size: 800000 (of 1006629 registrations; slot 18985 640x360 -> 60x33) [hos host] vk: 16 textures retired (1260 left, 2365 MB): budget 16, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 11460000 (6982606 pages from the captures) [hos host] vk: program picks: 16404134 kept (6762156 by the last picks' textures, a uniform buffer bound since), 26401 new, 6169464 with other textures (of 22600000) [hos host] vk: replay reads of guest memory as submitted: 11480000 (6995385 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 5280360 pages so far (0 held: may be written ahead in their submission) [hos BaseProcExecutor5] vcores: 152927 takes, 24313 waits (mean 77 us, longest 37.6 ms, 0 over HOS_VCORE_LOG_MS), 11692 kicks, 11916 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: sparse mirrors: 114, 1111 MB with memory of 1552 MB (440 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1242 binds (297.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 350 kept as spares, 0 destroyed; 51 spares now, 1268 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3040 dirty pages from their submits' captures, 46821 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 810 dispatches, 37661 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6192 runs, 616.4 MB compared; 70 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17414 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 4 retired still imported (164 MB), 47 retired given back (1132 MB) [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.24, present 0.17, draw 7.34, descriptors 3.07, texture lookups 0.08, uploads 0.00, pipelines 0.30, programs 1.28, commands 2.10, dispatch 0.81, memory sync 0.40, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 464 (1111 MB): 5.54 MB synced in 1072.2 copies, 17.61 MB written back [hos host] vk: 16440 presents, 59.3 fps; a frame: 2304 draws (0 skipped), 135 dispatches, 197 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1260 textures (2365 MB) [hos host] vk: host time in GPU resource waits: 6.61 ms a frame (fences 2.34, timelines 4.27, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 699.3 full barriers recorded, 308.6 left out (nothing recorded since the one before), 23.3 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1371.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.6 ms a frame (the longest submission 3.3 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15374 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 11500000 (7008092 pages from the captures) [hos sead::MainThread] vk: readbacks of GPU results: 1085 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1519 kept their virtual core (nothing there to wait for), 3089 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 352 mirrors given back (2397 MB in all); 1101 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: reaper: 124 images (157 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 33 (28 MB) [hos host] vk: late pages found at the replay (600 takes): 20 pages in 20 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 11520000 (7020741 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 602 ranges, 3356 MB), 241 pages written, 241 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 24 kept (112 MB), 12 spare buffers, 5189 texture pages copied (20 MB) [hos host] vk: write-backs left out (600 submits): 874, 3883.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1908 dirty pages from their submits' captures, 47678 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 890 dispatches, 37679 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6098 runs, 615.0 MB compared; 69 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17203 in captures, 0 only in the collection, 27 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 11540000 (7033469 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 11560000 (7046354 pages from the captures) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #295000: event 0x100d6c802c mode 0 value 213782 (now 213763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1243 binds (297.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 352 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2709 dirty pages from their submits' captures, 46749 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 808 dispatches, 37589 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5825 runs, 595.2 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17206 in captures, 0 only in the collection, 25 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 4 retired still imported (164 MB), 47 retired given back (1132 MB) [hos host] vk: replay reads of guest memory as submitted: 11580000 (7058958 pages from the captures) [hos host] vk: program picks: 16556232 kept (6831003 by the last picks' textures, a uniform buffer bound since), 26401 new, 6217366 with other textures (of 22800000) [hos host] vk: replay reads of guest memory as submitted: 11600000 (7072446 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 294 sync waits (1222 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 11620000 (7084963 pages from the captures) [hos host] vk: 13 textures retired (1247 left, 2359 MB): budget 13, drawn over 0, dead targets 0 [hos host] vk: late pages' fix (600 recordings with one): of 759 captures with late pages 759 fixed (75.1 pages, 75.1 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 503 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 5 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 15, of an earlier recording 483, not recorded yet 0, of this part 0; pages fixed from stamps 56039, from snapshots 927; the shadows: 288843 pages synced from captures, 287328 made known by it, 1515 left unknown because the replay or the collection wrote them since [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2749 dirty pages from their submits' captures, 46959 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 828 dispatches, 37620 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6101 runs, 629.6 MB compared; 71 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17177 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 11640000 (7097520 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 601 ranges, 3350 MB), 238 pages written, 238 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 24 kept (88 MB), 12 spare buffers, 5189 texture pages copied (20 MB) [hos host] vk: 28 textures retired (1219 left, 2322 MB): budget 28, drawn over 0, dead targets 0 [hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.81, present 0.14, draw 7.00, descriptors 2.88, texture lookups 0.08, uploads 0.00, pipelines 0.30, programs 1.25, commands 2.01, dispatch 0.77, memory sync 0.38, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 464 (1100 MB): 5.52 MB synced in 1073.0 copies, 17.52 MB written back [hos host] vk: 16560 presents, 60.1 fps; a frame: 2302 draws (0 skipped), 129 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1219 textures (2322 MB) [hos host] vk: host time in GPU resource waits: 5.37 ms a frame (fences 2.23, timelines 3.14, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 692.7 full barriers recorded, 308.5 left out (nothing recorded since the one before), 18.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.4 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #215000: event 0x100d6c8030 mode 0 value 214999 (lr 0x806612a8) [hos host] vk: replay reads of guest memory as submitted: 11660000 (7110282 pages from the captures) [hos host] vk: 4 textures retired (1215 left, 2321 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 685 dirty pages from their submits' captures, 50801 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 856 dispatches, 39629 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6296 runs, 647.3 MB compared; 89 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17744 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 4 retired still imported (164 MB), 47 retired given back (1132 MB) [hos host] vk: replay reads of guest memory as submitted: 11680000 (7122987 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11860 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 11700000 (7135709 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: program picks: 16708808 kept (6900163 by the last picks' textures, a uniform buffer bound since), 26401 new, 6264790 with other textures (of 23000000) [hos host] vk: late pages found at the replay (600 takes): 11 pages in 11 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 11720000 (7148533 pages from the captures) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1678 kept their virtual core (nothing there to wait for), 3186 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: readback pages refreshed from guest memory: 5347214 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 783 dirty pages from their submits' captures, 49416 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 38605 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6128 runs, 629.9 MB compared; 90 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17216 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 11740000 (7161267 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 11760000 (7174013 pages from the captures) [hos Presentation Thread] vi: 16640 frames presented (60.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 15, 15.5-17.8 262, 17.8-22 22, 22-30 2, 30-40 0, over 40 0 [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 600 ranges, 3346 MB), 256 pages written, 256 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 24 kept (84 MB), 10 spare buffers, 6288 texture pages copied (24 MB) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1512 dirty pages from their submits' captures, 49642 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 846 dispatches, 39066 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6129 runs, 629.7 MB compared; 88 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17899 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: replay reads of guest memory as submitted: 11780000 (7186732 pages from the captures) [hos ModuleSystemWorker2] nvn: 2200000 recordings made, 71930 live copies (24.5 MB), 6246 texture copies; 6 command buffers' segments (4.4 MB) [hos Presentation Thread] nvn: QueueSubmitCommands #100000: queue 000010100D6AA020 (its #16664), 1 handles (lr 0x805282f4) [hos host] vk: replay reads of guest memory as submitted: 11800000 (7199432 pages from the captures) [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.62, present 0.14, draw 6.99, descriptors 2.88, texture lookups 0.07, uploads 0.00, pipelines 0.29, programs 1.25, commands 2.00, dispatch 0.76, memory sync 0.42, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 464 (1100 MB): 5.45 MB synced in 1069.6 copies, 17.50 MB written back [hos host] vk: 16680 presents, 59.9 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1215 textures (2321 MB) [hos host] vk: host time in GPU resource waits: 3.30 ms a frame (fences 2.46, timelines 0.84, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 689.5 full barriers recorded, 308.2 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1366.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 11820000 (7211918 pages from the captures) [hos host] vk: program picks: 16861226 kept (6969208 by the last picks' textures, a uniform buffer bound since), 26401 new, 6312372 with other textures (of 23200000) [hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1354 dirty pages from their submits' captures, 49392 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 838 dispatches, 39097 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6274 runs, 647.2 MB compared; 77 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17454 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #300000: event 0x100d6c802c mode 0 value 217582 (now 217563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 11840000 (7224452 pages from the captures) [hos host] vk: recorder: 7345139 packets run in 10.0 s (busy 1.83 s, spun idle 0.14 s, slept 13218 times, 4912 of them under 200 us), 0 calls made directly (no room); 19838 batches over 598 frames (33.2 a frame, 370 packets a batch), 10308 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 26 times, 2.26 ms (0.0 times, 0.00 ms a frame); every thread 648 times, 61.96 ms in all, the longest 0.52 ms; vkWaitForFences 568/56.19 ms (max 0.52), the queue's lock 38/2.64 ms (max 0.22), vmaFreeMemory 9/1.69 ms (max 0.47), vkDestroyImageView 7/0.79 ms (max 0.21), vmaDestroyBuffer 6/0.30 ms (max 0.25), vkQueueBindSparse 10/0.16 ms (max 0.06) [hos host] vk: recorder: of its busy 1.83 s, 0.45 s in parts' ends and submits (the longest 1.45 ms), 0.13 s of that blocked on the queue's lock (199 times, the longest 1.38 ms), so 0.32 s in the ends, submits and holds; 1.38 s running the other packets; 12283 packets a frame [hos host] vk: replay reads of guest memory as submitted: 11860000 (7236942 pages from the captures) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1040 dirty pages from their submits' captures, 49247 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 834 dispatches, 38614 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6133 runs, 630.0 MB compared; 89 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17444 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11140 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 11880000 (7249701 pages from the captures) [hos ModuleSystemWorker2] vcores: 149193 takes, 21060 waits (mean 67 us, longest 15.7 ms, 0 over HOS_VCORE_LOG_MS), 10290 kicks, 10381 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 598 ranges, 3286 MB), 246 pages written, 246 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 6991 texture pages copied (27 MB) [hos host] vk: replay reads of guest memory as submitted: 11900000 (7262443 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 11920000 (7275116 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1028 dirty pages from their submits' captures, 49156 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 833 dispatches, 38543 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6124 runs, 629.8 MB compared; 99 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17262 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: presenter thread: the copy's fence 2.44 ms, the present 0.84 ms a present [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1580 kept their virtual core (nothing there to wait for), 3028 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: program picks: 17013594 kept (7038169 by the last picks' textures, a uniform buffer bound since), 26401 new, 6360004 with other textures (of 23400000) [hos host] vk: replay reads of guest memory as submitted: 11940000 (7287868 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 11960000 (7300572 pages from the captures) [hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 520000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040) [hos host] vk: readback pages refreshed from guest memory: 5414520 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1524 dirty pages from their submits' captures, 48773 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 832 dispatches, 38617 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6122 runs, 629.7 MB compared; 76 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17532 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: memory: working set 15317 MB (peak 16021), private 21494 MB, staging 48 MB in 3 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 848 MB; used 3304 of 15579 MB; heap 1 (device): 161 blocks 4855 MB, 19836 allocations 4095 MB; used 6978 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 18 blocks 576 MB, 133 images 306 MB; 70 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1076 images 1379 MB; 395 free ranges, the largest 32 MB [hos host] vk: render target images: 5911 reused from the pool, 4243 made new, 4108 destroyed from it, 1160 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1845 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 581 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3916 idle [hos host] vk: VRAM: 6978 MB used of 11474 MB allowed (759 MB of it free in the blocks); textures 2321 MB (budget 7158 MB), 1215 textures [hos host] vk: budget: VRAM 4855 of 11474 MB (6978 MB by the driver, 2123 MB unaccounted): textures 2321, mirrors 1100, blocks free 759, shadows 284, other buffers 229, other images 111, staging 48; RAM 3239 MB (the process's private memory 21494 MB): imports 2236, capture spares 742, retired imports 120, captures 102, capture table 35, other buffers 4, others 0 [hos host] vk: textures: decoded ASTC 1476 MB, rendered 516 MB (84 drawn in one frame, kept: 47616 KB), other 298 MB; used in the last 8 presents 1970 MB [hos host] vk: decoded ASTC used again: 429 textures (949 MB) copied from ones the game finalized, not decoded; 429 of those retired, 1 listed; 1941 MB of blocks hashed in 333 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 11653 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 68 (14164 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 479 waited for it, 3.27 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: replay reads of guest memory as submitted: 11980000 (7313211 pages from the captures) [hos host] vk: status: 600 frames, the slowest 25.4 ms; VRAM 6.8 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2312 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.70, present 0.13, draw 7.01, descriptors 2.87, texture lookups 0.07, uploads 0.00, pipelines 0.30, programs 1.25, commands 2.01, dispatch 0.76, memory sync 0.42, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 464 (1100 MB): 5.45 MB synced in 1069.9 copies, 17.50 MB written back [hos host] vk: 16800 presents, 60.0 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1215 textures (2321 MB) [hos host] vk: host time in GPU resource waits: 3.27 ms a frame (fences 2.33, timelines 0.93, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 690.3 full barriers recorded, 308.7 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1366.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.4 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: the render thread (600 presents, 16.74 ms a frame): its jobs 15.77 ms a frame (submits 2.85, presents 4.65, others 8.26), idle 0.97; the game waited for room at 73 presents, 0.09 ms each (0.01 a frame; 19.81 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 482 captures at a CPU signal; the submits' waits read in 0.042 ms a frame; readbacks: 2425 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.05 ms after its hand-over (20.03 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.45 + 0.13 (the window's copy is in the idle), after it 2.14 + 0.47; after a frame's last recording that reads or writes the game's memory come 1.23 recordings (0.00 ms of GPU time), after its last that reads it 1.23 (0.00 ms) [hos host] vk: frame ahead (600 presents): 9.91 parts a frame recorded in the signal's job, 70.5% of them took nothing that came with the signal; of the others, 2.92 a frame took pages of the signal's capture into a mirror (70.6 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 11 of 482 frames, its first three in 10 [hos host] vk: frame ahead, what came with the signal: 482 signals with a capture, each of 95.0 pages in 13.6 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.98+0.17, 0.09+0.03, 0.28+0.11, 0.46+0.07, 0.02+0.01, 0.07+0.01, 0.09+0.01, 0.06+0.02, 0.05+0.02, 0.03+0.01, 0.01+0.00, 0.00+0.00 [hos host] vk: queues ordered by their events: 108000 submissions stopped at a wait, 107936 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 302622, not met 0; captures at a CPU event's wait 10576, 0 of them at a wait met on arrival [hos host] vk: render targets drawn over from another start, let go: 10689 so far (50405 MB), 9614 of them (45466 MB) between the passes [hos host] vk: replay reads of guest memory as submitted: 12000000 (7326124 pages from the captures) [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #220000: event 0x100d6c803c mode 0 value 219999 (lr 0x80661bfc) [hos host] vk: (its written part: 1 of 1 KB readable, 759 bytes not zero; levels L0) [hos host] vk: texture 0x112a256a00 2448x3056x1 x1 levels 1 NVN 0x4b (7344 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 1 KB of it [hos host] vk: (its written part: 1 of 1 KB readable, 759 bytes not zero; levels L0) [hos host] vk: texture 0x112a256a00 2448x3056x1 x1 levels 1 NVN 0x4b (7344 KB of storage) re-uploaded (its memory was written; flags 0x4): what the CPU wrote, 1 KB of it [hos host] vk: replay reads of guest memory as submitted: 12020000 (7338934 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 594 ranges, 3216 MB), 247 pages written, 247 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (86 MB), 10 spare buffers, 7001 texture pages copied (27 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 150 sync waits (202 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1064 dirty pages from their submits' captures, 49440 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 840 dispatches, 38669 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6128 runs, 629.8 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17390 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos ModuleSystemWorker5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 14555558 lookups, 119244 missed (0.82 %) [hos host] vk: replay reads of guest memory as submitted: 12040000 (7351517 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5986 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 17166298 kept (7107258 by the last picks' textures, a uniform buffer bound since), 26402 new, 6407299 with other textures (of 23600000) [hos host] vk: replay reads of guest memory as submitted: 12060000 (7364301 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1244 binds (298.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 299 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 635 dirty pages from their submits' captures, 50841 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 849 dispatches, 39658 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6325 runs, 647.9 MB compared; 83 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17791 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 192 live (2236 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: replay reads of guest memory as submitted: 12080000 (7377173 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 12100000 (7389960 pages from the captures) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #305000: event 0x100d6c802c mode 0 value 221382 (now 221363) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 12120000 (7402824 pages from the captures) [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1690 dirty pages from their submits' captures, 47330 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 812 dispatches, 37530 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6041 runs, 613.8 MB compared; 75 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17548 in captures, 0 only in the collection, 27 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 12140000 (7415590 pages from the captures) [hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.63, present 0.15, draw 6.87, descriptors 2.79, texture lookups 0.07, uploads 0.01, pipelines 0.29, programs 1.23, commands 1.99, dispatch 0.77, memory sync 0.41, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 465 (1101 MB): 5.48 MB synced in 1071.6 copies, 17.52 MB written back [hos host] vk: 16920 presents, 60.0 fps; a frame: 2305 draws (0 skipped), 129 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1218 textures (2322 MB) [hos host] vk: host time in GPU resource waits: 3.91 ms a frame (fences 2.28, timelines 1.63, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 690.5 full barriers recorded, 308.1 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.4 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1083 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1349 kept their virtual core (nothing there to wait for), 2747 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 581 ranges, 3224 MB), 243 pages written, 243 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 7001 texture pages copied (27 MB) [hos host] vk: replay reads of guest memory as submitted: 12160000 (7428406 pages from the captures) [hos Presentation Thread] vi: 16940 frames presented (59.9 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 9, 15.5-17.8 280, 17.8-22 11, 22-30 0, 30-40 0, over 40 0 [hos host] vk: program picks: 17318891 kept (7176301 by the last picks' textures, a uniform buffer bound since), 26406 new, 6454702 with other textures (of 23800000) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2635 dirty pages from their submits' captures, 48375 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 831 dispatches, 38730 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6191 runs, 631.3 MB compared; 71 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17341 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: replay reads of guest memory as submitted: 12180000 (7441276 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 12200000 (7453992 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 5481596 pages so far (0 held: may be written ahead in their submission) [hos host] vk: replay reads of guest memory as submitted: 12220000 (7466827 pages from the captures) [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2550 dirty pages from their submits' captures, 46968 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 812 dispatches, 37584 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6038 runs, 613.8 MB compared; 72 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17232 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 12240000 (7479667 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13386 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 12260000 (7492344 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 576 ranges, 3186 MB), 246 pages written, 246 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (150 MB), 11 spare buffers, 6019 texture pages copied (23 MB) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1643 dirty pages from their submits' captures, 50427 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 854 dispatches, 39785 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6351 runs, 648.5 MB compared; 86 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18173 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: replay reads of guest memory as submitted: 12280000 (7504967 pages from the captures) [hos host] vk: program picks: 17471479 kept (7245400 by the last picks' textures, a uniform buffer bound since), 26406 new, 6502114 with other textures (of 24000000) [hos host] vk: late pages found at the replay (600 takes): 14 pages in 14 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 12300000 (7517712 pages from the captures) [hos ModuleSystemWorker1] vcores: 148897 takes, 20983 waits (mean 64 us, longest 7.6 ms, 0 over HOS_VCORE_LOG_MS), 10181 kicks, 10283 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.98, present 0.14, draw 7.14, descriptors 2.94, texture lookups 0.07, uploads 0.00, pipelines 0.30, programs 1.26, commands 2.05, dispatch 0.80, memory sync 0.41, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 465 (1101 MB): 5.52 MB synced in 1072.6 copies, 17.54 MB written back [hos host] vk: 17040 presents, 60.0 fps; a frame: 2308 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1220 textures (2322 MB) [hos host] vk: host time in GPU resource waits: 4.59 ms a frame (fences 2.38, timelines 2.21, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 692.7 full barriers recorded, 308.2 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2479 dirty pages from their submits' captures, 46903 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 37557 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6024 runs, 613.6 MB compared; 80 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17136 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 12320000 (7530495 pages from the captures) [hos ModuleSystemWorker4] nvn: 2250000 recordings made, 71929 live copies (24.6 MB), 6246 texture copies; 5 command buffers' segments (4.3 MB) [hos host] vk: replay reads of guest memory as submitted: 12340000 (7543294 pages from the captures) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1587 kept their virtual core (nothing there to wait for), 3021 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 12360000 (7556055 pages from the captures) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2235 dirty pages from their submits' captures, 48668 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 836 dispatches, 38703 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6195 runs, 631.2 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17697 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos sead::MainThread] nvn: CommandBufferSignalEvent #225000: event 0x100d6c8048 mode 0 value 224999 (lr 0x80666598) [hos host] vk: replay reads of guest memory as submitted: 12380000 (7568901 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 563 ranges, 3064 MB), 239 pages written, 239 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 24 kept (104 MB), 12 spare buffers, 5592 texture pages copied (21 MB) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #310000: event 0x100d6c802c mode 0 value 225182 (now 225163) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 12400000 (7581808 pages from the captures) [hos host] vk: program picks: 17624108 kept (7314397 by the last picks' textures, a uniform buffer bound since), 26406 new, 6549485 with other textures (of 24200000) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1673 dirty pages from their submits' captures, 47815 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 814 dispatches, 37668 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6038 runs, 613.9 MB compared; 77 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17147 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 12420000 (7595023 pages from the captures) [hos host] vk: reaper: 47 images (44 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 27 (37 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 222 sync waits (601 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 12440000 (7607695 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 12460000 (7620312 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 5549094 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2337 dirty pages from their submits' captures, 48525 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 38689 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6197 runs, 631.1 MB compared; 73 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17701 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.90, present 0.14, draw 7.03, descriptors 2.85, texture lookups 0.09, uploads 0.01, pipelines 0.31, programs 1.26, commands 2.02, dispatch 0.79, memory sync 0.43, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 465 (1101 MB): 5.52 MB synced in 1073.1 copies, 17.54 MB written back [hos host] vk: 17160 presents, 60.1 fps; a frame: 2310 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1226 textures (2323 MB) [hos host] vk: host time in GPU resource waits: 4.60 ms a frame (fences 2.41, timelines 2.19, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 693.3 full barriers recorded, 308.5 left out (nothing recorded since the one before), 18.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1366.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos ModuleSystemWorker3] barrier: 267027 looks: 103187 found the masks moved already, 161078 waited until they moved, 2762 for the whole 1 ms (a wait 337 us on average); 250213 wakes [hos host] vk: replay reads of guest memory as submitted: 12480000 (7633063 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11957 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 12500000 (7645673 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 6 pages in 6 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 562 ranges, 3062 MB), 256 pages written, 256 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (166 MB), 12 spare buffers, 5992 texture pages copied (23 MB) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1538 dirty pages from their submits' captures, 49520 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 845 dispatches, 38942 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6197 runs, 631.2 MB compared; 83 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17973 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 540000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040) [hos host] vk: program picks: 17776761 kept (7383482 by the last picks' textures, a uniform buffer bound since), 26406 new, 6596832 with other textures (of 24400000) [hos host] vk: replay reads of guest memory as submitted: 12520000 (7658128 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 12540000 (7670809 pages from the captures) [hos rmbl::VibrationThread] hid: 57344 vibration values, strongest 0.00 since the last count [hos host] vk: replay reads of guest memory as submitted: 12560000 (7683576 pages from the captures) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1353 dirty pages from their submits' captures, 48716 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 820 dispatches, 38361 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6206 runs, 631.4 MB compared; 77 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17032 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1082 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1560 kept their virtual core (nothing there to wait for), 3048 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 12580000 (7696276 pages from the captures) [hos Presentation Thread] vi: 17241 frames presented (60.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 4, 15.5-17.8 281, 17.8-22 16, 22-30 0, 30-40 0, over 40 0 [hos host] vk: replay reads of guest memory as submitted: 12600000 (7709089 pages from the captures) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1813 dirty pages from their submits' captures, 48907 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 857 dispatches, 38708 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6191 runs, 631.2 MB compared; 85 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17600 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 12620000 (7721778 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 553 ranges, 3018 MB), 246 pages written, 246 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 6039 texture pages copied (23 MB) [hos host] vk: program picks: 17929255 kept (7452566 by the last picks' textures, a uniform buffer bound since), 26406 new, 6644338 with other textures (of 24600000) [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.84, present 0.13, draw 6.92, descriptors 2.80, texture lookups 0.07, uploads 0.01, pipelines 0.30, programs 1.24, commands 2.01, dispatch 0.78, memory sync 0.44, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 465 (1101 MB): 5.50 MB synced in 1072.5 copies, 17.54 MB written back [hos host] vk: 17280 presents, 59.8 fps; a frame: 2308 draws (0 skipped), 130 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1228 textures (2326 MB) [hos host] vk: host time in GPU resource waits: 3.55 ms a frame (fences 2.43, timelines 1.12, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 692.8 full barriers recorded, 308.7 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1366.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 12640000 (7734472 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 12660000 (7747210 pages from the captures) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #315000: event 0x100d6c802c mode 0 value 228982 (now 228963) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2561 dirty pages from their submits' captures, 48378 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 833 dispatches, 38717 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6190 runs, 631.1 MB compared; 73 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17755 in captures, 0 only in the collection, 26 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: recorder: 7311847 packets run in 10.0 s (busy 1.84 s, spun idle 0.14 s, slept 13083 times, 4757 of them under 200 us), 0 calls made directly (no room); 19469 batches over 599 frames (32.5 a frame, 376 packets a batch), 10019 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 2 times, 0.29 ms (0.0 times, 0.00 ms a frame); every thread 612 times, 57.44 ms in all, the longest 0.59 ms; vkWaitForFences 567/55.06 ms (max 0.59), the queue's lock 35/1.55 ms (max 0.18), a part's submit 2/0.33 ms (max 0.18), vmaDestroyBuffer 3/0.21 ms (max 0.18), vkDestroyImageView 2/0.15 ms (max 0.09), vkQueueBindSparse 1/0.11 ms (max 0.11) [hos host] vk: recorder: of its busy 1.84 s, 0.49 s in parts' ends and submits (the longest 1.68 ms), 0.17 s of that blocked on the queue's lock (268 times, the longest 1.63 ms), so 0.32 s in the ends, submits and holds; 1.35 s running the other packets; 12207 packets a frame [hos host] vk: replay reads of guest memory as submitted: 12680000 (7759920 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 29341 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 12700000 (7772585 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 5616185 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2704 dirty pages from their submits' captures, 48222 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 832 dispatches, 38690 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6185 runs, 631.0 MB compared; 76 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17726 in captures, 0 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: late pages found at the replay (600 takes): 8 pages in 8 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos nn.npln.Worker] vcores: 148822 takes, 21324 waits (mean 64 us, longest 7.0 ms, 0 over HOS_VCORE_LOG_MS), 10481 kicks, 10603 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 12720000 (7785098 pages from the captures) [hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #230000: event 0x100d6c800c mode 0 value 230000 (lr 0x802915c8) [hos host] vk: late pages' fix (600 recordings with one): of 859 captures with late pages 859 fixed (75.4 pages, 75.4 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 533 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 70, of an earlier recording 463, not recorded yet 0, of this part 0; pages fixed from stamps 63620, from snapshots 1147; the shadows: 399955 pages synced from captures, 393578 made known by it, 6377 left unknown because the replay or the collection wrote them since [hos host] vk: replay reads of guest memory as submitted: 12740000 (7797512 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 539 ranges, 2920 MB), 253 pages written, 253 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.036; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 24 kept (90 MB), 12 spare buffers, 8439 texture pages copied (32 MB) [hos host] vk: presenter thread: the copy's fence 2.28 ms, the present 0.82 ms a present [hos host] vk: program picks: 18081408 kept (7521617 by the last picks' textures, a uniform buffer bound since), 26406 new, 6692185 with other textures (of 24800000) [hos host] vk: replay reads of guest memory as submitted: 12760000 (7810152 pages from the captures) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2431 dirty pages from their submits' captures, 46989 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 810 dispatches, 37613 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6027 runs, 613.8 MB compared; 76 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17220 in captures, 0 only in the collection, 25 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos host] vk: replay reads of guest memory as submitted: 12780000 (7822894 pages from the captures) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1679 kept their virtual core (nothing there to wait for), 3185 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 12800000 (7835625 pages from the captures) [hos host] vk: memory: working set 15323 MB (peak 16021), private 21499 MB, staging 32 MB in 2 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 848 MB; used 3308 of 15579 MB; heap 1 (device): 161 blocks 4855 MB, 19868 allocations 4116 MB; used 6978 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 18 blocks 576 MB, 133 images 306 MB; 70 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1097 images 1415 MB; 396 free ranges, the largest 32 MB [hos host] vk: render target images: 5911 reused from the pool, 4243 made new, 4108 destroyed from it, 1160 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1845 with the pool empty (0 of them with targets still in the dead queue), 334 with none in it a present yet, 37 with none of a kind they can use, 569 with all smaller (the largest 50% short on average), 581 with all bigger (over 1.5 times), 876 with some of each, 1 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 3916 idle [hos host] vk: VRAM: 6978 MB used of 11474 MB allowed (738 MB of it free in the blocks); textures 2358 MB (budget 7158 MB), 1236 textures [hos host] vk: budget: VRAM 4855 of 11474 MB (6978 MB by the driver, 2123 MB unaccounted): textures 2358, mirrors 1101, blocks free 738, shadows 284, other buffers 229, other images 111, staging 32; RAM 3243 MB (the process's private memory 21499 MB): imports 2240, capture spares 736, retired imports 120, captures 112, capture table 35, others 0 [hos host] vk: textures: decoded ASTC 1476 MB, rendered 516 MB (84 drawn in one frame, kept: 47616 KB), other 334 MB; used in the last 8 presents 1968 MB [hos host] vk: decoded ASTC used again: 437 textures (955 MB) copied from ones the game finalized, not decoded; 437 of those retired, 2 listed; 1943 MB of blocks hashed in 333 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 12253 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 68 (14164 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 105x33x1 NVN 0x25: 1 (13 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 169x97x1 NVN 0x1: 1 (16 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 474 waited for it, 2.72 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 24.8 ms; VRAM 6.8 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2315 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.07, present 0.16, draw 7.14, descriptors 2.96, texture lookups 0.15, uploads 0.07, pipelines 0.31, programs 1.24, commands 2.05, dispatch 0.79, memory sync 0.48, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 465 (1101 MB): 5.53 MB synced in 1073.0 copies, 17.54 MB written back [hos host] vk: 17400 presents, 60.0 fps; a frame: 2308 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.1 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1236 textures (2358 MB) [hos host] vk: host time in GPU resource waits: 6.76 ms a frame (fences 2.30, timelines 4.45, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 693.2 full barriers recorded, 308.4 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.26 MB unswizzled on the CPU in 0.03 ms [hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 15.72 ms a frame (submits 2.94, presents 4.60, others 8.18), idle 0.96; the game waited for room at 47 presents, 0.09 ms each (0.01 a frame; 14.49 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 477 captures at a CPU signal; the submits' waits read in 0.043 ms a frame; readbacks: 2352 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.95 ms after its hand-over (21.30 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.41 + 0.13 (the window's copy is in the idle), after it 2.16 + 0.47; after a frame's last recording that reads or writes the game's memory come 1.16 recordings (0.00 ms of GPU time), after its last that reads it 1.16 (0.00 ms) [hos host] vk: frame ahead (600 presents): 9.77 parts a frame recorded in the signal's job, 70.7% of them took nothing that came with the signal; of the others, 2.86 a frame took pages of the signal's capture into a mirror (69.9 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 19 of 477 frames, its first three in 19 [hos host] vk: frame ahead, what came with the signal: 477 signals with a capture, each of 88.0 pages in 13.4 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.98+0.17, 0.09+0.03, 0.30+0.11, 0.46+0.07, 0.02+0.01, 0.08+0.01, 0.10+0.01, 0.05+0.02, 0.05+0.02, 0.02+0.01, 0.01+0.00, 0.00+0.00 [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2896 dirty pages from their submits' captures, 46860 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 809 dispatches, 37644 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6017 runs, 613.6 MB compared; 70 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 104 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17831 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: render targets drawn over from another start, let go: 10689 so far (50405 MB), 9614 of them (45466 MB) between the passes [hos host] vk: replay reads of guest memory as submitted: 12820000 (7848301 pages from the captures) [hos sead::MainThread] nvn: 2300000 recordings made, 71990 live copies (25.8 MB), 6246 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: replay reads of guest memory as submitted: 12840000 (7861840 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 240 sync waits (738 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 14777399 lookups, 120813 missed (0.82 %) [hos host] vk: replay reads of guest memory as submitted: 12860000 (7874683 pages from the captures) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1927 dirty pages from their submits' captures, 49017 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 843 dispatches, 38762 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6190 runs, 631.3 MB compared; 77 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17074 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 3 retired still imported (120 MB), 48 retired given back (1176 MB) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 539 ranges, 2920 MB), 250 pages written, 250 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (90 MB), 11 spare buffers, 3450 texture pages copied (13 MB) [hos host] vk: program picks: 18233822 kept (7590642 by the last picks' textures, a uniform buffer bound since), 26406 new, 6739771 with other textures (of 25000000) [hos host] vk: replay reads of guest memory as submitted: 12880000 (7887298 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 12900000 (7899919 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1641 dirty pages from their submits' captures, 48775 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 824 dispatches, 38619 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6180 runs, 631.1 MB compared; 79 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17625 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10861 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 12920000 (7912541 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 11 pages in 11 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #320000: event 0x100d6c802c mode 0 value 232782 (now 232763) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos sead::MainThread] nvn: EventSignal (CPU) #35000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: replay reads of guest memory as submitted: 12940000 (7925173 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 5683644 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1435 dirty pages from their submits' captures, 49058 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 836 dispatches, 38625 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6204 runs, 631.3 MB compared; 84 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17475 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 1 retired still imported (60 MB), 50 retired given back (1236 MB) [hos host] vk: replay reads of guest memory as submitted: 12960000 (7937990 pages from the captures) [hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.74, present 0.14, draw 6.95, descriptors 2.84, texture lookups 0.07, uploads 0.00, pipelines 0.29, programs 1.23, commands 2.02, dispatch 0.77, memory sync 0.41, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 465 (1101 MB): 5.50 MB synced in 1072.6 copies, 17.54 MB written back [hos host] vk: 17520 presents, 60.2 fps; a frame: 2309 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1236 textures (2358 MB) [hos host] vk: host time in GPU resource waits: 4.09 ms a frame (fences 2.35, timelines 1.74, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 692.3 full barriers recorded, 308.2 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 12980000 (7950885 pages from the captures) [hos host] vk: program picks: 18386616 kept (7659725 by the last picks' textures, a uniform buffer bound since), 26406 new, 6786977 with other textures (of 25200000) [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 534 ranges, 2878 MB), 239 pages written, 239 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (108 MB), 11 spare buffers, 3450 texture pages copied (13 MB) [hos Presentation Thread] vi: 17541 frames presented (59.9 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 9, 15.5-17.8 270, 17.8-22 19, 22-30 2, 30-40 0, over 40 0 [hos sead::MainThread] vk: readbacks of GPU results: 1085 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1380 kept their virtual core (nothing there to wait for), 2716 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 13000000 (7963705 pages from the captures) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2647 dirty pages from their submits' captures, 47739 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 827 dispatches, 38128 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6037 runs, 613.8 MB compared; 78 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17746 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 13020000 (7976642 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 13040000 (7989432 pages from the captures) [hos host] vk: queues ordered by their events: 114000 submissions stopped at a wait, 113938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 321855, not met 0; captures at a CPU event's wait 11192, 0 of them at a wait met on arrival [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2705 dirty pages from their submits' captures, 46145 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 797 dispatches, 37093 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6026 runs, 613.8 MB compared; 72 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16773 in captures, 0 only in the collection, 26 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 13060000 (8002398 pages from the captures) [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 560000 so far (x1 0x11789aa000 x2 0x1 x3 0x11789aa000 x4 0x4040) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 13080000 (8015201 pages from the captures) [hos ModuleSystemWorker2] nvn: CommandBufferSignalEvent #235000: event 0x100d6c8004 mode 0 value 234995 (lr 0x80291504) [hos host] vk: replay reads of guest memory as submitted: 13100000 (8028049 pages from the captures) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2899 dirty pages from their submits' captures, 46766 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 810 dispatches, 37600 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6019 runs, 613.7 MB compared; 73 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17242 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: program picks: 18539114 kept (7728820 by the last picks' textures, a uniform buffer bound since), 26406 new, 6834479 with other textures (of 25400000) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 530 ranges, 2866 MB), 244 pages written, 244 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 24 kept (106 MB), 12 spare buffers, 3824 texture pages copied (14 MB) [hos host] vk: replay reads of guest memory as submitted: 13120000 (8040713 pages from the captures) [hos BaseProcExecutor1] vcores: 149385 takes, 21190 waits (mean 65 us, longest 6.8 ms, 0 over HOS_VCORE_LOG_MS), 10377 kicks, 10549 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.11, present 0.14, draw 7.20, descriptors 2.97, texture lookups 0.08, uploads 0.00, pipelines 0.31, programs 1.26, commands 2.07, dispatch 0.81, memory sync 0.37, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 465 (1101 MB): 5.54 MB synced in 1072.3 copies, 17.54 MB written back [hos host] vk: 17640 presents, 60.0 fps; a frame: 2306 draws (0 skipped), 130 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1236 textures (2358 MB) [hos host] vk: host time in GPU resource waits: 6.57 ms a frame (fences 2.55, timelines 4.02, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 694.3 full barriers recorded, 308.7 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.1, an op 1366.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.6 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 13140000 (8053310 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14061 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2026 dirty pages from their submits' captures, 50063 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 855 dispatches, 39741 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6352 runs, 648.6 MB compared; 82 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18105 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: late pages found at the replay (600 takes): 10 pages in 10 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 13160000 (8065933 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 13180000 (8078758 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 5750709 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1659 dirty pages from their submits' captures, 48171 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 826 dispatches, 38078 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6033 runs, 613.7 MB compared; 86 pages read into the shadows; 284 MB of shadows (52 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17512 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 13200000 (8091608 pages from the captures) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1079 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1519 kept their virtual core (nothing there to wait for), 2833 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #325000: event 0x100d6c802c mode 0 value 236582 (now 236563) (lr 0x8028c554) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.04 [hos host] vk: program picks: 18691331 kept (7797693 by the last picks' textures, a uniform buffer bound since), 26406 new, 6882262 with other textures (of 25600000) [hos host] vk: replay reads of guest memory as submitted: 13220000 (8104499 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.03 over 524 ranges, 2822 MB), 256 pages written, 256 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.030; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (108 MB), 10 spare buffers, 4527 texture pages copied (17 MB) [hos host] vk: replay reads of guest memory as submitted: 13240000 (8117929 pages from the captures) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2800 dirty pages from their submits' captures, 47604 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 816 dispatches, 38254 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6173 runs, 631.1 MB compared; 70 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17295 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 242 sync waits (822 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 13260000 (8130695 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 13280000 (8143166 pages from the captures) [hos host] vk: a submission's parts ended: 82794 where the game had ended its render pass, 17206 at 512 draws [hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.44, present 0.14, draw 7.34, descriptors 2.99, texture lookups 0.09, uploads 0.01, pipelines 0.32, programs 1.31, commands 2.12, dispatch 0.83, memory sync 0.48, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 465 (1101 MB): 5.50 MB synced in 1072.1 copies, 17.53 MB written back [hos host] vk: 17760 presents, 58.2 fps; a frame: 2303 draws (0 skipped), 130 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1238 textures (2365 MB) [hos host] vk: host time in GPU resource waits: 5.24 ms a frame (fences 2.35, timelines 2.89, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 692.9 full barriers recorded, 308.9 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.9; 0x8: 3.0; 0x9: 1.0; 0x10: 2.9; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1366.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.7 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.06 MB unswizzled on the CPU in 0.01 ms [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2568 dirty pages from their submits' captures, 46814 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 813 dispatches, 37512 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5994 runs, 613.1 MB compared; 74 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17166 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 13300000 (8155661 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 13320000 (8168140 pages from the captures) [hos host] vk: program picks: 18843726 kept (7866634 by the last picks' textures, a uniform buffer bound since), 26406 new, 6929867 with other textures (of 25800000) [hos host] vk: replay reads of guest memory as submitted: 13340000 (8180644 pages from the captures) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2514 dirty pages from their submits' captures, 48182 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 830 dispatches, 38627 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6122 runs, 629.7 MB compared; 73 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17703 in captures, 0 only in the collection, 24 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 523 ranges, 2804 MB), 240 pages written, 240 copied (0.9 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 24 kept (100 MB), 12 spare buffers, 3920 texture pages copied (15 MB) [hos host] vk: replay reads of guest memory as submitted: 13360000 (8193391 pages from the captures) [hos sead::MainThread] nvn: 2350000 recordings made, 71998 live copies (25.9 MB), 6246 texture copies; 2 command buffers' segments (5.0 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 23138 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 13380000 (8205889 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2993 dirty pages from their submits' captures, 46587 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 37459 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5963 runs, 612.4 MB compared; 71 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17165 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 13400000 (8218406 pages from the captures) [hos Presentation Thread] vi: 17839 frames presented (59.6 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 17, 15.5-17.8 240, 17.8-22 37, 22-30 4, 30-40 0, over 40 0 [hos host] vk: late pages found at the replay (600 takes): 19 pages in 19 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1085 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1586 kept their virtual core (nothing there to wait for), 3022 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 13420000 (8231137 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 13440000 (8243874 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 5817798 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 112, 1101 MB with memory of 1536 MB (434 MB spared); blocks given memory later 13741, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1245 binds (298.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 300 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 53 spares now, 1284 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2260 dirty pages from their submits' captures, 48449 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 830 dispatches, 38626 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5978 runs, 612.3 MB compared; 88 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17582 in captures, 1 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: program picks: 18996431 kept (7935813 by the last picks' textures, a uniform buffer bound since), 26406 new, 6977162 with other textures (of 26000000) [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #240000: event 0x100d6c8024 mode 0 value 240000 (lr 0x80665068) [hos host] vk: reaper: 8 images (6 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 4 (3 MB) [hos host] vk: a frame: 312 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.99, present 0.14, draw 7.29, descriptors 3.00, texture lookups 0.18, uploads 0.06, pipelines 0.30, programs 1.30, commands 2.09, dispatch 0.80, memory sync 0.39, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 465 (1102 MB): 5.49 MB synced in 1066.5 copies, 17.30 MB written back [hos host] vk: 17880 presents, 60.3 fps; a frame: 2287 draws (0 skipped), 125 dispatches, 193 passes, 134 barriers, 6 submits, 0.8 uploads (0.2 MB), 85 ub updates (58 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1325 textures (2431 MB) [hos host] vk: host time in GPU resource waits: 6.01 ms a frame (fences 2.37, timelines 3.64, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 687.2 full barriers recorded, 307.2 left out (nothing recorded since the one before), 19.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.1; 0x11: 26.7; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 15.9; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.6 from 29.8 batches at the game's barriers (48.6 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1355.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 25.8; 48: 7.9; [hos host] vk: GPU time 12.3 ms a frame (the longest submission 3.2 ms); 1.1 textures made a frame: colour targets 0.8 (0.7 MB), depth 0.0 (0.4 MB), ASTC or BC 0.3 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.12 MB unswizzled on the CPU in 0.02 ms [hos SceneLowPriority] vk: write-back of #16652 0x11636fd000: 652 KB the CPU wrote since the GPU kept [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 517 ranges, 3012 MB), 233 pages written, 233 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.017; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (118 MB), 11 spare buffers, 4476 texture pages copied (17 MB) [hos host] vk: 168 textures retired (1176 left, 2257 MB): budget 167, drawn over 1, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: recorder: 7041924 packets run in 10.0 s (busy 1.79 s, spun idle 0.14 s, slept 12963 times, 4922 of them under 200 us), 0 calls made directly (no room); 19264 batches over 599 frames (32.2 a frame, 366 packets a batch), 9899 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 7 times, 0.82 ms (0.0 times, 0.00 ms a frame); every thread 625 times, 68.40 ms in all, the longest 0.61 ms; vkWaitForFences 563/62.88 ms (max 0.61), the queue's lock 38/2.70 ms (max 0.27), vkDestroyImageView 9/1.25 ms (max 0.34), vmaDestroyImage 5/0.75 ms (max 0.57), vkDestroyBuffer 2/0.48 ms (max 0.47), vkQueueBindSparse 3/0.23 ms (max 0.20) [hos host] vk: recorder: of its busy 1.79 s, 0.43 s in parts' ends and submits (the longest 1.93 ms), 0.11 s of that blocked on the queue's lock (165 times, the longest 1.89 ms), so 0.32 s in the ends, submits and holds; 1.37 s running the other packets; 11756 packets a frame [hos host] vk: 8 textures retired (1168 left, 2225 MB): budget 8, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 13460000 (8257172 pages from the captures) [hos host] vk: texture #21381 0x10002094a00 240x135 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (11 so far) [hos host] vk: write-backs left out (600 submits): 463, 2147.9 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1093 dirty pages from their submits' captures, 48781 from earlier ones', 257 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 806 dispatches, 40984 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 2992 runs, 313.7 MB compared; 274 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 13949 in captures, 250 only in the collection, 49 only by the replay itself [hos host] vk: read live, by pool: 0x115c000000 (16 MB, captured, state 2): 133; 0x1163400000 (8 MB, captured, state 2): 107; 0x104e000000 (24 MB, captured, state 2): 14; 0x7ffe000000 (24 MB, captured, state 2): 1; 0x7f88000000 (12 MB, captured, state 2): 1; 0x1051800000 (4 MB, captured, state 2): 1; [hos sead::MainThread] nvn: CommandBufferWaitEvent #330000: event 0x100d6c8000 mode 0 value 240329 (now 240310) (lr 0x82c0301c) [hos host] vk: replay reads of guest memory as submitted: 13480000 (8269818 pages from the captures) [hos NavMeshCAStepDepWorker1] vcores: 141827 takes, 21600 waits (mean 61 us, longest 9.4 ms, 0 over HOS_VCORE_LOG_MS), 10584 kicks, 10743 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 13 textures retired (1172 left, 2225 MB): budget 0, drawn over 13, dead targets 0 [hos host] vk: texture #21337 0x10000000000 761x177 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 17945 (drawn at 17881) [hos host] vk: texture #21342 0x10000000000 725x220 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 17945 (drawn at 17881) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 18152 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: presenter thread: the copy's fence 2.60 ms, the present 0.82 ms a present [hos host] vk: replay reads of guest memory as submitted: 13500000 (8282634 pages from the captures) [hos host] vk: sparse mirrors: 113, 1104 MB with memory of 1548 MB (443 MB spared); blocks given memory later 13768, given back 16907; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1248 binds (298.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 353 kept as spares, 0 destroyed; 52 spares now, 1272 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 670, 3031.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2237 dirty pages from their submits' captures, 49063 from earlier ones', 122 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 818 dispatches, 40480 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 4255 runs, 452.3 MB compared; 92 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16175 in captures, 61 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x1163400000 (8 MB, captured, state 2): 102; 0x104e000000 (24 MB, captured, state 2): 15; 0x7ffe000000 (24 MB, captured, state 2): 1; 0x7fb7c00000 (12 MB, captured, state 2): 1; 0x7f88000000 (12 MB, captured, state 2): 1; 0x1159c00000 (36 MB, captured, state 2): 1; 0x1051800000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: 1 textures retired (1203 left, 2235 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 8 pages in 8 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.04 over 513 ranges, 3000 MB), 226 pages written, 226 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 24 kept (136 MB), 12 spare buffers, 7041 texture pages copied (27 MB) [hos BaseProcExecutor0] hid: game window not in front (foreground window: process 38376, class "TkTopLevel") [hos host] vk: replay reads of guest memory as submitted: 13520000 (8295361 pages from the captures) [hos host] vk: texture #21504 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (12 so far) [hos host] vk: replay reads of guest memory as submitted: 13540000 (8308090 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: memory: working set 15446 MB (peak 16021), private 21691 MB, staging 32 MB in 2 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 848 MB; used 3188 of 15579 MB; heap 1 (device): 160 blocks 4825 MB, 19872 allocations 3996 MB; used 6949 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 16 blocks 512 MB, 144 images 287 MB; 75 free ranges, the largest 29 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1053 images 1278 MB; 447 free ranges, the largest 32 MB [hos host] vk: render target images: 5916 reused from the pool, 4366 made new, 4220 destroyed from it, 1162 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1937 with the pool empty (0 of them with targets still in the dead queue), 340 with none in it a present yet, 46 with none of a kind they can use, 576 with all smaller (the largest 50% short on average), 588 with all bigger (over 1.5 times), 876 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4028 idle [hos host] vk: VRAM: 6949 MB used of 11474 MB allowed (829 MB of it free in the blocks); textures 2235 MB (budget 7158 MB), 1204 textures [hos host] vk: budget: VRAM 4825 of 11474 MB (6949 MB by the driver, 2123 MB unaccounted): textures 2235, mirrors 1103, blocks free 829, shadows 284, other buffers 229, other images 111, staging 32; RAM 3125 MB (the process's private memory 21691 MB): imports 2240, capture spares 744, captures 106, capture table 35, others 0 [hos host] vk: textures: decoded ASTC 1367 MB, rendered 537 MB (100 drawn in one frame, kept: 48482 KB), other 301 MB; used in the last 8 presents 1967 MB [hos host] vk: decoded ASTC used again: 437 textures (955 MB) copied from ones the game finalized, not decoded; 437 of those retired, 0 listed; 1956 MB of blocks hashed in 336 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 37 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 12853 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 60 (11073 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 496 waited for it, 3.59 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 57.9 ms; VRAM 6.8 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2315 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 170 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 7.43, present 0.16, draw 4.34, descriptors 1.84, texture lookups 0.21, uploads 0.09, pipelines 0.18, programs 0.73, commands 1.23, dispatch 0.47, memory sync 0.57, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 466 (1102 MB): 4.30 MB synced in 867.7 copies, 9.50 MB written back [hos host] vk: 18000 presents, 58.1 fps; a frame: 1347 draws (0 skipped), 67 dispatches, 166 passes, 111 barriers, 6 submits, 0.7 uploads (0.5 MB), 256 ub updates (122 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1204 textures (2235 MB) [hos host] vk: host time in GPU resource waits: 7.94 ms a frame (fences 4.14, timelines 3.80, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 470.6 full barriers recorded, 213.6 left out (nothing recorded since the one before), 43.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.0; 0x8: 1.6; 0x9: 0.5; 0x10: 2.9; 0x11: 15.0; 0x12: 69.1; 0x13: 1.6; 0x21: 3.1; 0x28: 8.7; 0x92: 3.5; 0x97: 0.5; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 24.9 from 16.2 batches at the game's barriers (26.5 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 104.3, an op 759.7, the end 4.7, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 1.1; 5: 14.1; 48: 4.3; [hos host] vk: GPU time 10.3 ms a frame (the longest submission 15.1 ms); 1.4 textures made a frame: colour targets 1.1 (1.8 MB), depth 0.2 (1.2 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.45 MB unswizzled on the CPU in 0.07 ms [hos host] vk: queues ordered by their events: 117168 submissions stopped at a wait, 117108 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 331436, not met 0; captures at a CPU event's wait 11549, 0 of them at a wait met on arrival [hos host] vk: the render thread (600 presents, 16.85 ms a frame): its jobs 15.10 ms a frame (submits 2.03, presents 4.51, others 8.55), idle 1.75; the game waited for room at 44 presents, 0.08 ms each (0.01 a frame; 13.61 jobs queued at a present); 13 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 3.75 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 499 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 2388 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.33 ms after its hand-over (25.35 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.08 + 0.15 (the window's copy is in the idle), after it 2.10 + 0.48; after a frame's last recording that reads or writes the game's memory come 1.14 recordings (0.00 ms of GPU time), after its last that reads it 1.14 (0.00 ms) [hos host] vk: frame ahead (600 presents): 9.62 parts a frame recorded in the signal's job, 70.4% of them took nothing that came with the signal; of the others, 2.85 a frame took pages of the signal's capture into a mirror (72.2 pages a frame) and 0.00 pages read as guest memory had them (0.3 pages a frame); a frame's first part took nothing in 22 of 499 frames, its first three in 22 [hos host] vk: frame ahead, what came with the signal: 499 signals with a capture, each of 94.6 pages in 13.5 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.94+0.20, 0.09+0.03, 0.29+0.10, 0.43+0.06, 0.02+0.01, 0.11+0.01, 0.11+0.01, 0.05+0.02, 0.05+0.02, 0.02+0.01, 0.01+0.00, 0.00+0.00 [hos host] vk: write-backs left out (600 submits): 851, 3776.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2727 dirty pages from their submits' captures, 46125 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 805 dispatches, 36925 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5814 runs, 595.1 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17159 in captures, 0 only in the collection, 26 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vi: real-time step on: ratio 1.002, game's frame rate 60 [hos host] vk: replay reads of guest memory as submitted: 13560000 (8320884 pages from the captures) [hos host] vk: render targets drawn over from another start, let go: 10833 so far (50877 MB), 9758 of them (45937 MB) between the passes [hos host] vk: program picks: 19144759 kept (8001344 by the last picks' textures, a uniform buffer bound since), 26406 new, 7028834 with other textures (of 26200000) [hos host] vk: replay reads of guest memory as submitted: 13580000 (8334398 pages from the captures) [hos sead::MainThread] nvn: GPU waits in 5 s: 235 sync waits (998 ms), 0 polls timed out (0 ms), 12 queue finishes (80 ms) [hos BaseProcExecutor5] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 13993789 lookups, 109195 missed (0.78 %) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 112, 1102 MB with memory of 1544 MB (441 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2861 dirty pages from their submits' captures, 46799 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 822 dispatches, 37725 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5951 runs, 612.4 MB compared; 70 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17222 in captures, 0 only in the collection, 24 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 13600000 (8347305 pages from the captures) [hos host] vk: late pages' fix (600 recordings with one): of 770 captures with late pages 770 fixed (75.1 pages, 75.1 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 522 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 21, of an earlier recording 501, not recorded yet 0, of this part 0; pages fixed from stamps 56772, from snapshots 1028; the shadows: 325831 pages synced from captures, 322688 made known by it, 3143 left unknown because the replay or the collection wrote them since [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 580000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040) [hos host] vk: replay reads of guest memory as submitted: 13620000 (8359943 pages from the captures) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1009 of 3000 waited for their write-back, 0.07 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1500 kept their virtual core (nothing there to wait for), 3108 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 513 ranges, 2796 MB), 237 pages written, 237 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (106 MB), 11 spare buffers, 7041 texture pages copied (27 MB) [hos host] vk: replay reads of guest memory as submitted: 13640000 (8372687 pages from the captures) [hos host] vk: 15 textures retired (1198 left, 2238 MB): budget 0, drawn over 15, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 5884792 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2919 dirty pages from their submits' captures, 47129 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 819 dispatches, 38045 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6115 runs, 629.8 MB compared; 71 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17153 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 13660000 (8385543 pages from the captures) [hos host] vk: 12 textures retired (1188 left, 2237 MB): budget 0, drawn over 12, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 13680000 (8398467 pages from the captures) [hos host] vk: program picks: 19296684 kept (8070234 by the last picks' textures, a uniform buffer bound since), 26406 new, 7076909 with other textures (of 26400000) [hos host] vk: sparse mirrors: 112, 1102 MB with memory of 1544 MB (441 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1936 dirty pages from their submits' captures, 48196 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 38131 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5957 runs, 612.5 MB compared; 77 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17518 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 13700000 (8411246 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 16263 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.93, present 0.17, draw 7.05, descriptors 2.89, texture lookups 0.09, uploads 0.01, pipelines 0.30, programs 1.26, commands 2.02, dispatch 0.79, memory sync 0.39, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 466 (1102 MB): 5.51 MB synced in 1071.6 copies, 17.50 MB written back [hos host] vk: 18120 presents, 61.0 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 192 passes, 134 barriers, 6 submits, 0.2 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1190 textures (2237 MB) [hos host] vk: host time in GPU resource waits: 5.80 ms a frame (fences 2.43, timelines 3.37, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 692.0 full barriers recorded, 308.5 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.3 ms a frame (the longest submission 4.1 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 13720000 (8424139 pages from the captures) [hos host] vk: texture #21526 0x10002199c00 120x67 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (13 so far) [hos host] vk: late pages found at the replay (600 takes): 21 pages in 21 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vi: 18137 frames presented (59.4 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 32, 15.5-17.8 227, 17.8-22 34, 22-30 4, 30-40 0, over 40 1 [hos host] vk: replay reads of guest memory as submitted: 13740000 (8437005 pages from the captures) [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #335000: event 0x100d6c8010 mode 0 value 244100 (now 244081) (lr 0x80660e38) [hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1068 dirty pages from their submits' captures, 49713 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 838 dispatches, 39079 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6289 runs, 647.2 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17280 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 514 ranges, 2800 MB), 252 pages written, 252 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (134 MB), 10 spare buffers, 7041 texture pages copied (27 MB) [hos host] vk: replay reads of guest memory as submitted: 13760000 (8449926 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 13780000 (8462782 pages from the captures) [hos host] vk: sparse mirrors: 112, 1102 MB with memory of 1544 MB (441 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1515 dirty pages from their submits' captures, 48948 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 839 dispatches, 38622 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6130 runs, 629.8 MB compared; 79 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17446 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: program picks: 19449006 kept (8139295 by the last picks' textures, a uniform buffer bound since), 26406 new, 7124587 with other textures (of 26600000) [hos host] vk: replay reads of guest memory as submitted: 13800000 (8475545 pages from the captures) [hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #245000: event 0x100d6c8028 mode 0 value 244999 (lr 0x80667410) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 13820000 (8488163 pages from the captures) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1084 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1421 kept their virtual core (nothing there to wait for), 2931 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 13840000 (8500756 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1426 dirty pages from their submits' captures, 48852 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 835 dispatches, 38548 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6135 runs, 629.9 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17383 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos ModuleSystemWorker2] nvn: 2400000 recordings made, 72061 live copies (24.8 MB), 6301 texture copies; 3 command buffers' segments (4.2 MB) [hos host] vk: replay reads of guest memory as submitted: 13860000 (8513504 pages from the captures) [hos BaseProcExecutor1] vcores: 144698 takes, 22135 waits (mean 63 us, longest 11.9 ms, 0 over HOS_VCORE_LOG_MS), 10868 kicks, 10973 yields, 0 starved grants, 0 escapes in 5 s [hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 515 ranges, 2806 MB), 242 pages written, 242 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (94 MB), 11 spare buffers, 7041 texture pages copied (27 MB) [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.77, present 0.13, draw 7.07, descriptors 2.90, texture lookups 0.08, uploads 0.00, pipelines 0.30, programs 1.25, commands 2.05, dispatch 0.78, memory sync 0.42, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 466 (1102 MB): 5.47 MB synced in 1070.2 copies, 17.50 MB written back [hos host] vk: 18240 presents, 59.9 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 193 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1193 textures (2238 MB) [hos host] vk: host time in GPU resource waits: 4.06 ms a frame (fences 2.70, timelines 1.36, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 690.7 full barriers recorded, 308.9 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.4, an op 1367.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.3 ms a frame (the longest submission 3.7 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 13880000 (8526372 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 5951888 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 112, 1102 MB with memory of 1544 MB (441 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2388 dirty pages from their submits' captures, 47113 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 811 dispatches, 37563 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5953 runs, 612.3 MB compared; 72 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17160 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10240 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 13900000 (8539216 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 19601708 kept (8208437 by the last picks' textures, a uniform buffer bound since), 26406 new, 7171885 with other textures (of 26800000) [hos host] vk: replay reads of guest memory as submitted: 13920000 (8552086 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 13940000 (8565010 pages from the captures) [hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1111 dirty pages from their submits' captures, 50479 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 857 dispatches, 39652 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6291 runs, 647.2 MB compared; 91 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17761 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 13960000 (8577879 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 13980000 (8590511 pages from the captures) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1251 binds (298.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 354 kept as spares, 0 destroyed; 53 spares now, 1276 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2417 dirty pages from their submits' captures, 48518 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 833 dispatches, 38700 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6118 runs, 629.9 MB compared; 76 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17519 in captures, 0 only in the collection, 29 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.00, scan 0.03 over 515 ranges, 2812 MB), 250 pages written, 250 copied (1.0 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 199 sync waits (590 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 24 kept (106 MB), 11 spare buffers, 7041 texture pages copied (27 MB) [hos host] vk: replay reads of guest memory as submitted: 14000000 (8603763 pages from the captures) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #340000: event 0x100d6c8010 mode 0 value 247900 (now 247881) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 14020000 (8616229 pages from the captures) [hos host] vk: program picks: 19754177 kept (8277363 by the last picks' textures, a uniform buffer bound since), 26406 new, 7219416 with other textures (of 27000000) [hos BaseProcExecutor4] hid: game window in front (foreground window: process 32080, class "Longshot") [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2895 dirty pages from their submits' captures, 46693 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 813 dispatches, 37522 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5974 runs, 612.5 MB compared; 71 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17221 in captures, 0 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 14040000 (8628769 pages from the captures) [hos host] vk: a frame: 315 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.08, present 0.14, draw 7.35, descriptors 3.06, texture lookups 0.10, uploads 0.01, pipelines 0.31, programs 1.30, commands 2.09, dispatch 0.78, memory sync 0.40, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 466 (1100 MB): 5.51 MB synced in 1070.9 copies, 17.50 MB written back [hos host] vk: 18360 presents, 58.4 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 192 passes, 134 barriers, 6 submits, 0.2 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1193 textures (2238 MB) [hos host] vk: host time in GPU resource waits: 6.05 ms a frame (fences 3.01, timelines 3.04, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 691.1 full barriers recorded, 308.2 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 13.0 ms a frame (the longest submission 14.8 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: queues ordered by their events: 120000 submissions stopped at a wait, 119938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 340545, not met 0; captures at a CPU event's wait 11828, 0 of them at a wait met on arrival [hos ModuleSystemWorker5] vk: readbacks of GPU results: 1087 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1663 kept their virtual core (nothing there to wait for), 3201 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 14060000 (8641481 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 14080000 (8654310 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1252 binds (298.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2901 dirty pages from their submits' captures, 46680 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 818 dispatches, 37545 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5952 runs, 612.3 MB compared; 73 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17669 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: replay reads of guest memory as submitted: 14100000 (8666984 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.03 over 516 ranges, 2816 MB), 236 pages written, 236 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 24 kept (96 MB), 12 spare buffers, 7041 texture pages copied (27 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13134 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 14120000 (8679761 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 6019375 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3025 dirty pages from their submits' captures, 46751 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 816 dispatches, 37625 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5944 runs, 612.3 MB compared; 70 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17248 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 14140000 (8692492 pages from the captures) [hos Presentation Thread] vi: 18436 frames presented (59.7 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 18, 15.5-17.8 248, 17.8-22 32, 22-30 0, 30-40 1, over 40 0 [hos host] vk: program picks: 19906557 kept (8346346 by the last picks' textures, a uniform buffer bound since), 26406 new, 7267036 with other textures (of 27200000) [hos host] vk: late pages found at the replay (600 takes): 17 pages in 17 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos rmbl::VibrationThread] hid: 61440 vibration values, strongest 0.00 since the last count [hos host] vk: replay reads of guest memory as submitted: 14160000 (8705180 pages from the captures) [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 600000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040) [hos sead::MainThread] nvn: CommandBufferSignalEvent #250000: event 0x100d6c8034 mode 0 value 249999 (lr 0x806616b0) [hos host] vk: replay reads of guest memory as submitted: 14180000 (8717932 pages from the captures) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1252 binds (298.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2175 dirty pages from their submits' captures, 49177 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 845 dispatches, 39023 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6118 runs, 629.6 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17470 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 14200000 (8730675 pages from the captures) [hos host] vk: a frame: 314 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.69, present 0.14, draw 6.95, descriptors 2.86, texture lookups 0.07, uploads 0.00, pipelines 0.29, programs 1.24, commands 1.99, dispatch 0.76, memory sync 0.39, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 466 (1100 MB): 5.52 MB synced in 1070.2 copies, 17.50 MB written back [hos host] vk: 18480 presents, 60.9 fps; a frame: 2303 draws (0 skipped), 128 dispatches, 192 passes, 134 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1196 textures (2240 MB) [hos host] vk: host time in GPU resource waits: 5.53 ms a frame (fences 2.46, timelines 3.07, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 691.2 full barriers recorded, 308.3 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 124.9, an op 1366.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.2 ms); 0.0 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 14220000 (8743268 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: recorder: 6913407 packets run in 10.0 s (busy 1.75 s, spun idle 0.14 s, slept 12749 times, 4711 of them under 200 us), 0 calls made directly (no room); 18978 batches over 595 frames (31.9 a frame, 364 packets a batch), 9761 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 9 times, 1.16 ms (0.0 times, 0.00 ms a frame); every thread 683 times, 73.77 ms in all, the longest 0.49 ms; vkWaitForFences 587/65.92 ms (max 0.49), vkDestroyImageView 44/4.40 ms (max 0.27), the queue's lock 32/1.99 ms (max 0.18), vmaFreeMemory 4/1.06 ms (max 0.42), vmaDestroyImage 7/0.16 ms (max 0.04), a part's submit 2/0.14 ms (max 0.10) [hos host] vk: recorder: of its busy 1.75 s, 0.43 s in parts' ends and submits (the longest 1.81 ms), 0.11 s of that blocked on the queue's lock (173 times, the longest 1.75 ms), so 0.32 s in the ends, submits and holds; 1.33 s running the other packets; 11619 packets a frame [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 952 dirty pages from their submits' captures, 48494 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 815 dispatches, 38063 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6124 runs, 630.0 MB compared; 83 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16911 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 517 ranges, 2818 MB), 249 pages written, 249 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 24 kept (92 MB), 12 spare buffers, 7212 texture pages copied (28 MB) [hos host] vk: replay reads of guest memory as submitted: 14240000 (8755799 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 14260000 (8768319 pages from the captures) [hos ModuleSystemWorker5] vk: readbacks of GPU results: 1083 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1473 kept their virtual core (nothing there to wait for), 2879 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: program picks: 20058700 kept (8415484 by the last picks' textures, a uniform buffer bound since), 26406 new, 7314893 with other textures (of 27400000) [hos aal::LowPrioWorkerThread] vcores: 148794 takes, 21046 waits (mean 66 us, longest 7.8 ms, 0 over HOS_VCORE_LOG_MS), 10232 kicks, 10382 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 14280000 (8780915 pages from the captures) [hos host] vk: sparse mirrors: 111, 1100 MB with memory of 1532 MB (431 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1252 binds (298.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 301 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 54 spares now, 1288 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1719 dirty pages from their submits' captures, 48871 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 844 dispatches, 38583 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6213 runs, 637.5 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17427 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 193 live (2240 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #345000: event 0x100d6c8010 mode 0 value 251700 (now 251681) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 14300000 (8794070 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14392 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: presenter thread: the copy's fence 2.71 ms, the present 0.86 ms a present [hos host] vk: replay reads of guest memory as submitted: 14320000 (8806517 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 367 dirty pages from their submits' captures, 51864 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 39997 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6351 runs, 650.1 MB compared; 85 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17954 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 14340000 (8819164 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 543 ranges, 2908 MB), 261 pages written, 261 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 7212 texture pages copied (28 MB) [hos host] vk: slow upload 22.2 ms: #21660 0x110aed5200 1024x1024x1 x1 levels 11 NVN 0x8d ASTC (BC3), 475 of 475 KB [hos host] vk: replay reads of guest memory as submitted: 14360000 (8832803 pages from the captures) [hos host] vk: memory: working set 15304 MB (peak 16021), private 21450 MB, staging 32 MB in 2 arenas, 42 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 12 blocks 912 MB, 36 allocations 840 MB; used 3176 of 15579 MB; heap 1 (device): 160 blocks 4825 MB, 20657 allocations 4152 MB; used 6949 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 16 blocks 512 MB, 133 images 284 MB; 69 free ranges, the largest 29 MB [hos host] vk: image blocks (textures, compacted): 52 blocks 1664 MB, 1168 images 1394 MB; 436 free ranges, the largest 32 MB [hos host] vk: render target images: 5916 reused from the pool, 4366 made new, 4231 destroyed from it, 1162 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 151 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 1937 with the pool empty (0 of them with targets still in the dead queue), 340 with none in it a present yet, 46 with none of a kind they can use, 576 with all smaller (the largest 50% short on average), 588 with all bigger (over 1.5 times), 876 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4039 idle [hos host] vk: VRAM: 6949 MB used of 11474 MB allowed (672 MB of it free in the blocks); textures 2349 MB (budget 7158 MB), 1308 textures [hos host] vk: budget: VRAM 4825 of 11474 MB (6949 MB by the driver, 2123 MB unaccounted): textures 2349, mirrors 1146, blocks free 672, shadows 284, other buffers 229, other images 111, staging 32; RAM 3119 MB (the process's private memory 21450 MB): imports 2244, capture spares 704, captures 86, other buffers 50, capture table 35, others 0 [hos host] vk: textures: decoded ASTC 1471 MB, rendered 533 MB (98 drawn in one frame, kept: 48302 KB), other 313 MB; used in the last 8 presents 2090 MB [hos host] vk: decoded ASTC used again: 466 textures (1013 MB) copied from ones the game finalized, not decoded; 466 of those retired, 1 listed; 1989 MB of blocks hashed in 341 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 34 MB: 0x7fc7773a00 6133x1467x1 x1 levels 1 NVN 0x87 ASTC, last used 637 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 13453 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 60 (11073 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 437 waited for it, 3.11 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 47.6 ms; VRAM 6.8 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6475 queued, 6475 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16357 ms and 348.1 MB read in all); 8780 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12294 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2320 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9631 (6475 ready, 0 loading), 17752 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13978 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 306 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.88, present 0.14, draw 7.06, descriptors 2.94, texture lookups 0.32, uploads 0.23, pipelines 0.29, programs 1.19, commands 2.10, dispatch 0.79, memory sync 0.45, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 476 (1146 MB): 5.91 MB synced in 1076.6 copies, 17.62 MB written back [hos host] vk: 18600 presents, 58.2 fps; a frame: 2262 draws (0 skipped), 133 dispatches, 188 passes, 134 barriers, 6 submits, 0.9 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1308 textures (2349 MB) [hos host] vk: host time in GPU resource waits: 4.02 ms a frame (fences 2.75, timelines 1.27, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 684.9 full barriers recorded, 303.7 left out (nothing recorded since the one before), 18.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1323.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.2 ms a frame (the longest submission 4.0 ms); 0.9 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.9 (0.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.07 MB unswizzled on the CPU in 0.01 ms [hos PreActorModelCompose] nvn: 2450000 recordings made, 73192 live copies (27.2 MB), 6301 texture copies; 2 command buffers' segments (4.1 MB) [hos host] vk: the render thread (600 presents, 16.76 ms a frame): its jobs 15.46 ms a frame (submits 3.43, presents 4.37, others 7.66), idle 1.30; the game waited for room at 18 presents, 0.15 ms each (0.00 a frame; 7.48 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 443 captures at a CPU signal; the submits' waits read in 0.044 ms a frame; readbacks: 2172 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 14.66 ms after its hand-over (41.13 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.42 + 0.13 (the window's copy is in the idle), after it 2.18 + 0.57; after a frame's last recording that reads or writes the game's memory come 1.18 recordings (0.00 ms of GPU time), after its last that reads it 1.18 (0.00 ms) [hos host] vk: frame ahead (600 presents): 9.85 parts a frame recorded in the signal's job, 71.0% of them took nothing that came with the signal; of the others, 2.86 a frame took pages of the signal's capture into a mirror (75.3 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 22 of 443 frames, its first three in 22 [hos host] vk: frame ahead, what came with the signal: 443 signals with a capture, each of 94.0 pages in 13.7 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.00+0.17, 0.11+0.03, 0.30+0.12, 0.41+0.06, 0.02+0.02, 0.09+0.03, 0.10+0.05, 0.06+0.06, 0.05+0.02, 0.02+0.01, 0.01+0.00, 0.00+0.00 [hos host] vk: readback pages refreshed from guest memory: 6086288 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 121, 1146 MB with memory of 1612 MB (465 MB spared); blocks given memory later 13768, given back 16934; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1262 binds (301.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 311 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 44 spares now, 1208 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 874, 3883.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2587 dirty pages from their submits' captures, 48438 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 38348 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6329 runs, 617.9 MB compared; 81 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17203 in captures, 0 only in the collection, 31 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: render targets drawn over from another start, let go: 10833 so far (50877 MB), 9758 of them (45937 MB) between the passes [hos host] vk: replay reads of guest memory as submitted: 14380000 (8847635 pages from the captures) [hos host] vk: program picks: 20210916 kept (8484766 by the last picks' textures, a uniform buffer bound since), 26416 new, 7362667 with other textures (of 27600000) [hos sead::MainThread] nvn: GPU waits in 5 s: 217 sync waits (741 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 14400000 (8860960 pages from the captures) [hos BaseProcExecutor1] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 14814359 lookups, 120760 missed (0.82 %) [hos host] vk: 88 textures retired (1232 left, 2288 MB): budget 88, drawn over 0, dead targets 0 [hos host] vk: reaper: 383 images (754 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 182 (244 MB) [hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1217 dirty pages from their submits' captures, 51041 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 891 dispatches, 39873 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6706 runs, 637.1 MB compared; 74 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17868 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 14420000 (8874338 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 14440000 (8887651 pages from the captures) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1114 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1635 kept their virtual core (nothing there to wait for), 2973 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: sparse mirrors: 121, 1145 MB with memory of 1612 MB (466 MB spared); blocks given memory later 13768, given back 16943; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1263 binds (301.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 311 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 44 spares now, 1208 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 500 dirty pages from their submits' captures, 57684 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 988 dispatches, 44350 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7196 runs, 705.2 MB compared; 76 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20105 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9583 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 556 ranges, 3139 MB), 262 pages written, 262 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (92 MB), 11 spare buffers, 6754 texture pages copied (26 MB) [hos host] vk: replay reads of guest memory as submitted: 14460000 (8901294 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: late pages found at the replay (600 takes): 2 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 20364946 kept (8557758 by the last picks' textures, a uniform buffer bound since), 26418 new, 7408635 with other textures (of 27800000) [hos host] vk: replay reads of guest memory as submitted: 14480000 (8913741 pages from the captures) [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #255000: event 0x100d6c8040 mode 0 value 254999 (lr 0x806621e8) [hos host] vk: a frame: 230 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.55, present 0.19, draw 6.52, descriptors 2.71, texture lookups 0.10, uploads 0.02, pipelines 0.28, programs 1.12, commands 1.88, dispatch 0.85, memory sync 0.55, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 476 (1146 MB): 5.63 MB synced in 1097.0 copies, 17.65 MB written back [hos host] vk: 18720 presents, 60.6 fps; a frame: 2046 draws (0 skipped), 138 dispatches, 167 passes, 135 barriers, 6 submits, 0.3 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14161 pipelines (3727 ms), 0 programs (10 failed), 1253 textures (2297 MB) [hos host] vk: host time in GPU resource waits: 2.91 ms a frame (fences 1.35, timelines 1.56, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 586.1 full barriers recorded, 280.7 left out (nothing recorded since the one before), 21.2 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.6, an op 1094.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.3 ms a frame (the longest submission 4.8 ms); 0.3 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.00 ms [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1405 dirty pages from their submits' captures, 56384 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 987 dispatches, 44360 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7358 runs, 722.6 MB compared; 86 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19667 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 14500000 (8927103 pages from the captures) [hos Presentation Thread] vi: 18734 frames presented (59.5 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 29, 15.5-17.8 240, 17.8-22 26, 22-30 2, 30-40 0, over 40 1 [hos sead::MainThread] nvn: CommandBufferWaitEvent #350000: event 0x100d6c800c mode 0 value 255499 (now 255480) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 14520000 (8940999 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 14540000 (8953888 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 123, 1151 MB with memory of 1624 MB (472 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1267 binds (301.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 313 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 42 spares now, 1196 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1590 dirty pages from their submits' captures, 55848 from earlier ones', 10 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 965 dispatches, 43802 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7224 runs, 705.9 MB compared; 64 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19706 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x1042000000 (4 MB, captured, state 2): 6; 0x1041c00000 (4 MB, captured, state 2): 3; 0x7fb1800000 (20 MB, captured, state 2): 1; [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 561 ranges, 3146 MB), 264 pages written, 264 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.035; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.019, bytes copied 0.000; 24 kept (98 MB), 11 spare buffers, 7453 texture pages copied (29 MB) [hos host] vk: replay reads of guest memory as submitted: 14560000 (8967351 pages from the captures) [hos host] vk: slow upload 16.2 ms: #21720 0x10ffefbe00 512x512x1 x1 levels 10 NVN 0x8e ASTC (BC3), 88 of 88 KB [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.07 [hos host] vk: texture #21786 0x1000a473400 1280x720 NVN 0x29 evicted, made again over storage render targets took: uploaded from the game's memory (14 so far) [hos host] vk: texture #21787 0x10001fe0800 480x270 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (15 so far) [hos host] vk: 1 textures retired (1310 left, 2423 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: readback pages refreshed from guest memory: 6153837 pages so far (0 held: may be written ahead in their submission) [hos host] vk: program picks: 20517721 kept (8627562 by the last picks' textures, a uniform buffer bound since), 26426 new, 7455852 with other textures (of 28000000) [hos host] vk: replay reads of guest memory as submitted: 14580000 (8981492 pages from the captures) [hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1861 dirty pages from their submits' captures, 53302 from earlier ones', 39 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 946 dispatches, 41958 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7198 runs, 674.6 MB compared; 100 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18810 in captures, 2 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fb4400000 (24 MB, captured, state 2): 39; [hos host] vk: 2 textures retired (1320 left, 2438 MB): budget 0, drawn over 0, dead targets 2 [hos aal::LowPrioWorkerThread] vcores: 152955 takes, 26089 waits (mean 76 us, longest 39.0 ms, 0 over HOS_VCORE_LOG_MS), 12520 kicks, 12679 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 14600000 (8993870 pages from the captures) [hos host] vk: 6 textures retired (1401 left, 2573 MB): budget 0, drawn over 0, dead targets 6 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 17044 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a frame: 220 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.03, present 0.21, draw 7.76, descriptors 3.62, texture lookups 0.87, uploads 0.54, pipelines 0.31, programs 1.15, commands 2.12, dispatch 0.96, memory sync 0.57, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 489 (1180 MB): 5.93 MB synced in 1097.4 copies, 17.71 MB written back [hos host] vk: 18840 presents, 55.1 fps; a frame: 1958 draws (0 skipped), 141 dispatches, 171 passes, 134 barriers, 6 submits, 2.0 uploads (1.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1467 textures (2679 MB) [hos host] vk: host time in GPU resource waits: 5.73 ms a frame (fences 1.75, timelines 3.99, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 638.9 full barriers recorded, 285.7 left out (nothing recorded since the one before), 18.2 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.0, an op 1063.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.5 ms a frame (the longest submission 4.5 ms); 5.2 textures made a frame: colour targets 2.7 (9.0 MB), depth 0.8 (4.6 MB), ASTC or BC 1.7 (2.8 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.44 MB unswizzled on the CPU in 0.07 ms [hos host] vk: sparse mirrors: 136, 1182 MB with memory of 1692 MB (509 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1280 binds (304.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 326 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 29 spares now, 1128 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1964 dirty pages from their submits' captures, 51658 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 923 dispatches, 40701 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6996 runs, 656.9 MB compared; 72 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17951 in captures, 1 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x7fe1400000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1100 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 13 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1446 kept their virtual core (nothing there to wait for), 3162 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 14620000 (9009288 pages from the captures) [hos host] vk: texture 0x110750be00 171x171x1 x1 levels 8 NVN 0x8d -> vk 37 (ASTC decoded) (1500 textures, 2714 MB) [hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 620000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040) [hos host] vk: 77 textures retired (1453 left, 2667 MB): budget 77, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 14640000 (9024977 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 125 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 590 ranges, 3384 MB), 256 pages written, 256 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (134 MB), 8 spare buffers, 7453 texture pages copied (29 MB) [hos host] vk: late pages' fix (600 recordings with one): of 818 captures with late pages 818 fixed (78.6 pages, 78.9 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 505 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 1 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 95, of an earlier recording 409, not recorded yet 0, of this part 0; pages fixed from stamps 63202, from snapshots 1084; the shadows: 407301 pages synced from captures, 398333 made known by it, 8968 left unknown because the replay or the collection wrote them since [hos host] vk: 8 textures retired (1449 left, 2665 MB): budget 8, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 14660000 (9040279 pages from the captures) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2285 dirty pages from their submits' captures, 50221 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 879 dispatches, 39571 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6478 runs, 637.0 MB compared; 73 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17946 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: 0x7fdfc00000 (12 MB, captured, state 2): 1; [hos host] vk: 1 textures retired (1451 left, 2669 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 14680000 (9055384 pages from the captures) [hos host] vk: texture 0x10ca831e00 1024x1024x1 x1 levels 11 NVN 0x8d -> vk 137 (ASTC decoded) (1500 textures, 2749 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: 5 textures retired (1590 left, 2911 MB): budget 4, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 14700000 (9069322 pages from the captures) [hos host] vk: program picks: 20668408 kept (8690778 by the last picks' textures, a uniform buffer bound since), 26433 new, 7505158 with other textures (of 28200000) [hos sead::MainThread] nvn: GPU waits in 5 s: 221 sync waits (895 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: 21 textures retired (1569 left, 2873 MB): budget 21, drawn over 0, dead targets 0 [hos host] vk: sparse mirrors: 138, 1186 MB with memory of 1700 MB (514 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1282 binds (304.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 328 taken from the spares, 165 made new; 355 kept as spares, 0 destroyed; 27 spares now, 1120 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 926, 4109.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2004 dirty pages from their submits' captures, 51629 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 905 dispatches, 40651 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6831 runs, 656.5 MB compared; 64 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18237 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 14720000 (9081733 pages from the captures) [hos host] vk: 40 textures retired (1530 left, 2789 MB): budget 40, drawn over 0, dead targets 0 [hos host] vk: 4 textures retired (1530 left, 2792 MB): budget 4, drawn over 0, dead targets 0 [hos ModuleSystemWorker5] nvn: 2500000 recordings made, 75229 live copies (26.1 MB), 6322 texture copies; 6 command buffers' segments (4.4 MB) [hos host] vk: replay reads of guest memory as submitted: 14740000 (9096421 pages from the captures) [hos host] vk: 6 textures retired (1524 left, 2765 MB): budget 6, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: CommandBufferWaitEvent #355000: event 0x100d6c800c mode 0 value 259299 (now 259280) (lr 0x80660e38) [hos host] vk: texture #21098 0x117103b000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 18945 (drawn at 16019) [hos host] vk: replay reads of guest memory as submitted: 14760000 (9111586 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 596 ranges, 3510 MB), 262 pages written, 262 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.033; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (90 MB), 10 spare buffers, 7873 texture pages copied (30 MB) [hos host] vk: 6 textures retired (1522 left, 2761 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2416 dirty pages from their submits' captures, 50022 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 876 dispatches, 39568 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6402 runs, 641.8 MB compared; 65 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17823 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 14780000 (9126981 pages from the captures) [hos host] vk: a frame: 288 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.27, present 0.22, draw 8.04, descriptors 3.79, texture lookups 0.80, uploads 0.48, pipelines 0.32, programs 1.26, commands 2.09, dispatch 0.89, memory sync 0.50, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 493 (1186 MB): 5.74 MB synced in 1097.1 copies, 17.74 MB written back [hos host] vk: 18960 presents, 55.4 fps; a frame: 2018 draws (0 skipped), 140 dispatches, 200 passes, 133 barriers, 6 submits, 2.0 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1533 textures (2769 MB) [hos host] vk: host time in GPU resource waits: 6.04 ms a frame (fences 1.85, timelines 4.18, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 755.2 full barriers recorded, 315.2 left out (nothing recorded since the one before), 16.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 25.9; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 124.1, an op 1252.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.2 ms a frame (the longest submission 4.8 ms); 2.2 textures made a frame: colour targets 0.2 (0.9 MB), depth 0.1 (0.4 MB), ASTC or BC 1.9 (3.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.13 MB unswizzled on the CPU in 0.16 ms [hos host] vk: 45 textures retired (1488 left, 2698 MB): budget 45, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 14800000 (9142325 pages from the captures) [hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #260000: event 0x100d6c8000 mode 0 value 259999 (lr 0x80291440) [hos host] vk: 40 textures retired (1449 left, 2656 MB): budget 40, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 6220788 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 135, 1173 MB with memory of 1680 MB (506 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1283 binds (304.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 328 taken from the spares, 165 made new; 358 kept as spares, 0 destroyed; 30 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 925, 4107.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2076 dirty pages from their submits' captures, 51286 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 40331 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6284 runs, 640.7 MB compared; 59 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18126 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: 6 textures retired (1443 left, 2655 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 14820000 (9157140 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.04 [hos host] vk: program picks: 20816396 kept (8749198 by the last picks' textures, a uniform buffer bound since), 26433 new, 7557170 with other textures (of 28400000) [hos host] vk: texture #21785 0x10005280e00 2560x1440x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21748 0x10005191c00 1600x1600x1 x3 NVN 0x32 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21755 0x10004741200 2560x1440x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21756 0x10004bf1600 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21757 0x10004ce1800 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21758 0x10004b01400 1280x720x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 184 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 39 textures retired (1404 left, 2589 MB): budget 39, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 6 finalized, 34 MB [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1100 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1471 kept their virtual core (nothing there to wait for), 2881 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 14840000 (9172227 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21715 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 66 textures retired (1348 left, 2531 MB): budget 66, drawn over 0, dead targets 0 [hos Presentation Thread] vi: 19015 frames presented (56.2 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 51, 15.5-17.8 152, 17.8-22 52, 22-30 17, 30-40 8, over 40 1 [hos host] vk: replay reads of guest memory as submitted: 14860000 (9187697 pages from the captures) [hos host] vk: 46 textures retired (1293 left, 2437 MB): budget 45, drawn over 0, dead targets 1 [hos host] vk: texture #22570 0x10008ca2e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19020 (drawn at 18959) [hos host] vk: write-backs left out (600 submits): 876, 3887.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2117 dirty pages from their submits' captures, 48220 from earlier ones', 1 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 851 dispatches, 38043 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6287 runs, 637.3 MB compared; 71 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17257 in captures, 0 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 1; [hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 5 textures retired (1346 left, 2512 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 14880000 (9202761 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 600 ranges, 3538 MB), 253 pages written, 253 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (148 MB), 9 spare buffers, 7873 texture pages copied (30 MB) [hos host] vk: 1 textures retired (1422 left, 2580 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: slow upload 15.8 ms: #22837 0x10bfa91e00 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB [hos host] vk: replay reads of guest memory as submitted: 14900000 (9217828 pages from the captures) [hos host] vk: 1 textures retired (1488 left, 2751 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: texture 0x10f24ecc00 128x128x1 x1 levels 8 NVN 0x49 -> vk 139 (1500 textures, 2763 MB) [hos host] vk: sparse mirrors: 135, 1173 MB with memory of 1680 MB (506 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1283 binds (304.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 328 taken from the spares, 165 made new; 358 kept as spares, 0 destroyed; 30 spares now, 1140 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2948 dirty pages from their submits' captures, 48467 from earlier ones', 6 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 916 dispatches, 38505 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6433 runs, 621.3 MB compared; 67 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17729 in captures, 0 only in the collection, 25 only by the replay itself [hos host] vk: read live, by pool: 0x7fb4400000 (24 MB, captured, state 2): 2; 0x7fdfc00000 (12 MB, captured, state 2): 1; 0x7fb7c00000 (12 MB, captured, state 2): 1; 0x7f90800000 (4 MB, captured, state 2): 1; 0x1044000000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: 9 textures retired (1520 left, 2795 MB): budget 1, drawn over 0, dead targets 8 [hos host] vk: recorder: 6319602 packets run in 10.0 s (busy 1.84 s, spun idle 0.13 s, slept 11675 times, 4042 of them under 200 us), 0 calls made directly (no room); 18017 batches over 567 frames (31.8 a frame, 351 packets a batch), 9199 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 54 times, 6.73 ms (0.1 times, 0.01 ms a frame); every thread 672 times, 57.97 ms in all, the longest 0.58 ms; vkWaitForFences 497/42.15 ms (max 0.47), vkDestroyImageView 68/5.36 ms (max 0.29), vkQueueBindSparse 32/3.34 ms (max 0.29), the queue's lock 34/3.07 ms (max 0.28), vkEndCommandBuffer 12/1.88 ms (max 0.28), vmaFreeMemory 4/1.08 ms (max 0.58) [hos host] vk: recorder: of its busy 1.84 s, 0.52 s in parts' ends and submits (the longest 1.80 ms), 0.20 s of that blocked on the queue's lock (350 times, the longest 1.75 ms), so 0.32 s in the ends, submits and holds; 1.33 s running the other packets; 11146 packets a frame [hos host] vk: replay reads of guest memory as submitted: 14920000 (9231034 pages from the captures) [hos host] vk: 5 textures retired (1581 left, 2936 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: 3 textures retired (1606 left, 2964 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: a frame: 256 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.98, present 0.19, draw 8.24, descriptors 4.26, texture lookups 1.25, uploads 0.91, pipelines 0.31, programs 1.13, commands 1.99, dispatch 0.83, memory sync 0.44, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 493 (1172 MB): 5.64 MB synced in 1090.9 copies, 17.74 MB written back [hos host] vk: 19080 presents, 54.7 fps; a frame: 2084 draws (0 skipped), 136 dispatches, 199 passes, 134 barriers, 6 submits, 3.0 uploads (1.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1606 textures (2964 MB) [hos host] vk: host time in GPU resource waits: 6.99 ms a frame (fences 2.66, timelines 4.32, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 759.9 full barriers recorded, 315.9 left out (nothing recorded since the one before), 18.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.3, an op 1326.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.8 ms a frame (the longest submission 4.0 ms); 3.6 textures made a frame: colour targets 0.5 (2.8 MB), depth 0.3 (1.5 MB), ASTC or BC 2.9 (5.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.47 MB unswizzled on the CPU in 0.08 ms [hos host] vk: replay reads of guest memory as submitted: 14940000 (9243939 pages from the captures) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2348 dirty pages from their submits' captures, 48868 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 888 dispatches, 39231 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6508 runs, 634.6 MB compared; 64 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17218 in captures, 0 only in the collection, 32 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: program picks: 20967720 kept (8815453 by the last picks' textures, a uniform buffer bound since), 26433 new, 7605846 with other textures (of 28600000) [hos tera::FileResourceMgr] vcores: 129856 takes, 20930 waits (mean 83 us, longest 10.4 ms, 0 over HOS_VCORE_LOG_MS), 10064 kicks, 10251 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 8 textures retired (1628 left, 3048 MB): budget 8, drawn over 0, dead targets 0 [hos host] vk: 79 textures retired (1557 left, 2933 MB): budget 79, drawn over 0, dead targets 0 [hos sead::MainThread] barrier: 295431 looks: 115207 found the masks moved already, 177257 waited until they moved, 2967 for the whole 1 ms (a wait 349 us on average); 271660 wakes [hos host] vk: 64 textures retired (1494 left, 2834 MB): budget 64, drawn over 0, dead targets 0 [hos Presentation Thread] vk: capture: a submit 0.13 ms (lock 0.01, scan 0.04 over 594 ranges, 3624 MB), 246 pages written, 246 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 4919 texture pages copied (19 MB) [hos host] vk: 4 textures retired (1490 left, 2826 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 14960000 (9252834 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11062 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: queues ordered by their events: 126000 submissions stopped at a wait, 125938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 359830, not met 0; captures at a CPU event's wait 12429, 0 of them at a wait met on arrival [hos host] vk: 18 textures retired (1472 left, 2763 MB): budget 18, drawn over 0, dead targets 0 [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #360000: event 0x100d6c800c mode 0 value 263099 (now 263080) (lr 0x80660e38) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 267 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1287 binds (305.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 331 taken from the spares, 165 made new; 359 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1225, 5439.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 167 dirty pages from their submits' captures, 65927 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1175 dispatches, 51851 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8142 runs, 853.2 MB compared; 113 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23476 in captures, 3 only in the collection, 49 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: texture #22127 0x10004bf1c00 640x360 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 19143 (drawn at 19082) [hos host] vk: texture #21473 0x10004a11800 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19143 (drawn at 19082) [hos host] vk: texture #21490 0x10002761000 960x540 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 19143 (drawn at 19082) [hos host] vk: late pages found at the replay (600 takes): 6 pages in 6 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 2 textures retired (1474 left, 2756 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: texture #23122 0x100089d2a00 768x768 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (16 so far) [hos host] vk: texture 0x1000ad13a00 1280x720x1 x1 levels 1 NVN 0x29 -> vk 97 (1500 textures, 2856 MB) [hos host] vk: 41 textures retired (1474 left, 2734 MB): budget 1, drawn over 0, dead targets 40 [hos sead::MainThread] vk: readbacks of GPU results: 1107 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 8 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1477 kept their virtual core (nothing there to wait for), 3131 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #23165 0x10008852400 384x384 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (17 so far) [hos host] vk: texture #23172 0x10002199c00 120x67 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (18 so far) [hos host] vk: presenter thread: the copy's fence 1.80 ms, the present 0.76 ms a present [hos host] vk: replay reads of guest memory as submitted: 14980000 (9263935 pages from the captures) [hos host] vk: 3 textures retired (1479 left, 2737 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 6288325 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1533 dirty pages from their submits' captures, 49694 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 891 dispatches, 38984 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6609 runs, 637.1 MB compared; 67 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17565 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: 0x7fb7c00000 (12 MB, captured, state 2): 1; [hos host] vk: 6 textures retired (1471 left, 2719 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15000000 (9277320 pages from the captures) [hos host] vk: 12 textures retired (1459 left, 2711 MB): budget 12, drawn over 0, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 15020000 (9290119 pages from the captures) [hos host] vk: memory: working set 14675 MB (peak 16021), private 21959 MB, staging 80 MB in 5 arenas, 45 made so far (205 ms) [hos host] vk: memory heaps: heap 0 (system): 12 blocks 912 MB, 36 allocations 840 MB; used 3176 of 15579 MB; heap 1 (device): 176 blocks 5335 MB, 21432 allocations 4655 MB; used 7458 of 8472 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 18 blocks 576 MB, 144 images 368 MB; 79 free ranges, the largest 23 MB [hos host] vk: image blocks (textures, compacted): 67 blocks 2144 MB, 1320 images 1760 MB; 533 free ranges, the largest 32 MB [hos host] vk: render target images: 6347 reused from the pool, 4556 made new, 4410 destroyed from it, 1171 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 155 missed after their image was taken by size; it holds 6 (45 MB) [hos host] vk: render targets not reused by size: 2086 with the pool empty (0 of them with targets still in the dead queue), 342 with none in it a present yet, 48 with none of a kind they can use, 582 with all smaller (the largest 50% short on average), 606 with all bigger (over 1.5 times), 889 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4218 idle [hos host] vk: VRAM: 7458 MB used of 8472 MB allowed (680 MB of it free in the blocks); textures 2712 MB (budget 2106 MB), 1462 textures [hos host] vk: budget: VRAM 5327 of 8472 MB (7458 MB by the driver, 2131 MB unaccounted): textures 2712, mirrors 1184, blocks free 680, shadows 284, other buffers 229, other images 111, staging 80, target pool 45; RAM 3119 MB (the process's private memory 21959 MB): imports 2244, capture spares 750, captures 90, capture table 35, others 0 [hos host] vk: textures: decoded ASTC 1778 MB, rendered 584 MB (100 drawn in one frame, kept: 49346 KB), other 314 MB; used in the last 8 presents 1996 MB [hos host] vk: decoded ASTC used again: 491 textures (1093 MB) copied from ones the game finalized, not decoded; 491 of those retired, 0 listed; 2225 MB of blocks hashed in 370 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 14053 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 64 (12501 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 1 (6 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 1 (19 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 1 (10 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 1 (17 KB); 138x96x1 NVN 0x1: 1 (12 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB); [hos host] vk: 8 textures retired (1454 left, 2706 MB): budget 8, drawn over 0, dead targets 0 [hos Presentation Thread] vk: presents after the CPU's signal (600): 463 waited for it, 3.73 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 44.9 ms; VRAM 7.3 of 8.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6476 queued, 6476 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16362 ms and 348.1 MB read in all); 8781 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12295 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2336 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9632 (6476 ready, 0 loading), 17767 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13980 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 182 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 8.37, present 0.20, draw 5.02, descriptors 2.18, texture lookups 0.28, uploads 0.06, pipelines 0.23, programs 0.79, commands 1.44, dispatch 0.82, memory sync 0.42, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 496 (1184 MB): 5.53 MB synced in 1073.0 copies, 17.57 MB written back [hos host] vk: 19200 presents, 60.7 fps; a frame: 1525 draws (0 skipped), 134 dispatches, 167 passes, 133 barriers, 6 submits, 0.7 uploads (0.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1454 textures (2706 MB) [hos host] vk: host time in GPU resource waits: 3.56 ms a frame (fences 2.78, timelines 0.78, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 604.4 full barriers recorded, 281.5 left out (nothing recorded since the one before), 16.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 2.8; 0x9: 1.0; 0x10: 3.0; 0x11: 25.8; 0x12: 69.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.7, an op 745.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.5 ms a frame (the longest submission 4.7 ms); 1.6 textures made a frame: colour targets 0.9 (3.8 MB), depth 0.3 (2.0 MB), ASTC or BC 0.4 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.27 MB unswizzled on the CPU in 0.04 ms [hos host] vk: the render thread (600 presents, 17.49 ms a frame): its jobs 16.40 ms a frame (submits 2.90, presents 5.12, others 8.38), idle 1.09; the game waited for room at 39 presents, 0.66 ms each (0.04 a frame; 14.15 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 466 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 2393 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.76 ms after its hand-over (42.72 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.45 + 0.22 (the window's copy is in the idle), after it 2.10 + 1.27; after a frame's last recording that reads or writes the game's memory come 1.02 recordings (0.00 ms of GPU time), after its last that reads it 1.02 (0.00 ms) [hos host] vk: frame ahead (600 presents): 9.34 parts a frame recorded in the signal's job, 71.0% of them took nothing that came with the signal; of the others, 2.69 a frame took pages of the signal's capture into a mirror (58.2 pages a frame) and 0.02 pages read as guest memory had them (0.1 pages a frame); a frame's first part took nothing in 12 of 466 frames, its first three in 12 [hos host] vk: frame ahead, what came with the signal: 466 signals with a capture, each of 90.5 pages in 12.7 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.93+0.17, 0.32+0.08, 0.28+0.09, 0.08+0.03, 0.09+0.05, 0.13+0.34, 0.17+0.38, 0.05+0.09, 0.03+0.02, 0.01+0.00, 0.01+0.00, 0.00+0.00 [hos host] vk: texture #20951 0x117047f000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 19201 (drawn at 15578) [hos ModuleSystemWorker1] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 41638293 lookups, 119840 missed (0.29 %) [hos sead::MainThread] nvn: GPU waits in 5 s: 192 sync waits (699 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: texture #23050 0x10008942600 640x360 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 19208 (drawn at 19147) [hos host] vk: texture #23060 0x10009932a00 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 19208 (drawn at 19147) [hos host] vk: render targets drawn over from another start, let go: 11509 so far (53914 MB), 10376 of them (48723 MB) between the passes [hos host] vk: 33 textures retired (1430 left, 2675 MB): budget 33, drawn over 0, dead targets 0 [hos host] vk: reaper: 943 images (2011 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 87 (130 MB) [hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1287 binds (305.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 331 taken from the spares, 165 made new; 359 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 575 dirty pages from their submits' captures, 54509 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 944 dispatches, 42365 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7008 runs, 688.8 MB compared; 69 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18949 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 595 ranges, 3628 MB), 270 pages written, 270 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 6820 texture pages copied (26 MB) [hos host] vk: program picks: 21118084 kept (8879861 by the last picks' textures, a uniform buffer bound since), 26433 new, 7655482 with other textures (of 28800000) [hos host] vk: texture #23034 0x10004380c00 1280x720 x1 NVN 0x5 the GPU drew, finalized by the game, sampled again at frame 19219 (drawn at 19158) [hos host] vk: texture #23059 0x10004831400 1280x720 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19219 (drawn at 19158) [hos host] vk: 54 textures retired (1391 left, 2644 MB): budget 51, drawn over 1, dead targets 2 [hos host] vk: replay reads of guest memory as submitted: 15040000 (9303550 pages from the captures) [hos host] vk: texture #23053 0x10004ce1800 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 19220 (drawn at 19158) [hos host] vk: texture #23044 0x10005191c00 1600x1600 x3 NVN 0x32 the GPU drew, finalized by the game, sampled again at frame 19220 (drawn at 19158) [hos host] vk: texture #23141 0x10008ca2e00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19227 (drawn at 19166) [hos host] vk: 6 textures retired (1394 left, 2703 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15060000 (9318008 pages from the captures) [hos host] vk: commit: 1824 MB left under Windows' limit as it is now, and the page files can grow by 79851 MB more: counted as room [hos host] vk: 6 textures retired (1394 left, 2745 MB): budget 6, drawn over 0, dead targets 0 [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #265000: event 0x100d6c8008 mode 0 value 264998 (lr 0x80660f34) [hos host] vk: texture #23051 0x10004741200 2560x1440 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 19246 (drawn at 19158) [hos host] vk: texture #23052 0x10004bf1600 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 19246 (drawn at 19158) [hos host] vk: texture #23054 0x10004b01400 1280x720 x1 NVN 0x1 the GPU drew, finalized by the game, sampled again at frame 19246 (drawn at 19158) [hos host] vk: texture #23078 0x10005280e00 2560x1440 x1 NVN 0x25 the GPU drew, finalized by the game, sampled again at frame 19246 (drawn at 19158) [hos host] vk: texture #23340 0x1000738e800 2400x2400 NVN 0x34 evicted, made again over storage render targets took: uploaded from the game's memory (19 so far) [hos host] [debug] vk: a staging arena of 22 MB took 6.5 ms to make [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 640000 so far (x1 0x11791f0000 x2 0x2 x3 0x11791f0000 x4 0x654040) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14041 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 1 textures retired (1378 left, 2689 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: texture #23140 0x10008c12c00 384x384 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19260 (drawn at 19199) [hos host] vk: late pages found at the replay (600 takes): 3 pages in 3 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 1125, 4995.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 206 dirty pages from their submits' captures, 61885 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1052 dispatches, 48030 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8215 runs, 793.4 MB compared; 96 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21914 in captures, 0 only in the collection, 46 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; 0x7f96800000 (4 MB, captured, state 2): 1; [hos host] vk: texture #23122 0x100089d2a00 768x768 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19265 (drawn at 19204) [hos host] vk: 4 textures retired (1374 left, 2689 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15080000 (9326305 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: texture #21307 0x1001e23cc00 1120x1333x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1421 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21306 0x1001e840e00 1120x1333x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1387 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 3 textures retired (1371 left, 2677 MB): budget 2, drawn over 1, dead targets 0; GPU-drawn with no copy kept: 0 live, 2 finalized, 11 MB [hos Presentation Thread] vi: 19304 frames presented (57.7 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 24, 15.5-17.8 227, 17.8-22 19, 22-30 14, 30-40 4, over 40 1 [hos host] vk: texture #23127 0x100020d0a00 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19306 (drawn at 19245) [hos host] vk: texture #23306 0x10008612200 1722x1722 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19308 (drawn at 19246) [hos host] vk: texture #23332 0x10008612200 1764x1764 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19309 (drawn at 19247) [hos host] vk: texture #21205 0x1001ef35200 640x360x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1433 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 11 textures retired (1360 left, 2604 MB): budget 7, drawn over 4, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.05 over 595 ranges, 3630 MB), 254 pages written, 254 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.012; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 7009 texture pages copied (27 MB) [hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1287 binds (305.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 331 taken from the spares, 165 made new; 359 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1300, 5772.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 215 dirty pages from their submits' captures, 69178 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1152 dispatches, 54443 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9332 runs, 915.8 MB compared; 91 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 25178 in captures, 0 only in the collection, 52 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1082 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1316 kept their virtual core (nothing there to wait for), 2780 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 1 textures retired (1359 left, 2590 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: a frame: 134 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 7.24, present 0.21, draw 4.00, descriptors 1.85, texture lookups 0.48, uploads 0.12, pipelines 0.17, programs 0.67, commands 1.02, dispatch 0.79, memory sync 0.43, queue submit 0.00, slot wait 0.21 [hos host] vk: a frame: mirrors 496 (1184 MB): 5.32 MB synced in 1058.8 copies, 17.63 MB written back [hos host] vk: 19320 presents, 60.2 fps; a frame: 967 draws (0 skipped), 141 dispatches, 171 passes, 133 barriers, 6 submits, 0.2 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1359 textures (2590 MB) [hos host] vk: host time in GPU resource waits: 2.65 ms a frame (fences 2.42, timelines 0.22, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 625.3 full barriers recorded, 286.5 left out (nothing recorded since the one before), 15.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 2.5; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 68.9; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 48.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.9, an op 592.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 9.7 ms a frame (the longest submission 5.6 ms); 1.2 textures made a frame: colour targets 0.5 (2.4 MB), depth 0.6 (3.9 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.30 MB unswizzled on the CPU in 0.06 ms [hos ModuleSystemWorker3] nvn: 2550000 recordings made, 74830 live copies (22.5 MB), 6411 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: 2 textures retired (1359 left, 2590 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: 17 textures retired (1343 left, 2571 MB): budget 17, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: CommandBufferWaitEvent #365000: event 0x100d6c800c mode 0 value 266899 (now 266880) (lr 0x80660e38) [hos host] vk: 37 textures retired (1306 left, 2544 MB): budget 37, drawn over 0, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11082 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 15100000 (9333992 pages from the captures) [hos host] vk: 64 textures retired (1242 left, 2466 MB): budget 64, drawn over 0, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1250, 5550.4 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 131 dirty pages from their submits' captures, 66764 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1106 dispatches, 52512 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8765 runs, 876.3 MB compared; 110 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 24399 in captures, 0 only in the collection, 50 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: texture #21220 0x1001f0ca400 105x33x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21224 0x1001f0bb000 133x42x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21219 0x1001f0d1600 272x106x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21223 0x1001f0c4200 169x97x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21227 0x1001f0b2e00 197x106x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1407 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 121 textures retired (1121 left, 2251 MB): budget 121, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 5 finalized, 0 MB [hos host] vk: texture #21243 0x1001f08c000 62x32x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21254 0x1001f07a000 54x46x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21246 0x1001f087e00 126x96x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21239 0x1001f092400 18x33x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21238 0x1001f094600 96x112x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21234 0x1001f098800 32x48x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21250 0x1001f084400 102x30x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21242 0x1001f08e200 82x97x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21233 0x1001f09aa00 160x160x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21253 0x1001f07e200 166x94x1 x1 NVN 0x1 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21230 0x1001f0a6c00 96x96x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1417 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 53 textures retired (1124 left, 2234 MB): budget 11, drawn over 0, dead targets 42; GPU-drawn with no copy kept: 0 live, 11 finalized, 0 MB [hos host] vk: slow upload 16.8 ms: #23505 0x10f4eb1600 768x768x1 x1 levels 10 NVN 0x7b ASTC, 528 of 528 KB [hos host] vk: texture #21400 0x1001ef51600 640x360x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1449 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21456 0x1001e24ba00 1120x1333x1 x1 NVN 0x3f the GPU drew goes with no copy of its texels (finalized, unused for 1447 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21258 0x1001f072c00 96x32x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1427 frames, over the budget): made again over this storage, it starts empty [hos host] vk: texture #21262 0x1001f069800 42x41x1 x1 NVN 0x25 the GPU drew goes with no copy of its texels (finalized, unused for 1427 frames, over the budget): made again over this storage, it starts empty [hos host] vk: 112 textures retired (1106 left, 2289 MB): budget 112, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 9 finalized, 12 MB [hos host] vk: replay reads of guest memory as submitted: 15120000 (9345507 pages from the captures) [hos tera::FileResourceMgr] vcores: 145649 takes, 20028 waits (mean 61 us, longest 9.1 ms, 0 over HOS_VCORE_LOG_MS), 9781 kicks, 9753 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos host] vk: 43 textures retired (1191 left, 2391 MB): budget 43, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 6355714 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1291 binds (305.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 333 taken from the spares, 165 made new; 361 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1535 dirty pages from their submits' captures, 56253 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 971 dispatches, 44320 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6919 runs, 716.3 MB compared; 79 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19928 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x7fa8400000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 15140000 (9360771 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 593 ranges, 3644 MB), 259 pages written, 259 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 7355 texture pages copied (28 MB) [hos host] vk: 2 textures retired (1280 left, 2495 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: program picks: 21250491 kept (8915218 by the last picks' textures, a uniform buffer bound since), 26457 new, 7723051 with other textures (of 29000000) [hos host] vk: 40 textures retired (1245 left, 2458 MB): budget 40, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15160000 (9375341 pages from the captures) [hos host] vk: 15 textures retired (1230 left, 2429 MB): budget 15, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15180000 (9389378 pages from the captures) [hos host] vk: ASTC texture #23788 0x1132ba6a00 384x384 x1 levels 9: the image of #23521 0x10cddbe000 (the same blocks, finalized by the game) copied, not decoded again (500 so far, 1101 MB) [hos host] vk: 1 textures retired (1238 left, 2429 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: a frame: 188 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.19, present 0.18, draw 5.76, descriptors 3.13, texture lookups 1.22, uploads 0.99, pipelines 0.21, programs 0.76, commands 1.30, dispatch 0.81, memory sync 0.46, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 498 (1184 MB): 5.45 MB synced in 1070.5 copies, 17.56 MB written back [hos host] vk: 19440 presents, 55.5 fps; a frame: 1237 draws (0 skipped), 133 dispatches, 181 passes, 132 barriers, 6 submits, 3.0 uploads (2.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1238 textures (2429 MB) [hos host] vk: host time in GPU resource waits: 6.19 ms a frame (fences 3.40, timelines 2.79, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 687.5 full barriers recorded, 296.9 left out (nothing recorded since the one before), 15.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 2.6; 0x9: 1.0; 0x10: 2.7; 0x11: 26.3; 0x12: 69.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.3, an op 761.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.3 ms a frame (the longest submission 4.6 ms); 3.7 textures made a frame: colour targets 0.4 (1.9 MB), depth 0.3 (2.4 MB), ASTC or BC 2.9 (4.8 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.17 MB unswizzled on the CPU in 0.17 ms [hos host] vk: write-backs left out (600 submits): 926, 4109.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1588 dirty pages from their submits' captures, 51085 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 865 dispatches, 40176 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6570 runs, 654.7 MB compared; 76 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17624 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 15200000 (9403810 pages from the captures) [hos host] vk: 1 textures retired (1242 left, 2425 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15220000 (9417858 pages from the captures) [hos sead::MainThread] vk: readbacks of GPU results: 1092 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1368 kept their virtual core (nothing there to wait for), 2984 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12344 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 137, 1184 MB with memory of 1696 MB (511 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1291 binds (305.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 333 taken from the spares, 165 made new; 361 kept as spares, 0 destroyed; 28 spares now, 1124 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 603 dirty pages from their submits' captures, 56178 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 929 dispatches, 43445 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6970 runs, 701.9 MB compared; 57 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19313 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 15240000 (9431880 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 15260000 (9445808 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 592 ranges, 3638 MB), 259 pages written, 259 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (136 MB), 10 spare buffers, 3629 texture pages copied (14 MB) [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #270000: event 0x100d6c801c mode 0 value 270000 (lr 0x80665068) [hos sead::MainThread] nvn: GPU waits in 5 s: 76 sync waits (364 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: 4 textures retired (1252 left, 2419 MB): budget 3, drawn over 0, dead targets 1 [hos host] vk: replay reads of guest memory as submitted: 15280000 (9459725 pages from the captures) [hos host] vk: texture slots registered again at another size: 1000000 (of 1248706 registrations; slot 28277 512x512 -> 120x67) [hos host] vk: program picks: 21395627 kept (8968541 by the last picks' textures, a uniform buffer bound since), 26457 new, 7777915 with other textures (of 29200000) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 199 dirty pages from their submits' captures, 56732 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 946 dispatches, 43576 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7063 runs, 703.4 MB compared; 64 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19458 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 15300000 (9474082 pages from the captures) [hos host] vk: texture #23826 0x100021ae000 30x16 NVN 0x3f evicted, made again over storage render targets took: uploaded from the game's memory (20 so far) [hos sead::MainThread] nvn: CommandBufferWaitEvent #370000: event 0x100d6c8010 mode 0 value 270700 (now 270681) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 15320000 (9488525 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 15340000 (9503437 pages from the captures) [hos host] vk: 6 textures retired (1271 left, 2430 MB): budget 6, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 19 MB [hos host] vk: a frame: 290 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.20, present 0.17, draw 6.38, descriptors 2.75, texture lookups 0.15, uploads 0.03, pipelines 0.28, programs 1.06, commands 1.79, dispatch 0.81, memory sync 0.45, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 498 (1166 MB): 5.58 MB synced in 1090.1 copies, 17.59 MB written back [hos host] vk: 19560 presents, 60.0 fps; a frame: 1851 draws (0 skipped), 136 dispatches, 196 passes, 132 barriers, 6 submits, 0.5 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1273 textures (2433 MB) [hos host] vk: host time in GPU resource waits: 1.68 ms a frame (fences 1.32, timelines 0.36, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 721.5 full barriers recorded, 309.4 left out (nothing recorded since the one before), 16.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.9, an op 1150.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.6 ms a frame (the longest submission 4.5 ms); 0.6 textures made a frame: colour targets 0.1 (0.4 MB), depth 0.1 (0.2 MB), ASTC or BC 0.4 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.04 MB unswizzled on the CPU in 0.01 ms [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 133, 1166 MB with memory of 1668 MB (501 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1294 binds (306.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 333 taken from the spares, 165 made new; 365 kept as spares, 0 destroyed; 32 spares now, 1152 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 382 dirty pages from their submits' captures, 57495 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 964 dispatches, 43959 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7144 runs, 705.4 MB compared; 54 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19599 in captures, 1 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2244 MB), 0 retired still imported (0 MB), 51 retired given back (1296 MB) [hos host] vk: 1 textures retired (1281 left, 2455 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15360000 (9517661 pages from the captures) [hos host] vk: 24 textures retired (1264 left, 2446 MB): budget 24, drawn over 0, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9616 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 15380000 (9531915 pages from the captures) [hos Presentation Thread] vi: 19595 frames presented (58.2 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 22, 15.5-17.8 242, 17.8-22 16, 22-30 6, 30-40 5, over 40 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 592 ranges, 3642 MB), 268 pages written, 268 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 25 kept (100 MB), 10 spare buffers, 4661 texture pages copied (18 MB) [hos host] vk: 28 textures retired (1253 left, 2482 MB): budget 28, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 6423220 pages so far (0 held: may be written ahead in their submission) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 400 dirty pages from their submits' captures, 56692 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 956 dispatches, 43621 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7292 runs, 707.7 MB compared; 47 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19674 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fdfc00000 (12 MB, captured, state 2): 1; [hos host] vk: replay reads of guest memory as submitted: 15400000 (9545975 pages from the captures) [hos host] vk: 368 mirrors given back (2441 MB in all); 1177 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: 25 textures retired (1292 left, 2605 MB): budget 21, drawn over 4, dead targets 0 [hos host] vk: texture #22534 0x1171b2b000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 19622 (drawn at 18893) [hos host] vk: replay reads of guest memory as submitted: 15420000 (9558583 pages from the captures) [hos host] vk: program picks: 21540426 kept (9021988 by the last picks' textures, a uniform buffer bound since), 26458 new, 7833115 with other textures (of 29400000) [hos rmbl::VibrationThread] hid: 65536 vibration values, strongest 0.00 since the last count [hos host] vk: 14 textures retired (1292 left, 2615 MB): budget 6, drawn over 8, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1117 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1535 kept their virtual core (nothing there to wait for), 3073 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 4 textures retired (1297 left, 2649 MB): budget 0, drawn over 4, dead targets 0 [hos host] vk: sparse mirrors: 127, 1134 MB with memory of 1620 MB (485 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1303 binds (306.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 336 taken from the spares, 165 made new; 374 kept as spares, 0 destroyed; 38 spares now, 1200 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1071 dirty pages from their submits' captures, 53780 from earlier ones', 1 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 932 dispatches, 41589 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7036 runs, 673.6 MB compared; 56 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18500 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 193 live (2244 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 15440000 (9569598 pages from the captures) [hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 660000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040) [hos host] vk: 9 textures retired (1294 left, 2648 MB): budget 1, drawn over 8, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 0 MB [hos host] vk: recorder: 5246760 packets run in 10.0 s (busy 1.59 s, spun idle 0.12 s, slept 10632 times, 3163 of them under 200 us), 0 calls made directly (no room); 17177 batches over 591 frames (29.1 a frame, 305 packets a batch), 8485 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 67 times, 10.21 ms (0.1 times, 0.02 ms a frame); every thread 742 times, 67.12 ms in all, the longest 0.59 ms; vkWaitForFences 521/43.27 ms (max 0.55), vkDestroyImageView 143/17.45 ms (max 0.57), vmaFreeMemory 13/2.59 ms (max 0.59), vkEndCommandBuffer 6/1.37 ms (max 0.45), vkQueueBindSparse 21/1.03 ms (max 0.30), the queue's lock 8/0.69 ms (max 0.18) [hos host] vk: recorder: of its busy 1.59 s, 0.41 s in parts' ends and submits (the longest 1.37 ms), 0.12 s of that blocked on the queue's lock (250 times, the longest 1.33 ms), so 0.30 s in the ends, submits and holds; 1.18 s running the other packets; 8878 packets a frame [hos host] vk: 11 textures retired (1290 left, 2643 MB): budget 0, drawn over 11, dead targets 0 [hos host] vk: texture #23902 0x100089d2a00 953x953 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19668 (drawn at 19606) [hos host] vk: 9 textures retired (1283 left, 2618 MB): budget 0, drawn over 9, dead targets 0 [hos host] vk: texture #23906 0x100089d2a00 1018x1018 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19672 (drawn at 19610) [hos host] vk: replay reads of guest memory as submitted: 15460000 (9580203 pages from the captures) [hos host] vk: texture #23909 0x100089d2a00 1046x1046 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19674 (drawn at 19612) [hos host] vk: 9 textures retired (1280 left, 2614 MB): budget 2, drawn over 7, dead targets 0 [hos host] vk: a frame: 274 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.40, present 0.20, draw 6.50, descriptors 2.90, texture lookups 0.42, uploads 0.09, pipelines 0.27, programs 1.07, commands 1.76, dispatch 0.82, memory sync 0.44, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 501 (1116 MB): 5.77 MB synced in 1088.3 copies, 17.72 MB written back [hos host] vk: 19680 presents, 60.2 fps; a frame: 1817 draws (0 skipped), 145 dispatches, 192 passes, 132 barriers, 6 submits, 0.5 uploads (0.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1281 textures (2619 MB) [hos host] vk: host time in GPU resource waits: 2.27 ms a frame (fences 1.30, timelines 0.97, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 719.8 full barriers recorded, 306.3 left out (nothing recorded since the one before), 15.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 25.0; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 123.0, an op 1137.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.1 ms a frame (the longest submission 4.5 ms); 1.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.8 (2.9 MB), ASTC or BC 0.4 (0.8 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.37 MB unswizzled on the CPU in 0.06 ms [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 281 dirty pages from their submits' captures, 57726 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 952 dispatches, 44454 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7514 runs, 710.0 MB compared; 62 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20163 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 6 textures retired (1279 left, 2619 MB): budget 2, drawn over 4, dead targets 0 [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 589 ranges, 3446 MB), 267 pages written, 267 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (102 MB), 10 spare buffers, 6194 texture pages copied (24 MB) [hos rmbl::VibrationThread] vcores: 140103 takes, 21119 waits (mean 72 us, longest 8.1 ms, 0 over HOS_VCORE_LOG_MS), 10093 kicks, 10095 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos sead::MainThread] nvn: 2600000 recordings made, 74825 live copies (24.1 MB), 6530 texture copies; 6 command buffers' segments (4.3 MB) [hos host] vk: replay reads of guest memory as submitted: 15480000 (9590690 pages from the captures) [hos host] vk: 8 textures retired (1276 left, 2608 MB): budget 1, drawn over 7, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11077 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 6 textures retired (1278 left, 2619 MB): budget 1, drawn over 5, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 15500000 (9602292 pages from the captures) [hos host] vk: 6 textures retired (1306 left, 2571 MB): budget 0, drawn over 0, dead targets 6 [hos host] vk: pool 0x10140a5600+0xf00000 imported in 4.0 ms as 0x1014000000..0x1015000000 (240 imports, 724 ms in all) [hos host] vk: sparse mirrors: 122, 1119 MB with memory of 1596 MB (476 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1309 binds (307.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 338 taken from the spares, 165 made new; 381 kept as spares, 0 destroyed; 43 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 729 dirty pages from their submits' captures, 54299 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 920 dispatches, 42340 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7336 runs, 709.0 MB compared; 63 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19494 in captures, 2 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x1041c00000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: program picks: 21684454 kept (9077147 by the last picks' textures, a uniform buffer bound since), 26458 new, 7889087 with other textures (of 29600000) [hos host] vk: 19 textures retired (1364 left, 2828 MB): budget 19, drawn over 0, dead targets 0 [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #375000: event 0x100d6c800c mode 0 value 274499 (now 274480) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 15520000 (9614241 pages from the captures) [hos host] vk: 1 textures retired (1392 left, 2926 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: 1 textures retired (1407 left, 2967 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1398 dirty pages from their submits' captures, 54501 from earlier ones', 1 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1009 dispatches, 42993 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8031 runs, 715.9 MB compared; 57 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18605 in captures, 0 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: 0x7fa8400000 (4 MB, captured, state 2): 1; [hos host] vk: presenter thread: the copy's fence 1.87 ms, the present 0.72 ms a present [hos host] vk: replay reads of guest memory as submitted: 15540000 (9624787 pages from the captures) [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #275000: event 0x100d6c801c mode 0 value 274997 (lr 0x802911b8) [hos host] vk: texture #23410 0x100020e5000 30x16 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 19778 (drawn at 19717) [hos host] vk: texture #23382 0x1000589ac00 640x360 x1 NVN 0x11 the GPU drew, finalized by the game, sampled again at frame 19778 (drawn at 19717) [hos host] vk: texture #23412 0x100020e5a00 2560x1440 x1 NVN 0x3f the GPU drew, finalized by the game, sampled again at frame 19778 (drawn at 19717) [hos host] vk: texture #23394 0x10004bf1c00 1280x720 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 19778 (drawn at 19717) [hos host] vk: 10 textures retired (1417 left, 3034 MB): budget 7, drawn over 3, dead targets 0 [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 584 ranges, 3522 MB), 256 pages written, 256 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.024; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (116 MB), 10 spare buffers, 8192 texture pages copied (32 MB) [hos host] vk: 52 textures retired (1371 left, 2913 MB): budget 44, drawn over 8, dead targets 0 [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1096 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 24 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1451 kept their virtual core (nothing there to wait for), 3157 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 15560000 (9633931 pages from the captures) [hos ModuleSystemWorker4] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 15444036 lookups, 120643 missed (0.78 %) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: memory: working set 14749 MB (peak 16021), private 22097 MB, staging 80 MB in 5 arenas, 47 made so far (212 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 944 MB, 35 allocations 864 MB; used 3232 of 15579 MB; heap 1 (device): 177 blocks 5399 MB, 20478 allocations 4790 MB; used 7522 of 8426 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 25 blocks 800 MB, 197 images 678 MB; 129 free ranges, the largest 15 MB [hos host] vk: image blocks (textures, compacted): 60 blocks 1920 MB, 1176 images 1639 MB; 510 free ranges, the largest 24 MB [hos host] vk: render target images: 6408 reused from the pool, 4918 made new, 4719 destroyed from it, 1224 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 164 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 2276 with the pool empty (0 of them with targets still in the dead queue), 363 with none in it a present yet, 59 with none of a kind they can use, 637 with all smaller (the largest 50% short on average), 629 with all bigger (over 1.5 times), 951 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4527 idle [hos host] vk: VRAM: 7522 MB used of 8426 MB allowed (605 MB of it free in the blocks); textures 2954 MB (budget 2106 MB), 1379 textures [hos host] vk: budget: VRAM 5395 of 8426 MB (7522 MB by the driver, 2127 MB unaccounted): textures 2954, mirrors 1130, blocks free 605, shadows 284, other buffers 229, other images 111, staging 80; RAM 3167 MB (the process's private memory 22097 MB): imports 2260, capture spares 728, captures 110, capture table 35, other buffers 26, retired imports 8, others 0 [hos host] vk: textures: decoded ASTC 1490 MB, rendered 840 MB (73 drawn in one frame, kept: 48107 KB), other 531 MB; used in the last 8 presents 1821 MB [hos host] vk: decoded ASTC used again: 528 textures (1160 MB) copied from ones the game finalized, not decoded; 528 of those retired, 4 listed; 2374 MB of blocks hashed in 390 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 14653 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34, last used 2 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 78 (13794 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB); [hos host] vk: 52 textures retired (1327 left, 2801 MB): budget 44, drawn over 8, dead targets 0 [hos Presentation Thread] vk: presents after the CPU's signal (600): 214 waited for it, 3.22 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 41.2 ms; VRAM 7.3 of 8.3 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: reaper: 1102 images (2840 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 120 (214 MB) [hos host] vk: program loads at create (play): 6476 queued, 6476 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16362 ms and 348.1 MB read in all); 8781 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12295 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2361 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9632 (6476 ready, 0 loading), 17767 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13980 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 238 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.51, present 0.20, draw 6.70, descriptors 3.24, texture lookups 0.92, uploads 0.57, pipelines 0.25, programs 1.08, commands 1.66, dispatch 0.84, memory sync 0.47, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 509 (1130 MB): 5.58 MB synced in 1074.0 copies, 17.82 MB written back [hos host] vk: 19800 presents, 59.2 fps; a frame: 1755 draws (0 skipped), 149 dispatches, 174 passes, 132 barriers, 6 submits, 2.1 uploads (2.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1328 textures (2806 MB) [hos host] vk: host time in GPU resource waits: 3.27 ms a frame (fences 1.50, timelines 1.77, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 671.6 full barriers recorded, 291.5 left out (nothing recorded since the one before), 13.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 25.0; 0x12: 68.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 122.8, an op 1010.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.4 ms a frame (the longest submission 4.0 ms); 2.8 textures made a frame: colour targets 0.5 (1.8 MB), depth 1.0 (5.4 MB), ASTC or BC 1.3 (2.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.66 MB unswizzled on the CPU in 0.23 ms [hos host] vk: the render thread (600 presents, 16.96 ms a frame): its jobs 14.63 ms a frame (submits 6.01, presents 4.91, others 3.70), idle 2.33; the game waited for room at 29 presents, 0.57 ms each (0.03 a frame; 9.69 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 222 captures at a CPU signal; the submits' waits read in 0.040 ms a frame; readbacks: 1267 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 13.84 ms after its hand-over (36.13 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.66 + 0.21 (the window's copy is in the idle), after it 1.56 + 0.90; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.33 parts a frame recorded in the signal's job, 71.3% of them took nothing that came with the signal; of the others, 2.38 a frame took pages of the signal's capture into a mirror (36.9 pages a frame) and 0.01 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 25 of 222 frames, its first three in 25 [hos host] vk: frame ahead, what came with the signal: 222 signals with a capture, each of 51.8 pages in 10.8 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.73+0.16, 0.20+0.10, 0.21+0.07, 0.08+0.03, 0.09+0.12, 0.11+0.21, 0.09+0.18, 0.04+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: readback pages refreshed from guest memory: 6490330 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 128, 1130 MB with memory of 1624 MB (493 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1315 binds (308.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 344 taken from the spares, 165 made new; 381 kept as spares, 0 destroyed; 37 spares now, 1196 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 155 dirty pages from their submits' captures, 59343 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1036 dispatches, 45790 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8356 runs, 766.0 MB compared; 57 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20654 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: render targets drawn over from another start, let go: 11824 so far (55528 MB), 10639 of them (50075 MB) between the passes [hos host] vk: 11 textures retired (1325 left, 2760 MB): budget 1, drawn over 10, dead targets 0 [hos host] vk: 10 textures retired (1323 left, 2741 MB): budget 2, drawn over 7, dead targets 1 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12610 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 15580000 (9642328 pages from the captures) [hos host] vk: 10 textures retired (1323 left, 2719 MB): budget 0, drawn over 10, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 6 textures retired (1322 left, 2716 MB): budget 0, drawn over 6, dead targets 0 [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 163 dirty pages from their submits' captures, 59838 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1028 dispatches, 45744 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8354 runs, 752.2 MB compared; 49 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21198 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: 0x7fa6800000 (4 MB, captured, state 2): 2; [hos host] vk: program picks: 21829563 kept (9142072 by the last picks' textures, a uniform buffer bound since), 26458 new, 7943978 with other textures (of 29800000) [hos host] vk: 4 textures retired (1330 left, 2741 MB): budget 0, drawn over 4, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15600000 (9652267 pages from the captures) [hos host] vk: 9 textures retired (1326 left, 2715 MB): budget 0, drawn over 9, dead targets 0 [hos sead::MainThread] nvn: GPU waits in 5 s: 94 sync waits (298 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: 8 textures retired (1328 left, 2718 MB): budget 0, drawn over 8, dead targets 0 [hos host] vk: 2 textures retired (1299 left, 2596 MB): budget 0, drawn over 2, dead targets 0 [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 582 ranges, 3521 MB), 271 pages written, 271 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (110 MB), 9 spare buffers, 2477 texture pages copied (9 MB) [hos host] vk: replay reads of guest memory as submitted: 15620000 (9661932 pages from the captures) [hos host] vk: sparse mirrors: 127, 1128 MB with memory of 1620 MB (491 MB spared); blocks given memory later 13773, given back 16947; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1316 binds (308.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 344 taken from the spares, 165 made new; 382 kept as spares, 0 destroyed; 38 spares now, 1200 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1075, 4773.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 207 dirty pages from their submits' captures, 59867 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1070 dispatches, 45981 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8721 runs, 773.7 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20865 in captures, 2 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7fa6800000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: 14 textures retired (1286 left, 2586 MB): budget 12, drawn over 2, dead targets 0 [hos Presentation Thread] vi: 19894 frames presented (59.7 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 41, 15.5-17.8 236, 17.8-22 14, 22-30 6, 30-40 2, over 40 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: 98 textures retired (1193 left, 2410 MB): budget 98, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 14 MB [hos host] vk: 69 textures retired (1130 left, 2295 MB): budget 61, drawn over 8, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15640000 (9671298 pages from the captures) [hos host] vk: 70 textures retired (1063 left, 2182 MB): budget 64, drawn over 6, dead targets 0 [hos host] vk: a frame: 214 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.48, present 0.22, draw 5.99, descriptors 2.70, texture lookups 0.27, uploads 0.26, pipelines 0.25, programs 1.01, commands 1.58, dispatch 0.85, memory sync 0.47, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 509 (1124 MB): 5.53 MB synced in 1079.0 copies, 18.00 MB written back [hos host] vk: 19920 presents, 60.0 fps; a frame: 1744 draws (0 skipped), 159 dispatches, 168 passes, 135 barriers, 6 submits, 1.1 uploads (1.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1064 textures (2185 MB) [hos host] vk: host time in GPU resource waits: 1.44 ms a frame (fences 1.42, timelines 0.02, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 651.0 full barriers recorded, 289.2 left out (nothing recorded since the one before), 13.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 28.6; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 125.5, an op 886.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.4 ms a frame (the longest submission 3.3 ms); 1.0 textures made a frame: colour targets 0.2 (1.1 MB), depth 0.7 (2.5 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.89 MB unswizzled on the CPU in 0.25 ms [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11067 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 311 dirty pages from their submits' captures, 56662 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1031 dispatches, 43351 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8372 runs, 724.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19984 in captures, 2 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 20 textures retired (1060 left, 2194 MB): budget 20, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15660000 (9679863 pages from the captures) [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #380000: event 0x100d6c800c mode 0 value 278299 (now 278280) (lr 0x80660e38) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: VRAM compaction 7 begins: the textures' blocks hold 1920 MB, 746 MB of it free space [hos host] vk: 2 textures retired (1040 left, 2125 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: program picks: 21975766 kept (9208829 by the last picks' textures, a uniform buffer bound since), 26458 new, 7997775 with other textures (of 30000000) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1150 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 4 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1412 kept their virtual core (nothing there to wait for), 2940 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: queues ordered by their events: 132000 submissions stopped at a wait, 131938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 380286, not met 0; captures at a CPU event's wait 12701, 0 of them at a wait met on arrival [hos host] vk: replay reads of guest memory as submitted: 15680000 (9689300 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 128, 1130 MB with memory of 1628 MB (497 MB spared); blocks given memory later 13773, given back 17008; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1319 binds (309.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 345 taken from the spares, 165 made new; 382 kept as spares, 0 destroyed; 37 spares now, 1192 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 129 dirty pages from their submits' captures, 58229 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1041 dispatches, 44371 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8451 runs, 740.7 MB compared; 53 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20380 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 581 ranges, 3531 MB), 284 pages written, 284 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (132 MB), 9 spare buffers, 3650 texture pages copied (14 MB) [hos BaseProcExecutor1] vcores: 152285 takes, 23128 waits (mean 72 us, longest 7.3 ms, 0 over HOS_VCORE_LOG_MS), 10974 kicks, 10935 yields, 0 starved grants, 0 escapes in 5 s [hos ModuleSystemWorker2] ocb: 28 choices of another neighbour held back, 13 taken, 0 updates refused, in 18000 updates [hos sead::MainThread] nvn: QueueSubmitCommands #120000: queue 000010100D695020 (its #19996), 1 handles (lr 0x80666be0) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: replay reads of guest memory as submitted: 15700000 (9699649 pages from the captures) [hos sead::MainThread] nvn: EventSignal (CPU) #40000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: readback pages refreshed from guest memory: 6557874 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 241 dirty pages from their submits' captures, 64070 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1061 dispatches, 45069 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8765 runs, 745.9 MB compared; 63 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20878 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2; [hos host] vk: VRAM compaction 7 over: 389 textures moved (280 MB) in 34 passes, 0 left where they were; 21 blocks (672 MB) given back to the driver, 4480 MB in all so far [hos host] vk: texture #20735 0x116ff5c000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 20017 (drawn at 15243) [hos host] vk: 384 mirrors given back (2517 MB in all); 1123 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: replay reads of guest memory as submitted: 15720000 (9709917 pages from the captures) [hos sead::MainThread] nvn: CommandBufferSignalEvent #280000: event 0x100d6c8030 mode 0 value 279999 (lr 0x806612a8) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5217 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a frame: 190 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.01, present 0.44, draw 6.20, descriptors 2.90, texture lookups 0.44, uploads 0.28, pipelines 0.25, programs 1.00, commands 1.60, dispatch 0.91, memory sync 0.52, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 510 (1116 MB): 5.81 MB synced in 1084.4 copies, 18.16 MB written back [hos host] vk: 20040 presents, 60.0 fps; a frame: 1723 draws (0 skipped), 170 dispatches, 171 passes, 144 barriers, 6 submits, 0.9 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1072 textures (2159 MB) [hos host] vk: host time in GPU resource waits: 1.78 ms a frame (fences 1.44, timelines 0.34, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 665.4 full barriers recorded, 300.4 left out (nothing recorded since the one before), 17.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.6 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.1, an op 895.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.7 ms a frame (the longest submission 4.4 ms); 1.5 textures made a frame: colour targets 0.4 (2.0 MB), depth 0.7 (2.4 MB), ASTC or BC 0.5 (0.8 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.55 MB unswizzled on the CPU in 0.23 ms [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 680000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: sparse mirrors: 124, 1116 MB with memory of 1604 MB (487 MB spared); blocks given memory later 13773, given back 17009; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1324 binds (309.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 345 taken from the spares, 165 made new; 386 kept as spares, 0 destroyed; 41 spares now, 1216 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1001, 4442.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 418 dirty pages from their submits' captures, 56919 from earlier ones', 28 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1043 dispatches, 43485 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8846 runs, 731.3 MB compared; 53 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20674 in captures, 0 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 28; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 15740000 (9719726 pages from the captures) [hos host] vk: program picks: 22124245 kept (9277010 by the last picks' textures, a uniform buffer bound since), 26461 new, 8049293 with other textures (of 30200000) [hos ModuleSystemWorker1] tex: the game's texture heap: 331 compaction rounds asked for in 18000 frames without one running [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 579 ranges, 3531 MB), 292 pages written, 292 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 25 kept (122 MB), 9 spare buffers, 6319 texture pages copied (24 MB) [hos host] vk: replay reads of guest memory as submitted: 15760000 (9726878 pages from the captures) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 277 dirty pages from their submits' captures, 59599 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1035 dispatches, 45638 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9360 runs, 770.6 MB compared; 72 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21839 in captures, 2 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7f86400000 (8 MB, captured, state 2): 1; [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: replay reads of guest memory as submitted: 15780000 (9736021 pages from the captures) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1140 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1303 kept their virtual core (nothing there to wait for), 3305 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture #24295 0x10009c02a00 1280x720 x1 NVN 0x29 the GPU drew, finalized by the game, sampled again at frame 20126 (drawn at 20065) [hos host] vk: texture #24291 0x10008c12600 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20126 (drawn at 20065) [hos host] vk: sparse mirrors: 120, 1062 MB with memory of 1528 MB (465 MB spared); blocks given memory later 13773, given back 17017; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1329 binds (310.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 345 taken from the spares, 165 made new; 390 kept as spares, 0 destroyed; 45 spares now, 1292 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 283 dirty pages from their submits' captures, 57059 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 971 dispatches, 43677 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8960 runs, 732.9 MB compared; 47 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20920 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos ModuleSystemWorker5] nvn: 2650000 recordings made, 74449 live copies (24.5 MB), 6593 texture copies; 6 command buffers' segments (4.3 MB) [hos sead::MainThread] nvn: CommandBufferWaitEvent #385000: event 0x100d6c800c mode 0 value 282099 (now 282080) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 15800000 (9744428 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14238 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a frame: 172 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.27, present 0.18, draw 6.38, descriptors 2.86, texture lookups 0.26, uploads 0.28, pipelines 0.27, programs 1.05, commands 1.73, dispatch 1.01, memory sync 0.48, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 511 (1054 MB): 5.60 MB synced in 1080.6 copies, 18.34 MB written back [hos host] vk: 20160 presents, 60.0 fps; a frame: 1682 draws (0 skipped), 180 dispatches, 168 passes, 144 barriers, 6 submits, 0.8 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1130 textures (2265 MB) [hos host] vk: host time in GPU resource waits: 1.90 ms a frame (fences 1.42, timelines 0.49, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 662.7 full barriers recorded, 297.6 left out (nothing recorded since the one before), 17.2 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.9, an op 897.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.2 ms); 1.1 textures made a frame: colour targets 0.3 (1.9 MB), depth 0.4 (1.7 MB), ASTC or BC 0.4 (0.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.53 MB unswizzled on the CPU in 0.24 ms [hos sead::MainThread] nvn: GPU waits in 5 s: 37 sync waits (68 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 578 ranges, 3535 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (126 MB), 10 spare buffers, 7624 texture pages copied (29 MB) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 764 dirty pages from their submits' captures, 54800 from earlier ones', 3 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 938 dispatches, 41514 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8718 runs, 714.3 MB compared; 59 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20212 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 3; [hos host] vk: replay reads of guest memory as submitted: 15820000 (9754637 pages from the captures) [hos host] vk: program picks: 22272273 kept (9344177 by the last picks' textures, a uniform buffer bound since), 26490 new, 8101236 with other textures (of 30400000) [hos Presentation Thread] vi: 20194 frames presented (59.9 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 26, 15.5-17.8 251, 17.8-22 20, 22-30 3, 30-40 0, over 40 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.06 [hos host] vk: texture #24398 0x10008942600 640x360 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20196 (drawn at 20135) [hos host] vk: replay reads of guest memory as submitted: 15840000 (9764386 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 6625116 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 115, 1045 MB with memory of 1496 MB (450 MB spared); blocks given memory later 13773, given back 17082; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1338 binds (312.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 347 taken from the spares, 165 made new; 397 kept as spares, 0 destroyed; 50 spares now, 1324 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 423 dirty pages from their submits' captures, 55479 from earlier ones', 82 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 933 dispatches, 42292 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9164 runs, 727.6 MB compared; 51 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21058 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 80; 0x7f8a000000 (4 MB, captured, state 2): 2; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 15860000 (9773230 pages from the captures) [hos host] vk: texture #24751 0x10008612200 860x860 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20248 (drawn at 20187) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 285 dirty pages from their submits' captures, 55374 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 933 dispatches, 42317 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9011 runs, 724.1 MB compared; 50 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20907 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15397 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: recorder: 5663800 packets run in 10.0 s (busy 1.70 s, spun idle 0.12 s, slept 10672 times, 3229 of them under 200 us), 0 calls made directly (no room); 17934 batches over 599 frames (29.9 a frame, 316 packets a batch), 8843 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 182 times, 20.98 ms (0.3 times, 0.04 ms a frame); every thread 790 times, 57.91 ms in all, the longest 0.85 ms; vkWaitForFences 510/29.61 ms (max 0.49), vkDestroyImageView 189/20.30 ms (max 0.53), vmaFreeMemory 23/2.64 ms (max 0.32), vkQueueBindSparse 33/2.25 ms (max 0.41), vmaDestroyBuffer 10/1.24 ms (max 0.85), a part's submit 5/0.73 ms (max 0.28) [hos host] vk: recorder: of its busy 1.70 s, 0.41 s in parts' ends and submits (the longest 1.59 ms), 0.08 s of that blocked on the queue's lock (164 times, the longest 1.53 ms), so 0.33 s in the ends, submits and holds; 1.29 s running the other packets; 9455 packets a frame [hos host] vk: 4 textures retired (1216 left, 2428 MB): budget 0, drawn over 4, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15880000 (9781472 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 576 ranges, 3538 MB), 278 pages written, 278 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.015, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 5954 texture pages copied (23 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 7 textures retired (1211 left, 2413 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: texture #24832 0x10008612200 885x885 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20275 (drawn at 20213) [hos host] vk: program picks: 22420382 kept (9407784 by the last picks' textures, a uniform buffer bound since), 26538 new, 8153079 with other textures (of 30600000) [hos host] vk: 3 textures retired (1212 left, 2416 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: a frame: 253 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.91, present 0.20, draw 6.83, descriptors 3.03, texture lookups 0.23, uploads 0.30, pipelines 0.28, programs 1.19, commands 1.80, dispatch 1.03, memory sync 0.54, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 512 (1045 MB): 5.57 MB synced in 1077.6 copies, 18.47 MB written back [hos host] vk: 20280 presents, 59.9 fps; a frame: 1815 draws (0 skipped), 183 dispatches, 165 passes, 145 barriers, 6 submits, 0.8 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1212 textures (2416 MB) [hos host] vk: host time in GPU resource waits: 1.46 ms a frame (fences 1.18, timelines 0.28, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 667.0 full barriers recorded, 299.6 left out (nothing recorded since the one before), 13.7 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 38.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.7, an op 932.4, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.0 ms); 0.9 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.5 (1.3 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.56 MB unswizzled on the CPU in 0.26 ms [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1150 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 11 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1056 kept their virtual core (nothing there to wait for), 2784 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos BaseProcExecutor1] vcores: 154079 takes, 22438 waits (mean 71 us, longest 7.6 ms, 0 over HOS_VCORE_LOG_MS), 10871 kicks, 10833 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 2 textures retired (1218 left, 2412 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15900000 (9791960 pages from the captures) [hos host] vk: sparse mirrors: 115, 1045 MB with memory of 1496 MB (450 MB spared); blocks given memory later 13774, given back 17083; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1340 binds (312.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 347 taken from the spares, 165 made new; 397 kept as spares, 0 destroyed; 50 spares now, 1324 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 473 dirty pages from their submits' captures, 57027 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 994 dispatches, 43820 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8754 runs, 743.6 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20583 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #285000: event 0x100d6c803c mode 0 value 284999 (lr 0x80661bfc) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: a submission's parts ended: 98860 where the game had ended its render pass, 21140 at 512 draws [hos host] vk: 8 textures retired (1219 left, 2402 MB): budget 0, drawn over 8, dead targets 0 [hos host] vk: 5 textures retired (1221 left, 2406 MB): budget 0, drawn over 5, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 15920000 (9801987 pages from the captures) [hos host] vk: 7 textures retired (1219 left, 2400 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: texture #20276 0x116fb36000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 20321 (drawn at 14859) [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 665 dirty pages from their submits' captures, 57417 from earlier ones', 12 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1000 dispatches, 44083 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8683 runs, 729.4 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21155 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 12; [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #390000: event 0x100d6c800c mode 0 value 285899 (now 285880) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 15940000 (9811727 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.05 over 576 ranges, 3540 MB), 277 pages written, 277 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.033; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.022, bytes copied 0.000; 25 kept (100 MB), 11 spare buffers, 4162 texture pages copied (16 MB) [hos host] vk: presenter thread: the copy's fence 1.33 ms, the present 0.73 ms a present [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 10276 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 117, 1050 MB with memory of 1504 MB (453 MB spared); blocks given memory later 13774, given back 17084; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1343 binds (313.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 349 taken from the spares, 165 made new; 397 kept as spares, 0 destroyed; 48 spares now, 1316 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 110 dirty pages from their submits' captures, 60353 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1017 dispatches, 45740 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9185 runs, 770.3 MB compared; 56 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22467 in captures, 3 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 3; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 15960000 (9820548 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 22568243 kept (9471445 by the last picks' textures, a uniform buffer bound since), 26538 new, 8205218 with other textures (of 30800000) [hos host] vk: 2 textures retired (1168 left, 2224 MB): budget 0, drawn over 0, dead targets 2 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos BaseProcExecutor1] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 15888477 lookups, 129796 missed (0.82 %) [hos host] vk: memory: working set 14733 MB (peak 16021), private 21447 MB, staging 80 MB in 5 arenas, 47 made so far (212 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 858 MB; used 3216 of 15579 MB; heap 1 (device): 157 blocks 4759 MB, 18934 allocations 4056 MB; used 6882 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 22 blocks 704 MB, 159 images 390 MB; 95 free ranges, the largest 31 MB [hos host] vk: image blocks (textures, compacted): 43 blocks 1376 MB, 1019 images 1278 MB; 321 free ranges, the largest 24 MB [hos host] vk: render target images: 6493 reused from the pool, 5216 made new, 5055 destroyed from it, 1309 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 166 missed after their image was taken by size; it holds 6 (51 MB) [hos host] vk: render targets not reused by size: 2467 with the pool empty (0 of them with targets still in the dead queue), 380 with none in it a present yet, 59 with none of a kind they can use, 651 with all smaller (the largest 50% short on average), 705 with all bigger (over 1.5 times), 951 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 4863 idle [hos host] vk: VRAM: 6882 MB used of 11474 MB allowed (703 MB of it free in the blocks); textures 2253 MB (budget 7158 MB), 1178 textures [hos host] vk: budget: VRAM 4759 of 11474 MB (6882 MB by the driver, 2123 MB unaccounted): textures 2253, mirrors 1046, blocks free 703, shadows 284, other buffers 229, other images 111, staging 80, target pool 51; RAM 3161 MB (the process's private memory 21447 MB): imports 2260, capture spares 750, captures 96, capture table 35, other buffers 12, retired imports 8, others 0 [hos host] vk: textures: decoded ASTC 1179 MB, rendered 552 MB (77 drawn in one frame, kept: 49159 KB), other 471 MB; used in the last 8 presents 1644 MB [hos host] vk: decoded ASTC used again: 577 textures (1277 MB) copied from ones the game finalized, not decoded; 577 of those retired, 4 listed; 2411 MB of blocks hashed in 397 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 15253 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34, last used 230 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 82 (15946 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 136 waited for it, 1.21 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 24.7 ms; VRAM 6.7 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6476 queued, 6476 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16362 ms and 348.1 MB read in all); 8781 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12295 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2395 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9632 (6476 ready, 0 loading), 17569 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13980 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 291 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.88, present 0.22, draw 6.81, descriptors 3.10, texture lookups 0.34, uploads 0.30, pipelines 0.28, programs 1.11, commands 1.80, dispatch 0.99, memory sync 0.55, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 514 (1046 MB): 5.62 MB synced in 1082.2 copies, 18.29 MB written back [hos host] vk: 20400 presents, 60.1 fps; a frame: 1854 draws (0 skipped), 176 dispatches, 166 passes, 144 barriers, 6 submits, 0.7 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1179 textures (2256 MB) [hos host] vk: host time in GPU resource waits: 1.68 ms a frame (fences 1.32, timelines 0.36, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 660.9 full barriers recorded, 298.8 left out (nothing recorded since the one before), 13.2 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.1, an op 930.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.7 ms a frame (the longest submission 3.0 ms); 1.2 textures made a frame: colour targets 0.4 (1.9 MB), depth 0.7 (2.3 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.62 MB unswizzled on the CPU in 0.26 ms [hos host] vk: the render thread (600 presents, 16.66 ms a frame): its jobs 14.50 ms a frame (submits 8.07, presents 3.98, others 2.46), idle 2.16; the game waited for room at 17 presents, 0.18 ms each (0.01 a frame; 7.25 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 142 captures at a CPU signal; the submits' waits read in 0.041 ms a frame; readbacks: 1006 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 13.62 ms after its hand-over (18.67 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.31 + 0.31 (the window's copy is in the idle), after it 1.08 + 0.35; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.57 parts a frame recorded in the signal's job, 79.9% of them took nothing that came with the signal; of the others, 1.72 a frame took pages of the signal's capture into a mirror (25.8 pages a frame) and 0.05 pages read as guest memory had them (0.7 pages a frame); a frame's first part took nothing in 18 of 142 frames, its first three in 18 [hos host] vk: frame ahead, what came with the signal: 142 signals with a capture, each of 65.7 pages in 8.2 pools; between a frame's last submit and that signal 0.01 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (1 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.13+0.06, 0.44+0.14, 0.04+0.02, 0.06+0.02, 0.12+0.03, 0.16+0.03, 0.08+0.02, 0.05+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: replay reads of guest memory as submitted: 15980000 (9830084 pages from the captures) [hos host] vk: render targets drawn over from another start, let go: 12228 so far (57365 MB), 11040 of them (51903 MB) between the passes [hos host] vk: 2 textures retired (1184 left, 2270 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 6692530 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 1001, 4442.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 217 dirty pages from their submits' captures, 57275 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1021 dispatches, 43817 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9070 runs, 752.0 MB compared; 53 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21633 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2; [hos host] vk: 2 textures retired (1189 left, 2274 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 16000000 (9840143 pages from the captures) [hos host] vk: reaper: 703 images (2217 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 86 (142 MB) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1160 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 7 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1164 kept their virtual core (nothing there to wait for), 2932 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos ModuleSystemWorker1] nvn: DrawMeshTasksIndirect: 700000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040) [hos host] vk: sparse mirrors: 116, 1046 MB with memory of 1500 MB (453 MB spared); blocks given memory later 13774, given back 17123; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1346 binds (313.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 349 taken from the spares, 165 made new; 398 kept as spares, 0 destroyed; 49 spares now, 1320 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 100 dirty pages from their submits' captures, 59660 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1035 dispatches, 45414 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8818 runs, 767.6 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21630 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 3; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 570 ranges, 3528 MB), 281 pages written, 281 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.015; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 6665 texture pages copied (26 MB) [hos host] vk: replay reads of guest memory as submitted: 16020000 (9849568 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13170 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] nvn: GPU waits in 5 s: 48 sync waits (70 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 22716214 kept (9536998 by the last picks' textures, a uniform buffer bound since), 26545 new, 8257240 with other textures (of 31000000) [hos host] vk: replay reads of guest memory as submitted: 16040000 (9858899 pages from the captures) [hos Presentation Thread] vi: 20495 frames presented (60.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 17, 15.5-17.8 269, 17.8-22 15, 22-30 0, 30-40 0, over 40 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos host] vk: write-backs left out (600 submits): 1049, 4660.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 163 dirty pages from their submits' captures, 58994 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1010 dispatches, 44988 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8769 runs, 770.5 MB compared; 70 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21866 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; 0x7fa9000000 (4 MB, captured, state 2): 1; [hos host] vk: replay reads of guest memory as submitted: 16060000 (9868323 pages from the captures) [hos host] vk: a frame: 253 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.66, present 0.22, draw 5.94, descriptors 2.61, texture lookups 0.32, uploads 0.12, pipelines 0.25, programs 1.01, commands 1.60, dispatch 0.90, memory sync 0.51, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 514 (1046 MB): 5.55 MB synced in 1077.0 copies, 18.33 MB written back [hos host] vk: 20520 presents, 59.5 fps; a frame: 1799 draws (0 skipped), 168 dispatches, 167 passes, 144 barriers, 6 submits, 0.3 uploads (0.5 MB), 81 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1208 textures (2296 MB) [hos host] vk: host time in GPU resource waits: 1.92 ms a frame (fences 1.44, timelines 0.48, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 656.1 full barriers recorded, 299.2 left out (nothing recorded since the one before), 13.6 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.8; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.3, an op 923.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.8 ms a frame (the longest submission 5.1 ms); 1.1 textures made a frame: colour targets 0.3 (1.2 MB), depth 0.7 (2.1 MB), ASTC or BC 0.1 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.54 MB unswizzled on the CPU in 0.10 ms [hos host] vk: sparse mirrors: 116, 1046 MB with memory of 1500 MB (453 MB spared); blocks given memory later 13774, given back 17127; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1348 binds (314.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 349 taken from the spares, 165 made new; 398 kept as spares, 0 destroyed; 49 spares now, 1320 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1169 dirty pages from their submits' captures, 53943 from earlier ones', 4 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 956 dispatches, 41809 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8223 runs, 715.7 MB compared; 88 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19892 in captures, 2 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: 0x7fa9000000 (4 MB, captured, state 2): 4; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #395000: event 0x100d6c800c mode 0 value 289699 (now 289680) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 16080000 (9878570 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.18 ms (lock 0.00, scan 0.05 over 568 ranges, 3566 MB), 309 pages written, 309 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.030; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.023, bytes copied 0.000; 25 kept (114 MB), 10 spare buffers, 6452 texture pages copied (25 MB) [hos sead::MainThread] nvn: CommandBufferSignalEvent #290000: event 0x100d6c8048 mode 0 value 289999 (lr 0x80666598) [hos host] vk: replay reads of guest memory as submitted: 16100000 (9889211 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 231 dirty pages from their submits' captures, 58426 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1015 dispatches, 44291 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8518 runs, 748.9 MB compared; 58 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21566 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11012 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos rmbl::VibrationThread] vcores: 154113 takes, 24103 waits (mean 77 us, longest 9.4 ms, 0 over HOS_VCORE_LOG_MS), 11632 kicks, 11559 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: late pages found at the replay (600 takes): 155 pages in 2 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 22862211 kept (9603557 by the last picks' textures, a uniform buffer bound since), 26549 new, 8311239 with other textures (of 31200000) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: replay reads of guest memory as submitted: 16120000 (9898877 pages from the captures) [hos sead::MainThread] nvn: 2700000 recordings made, 74154 live copies (24.8 MB), 6604 texture copies; 3 command buffers' segments (5.1 MB) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1116 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 2 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1123 kept their virtual core (nothing there to wait for), 2973 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: readback pages refreshed from guest memory: 6760273 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 118, 1045 MB with memory of 1508 MB (462 MB spared); blocks given memory later 13774, given back 17142; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1353 binds (315.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 351 taken from the spares, 165 made new; 398 kept as spares, 0 destroyed; 47 spares now, 1312 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 253 dirty pages from their submits' captures, 57246 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 996 dispatches, 44203 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8654 runs, 761.3 MB compared; 67 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21028 in captures, 2 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 16140000 (9907662 pages from the captures) [hos host] vk: a frame: 229 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.81, present 0.19, draw 6.16, descriptors 2.77, texture lookups 0.23, uploads 0.18, pipelines 0.25, programs 1.03, commands 1.63, dispatch 0.90, memory sync 0.50, queue submit 0.00, slot wait 0.22 [hos host] vk: a frame: mirrors 516 (1045 MB): 5.47 MB synced in 1063.9 copies, 18.19 MB written back [hos host] vk: 20640 presents, 60.5 fps; a frame: 1759 draws (0 skipped), 165 dispatches, 167 passes, 144 barriers, 6 submits, 0.5 uploads (1.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1209 textures (2270 MB) [hos host] vk: host time in GPU resource waits: 1.48 ms a frame (fences 1.20, timelines 0.29, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 655.9 full barriers recorded, 299.9 left out (nothing recorded since the one before), 13.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.0, an op 850.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.6 ms a frame (the longest submission 4.1 ms); 1.0 textures made a frame: colour targets 0.2 (1.5 MB), depth 0.6 (2.1 MB), ASTC or BC 0.1 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.03 MB unswizzled on the CPU in 0.17 ms [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 563 ranges, 3530 MB), 270 pages written, 270 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.021; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (162 MB), 9 spare buffers, 8065 texture pages copied (31 MB) [hos host] vk: replay reads of guest memory as submitted: 16160000 (9917325 pages from the captures) [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 195 dirty pages from their submits' captures, 58028 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1018 dispatches, 44090 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8656 runs, 728.1 MB compared; 58 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21721 in captures, 4 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 14894 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 16180000 (9927664 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: program picks: 23008893 kept (9666025 by the last picks' textures, a uniform buffer bound since), 26567 new, 8364539 with other textures (of 31400000) [hos host] vk: sparse mirrors: 118, 1045 MB with memory of 1508 MB (463 MB spared); blocks given memory later 13774, given back 17157; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1355 binds (315.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 351 taken from the spares, 165 made new; 398 kept as spares, 0 destroyed; 47 spares now, 1312 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 976, 4331.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 137 dirty pages from their submits' captures, 58027 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 976 dispatches, 43588 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9748 runs, 759.8 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22669 in captures, 2 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 3; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 16200000 (9937180 pages from the captures) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 227 dirty pages from their submits' captures, 57256 from earlier ones', 10 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 986 dispatches, 43744 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9052 runs, 746.9 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21764 in captures, 1 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 10; [hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #400000: event 0x100d6c800c mode 0 value 293499 (now 293480) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 16220000 (9946072 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.05 over 564 ranges, 3530 MB), 280 pages written, 280 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.015, bytes copied 0.000; 25 kept (108 MB), 10 spare buffers, 5013 texture pages copied (19 MB) [hos host] vk: a frame: 280 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.74, present 0.17, draw 6.03, descriptors 2.66, texture lookups 0.15, uploads 0.22, pipelines 0.26, programs 1.00, commands 1.62, dispatch 0.95, memory sync 0.47, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 517 (1042 MB): 5.66 MB synced in 1092.5 copies, 18.63 MB written back [hos host] vk: 20760 presents, 60.0 fps; a frame: 1794 draws (0 skipped), 185 dispatches, 166 passes, 144 barriers, 6 submits, 0.5 uploads (1.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1269 textures (2428 MB) [hos host] vk: host time in GPU resource waits: 1.38 ms a frame (fences 1.35, timelines 0.03, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 672.6 full barriers recorded, 300.3 left out (nothing recorded since the one before), 13.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.5, an op 812.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.5 ms a frame (the longest submission 3.6 ms); 0.9 textures made a frame: colour targets 0.1 (0.9 MB), depth 0.7 (2.0 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.44 MB unswizzled on the CPU in 0.22 ms [hos sead::MainThread] nvn: GPU waits in 5 s: 29 sync waits (84 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: 400 mirrors given back (2601 MB in all); 1040 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1140 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1195 kept their virtual core (nothing there to wait for), 3157 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 16240000 (9955329 pages from the captures) [hos host] vk: texture #25259 0x10008612200 843x843 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20777 (drawn at 20716) [hos host] vk: 1 textures retired (1289 left, 2528 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 117, 1040 MB with memory of 1500 MB (459 MB spared); blocks given memory later 13774, given back 17159; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1360 binds (316.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 352 taken from the spares, 165 made new; 400 kept as spares, 0 destroyed; 48 spares now, 1320 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 190 dirty pages from their submits' captures, 58337 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1023 dispatches, 44618 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8525 runs, 756.3 MB compared; 55 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21508 in captures, 4 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; 0x7fa9000000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: queues ordered by their events: 138000 submissions stopped at a wait, 137936 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 401012, not met 0; captures at a CPU event's wait 12896, 0 of them at a wait met on arrival [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15032 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: texture #25323 0x10008612200 822x822 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20790 (drawn at 20728) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.05 [hos Presentation Thread] vi: 20796 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 21, 15.5-17.8 270, 17.8-22 8, 22-30 2, 30-40 0, over 40 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 199 textures retired (1111 left, 2383 MB): budget 172, drawn over 27, dead targets 0; GPU-drawn with no copy kept: 0 live, 4 finalized, 6 MB [hos host] vk: replay reads of guest memory as submitted: 16260000 (9964133 pages from the captures) [hos host] vk: program picks: 23156325 kept (9731044 by the last picks' textures, a uniform buffer bound since), 26567 new, 8417107 with other textures (of 31600000) [hos host] vk: 46 textures retired (1075 left, 2364 MB): budget 37, drawn over 9, dead targets 0 [hos host] vk: texture #25291 0x10008612200 811x811 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 20815 (drawn at 20753) [hos sead::MainThread] nvn: CommandBufferSignalEvent #295000: event 0x100d6c8010 mode 0 value 295001 (lr 0x80660bcc) [hos host] vk: 9 textures retired (1057 left, 2226 MB): budget 0, drawn over 9, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 6827562 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 175 dirty pages from their submits' captures, 58259 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1029 dispatches, 44496 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8569 runs, 756.1 MB compared; 59 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21281 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; [hos host] vk: 10 textures retired (1056 left, 2232 MB): budget 0, drawn over 10, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 16280000 (9972653 pages from the captures) [hos host] vk: 8 textures retired (1055 left, 2216 MB): budget 0, drawn over 8, dead targets 0 [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 564 ranges, 3480 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.023; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.019, bytes copied 0.000; 24 kept (100 MB), 11 spare buffers, 8046 texture pages copied (31 MB) [hos ModuleSystemWorker2] nvn: DrawMeshTasksIndirect: 720000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040) [hos host] vk: 2 textures retired (1062 left, 2239 MB): budget 0, drawn over 2, dead targets 0 [hos host] vk: recorder: 5650235 packets run in 10.0 s (busy 1.68 s, spun idle 0.12 s, slept 10557 times, 3178 of them under 200 us), 0 calls made directly (no room); 17963 batches over 600 frames (29.9 a frame, 315 packets a batch), 8767 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 133 times, 17.73 ms (0.2 times, 0.03 ms a frame); every thread 719 times, 52.69 ms in all, the longest 0.56 ms; vkWaitForFences 505/29.97 ms (max 0.56), vkDestroyImageView 154/19.26 ms (max 0.44), vmaFreeMemory 18/1.57 ms (max 0.53), vkQueueBindSparse 24/1.40 ms (max 0.38), vmaDestroyBuffer 12/0.22 ms (max 0.05), the queue's lock 2/0.16 ms (max 0.13) [hos host] vk: recorder: of its busy 1.68 s, 0.37 s in parts' ends and submits (the longest 1.32 ms), 0.05 s of that blocked on the queue's lock (107 times, the longest 1.28 ms), so 0.32 s in the ends, submits and holds; 1.31 s running the other packets; 9417 packets a frame [hos rmbl::VibrationThread] hid: 69632 vibration values, strongest 0.02 since the last count [hos host] vk: 17 textures retired (1053 left, 2239 MB): budget 13, drawn over 4, dead targets 0 [hos host] vk: sparse mirrors: 118, 1043 MB with memory of 1508 MB (464 MB spared); blocks given memory later 13778, given back 17167; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1366 binds (317.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 353 taken from the spares, 165 made new; 400 kept as spares, 0 destroyed; 47 spares now, 1312 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: texture #21167 0x117144e000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 20866 (drawn at 16210) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 178 dirty pages from their submits' captures, 58894 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1031 dispatches, 44856 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8688 runs, 757.0 MB compared; 54 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21553 in captures, 3 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 16300000 (9981543 pages from the captures) [hos host] vk: 9 textures retired (1052 left, 2207 MB): budget 0, drawn over 9, dead targets 0 [hos host] vk: a frame: 240 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.54, present 0.20, draw 5.73, descriptors 2.52, texture lookups 0.40, uploads 0.18, pipelines 0.24, programs 0.96, commands 1.56, dispatch 0.89, memory sync 0.49, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 518 (1043 MB): 5.47 MB synced in 1057.9 copies, 18.01 MB written back [hos host] vk: 20880 presents, 60.0 fps; a frame: 1720 draws (0 skipped), 164 dispatches, 164 passes, 144 barriers, 6 submits, 0.4 uploads (1.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1071 textures (2251 MB) [hos host] vk: host time in GPU resource waits: 1.68 ms a frame (fences 1.47, timelines 0.21, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 640.8 full barriers recorded, 296.3 left out (nothing recorded since the one before), 13.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.8 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.9, an op 816.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.1 ms a frame (the longest submission 4.1 ms); 1.2 textures made a frame: colour targets 0.2 (1.2 MB), depth 0.9 (4.9 MB), ASTC or BC 0.1 (0.1 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.16 MB unswizzled on the CPU in 0.18 ms [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6656 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos NavMeshCAStepDepWorker1] vcores: 152759 takes, 23202 waits (mean 69 us, longest 7.1 ms, 0 over HOS_VCORE_LOG_MS), 11116 kicks, 10990 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 15 textures retired (1061 left, 2189 MB): budget 0, drawn over 15, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 16320000 (9991066 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 283 dirty pages from their submits' captures, 55335 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 964 dispatches, 42312 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8372 runs, 721.5 MB compared; 50 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20489 in captures, 1 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x1043000000 (4 MB, captured, state 2): 3; [hos host] vk: 14 textures retired (1069 left, 2195 MB): budget 0, drawn over 14, dead targets 0 [hos host] vk: program picks: 23302135 kept (9796634 by the last picks' textures, a uniform buffer bound since), 26569 new, 8471295 with other textures (of 31800000) [hos host] vk: replay reads of guest memory as submitted: 16340000 (9999977 pages from the captures) [hos ModuleSystemWorker2] vk: readbacks of GPU results: 1140 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 6 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 1148 kept their virtual core (nothing there to wait for), 2948 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #405000: event 0x100d6c800c mode 0 value 297299 (now 297280) (lr 0x80660e38) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.05 over 562 ranges, 3395 MB), 276 pages written, 276 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (150 MB), 10 spare buffers, 4312 texture pages copied (16 MB) [hos host] vk: sparse mirrors: 117, 1041 MB with memory of 1496 MB (454 MB spared); blocks given memory later 13778, given back 17167; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1367 binds (318.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 353 taken from the spares, 165 made new; 401 kept as spares, 0 destroyed; 48 spares now, 1324 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 999, 4438.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 187 dirty pages from their submits' captures, 56443 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 970 dispatches, 42926 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8351 runs, 721.1 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20310 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 16360000 (10009149 pages from the captures) [hos host] vk: 22 textures retired (1087 left, 2201 MB): budget 0, drawn over 22, dead targets 0 [hos host] vk: 8 textures retired (1087 left, 2193 MB): budget 0, drawn over 8, dead targets 0 [hos host] vk: presenter thread: the copy's fence 1.35 ms, the present 0.65 ms a present [hos host] vk: replay reads of guest memory as submitted: 16380000 (10019497 pages from the captures) [hos host] vk: 13 textures retired (1086 left, 2192 MB): budget 0, drawn over 13, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1001, 4442.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 302 dirty pages from their submits' captures, 56648 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1006 dispatches, 43240 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8568 runs, 725.2 MB compared; 46 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21443 in captures, 3 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 7 textures retired (1083 left, 2180 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 18676 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos sead::MainThread] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 40 made): 15527561 lookups, 126666 missed (0.82 %) [hos host] vk: memory: working set 14750 MB (peak 16021), private 21721 MB, staging 80 MB in 5 arenas, 47 made so far (212 ms) [hos host] vk: memory heaps: heap 0 (system): 14 blocks 944 MB, 34 allocations 856 MB; used 3232 of 15579 MB; heap 1 (device): 160 blocks 4855 MB, 19536 allocations 4016 MB; used 6978 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 24 blocks 768 MB, 193 images 504 MB; 124 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 44 blocks 1408 MB, 902 images 1082 MB; 370 free ranges, the largest 23 MB [hos host] vk: render target images: 6549 reused from the pool, 5581 made new, 5386 destroyed from it, 1386 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 172 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 2717 with the pool empty (0 of them with targets still in the dead queue), 394 with none in it a present yet, 59 with none of a kind they can use, 687 with all smaller (the largest 50% short on average), 753 with all bigger (over 1.5 times), 968 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 5194 idle [hos host] vk: VRAM: 6978 MB used of 11474 MB allowed (839 MB of it free in the blocks); textures 2221 MB (budget 7158 MB), 1101 textures [hos host] vk: budget: VRAM 4855 of 11474 MB (6978 MB by the driver, 2123 MB unaccounted): textures 2221, mirrors 1089, blocks free 839, shadows 284, other buffers 229, other images 111, staging 80; RAM 3167 MB (the process's private memory 21721 MB): imports 2260, capture spares 746, captures 118, capture table 35, retired imports 8, others 0 [hos host] vk: textures: decoded ASTC 1081 MB, rendered 687 MB (77 drawn in one frame, kept: 54095 KB), other 385 MB; used in the last 8 presents 1716 MB [hos host] vk: decoded ASTC used again: 581 textures (1288 MB) copied from ones the game finalized, not decoded; 581 of those retired, 1 listed; 2424 MB of blocks hashed in 399 ms [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 15853 presents ago [hos host] vk: 21 MB: 0x10006fce000 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d7e62a00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 446 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 87 (15377 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 111 waited for it, 1.30 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 26.7 ms; VRAM 6.8 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 12, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6476 queued, 6476 loaded (median 1.46 ms, 99th percentile 16.77 ms, worst 38.13 ms, 16362 ms and 348.1 MB read in all); 8781 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12295 graphics pipelines (1 not in a cache), 720 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2407 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9632 (6476 ready, 0 loading), 17567 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 13980 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 764 known, 7 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 216 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.77, present 0.19, draw 6.12, descriptors 2.59, texture lookups 0.19, uploads 0.02, pipelines 0.27, programs 1.06, commands 1.72, dispatch 0.90, memory sync 0.48, queue submit 0.00, slot wait 0.24 [hos host] vk: a frame: mirrors 519 (1089 MB): 5.90 MB synced in 1068.5 copies, 18.10 MB written back [hos host] vk: 21000 presents, 60.0 fps; a frame: 1828 draws (0 skipped), 169 dispatches, 166 passes, 144 barriers, 6 submits, 0.3 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14162 pipelines (3728 ms), 0 programs (10 failed), 1102 textures (2226 MB) [hos host] vk: host time in GPU resource waits: 1.70 ms a frame (fences 1.49, timelines 0.21, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 651.2 full barriers recorded, 297.9 left out (nothing recorded since the one before), 13.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.0, an op 895.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.4 ms a frame (the longest submission 3.7 ms); 0.9 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.6 (2.1 MB), ASTC or BC 0.3 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.07 MB unswizzled on the CPU in 0.01 ms [hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 13.90 ms a frame (submits 8.10, presents 3.99, others 1.82), idle 2.76; the game waited for room at 7 presents, 0.57 ms each (0.01 a frame; 4.32 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 114 captures at a CPU signal; the submits' waits read in 0.042 ms a frame; readbacks: 704 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 13.02 ms after its hand-over (25.09 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 4.71 + 0.30 (the window's copy is in the idle), after it 0.81 + 0.35; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.54 parts a frame recorded in the signal's job, 80.1% of them took nothing that came with the signal; of the others, 1.70 a frame took pages of the signal's capture into a mirror (22.8 pages a frame) and 0.02 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 18 of 114 frames, its first three in 18 [hos host] vk: frame ahead, what came with the signal: 114 signals with a capture, each of 37.5 pages in 7.9 pools; between a frame's last submit and that signal 0.03 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (2 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.14+0.06, 0.35+0.13, 0.03+0.02, 0.07+0.03, 0.05+0.03, 0.07+0.03, 0.06+0.02, 0.04+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos host] vk: replay reads of guest memory as submitted: 16400000 (10028853 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vi: real-time step on: ratio 1.122, game's frame rate 60 [hos host] vk: render targets drawn over from another start, let go: 12537 so far (58776 MB), 11348 of them (53314 MB) between the passes [hos host] vk: program picks: 23448414 kept (9864410 by the last picks' textures, a uniform buffer bound since), 26573 new, 8525012 with other textures (of 32000000) [hos host] vk: readback pages refreshed from guest memory: 6895265 pages so far (0 held: may be written ahead in their submission) [hos host] vk: replay reads of guest memory as submitted: 16420000 (10040469 pages from the captures) [hos host] vk: sparse mirrors: 119, 1090 MB with memory of 1560 MB (469 MB spared); blocks given memory later 13784, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1371 binds (318.6 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 355 taken from the spares, 165 made new; 401 kept as spares, 0 destroyed; 46 spares now, 1260 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1024, 4549.6 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 595 dirty pages from their submits' captures, 57557 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1047 dispatches, 44163 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9056 runs, 749.7 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21639 in captures, 3 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: 0x7fa6800000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos sead::MainThread] nvn: 2750000 recordings made, 74295 live copies (24.7 MB), 6657 texture copies; 6 command buffers' segments (4.3 MB) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.05 over 560 ranges, 3395 MB), 290 pages written, 290 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.028; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.020, bytes copied 0.000; 25 kept (104 MB), 10 spare buffers, 7312 texture pages copied (28 MB) [hos host] vk: 29 textures retired (1117 left, 2231 MB): budget 0, drawn over 29, dead targets 0 [hos host] vk: reaper: 565 images (1933 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 152 (281 MB) [hos host] vk: 10 textures retired (1114 left, 2215 MB): budget 0, drawn over 10, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 16440000 (10050102 pages from the captures) [hos host] vk: 7 textures retired (1113 left, 2210 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 101 dirty pages from their submits' captures, 56860 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1026 dispatches, 43334 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8659 runs, 727.1 MB compared; 48 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21362 in captures, 2 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 8 textures retired (1112 left, 2206 MB): budget 0, drawn over 8, dead targets 0 [hos sead::MainThread] nvn: GPU waits in 5 s: 26 sync waits (40 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: replay reads of guest memory as submitted: 16460000 (10061711 pages from the captures) [hos host] vk: 9 textures retired (1121 left, 2211 MB): budget 2, drawn over 7, dead targets 0 [hos ModuleSystemWorker3] nvn: CommandBufferSignalEvent #300000: event 0x100d6c8018 mode 0 value 300000 (lr 0x802915c8) [hos host] vk: 3 textures retired (1130 left, 2273 MB): budget 0, drawn over 3, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos Presentation Thread] vi: 21096 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 25, 15.5-17.8 259, 17.8-22 13, 22-30 3, 30-40 0, over 40 0 [hos host] vk: replay reads of guest memory as submitted: 16480000 (10072325 pages from the captures) [hos host] vk: 10 textures retired (1136 left, 2338 MB): budget 3, drawn over 7, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12055 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1128 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 9 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1104 kept their virtual core (nothing there to wait for), 2992 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: sparse mirrors: 120, 1091 MB with memory of 1564 MB (472 MB spared); blocks given memory later 13784, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1372 binds (318.7 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 356 taken from the spares, 165 made new; 401 kept as spares, 0 destroyed; 45 spares now, 1256 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 215 dirty pages from their submits' captures, 56560 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 959 dispatches, 43155 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8438 runs, 708.6 MB compared; 52 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21151 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: 15 textures retired (1130 left, 2379 MB): budget 7, drawn over 8, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture #25562 0x10008612200 933x933 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 21113 (drawn at 21051) [hos host] vk: program picks: 23594927 kept (9933142 by the last picks' textures, a uniform buffer bound since), 26582 new, 8578490 with other textures (of 32200000) [hos host] vk: 11 textures retired (1110 left, 2238 MB): budget 3, drawn over 8, dead targets 0 [hos ModuleSystemWorker3] barrier: 324307 looks: 127699 found the masks moved already, 193538 waited until they moved, 3070 for the whole 1 ms (a wait 360 us on average); 292555 wakes [hos ModuleSystemWorker5] barrier: 324307 looks: 127699 found the masks moved already, 193538 waited until they moved, 3070 for the whole 1 ms (a wait 360 us on average); 292555 wakes [hos host] vk: a frame: 226 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.31, present 0.20, draw 6.36, descriptors 2.75, texture lookups 0.33, uploads 0.08, pipelines 0.26, programs 1.11, commands 1.74, dispatch 0.93, memory sync 0.48, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 521 (1091 MB): 5.60 MB synced in 1084.3 copies, 18.21 MB written back [hos host] vk: 21120 presents, 60.0 fps; a frame: 1887 draws (0 skipped), 176 dispatches, 167 passes, 144 barriers, 6 submits, 0.4 uploads (0.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14189 pipelines (3740 ms), 0 programs (10 failed), 1110 textures (2238 MB) [hos host] vk: host time in GPU resource waits: 1.79 ms a frame (fences 1.39, timelines 0.40, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 662.5 full barriers recorded, 298.7 left out (nothing recorded since the one before), 13.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.9 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.9, an op 990.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.9 ms a frame (the longest submission 3.7 ms); 1.2 textures made a frame: colour targets 0.1 (0.4 MB), depth 0.8 (3.7 MB), ASTC or BC 0.3 (0.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.39 MB unswizzled on the CPU in 0.05 ms [hos host] vk: replay reads of guest memory as submitted: 16500000 (10082851 pages from the captures) [hos host] vk: texture #25681 0x10008612200 937x937 x1 NVN 0x34 the GPU drew, finalized by the game, sampled again at frame 21126 (drawn at 21065) [hos host] vk: 15 textures retired (1111 left, 2267 MB): budget 4, drawn over 6, dead targets 5 [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 559 ranges, 3393 MB), 279 pages written, 279 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (156 MB), 7 spare buffers, 8939 texture pages copied (34 MB) [hos host] vk: 13 textures retired (1110 left, 2239 MB): budget 2, drawn over 11, dead targets 0 [hos sead::MainThread] nvn: CommandBufferWaitEvent #410000: event 0x100d6c800c mode 0 value 301099 (now 301080) (lr 0x80660e38) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 193 dirty pages from their submits' captures, 56516 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 937 dispatches, 42830 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8792 runs, 730.4 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20813 in captures, 1 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 16520000 (10094154 pages from the captures) [hos host] vk: 7 textures retired (1111 left, 2225 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: 6 textures retired (1113 left, 2215 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: 8 textures retired (1114 left, 2215 MB): budget 4, drawn over 4, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 16540000 (10105063 pages from the captures) [hos host] vk: 7 textures retired (1116 left, 2208 MB): budget 0, drawn over 7, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 120, 1091 MB with memory of 1564 MB (472 MB spared); blocks given memory later 13788, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1375 binds (318.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 356 taken from the spares, 165 made new; 401 kept as spares, 0 destroyed; 45 spares now, 1256 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 111 dirty pages from their submits' captures, 57296 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 914 dispatches, 43293 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8949 runs, 720.4 MB compared; 51 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21168 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: 0x7fff800000 (8 MB, captured, state 2): 2; [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos ModuleSystemWorker1] vcores: 157468 takes, 27074 waits (mean 74 us, longest 28.1 ms, 0 over HOS_VCORE_LOG_MS), 12886 kicks, 12745 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: 11 textures retired (1109 left, 2179 MB): budget 3, drawn over 8, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: texture #21020 0x11708f8000 260x260 x1 NVN 0xd the GPU drew, finalized by the game, sampled again at frame 21197 (drawn at 15802) [hos host] vk: 7 textures retired (1114 left, 2180 MB): budget 1, drawn over 6, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 16560000 (10115140 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13225 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 10 textures retired (1116 left, 2186 MB): budget 0, drawn over 10, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 6962371 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 418 dirty pages from their submits' captures, 55819 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 42368 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8621 runs, 700.5 MB compared; 51 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20654 in captures, 4 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: program picks: 23737222 kept (9997105 by the last picks' textures, a uniform buffer bound since), 26584 new, 8636193 with other textures (of 32400000) [hos host] vk: replay reads of guest memory as submitted: 16580000 (10126875 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.05 over 556 ranges, 3222 MB), 293 pages written, 293 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 25 kept (104 MB), 10 spare buffers, 4675 texture pages copied (18 MB) [hos host] vk: a frame: 199 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.12, present 0.19, draw 6.11, descriptors 2.71, texture lookups 0.30, uploads 0.16, pipelines 0.26, programs 1.03, commands 1.64, dispatch 0.94, memory sync 0.51, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 521 (1091 MB): 5.75 MB synced in 1107.5 copies, 18.41 MB written back [hos host] vk: 21240 presents, 60.0 fps; a frame: 1729 draws (0 skipped), 182 dispatches, 164 passes, 144 barriers, 6 submits, 0.5 uploads (0.9 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14189 pipelines (3740 ms), 0 programs (10 failed), 1136 textures (2256 MB) [hos host] vk: host time in GPU resource waits: 1.75 ms a frame (fences 1.62, timelines 0.12, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 660.8 full barriers recorded, 297.1 left out (nothing recorded since the one before), 13.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.5, an op 908.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.3 ms a frame (the longest submission 4.7 ms); 1.1 textures made a frame: colour targets 0.1 (0.5 MB), depth 0.7 (2.8 MB), ASTC or BC 0.2 (0.2 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.87 MB unswizzled on the CPU in 0.12 ms [hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 740000 so far (x1 0x1178aa8000 x2 0x2 x3 0x1178aa8000 x4 0xf0040) [hos host] vk: replay reads of guest memory as submitted: 16600000 (10138535 pages from the captures) [hos host] vk: sparse mirrors: 119, 1091 MB with memory of 1560 MB (469 MB spared); blocks given memory later 13790, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1377 binds (318.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 356 taken from the spares, 165 made new; 402 kept as spares, 0 destroyed; 46 spares now, 1260 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 951, 4220.1 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 370 dirty pages from their submits' captures, 55726 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 907 dispatches, 42373 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8644 runs, 699.3 MB compared; 57 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20772 in captures, 2 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos ModuleSystemWorker4] vk: readbacks of GPU results: 1134 of 3000 waited for their write-back, 0.00 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 1072 kept their virtual core (nothing there to wait for), 2768 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 16620000 (10150104 pages from the captures) [hos host] vk: 5 textures retired (1102 left, 2096 MB): budget 0, drawn over 5, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: 9 textures retired (1104 left, 2126 MB): budget 7, drawn over 2, dead targets 0 [hos host] vk: write-backs left out (600 submits): 975, 4329.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 805 dirty pages from their submits' captures, 56711 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 919 dispatches, 43580 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8543 runs, 712.9 MB compared; 52 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21179 in captures, 0 only in the collection, 39 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: pipeline cache: play file 10 saved (3 KB): 1 pipelines, 1 from the driver's cache, 0 compiled; 58 more from the start's cache [hos host] vk: replay reads of guest memory as submitted: 16640000 (10161389 pages from the captures) [hos host] vk: 2 textures retired (1117 left, 2177 MB): budget 0, drawn over 0, dead targets 2 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 21767 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.04 over 555 ranges, 3232 MB), 295 pages written, 295 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.034; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.026, bytes copied 0.000; 25 kept (124 MB), 10 spare buffers, 9660 texture pages copied (37 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 2 textures retired (1137 left, 2368 MB): budget 0, drawn over 0, dead targets 2 [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #415000: event 0x100d6c800c mode 0 value 304899 (now 304880) (lr 0x80660e38) [hos host] vk: sparse mirrors: 118, 1090 MB with memory of 1556 MB (465 MB spared); blocks given memory later 13790, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1378 binds (318.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 356 taken from the spares, 165 made new; 403 kept as spares, 0 destroyed; 47 spares now, 1264 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1100, 4884.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 174 dirty pages from their submits' captures, 61202 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 990 dispatches, 46942 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8462 runs, 784.0 MB compared; 98 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22947 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 2 retired still imported (8 MB), 51 retired given back (1296 MB) [hos host] vk: replay reads of guest memory as submitted: 16660000 (10172126 pages from the captures) [hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #305000: event 0x100d6c8024 mode 0 value 305000 (lr 0x80665068) [hos host] vk: program picks: 23878660 kept (10055818 by the last picks' textures, a uniform buffer bound since), 26584 new, 8694755 with other textures (of 32600000) [hos host] vk: a frame: 189 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 8.89, present 0.18, draw 5.21, descriptors 2.38, texture lookups 0.49, uploads 0.22, pipelines 0.22, programs 0.85, commands 1.37, dispatch 0.89, memory sync 0.49, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 521 (1090 MB): 5.63 MB synced in 1093.6 copies, 18.08 MB written back [hos host] vk: 21360 presents, 60.1 fps; a frame: 1405 draws (0 skipped), 165 dispatches, 156 passes, 144 barriers, 6 submits, 0.4 uploads (1.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14189 pipelines (3740 ms), 0 programs (10 failed), 1152 textures (2542 MB) [hos host] vk: host time in GPU resource waits: 2.28 ms a frame (fences 1.76, timelines 0.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 605.9 full barriers recorded, 286.1 left out (nothing recorded since the one before), 13.4 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.3 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.9, an op 700.1, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.7 ms a frame (the longest submission 4.3 ms); 1.3 textures made a frame: colour targets 0.3 (1.8 MB), depth 0.8 (4.8 MB), ASTC or BC 0.2 (0.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 1.00 MB unswizzled on the CPU in 0.15 ms [hos host] vk: 120 textures retired (1040 left, 2427 MB): budget 89, drawn over 31, dead targets 0 [hos sead::MainThread] nvn: GPU waits in 5 s: 39 sync waits (82 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: 60 textures retired (988 left, 2395 MB): budget 50, drawn over 10, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 16680000 (10182298 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 230 dirty pages from their submits' captures, 59415 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 951 dispatches, 45430 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8304 runs, 752.1 MB compared; 80 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22138 in captures, 1 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 29 textures retired (961 left, 2292 MB): budget 19, drawn over 10, dead targets 0 [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05 [hos Presentation Thread] vi: 21397 frames presented (60.0 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 20, 15.5-17.8 269, 17.8-22 9, 22-30 3, 30-40 0, over 40 0 [hos host] vk: 9 textures retired (959 left, 2271 MB): budget 0, drawn over 9, dead targets 0 [hos host] vk: 11 textures retired (957 left, 2308 MB): budget 6, drawn over 5, dead targets 0 [hos host] vk: 17 textures retired (970 left, 2331 MB): budget 2, drawn over 7, dead targets 8 [hos host] vk: replay reads of guest memory as submitted: 16700000 (10193244 pages from the captures) [hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.04 over 551 ranges, 3142 MB), 269 pages written, 269 copied (1.0 MB) [hos host] vk: a capture taken (ms): all 0.014; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 25 kept (136 MB), 8 spare buffers, 5636 texture pages copied (22 MB) [hos host] vk: readback pages refreshed from guest memory: 7029903 pages so far (0 held: may be written ahead in their submission) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5597 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: sparse mirrors: 119, 1107 MB with memory of 1580 MB (472 MB spared); blocks given memory later 13790, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1379 binds (319.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 357 taken from the spares, 165 made new; 403 kept as spares, 0 destroyed; 46 spares now, 1240 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 446 dirty pages from their submits' captures, 59879 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 980 dispatches, 45584 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8300 runs, 752.9 MB compared; 72 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22356 in captures, 0 only in the collection, 43 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB) [hos host] vk: 16 textures retired (975 left, 2367 MB): budget 0, drawn over 6, dead targets 10 [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1168 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 1 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1197 kept their virtual core (nothing there to wait for), 3155 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: texture 0x10009516800 640x360x1 x1 levels 1 NVN 0x25 -> vk 37 (1000 textures, 2469 MB) [hos host] vk: texture 0x10009422400 62x62x1 x1 levels 1 NVN 0x34 -> vk 126 (1000 textures, 2494 MB) [hos host] vk: 17 textures retired (993 left, 2404 MB): budget 11, drawn over 6, dead targets 0 [hos host] vk: texture 0x1000a50cc00 2560x1440x1 x1 levels 1 NVN 0x3f -> vk 122 (1000 textures, 2450 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: texture 0x10008612200 1799x1799x1 x1 levels 1 NVN 0x34 -> vk 126 (1000 textures, 2462 MB) [hos host] vk: texture 0x10008612200 1832x1832x1 x1 levels 1 NVN 0x34 -> vk 126 (1000 textures, 2492 MB) [hos host] vk: 38 textures retired (970 left, 2359 MB): budget 32, drawn over 6, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 16720000 (10205981 pages from the captures) [hos ModuleSystemWorker1] nvn: 2800000 recordings made, 72259 live copies (23.4 MB), 6822 texture copies; 3 command buffers' segments (4.2 MB) [hos host] vk: VRAM compaction 8 begins: the textures' blocks hold 1376 MB, 638 MB of it free space [hos host] vk: recorder: 5384180 packets run in 10.0 s (busy 1.62 s, spun idle 0.12 s, slept 10653 times, 3265 of them under 200 us), 0 calls made directly (no room); 17927 batches over 600 frames (29.9 a frame, 300 packets a batch), 8944 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 152 times, 27.20 ms (0.3 times, 0.05 ms a frame); every thread 768 times, 60.53 ms in all, the longest 1.24 ms; vkDestroyImageView 216/32.71 ms (max 0.52), vkWaitForFences 508/24.91 ms (max 0.45), a part's submit 3/1.51 ms (max 1.24), vkQueueBindSparse 15/0.54 ms (max 0.22), vmaDestroyImage 16/0.39 ms (max 0.04), vmaFreeMemory 4/0.34 ms (max 0.23) [hos host] vk: recorder: of its busy 1.62 s, 0.39 s in parts' ends and submits (the longest 1.19 ms), 0.07 s of that blocked on the queue's lock (153 times, the longest 1.16 ms), so 0.32 s in the ends, submits and holds; 1.24 s running the other packets; 8974 packets a frame [hos host] vk: 59 textures retired (924 left, 2337 MB): budget 50, drawn over 9, dead targets 0 [hos host] vk: 20 textures retired (913 left, 2336 MB): budget 11, drawn over 9, dead targets 0 [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 278 dirty pages from their submits' captures, 61360 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 983 dispatches, 46298 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8758 runs, 768.8 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23003 in captures, 0 only in the collection, 42 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 16740000 (10218873 pages from the captures) [hos host] vk: 12 textures retired (915 left, 2324 MB): budget 0, drawn over 12, dead targets 0 [hos host] vk: a frame: 226 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 8.78, present 0.34, draw 4.75, descriptors 2.13, texture lookups 0.84, uploads 0.05, pipelines 0.20, programs 0.81, commands 1.25, dispatch 0.88, memory sync 0.48, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 523 (1054 MB): 5.85 MB synced in 1093.0 copies, 18.09 MB written back [hos host] vk: 21480 presents, 59.8 fps; a frame: 1316 draws (0 skipped), 159 dispatches, 156 passes, 144 barriers, 6 submits, 0.4 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14195 pipelines (3742 ms), 0 programs (10 failed), 916 textures (2335 MB) [hos host] vk: host time in GPU resource waits: 1.93 ms a frame (fences 1.68, timelines 0.25, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 595.5 full barriers recorded, 284.3 left out (nothing recorded since the one before), 13.3 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 38.0; 0x12: 69.0; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.5 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 135.5, an op 613.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.5 ms a frame (the longest submission 4.7 ms); 3.1 textures made a frame: colour targets 1.2 (7.4 MB), depth 1.6 (12.1 MB), ASTC or BC 0.4 (0.3 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.11 MB unswizzled on the CPU in 0.01 ms [hos sead::MainThread] vcores: 154245 takes, 22865 waits (mean 65 us, longest 6.8 ms, 0 over HOS_VCORE_LOG_MS), 10934 kicks, 10907 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: program picks: 24019710 kept (10105860 by the last picks' textures, a uniform buffer bound since), 26627 new, 8753662 with other textures (of 32800000) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: 23 textures retired (893 left, 1905 MB): budget 0, drawn over 15, dead targets 8 [hos host] vk: replay reads of guest memory as submitted: 16760000 (10232420 pages from the captures) [hos host] vk: sparse mirrors: 112, 1051 MB with memory of 1492 MB (440 MB spared); blocks given memory later 13804, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1390 binds (327.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 359 taken from the spares, 165 made new; 412 kept as spares, 0 destroyed; 53 spares now, 1328 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1000, 4440.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 791 dirty pages from their submits' captures, 59446 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 997 dispatches, 44410 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8103 runs, 729.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21641 in captures, 0 only in the collection, 40 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 194 live (2260 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB) [hos host] vk: VRAM compaction 8 over: 333 textures moved (239 MB) in 30 passes, 0 left where they were; 19 blocks (608 MB) given back to the driver, 5088 MB in all so far [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.05 over 558 ranges, 3138 MB), 290 pages written, 290 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.026; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (96 MB), 11 spare buffers, 8106 texture pages copied (31 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13505 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #420000: event 0x100d6c800c mode 0 value 308699 (now 308680) (lr 0x80660e38) [hos host] vk: 1 textures retired (974 left, 1947 MB): budget 0, drawn over 0, dead targets 1 [hos host] vk: replay reads of guest memory as submitted: 16780000 (10243369 pages from the captures) [hos host] vk: slow upload 18.3 ms: #26760 0x10df1aba00 1024x1024x1 x1 levels 11 NVN 0x8d ASTC (BC3), 475 of 475 KB [hos host] vk: texture 0x10d7bfc600 768x768x1 x1 levels 10 NVN 0x7b -> vk 37 (ASTC decoded) (1000 textures, 1992 MB) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 652 dirty pages from their submits' captures, 59065 from earlier ones', 9 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1008 dispatches, 45350 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8221 runs, 768.0 MB compared; 58 pages read into the shadows; 284 MB of shadows (53 MB of pages known, in 106 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23846 in captures, 2 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x1043400000 (4 MB, captured, state 2): 7; 0x7fb8800000 (52 MB, captured, state 2): 1; 0x7f9f800000 (4 MB, captured, state 2): 1; [hos host] vk: slow upload 15.1 ms: #26717 0x10d2577200 1024x1024x1 x1 levels 11 NVN 0x8d ASTC (BC3), 475 of 475 KB [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: presenter thread: the copy's fence 1.67 ms, the present 0.71 ms a present [hos host] vk: replay reads of guest memory as submitted: 16800000 (10254228 pages from the captures) [hos host] vk: texture storage read as submitted (4001 times; 166149 pages from the copies, 0 of them rewritten in guest memory since their submit; 8260 pages kept, 208 spans, 734 MB) [hos host] vk: slow upload 17.2 ms: #26878 0x11038fba00 768x768x1 x1 levels 10 NVN 0x7e ASTC (BC3), 324 of 324 KB [hos BaseProcExecutor2] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 41 made): 39253942 lookups, 125690 missed (0.32 %) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 120, 1118 MB with memory of 1584 MB (465 MB spared); blocks given memory later 13819, given back 17182; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1400 binds (335.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 367 taken from the spares, 165 made new; 412 kept as spares, 0 destroyed; 45 spares now, 1236 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1262 dirty pages from their submits' captures, 59881 from earlier ones', 362 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1072 dispatches, 45554 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8381 runs, 749.7 MB compared; 91 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22628 in captures, 2 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb4400000 (24 MB, captured, state 2): 355; 0x7fb8800000 (52 MB, captured, state 2): 3; 0x7fa0800000 (8 MB, captured, state 2): 2; 0x1043400000 (4 MB, captured, state 2): 2; [hos host] vk: imports: 194 live (2260 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.06 [hos host] vk: memory: working set 14753 MB (peak 16021), private 21338 MB, staging 64 MB in 4 arenas, 47 made so far (212 ms) [hos host] vk: memory heaps: heap 0 (system): 13 blocks 928 MB, 36 allocations 846 MB; used 3208 of 15579 MB; heap 1 (device): 154 blocks 4631 MB, 20005 allocations 3951 MB; used 6754 of 11474 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 26 blocks 832 MB, 155 images 344 MB; 96 free ranges, the largest 31 MB [hos host] vk: image blocks (textures, compacted): 37 blocks 1184 MB, 942 images 1163 MB; 242 free ranges, the largest 13 MB [hos host] vk: render target images: 6809 reused from the pool, 6046 made new, 5889 destroyed from it, 1530 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 201 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 2884 with the pool empty (0 of them with targets still in the dead queue), 421 with none in it a present yet, 61 with none of a kind they can use, 817 with all smaller (the largest 49% short on average), 827 with all bigger (over 1.5 times), 1033 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 5697 idle [hos host] vk: VRAM: 6754 MB used of 11474 MB allowed (680 MB of it free in the blocks); textures 2143 MB (budget 7158 MB), 1103 textures [hos host] vk: budget: VRAM 4631 of 11474 MB (6754 MB by the driver, 2123 MB unaccounted): textures 2143, mirrors 1118, blocks free 680, shadows 284, other buffers 229, other images 111, staging 64; RAM 3141 MB (the process's private memory 21338 MB): imports 2260, capture spares 746, captures 100, capture table 35, others 0 [hos host] vk: textures: decoded ASTC 1234 MB, rendered 567 MB (84 drawn in one frame, kept: 44665 KB), other 303 MB; used in the last 8 presents 1690 MB [hos host] vk: decoded ASTC used again: 622 textures (1399 MB) copied from ones the game finalized, not decoded; 622 of those retired, 0 listed; 2531 MB of blocks hashed in 413 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 16453 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d503ee00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 89 (15729 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB); [hos Presentation Thread] vk: presents after the CPU's signal (600): 180 waited for it, 2.08 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 53.4 ms; VRAM 6.6 of 11.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 13, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6517 queued, 6517 loaded (median 1.46 ms, 99th percentile 16.63 ms, worst 38.13 ms, 16444 ms and 350.3 MB read in all); 8822 programs (1630 of them compute variants without their code, 0 graphics programs no note names), 12331 graphics pipelines (1 not in a cache), 736 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2449 pipelines made at create taken by draws; 3156 create calls with nothing kept for their stages, 0 before the index; units 9673 (6517 ready, 0 loading), 17647 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 14030 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 781 known, 24 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 344 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 9.52, present 0.26, draw 5.96, descriptors 3.25, texture lookups 1.54, uploads 1.01, pipelines 0.21, programs 0.78, commands 1.34, dispatch 0.86, memory sync 0.53, queue submit 0.00, slot wait 0.23 [hos host] vk: a frame: mirrors 532 (1118 MB): 6.30 MB synced in 1097.2 copies, 18.29 MB written back [hos host] vk: 21600 presents, 56.9 fps; a frame: 1412 draws (0 skipped), 162 dispatches, 152 passes, 146 barriers, 6 submits, 2.1 uploads (1.6 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14199 pipelines (3744 ms), 0 programs (10 failed), 1103 textures (2143 MB) [hos host] vk: host time in GPU resource waits: 5.60 ms a frame (fences 3.36, timelines 2.24, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 597.5 full barriers recorded, 282.1 left out (nothing recorded since the one before), 15.3 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 39.8; 0x12: 69.2; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.4 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 137.5, an op 563.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 10.8 ms a frame (the longest submission 4.3 ms); 3.8 textures made a frame: colour targets 0.8 (3.9 MB), depth 1.0 (4.4 MB), ASTC or BC 2.0 (4.0 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.65 MB unswizzled on the CPU in 0.11 ms [hos host] vk: queues ordered by their events: 143941 submissions stopped at a wait, 143881 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 421436, not met 0; captures at a CPU event's wait 13122, 0 of them at a wait met on arrival [hos host] vk: the render thread (600 presents, 16.86 ms a frame): its jobs 14.36 ms a frame (submits 6.71, presents 4.61, others 3.04), idle 2.50; the game waited for room at 10 presents, 1.42 ms each (0.02 a frame; 6.34 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 186 captures at a CPU signal; the submits' waits read in 0.039 ms a frame; readbacks: 1213 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 13.50 ms after its hand-over (57.40 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 3.91 + 0.27 (the window's copy is in the idle), after it 1.13 + 0.71; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 8.50 parts a frame recorded in the signal's job, 78.6% of them took nothing that came with the signal; of the others, 1.81 a frame took pages of the signal's capture into a mirror (32.7 pages a frame) and 0.03 pages read as guest memory had them (0.1 pages a frame); a frame's first part took nothing in 20 of 186 frames, its first three in 20 [hos host] vk: frame ahead, what came with the signal: 186 signals with a capture, each of 59.0 pages in 8.9 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.41+0.11, 0.23+0.10, 0.18+0.05, 0.04+0.03, 0.06+0.30, 0.07+0.07, 0.09+0.03, 0.04+0.02, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00 [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1127 of 3000 waited for their write-back, 0.06 ms each; 0 looked at the stopped queues first, 14 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 1263 kept their virtual core (nothing there to wait for), 3089 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.05 over 587 ranges, 3198 MB), 270 pages written, 270 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.022; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 25 kept (120 MB), 10 spare buffers, 9914 texture pages copied (38 MB) [hos host] vk: queues ordered by their events: 144000 submissions stopped at a wait, 143938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 421630, not met 0; captures at a CPU event's wait 13125, 0 of them at a wait met on arrival [hos sead::MainThread] nvn: CommandBufferSignalEvent #310000: event 0x100d6c802c mode 0 value 309999 (lr 0x80667688) [hos host] vk: render targets drawn over from another start, let go: 13160 so far (62285 MB), 11935 of them (56613 MB) between the passes [hos host] vk: replay reads of guest memory as submitted: 16820000 (10265821 pages from the captures) [hos host] vk: program picks: 24164830 kept (10146003 by the last picks' textures, a uniform buffer bound since), 26652 new, 8808517 with other textures (of 33000000) [hos host] vk: readback pages refreshed from guest memory: 7097017 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 1025, 4551.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 991 dirty pages from their submits' captures, 59981 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1030 dispatches, 45649 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8709 runs, 766.1 MB compared; 58 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23960 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 1; [hos host] vk: replay reads of guest memory as submitted: 16840000 (10277016 pages from the captures) [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 760000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 13058 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: 1 textures retired (1130 left, 2162 MB): budget 0, drawn over 1, dead targets 0 [hos host] vk: reaper: 864 images (3779 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 88 (275 MB) [hos host] vk: replay reads of guest memory as submitted: 16860000 (10288143 pages from the captures) [hos host] vk: sparse mirrors: 123, 1123 MB with memory of 1596 MB (472 MB spared); blocks given memory later 13822, given back 17188; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1406 binds (336.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 370 taken from the spares, 165 made new; 412 kept as spares, 0 destroyed; 42 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 1050, 4662.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 422 dirty pages from their submits' captures, 65136 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 1038 dispatches, 47309 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8753 runs, 781.6 MB compared; 82 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 24230 in captures, 1 only in the collection, 41 only by the replay itself [hos host] vk: read live, by pool: 0x7fb8800000 (52 MB, captured, state 2): 2; [hos host] vk: imports: 195 live (2264 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 67 sync waits (239 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos Presentation Thread] vi: 21686 frames presented (57.6 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 66, 15.5-17.8 182, 17.8-22 25, 22-30 6, 30-40 7, over 40 3 [hos host] vk: replay reads of guest memory as submitted: 16880000 (10300666 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.06 [hos host] vk: 417 mirrors given back (2682 MB in all); 1132 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone [hos Presentation Thread] vk: capture: a submit 0.16 ms (lock 0.01, scan 0.04 over 600 ranges, 3294 MB), 291 pages written, 291 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.027; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (96 MB), 11 spare buffers, 5135 texture pages copied (20 MB) [hos host] vk: write-backs left out (600 submits): 899, 3994.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2650 dirty pages from their submits' captures, 53666 from earlier ones', 69 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 946 dispatches, 41642 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8383 runs, 676.9 MB compared; 70 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21175 in captures, 13 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: 0x7f9c400000 (32 MB, captured, state 2): 62; 0x7f98800000 (8 MB, captured, state 2): 2; 0x7f95000000 (4 MB, captured, state 2): 2; 0x7fa9000000 (4 MB, captured, state 2): 1; 0x7f88000000 (12 MB, captured, state 2): 1; 0x1041c00000 (4 MB, captured, state 2): 1; [hos host] vk: replay reads of guest memory as submitted: 16900000 (10313419 pages from the captures) [hos host] vk: a frame: 336 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.04, present 0.18, draw 6.04, descriptors 2.81, texture lookups 0.64, uploads 0.40, pipelines 0.26, programs 0.94, commands 1.57, dispatch 1.02, memory sync 0.60, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 540 (1132 MB): 6.28 MB synced in 1130.0 copies, 18.90 MB written back [hos host] vk: 21720 presents, 56.5 fps; a frame: 1671 draws (0 skipped), 181 dispatches, 166 passes, 148 barriers, 6 submits, 1.6 uploads (0.5 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14230 pipelines (3751 ms), 0 programs (10 failed), 1290 textures (2332 MB) [hos host] vk: host time in GPU resource waits: 5.16 ms a frame (fences 3.00, timelines 2.16, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 647.0 full barriers recorded, 296.1 left out (nothing recorded since the one before), 19.8 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 40.5; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 139.1, an op 708.2, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.7 ms a frame (the longest submission 5.0 ms); 1.6 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 1.6 (1.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.22 MB unswizzled on the CPU in 0.02 ms [hos host] vk: replay reads of guest memory as submitted: 16920000 (10325944 pages from the captures) [hos host] vk: program picks: 24314157 kept (10191167 by the last picks' textures, a uniform buffer bound since), 26671 new, 8859171 with other textures (of 33200000) [hos sead::MainThread] nvn: CommandBufferWaitEvent #425000: event 0x100d6c8010 mode 0 value 312500 (now 312481) (lr 0x80660e38) [hos host] vk: sparse mirrors: 124, 1133 MB with memory of 1608 MB (474 MB spared); blocks given memory later 13830, given back 17188; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1419 binds (337.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 376 taken from the spares, 165 made new; 417 kept as spares, 0 destroyed; 41 spares now, 1212 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2176 dirty pages from their submits' captures, 81481 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 960 dispatches, 43291 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8334 runs, 696.4 MB compared; 62 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20439 in captures, 0 only in the collection, 35 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 196 live (2268 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB) [hos host] vk: replay reads of guest memory as submitted: 16940000 (10338172 pages from the captures) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1127 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 15 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 995 kept their virtual core (nothing there to wait for), 2589 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 16960000 (10350547 pages from the captures) [hos host] vk: late pages' fix (600 recordings with one): of 1076 captures with late pages 1076 fixed (75.4 pages, 76.1 fixes a capture), waited as before: 0 no pool, 0 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 439 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 832 blocks (53248 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 153, of an earlier recording 286, not recorded yet 0, of this part 0; pages fixed from stamps 80084, from snapshots 1089; the shadows: 1527891 pages synced from captures, 1497967 made known by it, 29924 left unknown because the replay or the collection wrote them since [hos tera::FileResourceMgr] vcores: 152738 takes, 28588 waits (mean 85 us, longest 28.8 ms, 0 over HOS_VCORE_LOG_MS), 13177 kicks, 13160 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1828 dirty pages from their submits' captures, 51509 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 881 dispatches, 40342 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7797 runs, 661.9 MB compared; 58 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19334 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: capture: a submit 0.25 ms (lock 0.01, scan 0.04 over 583 ranges, 3283 MB), 371 pages written, 371 copied (1.4 MB) [hos host] vk: a capture taken (ms): all 0.042; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.021, bytes copied 0.000; 25 kept (110 MB), 10 spare buffers, 7738 texture pages copied (30 MB) [hos host] vk: replay reads of guest memory as submitted: 16980000 (10363020 pages from the captures) [hos host] vk: 138 textures retired (1163 left, 2128 MB): budget 138, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 5 finalized, 31 MB [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.07 [hos host] vk: 6 textures retired (1157 left, 2111 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 20817 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 17000000 (10375305 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 7164543 pages so far (0 held: may be written ahead in their submission) [hos host] vk: 7 textures retired (1156 left, 2108 MB): budget 7, drawn over 0, dead targets 0 [hos host] vk: sparse mirrors: 125, 1026 MB with memory of 1612 MB (585 MB spared); blocks given memory later 13831, given back 18910; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1422 binds (346.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 377 taken from the spares, 165 made new; 417 kept as spares, 0 destroyed; 40 spares now, 1208 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 924, 4105.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2223 dirty pages from their submits' captures, 55052 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 902 dispatches, 41567 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8035 runs, 680.9 MB compared; 61 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19631 in captures, 132 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 196 live (2268 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 17020000 (10387682 pages from the captures) [hos host] vk: program picks: 24466918 kept (10240470 by the last picks' textures, a uniform buffer bound since), 26677 new, 8906404 with other textures (of 33400000) [hos ModuleSystemWorker3] nvn: 2850000 recordings made, 68209 live copies (23.3 MB), 6920 texture copies; 3 command buffers' segments (4.1 MB) [hos host] vk: a frame: 348 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.94, present 0.24, draw 6.63, descriptors 2.73, texture lookups 0.21, uploads 0.06, pipelines 0.31, programs 1.12, commands 1.96, dispatch 1.00, memory sync 0.72, queue submit 0.00, slot wait 0.27 [hos host] vk: a frame: mirrors 543 (1028 MB): 6.92 MB synced in 1126.2 copies, 18.46 MB written back [hos host] vk: 21840 presents, 56.4 fps; a frame: 1917 draws (0 skipped), 177 dispatches, 177 passes, 148 barriers, 6 submits, 0.2 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14230 pipelines (3751 ms), 0 programs (10 failed), 1162 textures (2135 MB) [hos host] vk: host time in GPU resource waits: 5.31 ms a frame (fences 2.00, timelines 3.32, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 676.7 full barriers recorded, 306.0 left out (nothing recorded since the one before), 20.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 40.6; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 138.8, an op 876.6, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.4 ms a frame (the longest submission 12.3 ms); 0.2 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.2 (0.5 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.23 MB unswizzled on the CPU in 0.02 ms [hos host] vk: replay reads of guest memory as submitted: 17040000 (10400194 pages from the captures) [hos host] vk: 13 textures retired (1154 left, 2110 MB): budget 13, drawn over 0, dead targets 0 [hos host] vk: write-backs left out (600 submits): 901, 3998.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2799 dirty pages from their submits' captures, 52300 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 885 dispatches, 40277 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7915 runs, 664.2 MB compared; 51 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19497 in captures, 0 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 17060000 (10412701 pages from the captures) [hos host] vk: 7 textures retired (1152 left, 2104 MB): budget 7, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: CommandBufferSignalEvent #315000: event 0x100d6c8038 mode 0 value 314999 (lr 0x8066189c) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 576 ranges, 3290 MB), 284 pages written, 284 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.041; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.022, bytes copied 0.000; 25 kept (110 MB), 10 spare buffers, 12554 texture pages copied (49 MB) [hos host] vk: replay reads of guest memory as submitted: 17080000 (10425233 pages from the captures) [hos host] vk: sparse mirrors: 123, 1026 MB with memory of 1592 MB (565 MB spared); blocks given memory later 13832, given back 18911; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1430 binds (347.4 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 380 taken from the spares, 165 made new; 422 kept as spares, 0 destroyed; 42 spares now, 1228 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1565 dirty pages from their submits' captures, 53179 from earlier ones', 81 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 896 dispatches, 41047 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8423 runs, 681.7 MB compared; 52 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20227 in captures, 1 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: 0x1065800000 (8 MB, captured, state 2): 79; 0x7f90400000 (4 MB, captured, state 2): 2; [hos host] vk: imports: 198 live (2276 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 17100000 (10437594 pages from the captures) [hos host] vk: 4 textures retired (1152 left, 2108 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1447 dirty pages from their submits' captures, 52658 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 41013 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7908 runs, 672.0 MB compared; 57 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20230 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos sead::MainThread] vk: readbacks of GPU results: 1123 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 16 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1205 kept their virtual core (nothing there to wait for), 3147 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 17120000 (10449811 pages from the captures) [hos host] vk: program picks: 24620868 kept (10289533 by the last picks' textures, a uniform buffer bound since), 26686 new, 8952445 with other textures (of 33600000) [hos sead::MainThread] nvn: CommandBufferWaitEvent #430000: event 0x100d6c8010 mode 0 value 316300 (now 316281) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 17140000 (10462503 pages from the captures) [hos host] vk: a frame: 378 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.31, present 0.17, draw 6.11, descriptors 2.59, texture lookups 0.22, uploads 0.11, pipelines 0.29, programs 0.96, commands 1.77, dispatch 0.97, memory sync 0.66, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 545 (1014 MB): 5.85 MB synced in 1125.2 copies, 18.70 MB written back [hos host] vk: 21960 presents, 60.5 fps; a frame: 1895 draws (0 skipped), 183 dispatches, 176 passes, 147 barriers, 6 submits, 0.3 uploads (0.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14230 pipelines (3751 ms), 0 programs (10 failed), 1157 textures (2129 MB) [hos host] vk: host time in GPU resource waits: 4.01 ms a frame (fences 1.56, timelines 2.45, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 684.3 full barriers recorded, 304.9 left out (nothing recorded since the one before), 19.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 40.1; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 138.0, an op 820.3, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.8 ms a frame (the longest submission 4.3 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.3 (0.9 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.33 MB unswizzled on the CPU in 0.04 ms [hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.04 over 574 ranges, 3304 MB), 312 pages written, 312 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.056; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.039, bytes copied 0.000; 25 kept (234 MB), 11 spare buffers, 11683 texture pages copied (45 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 252 sync waits (841 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 120, 1014 MB with memory of 1568 MB (553 MB spared); blocks given memory later 13832, given back 18911; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1432 binds (347.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 380 taken from the spares, 165 made new; 425 kept as spares, 0 destroyed; 45 spares now, 1252 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 949, 4216.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1116 dirty pages from their submits' captures, 54234 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 921 dispatches, 41997 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8003 runs, 686.5 MB compared; 69 pages read into the shadows; 284 MB of shadows (54 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19712 in captures, 1 only in the collection, 36 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 198 live (2276 MB), 0 retired still imported (0 MB), 53 retired given back (1304 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 62681 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: 8 textures retired (1151 left, 2111 MB): budget 8, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17160000 (10474471 pages from the captures) [hos Presentation Thread] vi: 21977 frames presented (58.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 72, 15.5-17.8 166, 17.8-22 37, 22-30 11, 30-40 1, over 40 4 [hos host] vk: 1 textures retired (1154 left, 2110 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 17180000 (10487405 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.08, writes collected 0.05 [hos host] vk: readback pages refreshed from guest memory: 7231646 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 926, 4109.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1329 dirty pages from their submits' captures, 71189 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 904 dispatches, 42600 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8062 runs, 703.2 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19622 in captures, 0 only in the collection, 37 only by the replay itself [hos host] vk: read live, by pool: 0x7f90400000 (4 MB, captured, state 2): 3; [hos host] vk: replay reads of guest memory as submitted: 17200000 (10499496 pages from the captures) [hos host] vk: 95 textures retired (1087 left, 2109 MB): budget 95, drawn over 0, dead targets 0; GPU-drawn with no copy kept: 0 live, 1 finalized, 3 MB [hos host] vk: program picks: 24774240 kept (10340971 by the last picks' textures, a uniform buffer bound since), 26686 new, 8999073 with other textures (of 33800000) [hos host] vk: recorder: 5395212 packets run in 10.0 s (busy 1.68 s, spun idle 0.12 s, slept 10782 times, 3681 of them under 200 us), 0 calls made directly (no room); 17463 batches over 577 frames (30.3 a frame, 309 packets a batch), 9058 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 161 times, 21.55 ms (0.3 times, 0.04 ms a frame); every thread 767 times, 65.99 ms in all, the longest 0.82 ms; vkWaitForFences 507/37.22 ms (max 0.57), vkDestroyImageView 107/11.75 ms (max 0.52), vkQueueBindSparse 56/4.88 ms (max 0.48), vmaFreeMemory 20/4.64 ms (max 0.82), vkEndCommandBuffer 13/2.90 ms (max 0.44), vmaDestroyBuffer 21/1.94 ms (max 0.63) [hos host] vk: recorder: of its busy 1.68 s, 0.49 s in parts' ends and submits (the longest 1.40 ms), 0.17 s of that blocked on the queue's lock (310 times, the longest 1.36 ms), so 0.31 s in the ends, submits and holds; 1.20 s running the other packets; 9350 packets a frame [hos host] vk: replay reads of guest memory as submitted: 17220000 (10511590 pages from the captures) [hos host] vk: sparse mirrors: 121, 1024 MB with memory of 1596 MB (571 MB spared); blocks given memory later 13832, given back 18919; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1439 binds (349.8 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 383 taken from the spares, 165 made new; 427 kept as spares, 0 destroyed; 44 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 950, 4218.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1858 dirty pages from their submits' captures, 55992 from earlier ones', 3 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 924 dispatches, 42572 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8040 runs, 700.9 MB compared; 61 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 20343 in captures, 0 only in the collection, 38 only by the replay itself [hos host] vk: read live, by pool: 0x7f9c400000 (32 MB, captured, state 2): 3; [hos host] vk: imports: 197 live (2276 MB), 2 retired still imported (8 MB), 53 retired given back (1304 MB) [hos sead::MainThread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.04 over 572 ranges, 3314 MB), 339 pages written, 339 copied (1.3 MB) [hos host] vk: a capture taken (ms): all 0.050; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.032, bytes copied 0.000; 25 kept (102 MB), 11 spare buffers, 12270 texture pages copied (47 MB) [hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 780000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040) [hos host] vk: replay reads of guest memory as submitted: 17240000 (10524052 pages from the captures) [hos host] vk: 5 textures retired (1098 left, 2110 MB): budget 5, drawn over 0, dead targets 0 [hos rmbl::VibrationThread] hid: 73728 vibration values, strongest 0.02 since the last count [hos BaseProcExecutor3] vcores: 151872 takes, 23883 waits (mean 82 us, longest 12.4 ms, 0 over HOS_VCORE_LOG_MS), 11102 kicks, 11349 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 17260000 (10536900 pages from the captures) [hos host] vk: write-backs left out (600 submits): 900, 3996.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1826 dirty pages from their submits' captures, 52584 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 928 dispatches, 40871 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7729 runs, 667.5 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19914 in captures, 1 only in the collection, 34 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 9 textures retired (1139 left, 2152 MB): budget 9, drawn over 0, dead targets 0 [hos host] vk: a frame: 332 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 10.83, present 0.18, draw 6.78, descriptors 2.90, texture lookups 0.37, uploads 0.22, pipelines 0.29, programs 1.12, commands 1.97, dispatch 0.95, memory sync 0.66, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 548 (1024 MB): 6.58 MB synced in 1122.2 copies, 18.50 MB written back [hos host] vk: 22080 presents, 57.7 fps; a frame: 2003 draws (0 skipped), 174 dispatches, 191 passes, 147 barriers, 6 submits, 0.9 uploads (0.4 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14231 pipelines (3752 ms), 0 programs (10 failed), 1139 textures (2152 MB) [hos host] vk: host time in GPU resource waits: 4.55 ms a frame (fences 1.72, timelines 2.83, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 720.3 full barriers recorded, 314.1 left out (nothing recorded since the one before), 24.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 40.2; 0x12: 69.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 45.2 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 138.1, an op 965.7, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 11.9 ms a frame (the longest submission 5.2 ms); 0.9 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.8 (1.6 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.10 MB unswizzled on the CPU in 0.01 ms [hos host] vk: replay reads of guest memory as submitted: 17280000 (10549529 pages from the captures) [hos host] vk: 8 textures retired (1158 left, 2177 MB): budget 8, drawn over 0, dead targets 0 [hos ModuleSystemWorker1] vk: readbacks of GPU results: 1139 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker1] vk: invalidates since the last readbacks line: 1264 kept their virtual core (nothing there to wait for), 3088 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 17300000 (10562696 pages from the captures) [hos host] vk: 3 textures retired (1170 left, 2215 MB): budget 3, drawn over 0, dead targets 0 [hos host] vk: 6 textures retired (1164 left, 2214 MB): budget 6, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17320000 (10575029 pages from the captures) [hos host] vk: sparse mirrors: 120, 1021 MB with memory of 1592 MB (570 MB spared); blocks given memory later 13832, given back 18930; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1445 binds (351.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 385 taken from the spares, 165 made new; 429 kept as spares, 0 destroyed; 44 spares now, 1224 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1966 dirty pages from their submits' captures, 51582 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 924 dispatches, 39986 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7580 runs, 650.4 MB compared; 60 pages read into the shadows; 284 MB of shadows (55 MB of pages known, in 109 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19561 in captures, 0 only in the collection, 33 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 197 live (2276 MB), 2 retired still imported (8 MB), 53 retired given back (1304 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 22394 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 24926062 kept (10394863 by the last picks' textures, a uniform buffer bound since), 26713 new, 9047224 with other textures (of 34000000) [hos host] vk: late pages' fix: stamps for 0x1038bf0000 (pool 0x1038400000; 896 blocks, 57344 KB) [hos host] vk: 433 mirrors given back (2737 MB in all); 1006 MB of mirrors now, 136 MB of them idle and kept for what the GPU wrote there alone [hos host] vk: replay reads of guest memory as submitted: 17340000 (10587185 pages from the captures) [hos host] vk: 3 textures retired (1161 left, 2213 MB): budget 3, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: CommandBufferSignalEvent #320000: event 0x100d6c8044 mode 0 value 319999 (lr 0x802742f0) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 579 ranges, 3363 MB), 278 pages written, 278 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.037; GPU wait for late pages 0.000 (1 of 600 takes), pools lock 0.000, retires and stamps 0.018, bytes copied 0.000; 25 kept (156 MB), 10 spare buffers, 7267 texture pages copied (28 MB) [hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: pool 0x7fb8800000..0x7fbbc00000 (52 MB) trimmed to its 1 live part(s), 8 MB: the rest outside every game pool for 1856 submissions (5 trims, 204 MB in all) [hos host] vk: 554 pools mirrored in VRAM (1008 MB) [hos host] vk: pool 0x7fbd800000..0x7fc1400000 (60 MB) trimmed to its 1 live part(s), 20 MB: the rest outside every game pool for 1920 submissions (6 trims, 244 MB in all) [hos host] vk: pool 0x7fbf400000..0x7fc0800000: 16 KB the CPU wrote since retired pool 0x7fbd800000's last sync kept from guest memory, not its mirror [hos host] vk: 5 textures retired (1180 left, 2226 MB): budget 5, drawn over 0, dead targets 0 [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #435000: event 0x100d6c8010 mode 0 value 320100 (now 320081) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 17360000 (10599530 pages from the captures) [hos host] vk: pool 0x7fd6400000..0x7fd9400000 (48 MB) trimmed to its 1 live part(s), 20 MB: the rest outside every game pool for 1984 submissions (7 trims, 272 MB in all) [hos host] vk: pool 0x7fd7000000..0x7fd8400000: 4 KB the CPU wrote since retired pool 0x7fd6400000's last sync kept from guest memory, not its mirror [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2114 dirty pages from their submits' captures, 53904 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 929 dispatches, 41812 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8664 runs, 669.8 MB compared; 392 pages read into the shadows; 288 MB of shadows (54 MB of pages known, in 110 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21303 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 4 textures retired (1176 left, 2224 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17380000 (10611243 pages from the captures) [hos host] vk: texture slots registered again at another size: 1200000 (of 1494067 registrations; slot 9736 480x270 -> 640x360) [hos host] vk: 2 textures retired (1174 left, 2224 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: presenter thread: the copy's fence 1.94 ms, the present 0.80 ms a present [hos host] vk: 1 textures retired (1173 left, 2223 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17400000 (10623476 pages from the captures) [hos ModuleSystemWorker3] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 41 made): 20617862 lookups, 135034 missed (0.65 %) [hos host] vk: 4 textures retired (1181 left, 2211 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 7299160 pages so far (0 held: may be written ahead in their submission) [hos host] vk: sparse mirrors: 120, 1007 MB with memory of 1464 MB (456 MB spared); blocks given memory later 13840, given back 18933; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1459 binds (352.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 390 taken from the spares, 166 made new; 436 kept as spares, 0 destroyed; 46 spares now, 1376 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1688 dirty pages from their submits' captures, 54613 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 918 dispatches, 41620 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9865 runs, 676.6 MB compared; 47 pages read into the shadows; 288 MB of shadows (52 MB of pages known, in 107 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23064 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 198 live (2172 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos host] vk: replay reads of guest memory as submitted: 17420000 (10635479 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 17440000 (10647732 pages from the captures) [hos host] vk: program picks: 25079019 kept (10447532 by the last picks' textures, a uniform buffer bound since), 26735 new, 9094245 with other textures (of 34200000) [hos host] vk: memory: working set 14962 MB (peak 16021), private 21905 MB, staging 64 MB in 4 arenas, 47 made so far (212 ms) [hos host] vk: memory heaps: heap 0 (system): 17 blocks 960 MB, 36 allocations 858 MB; used 3332 of 15579 MB; heap 1 (device): 165 blocks 4983 MB, 18449 allocations 3951 MB; used 7106 of 8475 MB; [hos host] vk: image blocks: 3 blocks 96 MB, 4 images 64 MB; 2 free ranges, the largest 31 MB [hos host] vk: image blocks (render targets): 26 blocks 832 MB, 150 images 303 MB; 92 free ranges, the largest 32 MB [hos host] vk: image blocks (textures, compacted): 48 blocks 1536 MB, 1069 images 1305 MB; 343 free ranges, the largest 12 MB [hos host] vk: render target images: 6809 reused from the pool, 6048 made new, 5896 destroyed from it, 1530 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=1), 201 missed after their image was taken by size; it holds 0 (0 MB) [hos host] vk: render targets not reused by size: 2886 with the pool empty (0 of them with targets still in the dead queue), 421 with none in it a present yet, 61 with none of a kind they can use, 817 with all smaller (the largest 49% short on average), 827 with all bigger (over 1.5 times), 1033 with some of each, 3 over a block, 0 refused by the driver, 0 not bound; the pool let go of 192 over its bound and 5704 idle [hos host] vk: VRAM: 7106 MB used of 8475 MB allowed (1032 MB of it free in the blocks); textures 2244 MB (budget 2107 MB), 1225 textures [hos host] vk: budget: VRAM 4983 of 8475 MB (7106 MB by the driver, 2123 MB unaccounted): textures 2244, blocks free 1032, mirrors 1009, shadows 288, other buffers 233, other images 111, staging 64; RAM 3253 MB (the process's private memory 21905 MB): imports 2176, capture spares 732, retired imports 176, captures 134, capture table 35, others 0 [hos host] vk: textures: decoded ASTC 1398 MB, rendered 508 MB (82 drawn in one frame, kept: 44485 KB), other 303 MB; used in the last 8 presents 2082 MB [hos host] vk: decoded ASTC used again: 686 textures (1539 MB) copied from ones the game finalized, not decoded; 686 of those retired, 2 listed; 2655 MB of blocks hashed in 430 ms [hos host] vk: 214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago [hos host] vk: 214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago [hos host] vk: 64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago [hos host] vk: 64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 37 MB: 0x112a982c00 1800x1380x1 x8 levels 1 NVN 0x6, last used 1 presents ago [hos host] vk: 32 MB: 0x1159c17000 4096x4096x1 x1 levels 1 NVN 0x32 rendered, last used 17053 presents ago [hos host] vk: 21 MB: 0x1000738e800 2400x2400x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago [hos host] vk: 16 MB: 0x10d503ee00 4096x4096x1 x1 levels 1 NVN 0x4b, last used 1 presents ago [hos host] vk: finalized textures the GPU drew once, kept for being made again: 91 (16081 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 18x33x1 NVN 0x25: 1 (2 KB); 42x41x1 NVN 0x25: 3 (20 KB); 62x32x1 NVN 0x25: 1 (7 KB); 73x70x1 NVN 0x25: 3 (59 KB); 74x32x1 NVN 0x25: 1 (9 KB); 82x97x1 NVN 0x1: 1 (7 KB); 96x112x1 NVN 0x1: 1 (10 KB); 102x30x1 NVN 0x25: 1 (11 KB); 106x105x1 NVN 0x1: 3 (32 KB); 126x96x1 NVN 0x1: 1 (11 KB); 133x42x1 NVN 0x25: 1 (21 KB); 137x134x1 NVN 0x1: 2 (35 KB); 138x96x1 NVN 0x1: 3 (38 KB); 197x106x1 NVN 0x1: 1 (20 KB); 202x38x1 NVN 0x25: 1 (29 KB); [hos host] vk: 6 textures retired (1219 left, 2243 MB): budget 6, drawn over 0, dead targets 0 [hos Presentation Thread] vk: presents after the CPU's signal (600): 502 waited for it, 3.26 ms each, 0 let go after RX_VK_FRAME_GATE_MS [hos host] vk: status: 600 frames, the slowest 48.3 ms; VRAM 6.9 of 8.2 GB [hos host] vk: program census (play): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 29 program records loaded (17.4 ms), 10 new graphics pipelines, pipelines from the driver's cache 13, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (play): 6554 queued, 6554 loaded (median 1.46 ms, 99th percentile 16.63 ms, worst 38.13 ms, 16489 ms and 351.4 MB read in all); 8864 programs (1634 of them compute variants without their code, 0 graphics programs no note names), 12363 graphics pipelines (1 not in a cache), 744 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2469 pipelines made at create taken by draws; 3157 create calls with nothing kept for their stages, 0 before the index; units 9711 (6554 ready, 0 loading), 17807 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (play): 14090 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (play): 789 known, 32 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: a frame: 340 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.36, present 0.18, draw 7.74, descriptors 3.31, texture lookups 0.31, uploads 0.12, pipelines 0.33, programs 1.36, commands 2.15, dispatch 1.13, memory sync 0.57, queue submit 0.00, slot wait 0.26 [hos host] vk: a frame: mirrors 557 (1009 MB): 6.53 MB synced in 1194.8 copies, 19.40 MB written back [hos host] vk: 22200 presents, 56.3 fps; a frame: 2316 draws (0 skipped), 230 dispatches, 213 passes, 148 barriers, 6 submits, 1.3 uploads (0.3 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 14231 pipelines (3752 ms), 0 programs (10 failed), 1221 textures (2243 MB) [hos host] vk: host time in GPU resource waits: 5.08 ms a frame (fences 1.48, timelines 3.60, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 815.1 full barriers recorded, 328.4 left out (nothing recorded since the one before), 35.1 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 5.0; 0x8: 3.0; 0x9: 1.0; 0x10: 3.0; 0x11: 40.9; 0x12: 69.7; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 139.5, an op 1278.0, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.4 ms a frame (the longest submission 4.2 ms); 1.4 textures made a frame: colour targets 0.1 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 1.3 (1.8 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.15 MB unswizzled on the CPU in 0.02 ms [hos host] vk: the render thread (600 presents, 17.41 ms a frame): its jobs 16.39 ms a frame (submits 2.61, presents 5.23, others 8.55), idle 1.02; the game waited for room at 30 presents, 0.52 ms each (0.03 a frame; 15.61 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 503 captures at a CPU signal; the submits' waits read in 0.038 ms a frame; readbacks: 2590 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 15.68 ms after its hand-over (40.32 at most) [hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 2.79 + 0.26 (the window's copy is in the idle), after it 1.55 + 0.77; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.00 ms of GPU time), after its last that reads it 1.00 (0.00 ms) [hos host] vk: frame ahead (600 presents): 9.67 parts a frame recorded in the signal's job, 77.2% of them took nothing that came with the signal; of the others, 2.19 a frame took pages of the signal's capture into a mirror (55.2 pages a frame) and 0.02 pages read as guest memory had them (0.2 pages a frame); a frame's first part took nothing in 15 of 503 frames, its first three in 15 [hos host] vk: frame ahead, what came with the signal: 503 signals with a capture, each of 85.9 pages in 10.7 pools; between a frame's last submit and that signal 0.03 texture and sampler registrations, 0.21 releases and 0.00 other CPU signals a signal (9 of them with any) [hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.62+0.14, 0.24+0.11, 0.16+0.05, 0.03+0.02, 0.03+0.02, 0.12+0.16, 0.19+0.19, 0.09+0.05, 0.05+0.02, 0.02+0.01, 0.00+0.00, 0.00+0.00 [hos host] vk: render targets drawn over from another start, let go: 13163 so far (62304 MB), 11938 of them (56633 MB) between the passes [hos host] vk: 4 textures retired (1222 left, 2242 MB): budget 4, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: 2900000 recordings made, 65646 live copies (27.5 MB), 6922 texture copies; 1 command buffers' segments (4.0 MB) [hos host] vk: replay reads of guest memory as submitted: 17460000 (10659529 pages from the captures) [hos host] vk: write-backs left out (600 submits): 824, 3661.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1744 dirty pages from their submits' captures, 54337 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 903 dispatches, 41770 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 10295 runs, 672.0 MB compared; 73 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 23053 in captures, 0 only in the collection, 25 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 4 textures retired (1229 left, 2255 MB): budget 4, drawn over 0, dead targets 0 [hos sead::MainThread] vk: capture: a submit 0.16 ms (lock 0.00, scan 0.03 over 578 ranges, 3288 MB), 296 pages written, 296 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.040; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.014, bytes copied 0.000; 25 kept (138 MB), 12 spare buffers, 3802 texture pages copied (14 MB) [hos host] vk: replay reads of guest memory as submitted: 17480000 (10671203 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 17500000 (10682685 pages from the captures) [hos host] vk: 7 textures retired (1226 left, 2252 MB): budget 7, drawn over 0, dead targets 0 [hos host] vk: sparse mirrors: 122, 1009 MB with memory of 1472 MB (462 MB spared); blocks given memory later 13843, given back 18933; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1462 binds (353.5 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 392 taken from the spares, 166 made new; 436 kept as spares, 0 destroyed; 44 spares now, 1368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2168 dirty pages from their submits' captures, 50590 from earlier ones', 916 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 873 dispatches, 39617 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 10278 runs, 641.1 MB compared; 69 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22800 in captures, 0 only in the collection, 23 only by the replay itself [hos host] vk: read live, by pool: 0x1055800000 (4 MB, captured, state 2): 916; [hos host] vk: imports: 199 live (2176 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos sead::MainThread] nvn: GPU waits in 5 s: 251 sync waits (842 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: reaper: 422 images (745 MB) destroyed in 10 s, 0 of them where they were handed over (its cap or VRAM pressure); the queue at most 129 (175 MB) [hos ModuleSystemWorker5] vk: readbacks of GPU results: 1085 of 3000 waited for their write-back, 0.04 ms each; 0 looked at the stopped queues first, 21 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker5] vk: invalidates since the last readbacks line: 1100 kept their virtual core (nothing there to wait for), 3252 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: 4 textures retired (1224 left, 2245 MB): budget 4, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17520000 (10694502 pages from the captures) [hos host] vk: 10 textures retired (1218 left, 2247 MB): budget 10, drawn over 0, dead targets 0 [hos Presentation Thread] vi: 22263 frames presented (57.1 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 21, 15.5-17.8 200, 17.8-22 49, 22-30 12, 30-40 3, over 40 1 [hos host] vk: replay reads of guest memory as submitted: 17540000 (10705862 pages from the captures) [hos host] vk: 11 textures retired (1208 left, 2234 MB): budget 11, drawn over 0, dead targets 0 [hos host] vk: program picks: 25233330 kept (10500052 by the last picks' textures, a uniform buffer bound since), 26735 new, 9139934 with other textures (of 34400000) [hos host] vk: write-backs left out (600 submits): 799, 3550.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2325 dirty pages from their submits' captures, 50932 from earlier ones', 208 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 39594 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9370 runs, 623.7 MB compared; 63 pages read into the shadows; 288 MB of shadows (55 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 22282 in captures, 0 only in the collection, 23 only by the replay itself [hos host] vk: read live, by pool: 0x7fb7c00000 (12 MB, captured, state 2): 118; 0x7fe2800000 (4 MB, captured, state 2): 90; [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 19477 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 17560000 (10718470 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 17580000 (10730361 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05 [hos host] vk: late pages found at the replay (600 takes): 15 pages in 15 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 575 ranges, 3289 MB), 290 pages written, 290 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.042; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.015, bytes copied 0.000; 25 kept (142 MB), 11 spare buffers, 5629 texture pages copied (21 MB) [hos host] vk: replay reads of guest memory as submitted: 17600000 (10741613 pages from the captures) [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 123, 1005 MB with memory of 1472 MB (466 MB spared); blocks given memory later 13843, given back 18944; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1468 binds (354.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 394 taken from the spares, 166 made new; 437 kept as spares, 0 destroyed; 43 spares now, 1368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1771 dirty pages from their submits' captures, 49583 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 851 dispatches, 38969 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 9130 runs, 628.1 MB compared; 61 pages read into the shadows; 288 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21597 in captures, 0 only in the collection, 23 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 200 live (2180 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos host] vk: 2 textures retired (1223 left, 2244 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: 17 textures retired (1214 left, 2251 MB): budget 17, drawn over 0, dead targets 0 [hos host] vk: a frame: 358 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.38, present 0.19, draw 7.75, descriptors 3.16, texture lookups 0.19, uploads 0.05, pipelines 0.33, programs 1.41, commands 2.25, dispatch 1.21, memory sync 0.57, queue submit 0.00, slot wait 0.25 [hos host] vk: a frame: mirrors 560 (1005 MB): 6.59 MB synced in 1253.7 copies, 20.08 MB written back [hos host] vk: 22320 presents, 57.9 fps; a frame: 2424 draws (0 skipped), 271 dispatches, 221 passes, 148 barriers, 6 submits, 0.5 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14232 pipelines (3752 ms), 0 programs (10 failed), 1215 textures (2251 MB) [hos host] vk: host time in GPU resource waits: 5.23 ms a frame (fences 1.84, timelines 3.39, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 858.8 full barriers recorded, 330.2 left out (nothing recorded since the one before), 43.5 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 40.5; 0x12: 70.4; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 46.1 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 139.4, an op 1346.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.6 ms a frame (the longest submission 3.8 ms); 0.6 textures made a frame: colour targets 0.1 (0.3 MB), depth 0.0 (0.1 MB), ASTC or BC 0.5 (0.5 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.12 MB unswizzled on the CPU in 0.02 ms [hos host] vk: replay reads of guest memory as submitted: 17620000 (10752559 pages from the captures) [hos host] vk: 1 textures retired (1214 left, 2251 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: pool 0x7fb3db5d00+0x74b8 imported in 2.4 ms as 0x7fb3c00000..0x7fb4000000 (251 imports, 737 ms in all) [hos host] vk: replay reads of guest memory as submitted: 17640000 (10764321 pages from the captures) [hos ModuleSystemWorker1] nvn: CommandBufferWaitEvent #440000: event 0x100d6c8010 mode 0 value 323900 (now 323881) (lr 0x80660e38) [hos host] vk: readback pages refreshed from guest memory: 7368204 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1881 dirty pages from their submits' captures, 52335 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 913 dispatches, 40935 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8757 runs, 651.6 MB compared; 69 pages read into the shadows; 288 MB of shadows (53 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 21083 in captures, 0 only in the collection, 28 only by the replay itself [hos host] vk: read live, by pool: [hos aal::LowPrioWorkerThread] vcores: 146913 takes, 24520 waits (mean 80 us, longest 10.2 ms, 0 over HOS_VCORE_LOG_MS), 11622 kicks, 11855 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: replay reads of guest memory as submitted: 17660000 (10776300 pages from the captures) [hos host] vk: program picks: 25388129 kept (10555272 by the last picks' textures, a uniform buffer bound since), 26745 new, 9185125 with other textures (of 34600000) [hos host] vk: 7 textures retired (1238 left, 2272 MB): budget 7, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17680000 (10788473 pages from the captures) [hos host] vk: 17 textures retired (1226 left, 2270 MB): budget 17, drawn over 0, dead targets 0 [hos host] vk: queues ordered by their events: 150000 submissions stopped at a wait, 149938 went on once it was reached, 60 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 5 (stopped 50.30 ms at most); waits met 440670, not met 0; captures at a CPU event's wait 13793, 0 of them at a wait met on arrival [hos host] vk: sparse mirrors: 123, 1004 MB with memory of 1472 MB (467 MB spared); blocks given memory later 13843, given back 18946; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1476 binds (356.0 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 398 taken from the spares, 166 made new; 441 kept as spares, 0 destroyed; 43 spares now, 1368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2460 dirty pages from their submits' captures, 50474 from earlier ones', 45 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 910 dispatches, 39771 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8372 runs, 630.5 MB compared; 73 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19628 in captures, 1 only in the collection, 25 only by the replay itself [hos host] vk: read live, by pool: 0x7fa8400000 (4 MB, captured, state 2): 45; [hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos host] vk: replay reads of guest memory as submitted: 17700000 (10800966 pages from the captures) [hos host] vk: 20 textures retired (1210 left, 2267 MB): budget 20, drawn over 0, dead targets 0 [hos host] vk: 1 textures retired (1209 left, 2267 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17720000 (10812909 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.01, scan 0.04 over 564 ranges, 3332 MB), 280 pages written, 280 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.029; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 25 kept (170 MB), 10 spare buffers, 5896 texture pages copied (23 MB) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05 [hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #325000: event 0x100d6c8004 mode 0 value 324999 (lr 0x80291504) [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1043 of 3000 waited for their write-back, 0.01 ms each; 0 looked at the stopped queues first, 15 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 851 kept their virtual core (nothing there to wait for), 2733 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 17740000 (10825151 pages from the captures) [hos host] vk: write-backs left out (600 submits): 799, 3550.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1852 dirty pages from their submits' captures, 50125 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 882 dispatches, 38811 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8140 runs, 594.9 MB compared; 85 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19605 in captures, 0 only in the collection, 22 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 8 textures retired (1204 left, 2261 MB): budget 8, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17760000 (10836493 pages from the captures) [hos host] vk: 13 textures retired (1192 left, 2244 MB): budget 13, drawn over 0, dead targets 0 [hos host] vk: program picks: 25544529 kept (10613961 by the last picks' textures, a uniform buffer bound since), 26745 new, 9228725 with other textures (of 34800000) [hos host] vk: 10 textures retired (1193 left, 2253 MB): budget 10, drawn over 0, dead targets 0 [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15283 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 17780000 (10848295 pages from the captures) [hos host] vk: sparse mirrors: 123, 1004 MB with memory of 1472 MB (468 MB spared); blocks given memory later 13845, given back 18953; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1480 binds (356.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 398 taken from the spares, 166 made new; 441 kept as spares, 0 destroyed; 43 spares now, 1368 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 775, 3441.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1547 dirty pages from their submits' captures, 47350 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 834 dispatches, 36972 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 8152 runs, 616.5 MB compared; 62 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 19097 in captures, 0 only in the collection, 19 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos host] vk: a frame: 345 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 12.64, present 0.17, draw 7.99, descriptors 3.27, texture lookups 0.14, uploads 0.02, pipelines 0.35, programs 1.43, commands 2.31, dispatch 1.04, memory sync 0.57, queue submit 0.00, slot wait 0.27 [hos host] vk: a frame: mirrors 564 (1004 MB): 6.32 MB synced in 1201.2 copies, 19.18 MB written back [hos host] vk: 22440 presents, 58.0 fps; a frame: 2513 draws (0 skipped), 216 dispatches, 228 passes, 135 barriers, 6 submits, 0.5 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14233 pipelines (3752 ms), 0 programs (10 failed), 1193 textures (2253 MB) [hos host] vk: host time in GPU resource waits: 5.04 ms a frame (fences 1.44, timelines 3.61, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 808.6 full barriers recorded, 317.4 left out (nothing recorded since the one before), 48.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.7; 0x8: 3.0; 0x9: 1.0; 0x10: 2.7; 0x11: 27.0; 0x12: 70.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 126.0, an op 1357.8, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.4 ms a frame (the longest submission 3.8 ms); 0.5 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.5 (0.4 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms [hos host] vk: replay reads of guest memory as submitted: 17800000 (10859974 pages from the captures) [hos sead::MainThread] nvn: DrawMeshTasksIndirect: 800000 so far (x1 0x117c952010 x2 0 x3 0x117c952000 x4 0x168040) [hos host] vk: late pages found at the replay (600 takes): 20 pages in 20 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: replay reads of guest memory as submitted: 17820000 (10873649 pages from the captures) [hos host] vk: 11 textures retired (1186 left, 2221 MB): budget 11, drawn over 0, dead targets 0 [hos host] vk: 5 textures retired (1181 left, 2211 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: write-backs left out (600 submits): 776, 3443.0 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2103 dirty pages from their submits' captures, 47461 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 852 dispatches, 37422 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7770 runs, 601.1 MB compared; 61 pages read into the shadows; 288 MB of shadows (55 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18819 in captures, 0 only in the collection, 23 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 17840000 (10885931 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.03 over 566 ranges, 3358 MB), 274 pages written, 274 copied (1.1 MB) [hos host] vk: a capture taken (ms): all 0.036; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.011, bytes copied 0.000; 24 kept (118 MB), 11 spare buffers, 5982 texture pages copied (23 MB) [hos host] vk: 7 textures retired (1188 left, 2228 MB): budget 7, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 17860000 (10898607 pages from the captures) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.05, writes collected 0.05 [hos host] vk: replay reads of guest memory as submitted: 17880000 (10910585 pages from the captures) [hos host] vk: readback pages refreshed from guest memory: 7438107 pages so far (0 held: may be written ahead in their submission) [hos sead::MainThread] nvn: EventSignal (CPU) #45000: event 0x1006e38004 mode 0 value 1 (lr 0x8348c78c) [hos host] vk: program picks: 25701719 kept (10674154 by the last picks' textures, a uniform buffer bound since), 26746 new, 9271534 with other textures (of 35000000) [hos host] vk: sparse mirrors: 127, 1014 MB with memory of 1496 MB (481 MB spared); blocks given memory later 13845, given back 18954; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1485 binds (358.2 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 402 taken from the spares, 166 made new; 441 kept as spares, 0 destroyed; 39 spares now, 1344 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 775, 3441.2 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 3353 dirty pages from their submits' captures, 50299 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 877 dispatches, 37835 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 7320 runs, 581.3 MB compared; 59 pages read into the shadows; 288 MB of shadows (55 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18659 in captures, 0 only in the collection, 30 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos host] vk: replay reads of guest memory as submitted: 17900000 (10923077 pages from the captures) [hos host] vk: replay reads of guest memory as submitted: 17920000 (10935147 pages from the captures) [hos host] vk: 3 textures retired (1204 left, 2237 MB): budget 3, drawn over 0, dead targets 0 [hos sead::MainThread] nvn: GPU waits in 5 s: 278 sync waits (955 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms) [hos host] vk: write-backs left out (600 submits): 825, 3663.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1693 dirty pages from their submits' captures, 47888 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 892 dispatches, 37601 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6859 runs, 621.3 MB compared; 64 pages read into the shadows; 288 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18238 in captures, 0 only in the collection, 25 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: 7 textures retired (1207 left, 2236 MB): budget 7, drawn over 0, dead targets 0 [hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #445000: event 0x100d6c800c mode 0 value 327699 (now 327680) (lr 0x80660e38) [hos host] vk: replay reads of guest memory as submitted: 17940000 (10947248 pages from the captures) [hos Presentation Thread] vi: 22546 frames presented (56.5 fps over the last 5.0 s, interval 1) [hos Presentation Thread] vi: present intervals: <15.5 ms 12, 15.5-17.8 169, 17.8-22 91, 22-30 11, 30-40 0, over 40 0 [hos ModuleSystemWorker3] vk: readbacks of GPU results: 1034 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 12 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos ModuleSystemWorker3] vk: invalidates since the last readbacks line: 944 kept their virtual core (nothing there to wait for), 3152 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: replay reads of guest memory as submitted: 17960000 (10959457 pages from the captures) [hos host] vk: 5 textures retired (1224 left, 2251 MB): budget 5, drawn over 0, dead targets 0 [hos host] vk: a frame: 363 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 13.66, present 0.17, draw 8.65, descriptors 3.54, texture lookups 0.24, uploads 0.07, pipelines 0.36, programs 1.57, commands 2.51, dispatch 1.00, memory sync 0.57, queue submit 0.00, slot wait 0.29 [hos host] vk: a frame: mirrors 570 (1017 MB): 6.33 MB synced in 1175.7 copies, 18.78 MB written back [hos host] vk: 22560 presents, 54.6 fps; a frame: 2676 draws (0 skipped), 174 dispatches, 230 passes, 135 barriers, 6 submits, 0.7 uploads (0.2 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 14241 pipelines (3757 ms), 0 programs (10 failed), 1224 textures (2251 MB) [hos host] vk: host time in GPU resource waits: 5.12 ms a frame (fences 1.32, timelines 3.80, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 769.4 full barriers recorded, 317.7 left out (nothing recorded since the one before), 49.0 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 4.8; 0x8: 3.0; 0x9: 1.0; 0x10: 2.8; 0x11: 27.0; 0x12: 70.6; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; 0x9b: 0.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 126.1, an op 1472.5, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.5 ms a frame (the longest submission 3.5 ms); 0.7 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.7 (0.8 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.11 MB unswizzled on the CPU in 0.01 ms [hos Presentation Thread] vk: capture: a submit 0.20 ms (lock 0.00, scan 0.04 over 568 ranges, 3376 MB), 307 pages written, 307 copied (1.2 MB) [hos host] vk: a capture taken (ms): all 0.035; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.013, bytes copied 0.000; 24 kept (88 MB), 12 spare buffers, 7123 texture pages copied (27 MB) [hos sead::MainThread] nvn: 2950000 recordings made, 66726 live copies (25.3 MB), 6923 texture copies; 5 command buffers' segments (4.3 MB) [hos host] vk: sparse mirrors: 129, 1017 MB with memory of 1504 MB (486 MB spared); blocks given memory later 13846, given back 18957; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1489 binds (358.9 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 404 taken from the spares, 166 made new; 441 kept as spares, 0 destroyed; 37 spares now, 1336 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 850, 3774.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2242 dirty pages from their submits' captures, 49540 from earlier ones', 2 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 956 dispatches, 38987 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6605 runs, 617.1 MB compared; 83 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 17465 in captures, 0 only in the collection, 20 only by the replay itself [hos host] vk: read live, by pool: 0x1042000000 (4 MB, captured, state 2): 1; 0x1041c00000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos host] vk: 5 textures retired (1247 left, 2266 MB): budget 0, drawn over 0, dead targets 5 [hos host] vk: replay reads of guest memory as submitted: 17980000 (10972355 pages from the captures) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 11948 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: program picks: 25858579 kept (10736833 by the last picks' textures, a uniform buffer bound since), 26751 new, 9314669 with other textures (of 35200000) [hos host] vk: replay reads of guest memory as submitted: 18000000 (10985186 pages from the captures) [hos host] vk: late pages found at the replay (600 takes): 13 pages in 13 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.04, writes collected 0.05 [hos host] vk: recorder: 7589637 packets run in 10.0 s (busy 2.02 s, spun idle 0.13 s, slept 12026 times, 4084 of them under 200 us), 0 calls made directly (no room); 19492 batches over 567 frames (34.4 a frame, 389 packets a batch), 10300 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 88 times, 8.64 ms (0.2 times, 0.02 ms a frame); every thread 869 times, 86.17 ms in all, the longest 1.03 ms; vkWaitForFences 515/45.41 ms (max 0.51), vmaDestroyBuffer 103/17.22 ms (max 1.03), the queue's lock 74/8.42 ms (max 0.47), vkDestroyImageView 50/5.41 ms (max 0.43), vmaFreeMemory 26/4.10 ms (max 0.82), vkQueueBindSparse 51/2.88 ms (max 0.30) [hos host] vk: recorder: of its busy 2.02 s, 0.59 s in parts' ends and submits (the longest 1.50 ms), 0.24 s of that blocked on the queue's lock (373 times, the longest 1.33 ms), so 0.35 s in the ends, submits and holds; 1.43 s running the other packets; 13386 packets a frame [hos host] vk: replay reads of guest memory as submitted: 18020000 (10996711 pages from the captures) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1904 dirty pages from their submits' captures, 49693 from earlier ones', 1 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 943 dispatches, 39089 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6555 runs, 637.4 MB compared; 74 pages read into the shadows; 288 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18941 in captures, 1 only in the collection, 17 only by the replay itself [hos host] vk: read live, by pool: 0x7fb6000000 (4 MB, captured, state 2): 1; [hos host] vk: replay reads of guest memory as submitted: 18040000 (11008256 pages from the captures) [hos host] vk: 6 textures retired (1265 left, 2274 MB): budget 0, drawn over 6, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 18060000 (11019795 pages from the captures) [hos aal::LowPrioWorkerThread] vcores: 146564 takes, 24783 waits (mean 82 us, longest 21.7 ms, 0 over HOS_VCORE_LOG_MS), 11725 kicks, 11942 yields, 0 starved grants, 0 escapes in 5 s [hos host] vk: the recording streams: 48 of 1 MB, the most one held 268 KB; 0 made on demand (0 MB kept now); c1 stages: at most 135 a recording; 10 descriptor pools made when a recording's own ran out (10 slots' pools made whole); a recording's pool took at most 484 sets (ub 5493, sb 2476, cis 41232, si 0, utb 0, stb 0) [hos host] vk: sparse mirrors: 129, 1014 MB with memory of 1508 MB (493 MB spared); blocks given memory later 13846, given back 18970; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1492 binds (359.1 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 405 taken from the spares, 166 made new; 442 kept as spares, 0 destroyed; 37 spares now, 1332 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2204 dirty pages from their submits' captures, 49678 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 926 dispatches, 39393 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6475 runs, 627.2 MB compared; 70 pages read into the shadows; 288 MB of shadows (54 MB of pages known, in 108 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18339 in captures, 0 only in the collection, 18 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos host] vk: 7 textures retired (1258 left, 2268 MB): budget 6, drawn over 1, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 18080000 (11031331 pages from the captures) [hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.03 over 568 ranges, 3611 MB), 240 pages written, 240 copied (0.9 MB) [hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 24 kept (128 MB), 12 spare buffers, 7123 texture pages copied (27 MB) [hos host] vk: program picks: 26015220 kept (10800824 by the last picks' textures, a uniform buffer bound since), 26751 new, 9358028 with other textures (of 35400000) [hos host] vk: 7 textures retired (1275 left, 2254 MB): budget 7, drawn over 0, dead targets 0 [hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #330000: event 0x100d6c8008 mode 0 value 329997 (lr 0x80660f34) [hos host] vk: replay reads of guest memory as submitted: 18100000 (11042877 pages from the captures) [hos host] vk: 13 textures retired (1262 left, 2245 MB): budget 13, drawn over 0, dead targets 0 [hos host] vk: readback pages refreshed from guest memory: 7508580 pages so far (0 held: may be written ahead in their submission) [hos host] vk: write-backs left out (600 submits): 875, 3885.3 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 1963 dirty pages from their submits' captures, 48468 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 894 dispatches, 38797 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 6375 runs, 630.3 MB compared; 75 pages read into the shadows; 288 MB of shadows (55 MB of pages known, in 111 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 18061 in captures, 0 only in the collection, 14 only by the replay itself [hos host] vk: read live, by pool: [hos host] vk: replay reads of guest memory as submitted: 18120000 (11054422 pages from the captures) [hos host] vk: 2 textures retired (1268 left, 2250 MB): budget 2, drawn over 0, dead targets 0 [hos host] vk: a frame: 350 draws kept their neighbour's descriptors (0 of them through other slots' registrations) [hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 102 ms in all, 0 jobs queued [hos host] vk: a frame's replay time (ms): submit 11.85, present 0.16, draw 7.65, descriptors 3.11, texture lookups 0.18, uploads 0.05, pipelines 0.33, programs 1.37, commands 2.24, dispatch 0.89, memory sync 0.40, queue submit 0.00, slot wait 0.27 [hos host] vk: a frame: mirrors 571 (1014 MB): 5.73 MB synced in 1112.1 copies, 18.07 MB written back [hos host] vk: 22680 presents, 58.9 fps; a frame: 2557 draws (0 skipped), 146 dispatches, 226 passes, 143 barriers, 6 submits, 1.1 uploads (0.1 MB), 81 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 14241 pipelines (3757 ms), 0 programs (10 failed), 1280 textures (2259 MB) [hos host] vk: host time in GPU resource waits: 6.27 ms a frame (fences 1.67, timelines 4.60, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking) [hos host] vk: a frame: 730.8 full barriers recorded, 323.9 left out (nothing recorded since the one before), 48.9 pass starts ordered in their fragment stages only (draw passes after draw passes) [hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.6; 0x8: 3.0; 0x9: 1.0; 0x10: 0.6; 0x11: 39.6; 0x12: 70.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0; [hos host] vk: a frame: indirect mesh draws' arguments (RX_VK_MESH_CONV=2): 0.0 converted one by one, 47.0 from 30.0 batches at the game's barriers (49.0 converted in them; 0.0 dropped after a write, 0.0 the ring came round), 0.0 for the compute fallback (no mesh shaders in use: one by one); scans ended at a barrier 134.2, an op 1439.9, the end 7.0, full 0.0; the draws batched by a scan right after an op that ended the one before, by op: 4: 2.0; 5: 26.0; 48: 8.0; [hos host] vk: GPU time 12.2 ms a frame (the longest submission 3.2 ms); 1.1 textures made a frame: colour targets 0.2 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.9 (0.7 MB), other 0.0 (0.0 MB) [hos host] vk: a frame: 0.08 MB unswizzled on the CPU in 0.01 ms [hos host] vk: 1 textures retired (1291 left, 2269 MB): budget 1, drawn over 0, dead targets 0 [hos host] vk: replay reads of guest memory as submitted: 18140000 (11066488 pages from the captures) [hos sead::MainThread] vk: readbacks of GPU results: 1020 of 3000 waited for their write-back, 0.09 ms each; 0 looked at the stopped queues first, 9 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror) [hos sead::MainThread] vk: invalidates since the last readbacks line: 1039 kept their virtual core (nothing there to wait for), 3057 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept) [hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.03, writes collected 0.04 [hos host] vk: a submission's parts ended: 116043 where the game had ended its render pass, 23957 at 512 draws [hos host] vk: late pages' fix (600 recordings with one): of 1139 captures with late pages 1136 fixed (91.6 pages, 92.2 fixes a capture), waited as before: 0 no pool, 3 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 444 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 896 blocks (57344 KB); of the reads on the CPU, 5 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 163, of an earlier recording 276, not recorded yet 0, of this part 0; pages fixed from stamps 103080, from snapshots 1009; the shadows: 555761 pages synced from captures, 542355 made known by it, 13406 left unknown because the replay or the collection wrote them since [hos host] vk: replay reads of guest memory as submitted: 18160000 (11079306 pages from the captures) [hos host] vk: sparse mirrors: 130, 1019 MB with memory of 1516 MB (496 MB spared); blocks given memory later 13846, given back 18970; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 1493 binds (359.3 ms); 0 made whole for want of memory [hos host] vk: sparse mirrors' buffers: 406 taken from the spares, 166 made new; 442 kept as spares, 0 destroyed; 36 spares now, 1324 MB of address space (RX_VK_SPARSE_SPARE_MB=4096) [hos host] vk: write-backs left out (600 submits): 849, 3772.5 MB; 29 bindings followed, 40 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1061 so far [hos host] vk: mirror syncs (600 submits): 2356 dirty pages from their submits' captures, 46235 from earlier ones', 1 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE) [hos host] vk: mirror syncs' pass (600 submits): 869 dispatches, 37332 chunks; 0 syncs as copies (no table room) [hos host] vk: write-back diff (600 submits): 5948 runs, 604.1 MB compared; 96 pages read into the shadows; 288 MB of shadows (56 MB of pages known, in 112 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 16838 in captures, 0 only in the collection, 13 only by the replay itself [hos host] vk: read live, by pool: 0x7fd3800000 (4 MB, captured, state 2): 1; [hos host] vk: imports: 202 live (2188 MB), 7 retired still imported (176 MB), 53 retired given back (1304 MB) [hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9978 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged) [hos host] vk: replay reads of guest memory as submitted: 18180000 (11092354 pages from the captures) [hos host] vk: 10 textures retired (1369 left, 2417 MB): budget 10, drawn over 0, dead targets 0 [hos host] vk: program picks: 26170057 kept (10862153 by the last picks' textures, a uniform buffer bound since), 26761 new, 9403181 with other textures (of 35600000) [hos host] vk: replay reads of guest memory as submitted: 18200000 (11105379 pages from the captures) [hos host] vk: 41 textures retired (1396 left, 2501 MB): budget 3, drawn over 0, dead targets 38 [hos host] vk: late pages found at the replay (600 takes): 19 pages in 19 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found) [hos host] vk: slow upload 15.0 ms: #27966 0x10c9e7e200 1024x1024x1 x1 levels 11 NVN 0x7b ASTC, 912 of 912 KB [hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.01, scan 0.03 over 568 ranges, 3663 MB), 252 pages written, 252 copied (1.0 MB) [hos PlacementMgr] sync: object cache (HOS_OBJ_CACHE=16384 a thread, 42 made): 15235183 lookups, 123849 missed (0.81 %) [hos host] vk: a capture taken (ms): all 0.030; GPU wait for late pages 0.009 (2 of 600 takes), pools lock 0.000, retires and stamps 0.008, bytes copied 0.000; 25 kept (154 MB), 9 spare buffers, 7777 texture pages copied (30 MB) [hos host] vk: program census (close): created 0 (0 with stages; 0 with no present for 100 ms before, 0 in a frame of 100 ms or more; 0.0 ms of the game's threads), freed 0 (0 never used); first uses 0: program ready 0, loaded 0, translated 0, failed 0; pipeline ready 0, from a cache 0, compiled 0, made 0, left out 0, none 0; 0 with no create call seen, 0 whose stages are not the create call's; after the create call: under 16.7 ms 0, under 100 ms 0, under 1 s 0, longer 0; needing work 0 (0 of them 100 ms or more after the create call, 0 created in a load or a long frame); in play: 31 program records loaded (30.6 ms), 10 new graphics pipelines, pipelines from the driver's cache 14, from the start's cache 0, compiled 2; 10 draws and dispatches left out, 0 waits (0.0 ms) [hos host] vk: program loads at create (close): 6566 queued, 6566 loaded (median 1.46 ms, 99th percentile 16.63 ms, worst 38.13 ms, 16513 ms and 351.7 MB read in all); 8876 programs (1634 of them compute variants without their code, 0 graphics programs no note names), 12373 graphics pipelines (1 not in a cache), 746 compute pipelines (34 not in a cache); a first use ran 1 itself and waited for 1 (2.8 ms); 2494 pipelines made at create taken by draws; 3157 create calls with nothing kept for their stages, 0 before the index; units 9723 (6566 ready, 0 loading), 17816 programs alive; 4 noted graphics pipeline probes deferred to first use [hos host] vk: program loads at create (close): 14110 shader modules let go once their programs' noted pipelines were built, 0 programs' modules made again for a new pipeline state (RX_VK_DROP_LOADED_MODULES=0 keeps them) [hos host] vk: module identifiers (close): 792 known, 35 learned in this run; 715 compute pipelines made by them at a create call, 32 asked that way were in no cache (their code read, as before) [hos host] vk: indexed descriptor arrays (close): 1343 in 1343 of 20151 program stages, 0 of them near a limit [hos host] vk: indexed descriptor arrays (close), fragment stages: 1343 texture (17-64: 1343); the longest 64 [hos host] vk: page table census (close): no pool outside its windows, 64-72 GB and 508-512 GB (4456 pools in the table now) [hos host] vk: pipeline cache: play file 11 saved (4 KB): 1 pipelines, 1 from the driver's cache, 0 compiled; 68 more from the start's cache [hos host] vk: the window is closing: the swapchain is let go with its images (what the compositor needs back for its final frames) [hos host] display: window closed, exiting [hos host] stop: close work done (a file the game was writing waited for, at most 3 s) ### Screenshots or video _No response_ ### Before you send it - [x] The problem is still there on the newest release or nightly. - [x] I searched the open issues for it. - [x] I'm not attaching game files, keys, the game's executable, its .bin caches or saves.
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
MagicBeanDev/Longshot#2077
No description provided.