WIP: perf(vk): programs with the same bindings share one set layout and pipeline layout, so a program change pushes only what changed #800

Draft
dyst wants to merge 1 commit from dyst/Longshot:perf/vk-shared-pipeline-layouts into main
Member

Type: perf · Risk: low · Tested on: Windows 11 (the game-free checks; the game A/B is next) · Needs: a game run on NVIDIA, AMD and Intel

Why

Each shader program had its own descriptor set layout and pipeline layout, and the replay pushes only the descriptors that changed only while the layout is the one it pushed with: every program change pushed every descriptor again, even between programs whose bindings are laid out the same way. Programs change often within a pass. Found in an audit of the renderer's per-draw path; no issue on the tracker.

What changed

Programs whose layouts' create infos are the same now share one set layout and pipeline layout, kept for the device's life, so a program change between them pushes only the writes that differ. Pipelines are made from the same create infos as before and no pipeline key holds a layout handle.

  • draw: layout_signature, layouts_for, SharedLayouts and ensure_layout; the preparation's compute warm leaves shared layouts alone (gpu/vk/vk_draw.cpp)
  • program cache: program_cache_borrow leaves shared layouts alone (gpu/vk/vk_progcache.cpp); Program::layout_shared (gpu/vk/vk_shader.h, not a stamped translator file)
  • docs: the RX_VK_LAYOUT_EACH=1 row in docs/development.md
  • tests: tests/test_shared_layouts.py

Measurements

Verdict: not measured yet: the measurements are pending from the RX_VK_BENCH A/B of this PR @ e2a24c1a against main @ b9e95173 (under way), which will replace this section.

Flags and settings

One switch for an A/B.

Flag What it is Before this PR After this PR Effect Helps / doesn't help / untested Recommendation
RX_VK_LAYOUT_EACH=1 A set layout and pipeline layout for every program, as before - unset (shared) On: every program change pushes everything measurement pending leave unset; =1 for an A/B

Testing

Tested: Windows 11 (Ryzen 9 3900X), llvm-mingw and the Vulkan headers of the builder's v2.0.14 tools: rebased onto main @ b9e95173 (this PR @ e2a24c1a; it stands alone on main, on no other PR). The 135 checks that read this diff (listed below: those under tests/ and nx2/ that name a changed file, and the ones this PR adds or changes), not the whole suite: python -I -B tests/run_all.py --jobs 6 <those checks> with RX_CPM at the builder's CPM cache: 127 passed, 7 failed; each failure (tests/test_crash_dialog.py, tests/test_linux_driver_fault_guard.py, tests/test_oot_state_tap.py, tests/test_pipeline_tour.py, tests/test_vk_occlusion_queries.py, tests/test_vk_pgo.py, tests/test_walk_key.py) also fails on main @ b9e95173 in the same checks, on the same PC and the same day.

The 135 checks

nx2/test_cbuf_table_extent.py nx2/test_dump_ssbo_read.py nx2/test_mesh_dynamic_state.py tests/test_amd_fragment_warp.py tests/test_bake_ds.py tests/test_bank_gen.py tests/test_bc_raw_targets.py tests/test_check_pr_description.py tests/test_commit_room.py tests/test_crash_dialog.py tests/test_desc_clock.py tests/test_desc_counts.py tests/test_dev_packs.py tests/test_device_lost_exit.py tests/test_dispatch_clock.py tests/test_dlss_checks.py tests/test_draw_binds.py tests/test_draw_side_effects.py tests/test_draw_stream_visibility.py tests/test_drawn_page_map.py tests/test_driver_cache_room.py tests/test_driver_faults.py tests/test_dummy_views.py tests/test_equal_depth_margin.py tests/test_fallback_buffer_cleanup.py tests/test_fallback_depth_hold.py tests/test_fallback_loaded.py tests/test_fallback_pipeline_state.py tests/test_fallback_size_policy.py tests/test_flat_map.py tests/test_framegen_preset.py tests/test_fsr_uniform_ownership.py tests/test_global_table_loop.py tests/test_gpl_cold_library.py tests/test_gpl_library_cache.py tests/test_handle_lookup.py tests/test_handle_table_arrays.py tests/test_hold_gpu.py tests/test_intel_compiler_gate.py tests/test_intel_quick_compute.py tests/test_jobs_failure.py tests/test_jobs_wait_runs_queued.py tests/test_linux_driver_fault_guard.py tests/test_linux_pressure.py tests/test_load_default.py tests/test_masked_clear_failure.py tests/test_mem_trim.py tests/test_menu_env_on.py tests/test_mesh_args_cleanup.py tests/test_mesh_fallback.py tests/test_mesh_fallback_failed_once.py tests/test_mesh_query_feature.py tests/test_met_wait_capture.py tests/test_module_ids.py tests/test_native_float_atomic_add.py tests/test_oot_state_tap.py tests/test_parallel_dry_marks.py tests/test_parallel_rb_refresh.py tests/test_pass_hazard.py tests/test_pcache_fetch_lock.py tests/test_pcache_header_read.py tests/test_pcache_stamp.py tests/test_pipeline_cache_save.py tests/test_pipeline_notes_open.py tests/test_pipeline_play_cache.py tests/test_pipeline_share.py tests/test_pipeline_state_computes.py tests/test_pipeline_state_counts.py tests/test_pipeline_state_portable.py tests/test_pipeline_states.py tests/test_pipeline_tour.py tests/test_precompile_compute_guard.py tests/test_precompile_pipeline_cleanup.py tests/test_predict_resume.py tests/test_predict_sees_loaded.py tests/test_predict_types.py tests/test_predicted_check.py tests/test_prepare_cache_save.py tests/test_prepare_cpus.py tests/test_prepare_seed_stamp.py tests/test_prepare_stall.py tests/test_prepare_workers.py tests/test_prepared_stamps.py tests/test_prof_sampling.py tests/test_program_answer_domains.py tests/test_program_cache_cleanup.py tests/test_program_cache_format.py tests/test_program_cache_index.py tests/test_program_cache_one_variant.py tests/test_program_cache_reader.py tests/test_program_load_wait.py tests/test_program_record_head.py tests/test_program_stamp_knobs.py tests/test_psychov35_dlss.py tests/test_psychov35_taau.py tests/test_push_state.py tests/test_quick_pipelines.py tests/test_radv_mesh_optimizer.py tests/test_reactive_mask.py tests/test_recomplauncher_json.py tests/test_recorder_own_idle.py tests/test_recorder_sync.py tests/test_recording_admission.py tests/test_replay_copies.py tests/test_rumble_menu.py tests/test_same_draw.py tests/test_same_dyn.py tests/test_sampler_admission.py tests/test_seed_variant_cleanup.py tests/test_shader_module_guard.py tests/test_shader_stamp_cache.py tests/test_shader_stamp_inputs.py tests/test_shadow_mask_guard.py tests/test_shared_layouts.py tests/test_split_before_draw.py tests/test_startup_progress.py tests/test_state_list_plain.py tests/test_taau_env_lookup.py tests/test_taau_fxaa.py tests/test_taau_typed_dispatch.py tests/test_table_track.py tests/test_texture_handle_banks.py tests/test_texture_squeeze.py tests/test_third_party_objects.py tests/test_threaded_rendering.py tests/test_tls_objects.py tests/test_tree_shadow_mask.py tests/test_ub_threads.py tests/test_upscale_fallback.py tests/test_upscaler_rows.py tests/test_vertex_bind.py tests/test_vk_occlusion_queries.py tests/test_vk_pgo.py tests/test_walk_key.py tests/test_warm_bounded.py

Not tested: the game yet. This PR is WIP until its A/B is in: the RX_VK_BENCH A/B against main is running on an RTX 3080 with a Ryzen 9 3900X, Windows 11; AMD and Intel need other testers.

Must be tested:

  • Any OS and GPU, defaults and once with RX_VK_LAYOUT_EACH=1: Kokiri Forest and Castle Town a few minutes, expecting the same picture and, with RX_VK_PROF_FINE=1, fewer full pushes a draw (Test Farm: @Beanstalk any, those scenes with and without RX_VK_LAYOUT_EACH=1)

Risks

What could break: A push elsewhere in the same command buffer at the same bind point that does not mark the game's pushed descriptors as disturbed (the temporal actor pass, frame generation, FSR, ASTC decode): the next game draw's delta push would leave the other bindings as that push set them. Two draws of the same program already relied on this; shared layouts make it more frequent.

Turning it off: RX_VK_LAYOUT_EACH=1

Compatibility: Nothing changes: no caches, settings or vk ABI involved.

Checklist

Everyone

  • One topic.
  • The title is a Conventional Commit (type(scope): description) of the status line's type.
  • Why gives the reason plainly and links the issues that fit (the tracker searched).
  • Tested on my own PC where it runs (in the game, for a change the game runs), as Testing's Tested says; what
    can't be run by hand is in Testing's Not testable by hand, with why. Not ticked yet: Testing says the game is not tested yet; only game-free checks are named, and game files changed.
  • Measurements has a verdict, or one line saying why there is nothing to measure.
  • Nothing gets worse that I know of: any cost (speed, latency, memory, a setup that worked before) is measured and
    in Risks. Not ticked yet: Measurements only says not measured yet, and Risks names a possible missed descriptor disturbance that is unmeasured.
  • Testing names what was tested, what wasn't, and what must still be tested, where and how.
  • Rebased on current main.
  • python -I -B tests/run_all.py passes (any failure that also fails on main is named).
  • No game files, dumps, build output or generated files.
  • third_party/ untouched (a new folder with a PROVENANCE.md instead).
  • The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled.

When it applies

  • AI assistance recorded with Assisted-by: / Co-Authored-By: trailers.
  • Switches documented (docs/development.md; the README for a menu row) and shipped at their recommended value.
  • A game-free test added for the fix, failing on main and passing here.
  • Docs updated.

🤖 Generated with Claude Code

**Type:** perf · **Risk:** low · **Tested on:** Windows 11 (the game-free checks; the game A/B is next) · **Needs:** a game run on NVIDIA, AMD and Intel ## Why Each shader program had its own descriptor set layout and pipeline layout, and the replay pushes only the descriptors that changed only while the layout is the one it pushed with: every program change pushed every descriptor again, even between programs whose bindings are laid out the same way. Programs change often within a pass. Found in an audit of the renderer's per-draw path; no issue on the tracker. ## What changed Programs whose layouts' create infos are the same now share one set layout and pipeline layout, kept for the device's life, so a program change between them pushes only the writes that differ. Pipelines are made from the same create infos as before and no pipeline key holds a layout handle. - draw: `layout_signature`, `layouts_for`, `SharedLayouts` and `ensure_layout`; the preparation's compute warm leaves shared layouts alone (`gpu/vk/vk_draw.cpp`) - program cache: `program_cache_borrow` leaves shared layouts alone (`gpu/vk/vk_progcache.cpp`); `Program::layout_shared` (`gpu/vk/vk_shader.h`, not a stamped translator file) - docs: the `RX_VK_LAYOUT_EACH=1` row in docs/development.md - tests: `tests/test_shared_layouts.py` ## Measurements **Verdict:** not measured yet: the measurements are pending from the RX_VK_BENCH A/B of this PR @ e2a24c1a against main @ b9e95173 (under way), which will replace this section. ### Flags and settings One switch for an A/B. | Flag | What it is | Before this PR | After this PR | Effect | Helps / doesn't help / untested | Recommendation | |---|---|---|---|---|---|---| | `RX_VK_LAYOUT_EACH=1` | A set layout and pipeline layout for every program, as before | - | unset (shared) | On: every program change pushes everything | measurement pending | leave unset; `=1` for an A/B | ## Testing **Tested:** Windows 11 (Ryzen 9 3900X), llvm-mingw and the Vulkan headers of the builder's v2.0.14 tools: rebased onto main @ b9e95173 (this PR @ e2a24c1a; it stands alone on main, on no other PR). The 135 checks that read this diff (listed below: those under tests/ and nx2/ that name a changed file, and the ones this PR adds or changes), not the whole suite: `python -I -B tests/run_all.py --jobs 6 <those checks>` with `RX_CPM` at the builder's CPM cache: 127 passed, 7 failed; each failure (`tests/test_crash_dialog.py`, `tests/test_linux_driver_fault_guard.py`, `tests/test_oot_state_tap.py`, `tests/test_pipeline_tour.py`, `tests/test_vk_occlusion_queries.py`, `tests/test_vk_pgo.py`, `tests/test_walk_key.py`) also fails on main @ b9e95173 in the same checks, on the same PC and the same day. <details><summary>The 135 checks</summary> `nx2/test_cbuf_table_extent.py` `nx2/test_dump_ssbo_read.py` `nx2/test_mesh_dynamic_state.py` `tests/test_amd_fragment_warp.py` `tests/test_bake_ds.py` `tests/test_bank_gen.py` `tests/test_bc_raw_targets.py` `tests/test_check_pr_description.py` `tests/test_commit_room.py` `tests/test_crash_dialog.py` `tests/test_desc_clock.py` `tests/test_desc_counts.py` `tests/test_dev_packs.py` `tests/test_device_lost_exit.py` `tests/test_dispatch_clock.py` `tests/test_dlss_checks.py` `tests/test_draw_binds.py` `tests/test_draw_side_effects.py` `tests/test_draw_stream_visibility.py` `tests/test_drawn_page_map.py` `tests/test_driver_cache_room.py` `tests/test_driver_faults.py` `tests/test_dummy_views.py` `tests/test_equal_depth_margin.py` `tests/test_fallback_buffer_cleanup.py` `tests/test_fallback_depth_hold.py` `tests/test_fallback_loaded.py` `tests/test_fallback_pipeline_state.py` `tests/test_fallback_size_policy.py` `tests/test_flat_map.py` `tests/test_framegen_preset.py` `tests/test_fsr_uniform_ownership.py` `tests/test_global_table_loop.py` `tests/test_gpl_cold_library.py` `tests/test_gpl_library_cache.py` `tests/test_handle_lookup.py` `tests/test_handle_table_arrays.py` `tests/test_hold_gpu.py` `tests/test_intel_compiler_gate.py` `tests/test_intel_quick_compute.py` `tests/test_jobs_failure.py` `tests/test_jobs_wait_runs_queued.py` `tests/test_linux_driver_fault_guard.py` `tests/test_linux_pressure.py` `tests/test_load_default.py` `tests/test_masked_clear_failure.py` `tests/test_mem_trim.py` `tests/test_menu_env_on.py` `tests/test_mesh_args_cleanup.py` `tests/test_mesh_fallback.py` `tests/test_mesh_fallback_failed_once.py` `tests/test_mesh_query_feature.py` `tests/test_met_wait_capture.py` `tests/test_module_ids.py` `tests/test_native_float_atomic_add.py` `tests/test_oot_state_tap.py` `tests/test_parallel_dry_marks.py` `tests/test_parallel_rb_refresh.py` `tests/test_pass_hazard.py` `tests/test_pcache_fetch_lock.py` `tests/test_pcache_header_read.py` `tests/test_pcache_stamp.py` `tests/test_pipeline_cache_save.py` `tests/test_pipeline_notes_open.py` `tests/test_pipeline_play_cache.py` `tests/test_pipeline_share.py` `tests/test_pipeline_state_computes.py` `tests/test_pipeline_state_counts.py` `tests/test_pipeline_state_portable.py` `tests/test_pipeline_states.py` `tests/test_pipeline_tour.py` `tests/test_precompile_compute_guard.py` `tests/test_precompile_pipeline_cleanup.py` `tests/test_predict_resume.py` `tests/test_predict_sees_loaded.py` `tests/test_predict_types.py` `tests/test_predicted_check.py` `tests/test_prepare_cache_save.py` `tests/test_prepare_cpus.py` `tests/test_prepare_seed_stamp.py` `tests/test_prepare_stall.py` `tests/test_prepare_workers.py` `tests/test_prepared_stamps.py` `tests/test_prof_sampling.py` `tests/test_program_answer_domains.py` `tests/test_program_cache_cleanup.py` `tests/test_program_cache_format.py` `tests/test_program_cache_index.py` `tests/test_program_cache_one_variant.py` `tests/test_program_cache_reader.py` `tests/test_program_load_wait.py` `tests/test_program_record_head.py` `tests/test_program_stamp_knobs.py` `tests/test_psychov35_dlss.py` `tests/test_psychov35_taau.py` `tests/test_push_state.py` `tests/test_quick_pipelines.py` `tests/test_radv_mesh_optimizer.py` `tests/test_reactive_mask.py` `tests/test_recomplauncher_json.py` `tests/test_recorder_own_idle.py` `tests/test_recorder_sync.py` `tests/test_recording_admission.py` `tests/test_replay_copies.py` `tests/test_rumble_menu.py` `tests/test_same_draw.py` `tests/test_same_dyn.py` `tests/test_sampler_admission.py` `tests/test_seed_variant_cleanup.py` `tests/test_shader_module_guard.py` `tests/test_shader_stamp_cache.py` `tests/test_shader_stamp_inputs.py` `tests/test_shadow_mask_guard.py` `tests/test_shared_layouts.py` `tests/test_split_before_draw.py` `tests/test_startup_progress.py` `tests/test_state_list_plain.py` `tests/test_taau_env_lookup.py` `tests/test_taau_fxaa.py` `tests/test_taau_typed_dispatch.py` `tests/test_table_track.py` `tests/test_texture_handle_banks.py` `tests/test_texture_squeeze.py` `tests/test_third_party_objects.py` `tests/test_threaded_rendering.py` `tests/test_tls_objects.py` `tests/test_tree_shadow_mask.py` `tests/test_ub_threads.py` `tests/test_upscale_fallback.py` `tests/test_upscaler_rows.py` `tests/test_vertex_bind.py` `tests/test_vk_occlusion_queries.py` `tests/test_vk_pgo.py` `tests/test_walk_key.py` `tests/test_warm_bounded.py` </details> **Not tested:** the game yet. This PR is WIP until its A/B is in: the RX_VK_BENCH A/B against main is running on an RTX 3080 with a Ryzen 9 3900X, Windows 11; AMD and Intel need other testers. **Must be tested:** - Any OS and GPU, defaults and once with `RX_VK_LAYOUT_EACH=1`: Kokiri Forest and Castle Town a few minutes, expecting the same picture and, with `RX_VK_PROF_FINE=1`, fewer `full pushes` a draw (Test Farm: @Beanstalk any, those scenes with and without RX_VK_LAYOUT_EACH=1) ## Risks **What could break:** A push elsewhere in the same command buffer at the same bind point that does not mark the game's pushed descriptors as disturbed (the temporal actor pass, frame generation, FSR, ASTC decode): the next game draw's delta push would leave the other bindings as that push set them. Two draws of the same program already relied on this; shared layouts make it more frequent. **Turning it off:** `RX_VK_LAYOUT_EACH=1` **Compatibility:** Nothing changes: no caches, settings or vk ABI involved. ## Checklist ### Everyone - [x] One topic. - [x] The title is a Conventional Commit (`type(scope): description`) of the status line's type. - [x] Why gives the reason plainly and links the issues that fit (the tracker searched). - [ ] Tested on my own PC where it runs (in the game, for a change the game runs), as Testing's Tested says; what can't be run by hand is in Testing's Not testable by hand, with why. Not ticked yet: Testing says the game is not tested yet; only game-free checks are named, and game files changed. - [x] Measurements has a verdict, or one line saying why there is nothing to measure. - [ ] Nothing gets worse that I know of: any cost (speed, latency, memory, a setup that worked before) is measured and in Risks. Not ticked yet: Measurements only says not measured yet, and Risks names a possible missed descriptor disturbance that is unmeasured. - [x] Testing names what was tested, what wasn't, and what must still be tested, where and how. - [x] Rebased on current `main`. - [x] `python -I -B tests/run_all.py` passes (any failure that also fails on `main` is named). - [x] No game files, dumps, build output or generated files. - [x] `third_party/` untouched (a new folder with a `PROVENANCE.md` instead). - [x] The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled. ### When it applies - [x] AI assistance recorded with `Assisted-by:` / `Co-Authored-By:` trailers. - [x] Switches documented (docs/development.md; the README for a menu row) and shipped at their recommended value. - [x] A game-free test added for the fix, failing on `main` and passing here. - [x] Docs updated. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Author
Member

The checks on this PC (Windows 11, RTX 3080, Ryzen 9 3900X, llvm-mingw from the builder's v2.0.14 tools): python -I -B tests/run_all.py on main @ aa95d35f fails these 19 here, and so does this branch: test_appimage_player_build.py, test_bake_portability.py, test_cpm_discovery.py, test_cpp_preflight.py, test_crash_dialog.py, test_env_on.py, test_fpx_sticky.py, test_lift_cleanup_error.py, test_linux_preflight.py, test_make_game_cache.py, test_make_game_prepare.py, test_missing_nup.py, test_oot_state_tap.py, test_options_budget_once.py, test_readback_late.py, test_spawn_emit.py, test_tested_configurations.py, test_vk_pgo.py, test_walk_key.py. Any other failure in a full run on this branch was rerun on its own and passed (the full runs shared the PC with dev_packs builds: timing checks).

Next: an interleaved A/B in the game against main on one dev_packs base of aa95d35f; the numbers will replace the Measurements section.

The checks on this PC (Windows 11, RTX 3080, Ryzen 9 3900X, llvm-mingw from the builder's v2.0.14 tools): `python -I -B tests/run_all.py` on main @ aa95d35f fails these 19 here, and so does this branch: `test_appimage_player_build.py`, `test_bake_portability.py`, `test_cpm_discovery.py`, `test_cpp_preflight.py`, `test_crash_dialog.py`, `test_env_on.py`, `test_fpx_sticky.py`, `test_lift_cleanup_error.py`, `test_linux_preflight.py`, `test_make_game_cache.py`, `test_make_game_prepare.py`, `test_missing_nup.py`, `test_oot_state_tap.py`, `test_options_budget_once.py`, `test_readback_late.py`, `test_spawn_emit.py`, `test_tested_configurations.py`, `test_vk_pgo.py`, `test_walk_key.py`. Any other failure in a full run on this branch was rerun on its own and passed (the full runs shared the PC with dev_packs builds: timing checks). Next: an interleaved A/B in the game against main on one dev_packs base of aa95d35f; the numbers will replace the Measurements section.
dyst force-pushed perf/vk-shared-pipeline-layouts from 67156a1f65
Some checks failed
CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Has been skipped
CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Has been skipped
CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Has been skipped
Third-party parts / The same as upstream (pull_request) Has been skipped
CI / Conventional Commit title (pull_request) Successful in 5s
PR description / Description follows the template (pull_request_target) Successful in 7s
CI / Workflow lint (actionlint) (pull_request) Successful in 6s
Guide screenshots / Guide images and local UI source review (pull_request) Failing after 5s
Third-party parts / As their notes say (pull_request) Successful in 9s
Website / Website on Cloudflare (pull_request_target) Has been skipped
to e2a24c1a90
Some checks failed
CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Has been skipped
CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Has been skipped
CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Has been skipped
Third-party parts / The same as upstream (pull_request) Has been skipped
Website / Website on Cloudflare (pull_request) Has been skipped
CI / Conventional Commit title (pull_request) Successful in 3s
CI / Workflow lint (actionlint) (pull_request) Successful in 6s
Guide screenshots / Guide images and local UI source review (pull_request) Failing after 6s
Third-party parts / As their notes say (pull_request) Successful in 11s
Website / Website on Cloudflare (pull_request_target) Has been skipped
PR description / Description follows the template (pull_request_target) Successful in 8s
2026-11-15 19:28:12 +00:00
Compare
Author
Member

Rebased onto main @ b9e95173 (now e2a24c1a; stands alone on main, not stacked); the 135 checks that read this diff, not the whole suite: 127 passed, 7 failed, each failure also failing on main @ b9e95173 in the same checks. Stays WIP: measurements pending from the RX_VK_BENCH A/B.

Rebased onto main @ b9e95173 (now e2a24c1a; stands alone on main, not stacked); the 135 checks that read this diff, not the whole suite: 127 passed, 7 failed, each failure also failing on main @ b9e95173 in the same checks. Stays WIP: measurements pending from the RX_VK_BENCH A/B.
Owner

Brought up to the ready-for-review standard (#1454): the checklist's Everyone list has two new boxes, added here from what this PR's description and comments show (please check they match what you know).

  • Tested on my own PC: not ticked: Testing says the game is not tested yet; only game-free checks are named, and game files changed. Run it yourself on your PC (in the game, for a change the game runs) and say where and what you saw in Testing's Tested; or, when the fix's own case can't be run by hand, add a **Not testable by hand:** line saying why and what stands in for it. Then tick it.
  • Nothing gets worse that I know of: not ticked: Measurements only says not measured yet, and Risks names a possible missed descriptor disturbance that is unmeasured. Measure that cost and state it in Risks, then tick it.
Brought up to the [ready-for-review standard](https://git.magicbean.dev/MagicBeanDev/Longshot/src/branch/main/CONTRIBUTING.md#ready-for-review-out-of-wip) (#1454): the checklist's Everyone list has two new boxes, added here from what this PR's description and comments show (please check they match what you know). - *Tested on my own PC*: not ticked: Testing says the game is not tested yet; only game-free checks are named, and game files changed. Run it yourself on your PC (in the game, for a change the game runs) and say where and what you saw in Testing's *Tested*; or, when the fix's own case can't be run by hand, add a `**Not testable by hand:**` line saying why and what stands in for it. Then tick it. - *Nothing gets worse that I know of*: not ticked: Measurements only says not measured yet, and Risks names a possible missed descriptor disturbance that is unmeasured. Measure that cost and state it in Risks, then tick it.
Member

#800 on today's main: the standard test, +1.6 % fps (diagnostic); it also invalidates the shader caches

I measured this PR on a second PC, on today's main.

Verdict: a small gain that holds in each pair, but it is under the 5 % that three pairs can call. Medians: 66.99
against 68.09 fps (+1.6 %), with the PR ahead in every pair (+3.7 %, +1.6 %, +2.6 %). CPU time a frame is 70.6
against 71.3 ms, unchanged within noise. The render thread's own line agrees with a small cut: its jobs a frame go from
13.79 to 13.54 ms, mostly in others (7.22 to 6.88), and descriptors go from 2.53 to 2.38 ms. One caution: the PR
arm's GPU clock read higher (2,579 against 2,495 MHz median), with the same gpu_ms (7.30 against 7.35). Some of the
fps difference may be the GPU's state rather than the change.

One thing to fix before merge: this PR invalidates every player's shader caches. The description says "no caches
... involved". But Program gains bool layout_shared (gpu/vk/vk_shader.h), and translator_stamp()
(gpu/vk/vk_shader.cpp) folds sizeof(Program) into the stamp, so a translator stamp change follows. On its first
start the PR's runtime logged vk: pipeline cache: 1 files made for another translator dropped (1435 MB); it starts empty and program cache ...: 0 program keys kept. It then began translating all 38,809 programs at start-up, and
the harness stopped that first start at its 240 s limit. Preparing the runtime again with the game's own
--prepare-shaders took 28 min here. The player's cost is the same as for a change to a stamped translator file. If
that isn't wanted, the flag could live outside Program (for example, a set of shared layout handles kept beside
layouts_for's cache), so that sizeof(Program) stays the same.

What was run

  • Base: Longshot main @ ef6038ab (2026-11-18 03:06Z), built as one development base with llvm-mingw 20261104, as
    the official builds are. Control and PR share the same longshot.exe (0a53b387). The renderer DLL is the base's
    (ed1771bb) for the control and the PR's (afa43e5d) for the PR arm, built over the same base (scripts/dev_packs.py).

  • How the PR was applied: your commit e2a24c1a (the PR's only commit) was cherry-picked onto ef6038ab. There were
    two conflicts, both text only:

    • docs/development.md: the PR's row sits next to rows main has changed since. I kept main's side, including its new
      RX_VK_STAGING_BURST_S row and without the RX_VK_DESC_COUNTS=0 row, which main has deleted. I added the PR's one
      new row, RX_VK_LAYOUT_EACH=1, right after it.
    • gpu/vk/vk_progcache.cpp: main added program_cache_borrow_all (the preparation's compute warm, run on all CPUs)
      directly above program_cache_borrow, whose doc comment the PR rewrites. I kept main's new function unchanged and
      took the PR's two-line comment for program_cache_borrow.

    The result has exactly the PR's added and removed lines (320+/7-). I wrote no code of my own.

  • A merge point for you: main's new program_cache_borrow_all destroys each variant's layout and set_layout
    without your layout_shared guard. Your own tests/test_shared_layouts.py catches it on the merged tree:
    "vk_progcache.cpp:857 destroys a program's layout without leaving a shared one". It looks harmless as written, because
    its only caller passes warm_compute, which destroys under your guard and then nulls both handles before
    borrow_all's loop runs. Still, the guard belongs there when you rebase (if (p->layout_shared) continue;, as in
    program_cache_borrow). It is in the preparation only, not in the frames measured here.

  • PC: Ryzen 7 5700X3D (8 cores, 16 threads), Radeon RX 9070 XT, AMD's Windows driver 26.11.1, Windows 11.

  • Harness: our standard test (std_test.py 604af9c9, run_profile.py 6d998641). It runs Hyrule Field from a fixed
    save, with the camera turning at full speed (right stick held from the load). The game runs in a 1280x720 window
    (16:9), with the frame rate unlocked (HOS_FPS_MAX=240) and the game clock checked (HOS_FPS_LOG=1, 1.000 ±
    0.01). The render thread is on (RX_VK_ASYNC=1, as Threaded rendering Auto gives a 16-thread CPU), and the fixed
    test settings file is used. The window is 20 s after the load, with 3 alternating pairs, no warm-up and no
    replacements.

  • The PR arm sets no switch: the change is on by default, and RX_VK_LAYOUT_EACH=1 is its off switch. The control
    is main without the PR.

  • Background load: a small fixed GPU load runs beside the game in every run of both arms: 500 blended quads a
    frame at 100 fps, in its own window. On this PC, AMD's driver otherwise keeps the RX 9070 XT at ~1,550-1,850 MHz
    because the game's bursty load never reaches its boost threshold. Runs then measure a slower GPU and compare
    poorly. With the load, both arms run at ~2,440-2,580 MHz.

  • Order: each runtime had one unmeasured first start before the A/B; the PR runtime's came after its re-prepare.
    The control runtime is the one the #795 A/B used just before. No runtime copy, pack or preparation ran within 15 min
    before the runs, and a disk-write probe was clean.

Numbers

Per run, in run order. RT jobs and RT idle are the render thread's own line (vk: the render thread (600 presents ...): its jobs a frame, split into submits, presents and others). draw, descriptors and programs come from
vk: a frame's replay time (ms). Each is the mean of the last two lines inside the window. error lines counts lines
with error, fail, VUID, lost or validation. In every run of both arms it is one line, the crash watcher's start line.

run fps CPU ms p99 ms worst ms >50 ms >100 ms clock gpu_ms gpu_mhz RT jobs ms submits presents others RT idle draw ms descriptors ms programs ms error lines
ctl-0 65.64 70.6 21.5 83.8 1 0 1.000 7.40 2441 14.05 3.09 3.75 7.21 0.93 5.99 2.50 0.93 1
pr-0 68.09 71.9 20.1 85.2 1 0 1.000 7.30 2579 13.54 3.12 3.67 6.75 0.95 5.93 2.38 0.93 1
pr-1 68.09 70.4 20.4 91.6 1 0 1.000 7.30 2533 13.59 2.95 3.69 6.94 0.92 6.14 2.42 1.02 1
ctl-1 66.99 72.7 20.1 83.8 1 0 1.000 7.30 2495 13.79 2.85 3.71 7.22 0.90 6.10 2.54 0.97 1
ctl-2 67.34 69.5 22.2 85.2 1 0 1.000 7.35 2533 13.75 2.76 3.63 7.35 0.84 6.02 2.53 0.95 1
pr-2 69.09 71.3 20.6 48.5 0 0 1.001 7.10 2582 13.38 2.88 3.62 6.88 0.91 5.87 2.31 0.93 1
control, median 66.99 70.6 21.5 85.2 (max) 3 (sum) 0 1.000 7.35 2495 13.79 2.85 3.71 7.22 0.90 6.02 2.53 0.95
PR, median 68.09 71.3 20.4 91.6 (max) 2 (sum) 0 1.000 7.30 2579 13.54 2.95 3.67 6.88 0.92 5.93 2.38 0.93

No run was rejected, and no frame went over 100 ms in any run. The PR's own line in every PR run: vk: 4096 programs have 548 pipeline layouts between them (1,024 have 171; 2,048 have 280-300).

Checked, not checked

  • Checked: the checks that read the changed files (tests/run_all.py --changed ef6038ab, 171) on the cherry-picked
    tree, and the same checks on ef6038ab. The only failure the base doesn't have is tests/test_shared_layouts.py, for
    the program_cache_borrow_all point above. The other 12 failures fail on ef6038ab too, on this PC.
    test_switch_table.py (failing on main here already) gains one line here, "RX_VK_LAYOUT_EACH is read
    (gpu/vk/vk_draw.cpp) and has no row in gpu/vk/switches.def". Main's switch table (#1300) wants a row for it.
  • The preparation: after the re-prepare, the PR runtime's counts match the base's: 32,825 noted pipelines of 20,529
    programs and 9,689 compute pipelines of 4,059 compute programs, none failed to make.
  • Correctness glance: no new error lines in any run (see above). I also made one extra unmeasured run per runtime
    with RX_VK_SHOTS (a picture every 600 presents) and looked at the pictures. They showed the menu, the load and the
    turning field with the HUD, and nothing was wrong: no missing or wrong textures, no stale descriptors showing as
    another object's texture. This is a glance, not an image comparison: the camera is at a different angle in each run.
  • Not checked: what the Risks name, a push elsewhere in the command buffer not marking the game's pushed descriptors
    disturbed (FSR, frame generation, the temporal actor pass, ASTC decode). None of those were on in this test. Also not
    checked: RX_VK_LAYOUT_EACH=1, RX_VK_PROF_FINE=1's push counts, other scenes, NVIDIA, Intel, Linux and the Steam
    Deck. No validation layers, no image comparison, no code review beyond the merge.

How to reproduce

git -C <main checkout> worktree add --detach wt-m-800 ef6038ab
git -C wt-m-800 cherry-pick e2a24c1a   # resolve docs/development.md and vk_progcache.cpp as described above
./with_lock.sh 60 "pack" ./devpacks.sh pack base/ef6038ab dyst-800 --tree wt-m-800
./with_lock.sh 30 "rt" ./devpacks.sh rt base/ef6038ab rt-dyst-800 dyst-800
# the pack is not marked as needing a prepare (no stamped file changed), but the stamp changes: prepare it as dev_packs does
./with_lock.sh 50 "rt" python -I prep_rt.py rt-dyst-800   # longshot.exe --prepare-shaders --compile-shaders most, 28 min here
./with_lock.sh 30 "rt" ./devpacks.sh rt base/ef6038ab rt-dyst-ctl
python -I std_test.py --runtime rt-dyst-ctl --name dyab800 --arm "pr@rt-dyst-800=" --options-ini test-profile-169.ini

(devpacks.sh is this PC's front for scripts/dev_packs.py; std_test.py is our standard-test harness.)

## #800 on today's main: the standard test, +1.6 % fps (diagnostic); it also invalidates the shader caches I measured this PR on a second PC, on today's main. **Verdict:** a small gain that holds in each pair, but it is under the 5 % that three pairs can call. Medians: 66.99 against 68.09 fps (+1.6 %), with the PR ahead in every pair (+3.7 %, +1.6 %, +2.6 %). CPU time a frame is 70.6 against 71.3 ms, unchanged within noise. The render thread's own line agrees with a small cut: its jobs a frame go from 13.79 to 13.54 ms, mostly in `others` (7.22 to 6.88), and `descriptors` go from 2.53 to 2.38 ms. One caution: the PR arm's GPU clock read higher (2,579 against 2,495 MHz median), with the same `gpu_ms` (7.30 against 7.35). Some of the fps difference may be the GPU's state rather than the change. **One thing to fix before merge: this PR invalidates every player's shader caches.** The description says "no caches ... involved". But `Program` gains `bool layout_shared` (`gpu/vk/vk_shader.h`), and `translator_stamp()` (`gpu/vk/vk_shader.cpp`) folds `sizeof(Program)` into the stamp, so a translator stamp change follows. On its first start the PR's runtime logged `vk: pipeline cache: 1 files made for another translator dropped (1435 MB); it starts empty` and `program cache ...: 0 program keys kept`. It then began translating all 38,809 programs at start-up, and the harness stopped that first start at its 240 s limit. Preparing the runtime again with the game's own `--prepare-shaders` took 28 min here. The player's cost is the same as for a change to a stamped translator file. If that isn't wanted, the flag could live outside `Program` (for example, a set of shared layout handles kept beside `layouts_for`'s cache), so that `sizeof(Program)` stays the same. ### What was run - **Base:** Longshot main @ ef6038ab (2026-11-18 03:06Z), built as one development base with llvm-mingw 20261104, as the official builds are. Control and PR share the same `longshot.exe` (0a53b387). The renderer DLL is the base's (ed1771bb) for the control and the PR's (afa43e5d) for the PR arm, built over the same base (`scripts/dev_packs.py`). - **How the PR was applied:** your commit e2a24c1a (the PR's only commit) was cherry-picked onto ef6038ab. There were two conflicts, both text only: - `docs/development.md`: the PR's row sits next to rows main has changed since. I kept main's side, including its new `RX_VK_STAGING_BURST_S` row and without the `RX_VK_DESC_COUNTS=0` row, which main has deleted. I added the PR's one new row, `RX_VK_LAYOUT_EACH=1`, right after it. - `gpu/vk/vk_progcache.cpp`: main added `program_cache_borrow_all` (the preparation's compute warm, run on all CPUs) directly above `program_cache_borrow`, whose doc comment the PR rewrites. I kept main's new function unchanged and took the PR's two-line comment for `program_cache_borrow`. The result has exactly the PR's added and removed lines (320+/7-). I wrote no code of my own. - **A merge point for you:** main's new `program_cache_borrow_all` destroys each variant's `layout` and `set_layout` without your `layout_shared` guard. Your own `tests/test_shared_layouts.py` catches it on the merged tree: "vk_progcache.cpp:857 destroys a program's layout without leaving a shared one". It looks harmless as written, because its only caller passes `warm_compute`, which destroys under your guard and then nulls both handles before `borrow_all`'s loop runs. Still, the guard belongs there when you rebase (`if (p->layout_shared) continue;`, as in `program_cache_borrow`). It is in the preparation only, not in the frames measured here. - **PC:** Ryzen 7 5700X3D (8 cores, 16 threads), Radeon RX 9070 XT, AMD's Windows driver 26.11.1, Windows 11. - **Harness:** our standard test (`std_test.py` 604af9c9, `run_profile.py` 6d998641). It runs Hyrule Field from a fixed save, with the camera turning at full speed (right stick held from the load). The game runs in a 1280x720 window (16:9), with the frame rate unlocked (`HOS_FPS_MAX=240`) and the game clock checked (`HOS_FPS_LOG=1`, 1.000 ± 0.01). The render thread is on (`RX_VK_ASYNC=1`, as Threaded rendering Auto gives a 16-thread CPU), and the fixed test settings file is used. The window is 20 s after the load, with 3 alternating pairs, no warm-up and no replacements. - **The PR arm sets no switch:** the change is on by default, and `RX_VK_LAYOUT_EACH=1` is its off switch. The control is main without the PR. - **Background load:** a small fixed GPU load runs beside the game in every run of both arms: 500 blended quads a frame at 100 fps, in its own window. On this PC, AMD's driver otherwise keeps the RX 9070 XT at ~1,550-1,850 MHz because the game's bursty load never reaches its boost threshold. Runs then measure a slower GPU and compare poorly. With the load, both arms run at ~2,440-2,580 MHz. - **Order:** each runtime had one unmeasured first start before the A/B; the PR runtime's came after its re-prepare. The control runtime is the one the #795 A/B used just before. No runtime copy, pack or preparation ran within 15 min before the runs, and a disk-write probe was clean. ### Numbers Per run, in run order. `RT jobs` and `RT idle` are the render thread's own line (`vk: the render thread (600 presents ...)`: its jobs a frame, split into submits, presents and others). `draw`, `descriptors` and `programs` come from `vk: a frame's replay time (ms)`. Each is the mean of the last two lines inside the window. `error lines` counts lines with error, fail, VUID, lost or validation. In every run of both arms it is one line, the crash watcher's start line. | run | fps | CPU ms | p99 ms | worst ms | >50 ms | >100 ms | clock | gpu_ms | gpu_mhz | RT jobs ms | submits | presents | others | RT idle | draw ms | descriptors ms | programs ms | error lines | |---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---| | ctl-0 | 65.64 | 70.6 | 21.5 | 83.8 | 1 | 0 | 1.000 | 7.40 | 2441 | 14.05 | 3.09 | 3.75 | 7.21 | 0.93 | 5.99 | 2.50 | 0.93 | 1 | | pr-0 | 68.09 | 71.9 | 20.1 | 85.2 | 1 | 0 | 1.000 | 7.30 | 2579 | 13.54 | 3.12 | 3.67 | 6.75 | 0.95 | 5.93 | 2.38 | 0.93 | 1 | | pr-1 | 68.09 | 70.4 | 20.4 | 91.6 | 1 | 0 | 1.000 | 7.30 | 2533 | 13.59 | 2.95 | 3.69 | 6.94 | 0.92 | 6.14 | 2.42 | 1.02 | 1 | | ctl-1 | 66.99 | 72.7 | 20.1 | 83.8 | 1 | 0 | 1.000 | 7.30 | 2495 | 13.79 | 2.85 | 3.71 | 7.22 | 0.90 | 6.10 | 2.54 | 0.97 | 1 | | ctl-2 | 67.34 | 69.5 | 22.2 | 85.2 | 1 | 0 | 1.000 | 7.35 | 2533 | 13.75 | 2.76 | 3.63 | 7.35 | 0.84 | 6.02 | 2.53 | 0.95 | 1 | | pr-2 | 69.09 | 71.3 | 20.6 | 48.5 | 0 | 0 | 1.001 | 7.10 | 2582 | 13.38 | 2.88 | 3.62 | 6.88 | 0.91 | 5.87 | 2.31 | 0.93 | 1 | | **control, median** | **66.99** | **70.6** | 21.5 | 85.2 (max) | 3 (sum) | 0 | 1.000 | 7.35 | 2495 | 13.79 | 2.85 | 3.71 | 7.22 | 0.90 | 6.02 | 2.53 | 0.95 | | | **PR, median** | **68.09** | **71.3** | 20.4 | 91.6 (max) | 2 (sum) | 0 | 1.000 | 7.30 | 2579 | 13.54 | 2.95 | 3.67 | 6.88 | 0.92 | 5.93 | 2.38 | 0.93 | | No run was rejected, and no frame went over 100 ms in any run. The PR's own line in every PR run: `vk: 4096 programs have 548 pipeline layouts between them` (1,024 have 171; 2,048 have 280-300). ### Checked, not checked - **Checked:** the checks that read the changed files (`tests/run_all.py --changed ef6038ab`, 171) on the cherry-picked tree, and the same checks on ef6038ab. The only failure the base doesn't have is `tests/test_shared_layouts.py`, for the `program_cache_borrow_all` point above. The other 12 failures fail on ef6038ab too, on this PC. `test_switch_table.py` (failing on main here already) gains one line here, "RX_VK_LAYOUT_EACH is read (gpu/vk/vk_draw.cpp) and has no row in gpu/vk/switches.def". Main's switch table (#1300) wants a row for it. - **The preparation:** after the re-prepare, the PR runtime's counts match the base's: 32,825 noted pipelines of 20,529 programs and 9,689 compute pipelines of 4,059 compute programs, none failed to make. - **Correctness glance:** no new error lines in any run (see above). I also made one extra unmeasured run per runtime with `RX_VK_SHOTS` (a picture every 600 presents) and looked at the pictures. They showed the menu, the load and the turning field with the HUD, and nothing was wrong: no missing or wrong textures, no stale descriptors showing as another object's texture. This is a glance, not an image comparison: the camera is at a different angle in each run. - **Not checked:** what the Risks name, a push elsewhere in the command buffer not marking the game's pushed descriptors disturbed (FSR, frame generation, the temporal actor pass, ASTC decode). None of those were on in this test. Also not checked: `RX_VK_LAYOUT_EACH=1`, `RX_VK_PROF_FINE=1`'s push counts, other scenes, NVIDIA, Intel, Linux and the Steam Deck. No validation layers, no image comparison, no code review beyond the merge. ### How to reproduce ``` git -C <main checkout> worktree add --detach wt-m-800 ef6038ab git -C wt-m-800 cherry-pick e2a24c1a # resolve docs/development.md and vk_progcache.cpp as described above ./with_lock.sh 60 "pack" ./devpacks.sh pack base/ef6038ab dyst-800 --tree wt-m-800 ./with_lock.sh 30 "rt" ./devpacks.sh rt base/ef6038ab rt-dyst-800 dyst-800 # the pack is not marked as needing a prepare (no stamped file changed), but the stamp changes: prepare it as dev_packs does ./with_lock.sh 50 "rt" python -I prep_rt.py rt-dyst-800 # longshot.exe --prepare-shaders --compile-shaders most, 28 min here ./with_lock.sh 30 "rt" ./devpacks.sh rt base/ef6038ab rt-dyst-ctl python -I std_test.py --runtime rt-dyst-ctl --name dyab800 --arm "pr@rt-dyst-800=" --options-ini test-profile-169.ini ``` (`devpacks.sh` is this PC's front for `scripts/dev_packs.py`; `std_test.py` is our standard-test harness.)
Contributor

Steam Deck OLED results for this PR.

Setup: Steam Deck OLED (SteamOS, RADV Mesa 26.2.0-devel (git-5c65d4fc31)). Main 544229a2 vs that main plus this PR (e2a24c1a, rebased as d8c55aa). Same game build, renderer-only difference, the same pinned Minimum-derived settings with overrides in both arms: fixed 50% render resolution, 1280×800 output on the 90 Hz panel, Handheld Profile off, plain window scaling and sharpness off, frame generation off, 15 W, main's clock policy, the FPS counter at its Dev level.

Method: four distinct-launch round pairs, M P / P M / M P / P M: Kokiri standing (45 s) and Hyrule Field turning (30 s). The pairs are time-ordered and descriptive (n = 4); this doesn't claim statistical independence. Arrows are M → P; percentage changes are 100*(P/M−1). Renderer frame tracing was enabled in both arms.

Kokiri standing: no consistent throughput benefit resolved.

  • Arm medians: 25.264 → 25.258 fps (−0.02%; median paired +0.04%).
  • Pair changes: +0.03, +0.34, −2.40 and +0.05%.
  • Game-process CPU time per renderer present (all threads) was a little higher with the PR in every pair: +0.08, +0.23, +1.74 and +0.43% (small, mechanism not measured).
  • Tails were mixed: p99 was higher with the PR in 3 of 4 pairs, the longest frame shorter in 3 of 4, and frames over 50 ms fewer in 2, equal in 1, more in 1.

Hyrule Field turning: effect unresolved. Only one pair was admitted (round 3: 27.88 → 30.97 fps), so no general Field benefit is claimed. Main's Field window was refused in 3 of its 4 measured runs by our "1 jobs queued" admission rule, each time when its window crossed present 7200; the PR's windows started later and crossed no 1800-present boundary. The admission imbalance coincided with different present phases; it does not establish a PR effect or a fix for queued jobs.

Safety: No GPU fault, reset or timeout was recorded in this block. The covered Kokiri and Field safety windows had zero flash flags, and the views we inspected showed no obvious new artifact. Zero detector flags and sampled views are not universal correctness proof.

Limits:

  • The fps and tail numbers include our /proc memory sampler, whose periodic read lines up with recurring long frames on this Deck; its observer effect, and any difference between the arms, is unresolved.
  • The PR's caches were prepared fresh for this run, main's are older; first starts aren't compared.
  • amdgpu "shaper LUT" precision warnings appeared in every run of both arms (no fault); impact unknown.
Steam Deck OLED results for this PR. **Setup:** Steam Deck OLED (SteamOS, RADV Mesa 26.2.0-devel (git-5c65d4fc31)). Main `544229a2` vs that main plus this PR (`e2a24c1a`, rebased as `d8c55aa`). Same game build, renderer-only difference, the same pinned Minimum-derived settings with overrides in both arms: fixed 50% render resolution, 1280×800 output on the 90 Hz panel, Handheld Profile off, plain window scaling and sharpness off, frame generation off, 15 W, main's clock policy, the FPS counter at its Dev level. **Method:** four distinct-launch round pairs, M P / P M / M P / P M: Kokiri standing (45 s) and Hyrule Field turning (30 s). The pairs are time-ordered and descriptive (n = 4); this doesn't claim statistical independence. Arrows are M → P; percentage changes are 100*(P/M−1). Renderer frame tracing was enabled in both arms. **Kokiri standing: no consistent throughput benefit resolved.** - Arm medians: 25.264 → 25.258 fps (−0.02%; median paired +0.04%). - Pair changes: +0.03, +0.34, −2.40 and +0.05%. - Game-process CPU time per renderer present (all threads) was a little higher with the PR in every pair: +0.08, +0.23, +1.74 and +0.43% (small, mechanism not measured). - Tails were mixed: p99 was higher with the PR in 3 of 4 pairs, the longest frame shorter in 3 of 4, and frames over 50 ms fewer in 2, equal in 1, more in 1. **Hyrule Field turning: effect unresolved.** Only one pair was admitted (round 3: 27.88 → 30.97 fps), so no general Field benefit is claimed. Main's Field window was refused in 3 of its 4 measured runs by our "1 jobs queued" admission rule, each time when its window crossed present 7200; the PR's windows started later and crossed no 1800-present boundary. The admission imbalance coincided with different present phases; it does not establish a PR effect or a fix for queued jobs. **Safety:** No GPU fault, reset or timeout was recorded in this block. The covered Kokiri and Field safety windows had zero flash flags, and the views we inspected showed no obvious new artifact. Zero detector flags and sampled views are not universal correctness proof. **Limits:** - The fps and tail numbers include our /proc memory sampler, whose periodic read lines up with recurring long frames on this Deck; its observer effect, and any difference between the arms, is unresolved. - The PR's caches were prepared fresh for this run, main's are older; first starts aren't compared. - amdgpu "shaper LUT" precision warnings appeared in every run of both arms (no fault); impact unknown.
Some checks failed
CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Has been skipped
CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Has been skipped
CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Has been skipped
Third-party parts / The same as upstream (pull_request) Has been skipped
Website / Website on Cloudflare (pull_request) Has been skipped
CI / Conventional Commit title (pull_request) Successful in 3s
CI / Workflow lint (actionlint) (pull_request) Successful in 6s
Guide screenshots / Guide images and local UI source review (pull_request) Failing after 6s
Third-party parts / As their notes say (pull_request) Successful in 11s
Website / Website on Cloudflare (pull_request_target) Has been skipped
PR description / Description follows the template (pull_request_target) Successful in 8s
This pull request has changes conflicting with the target branch.
  • docs/development.md
  • gpu/vk/vk_draw.cpp
  • gpu/vk/vk_progcache.cpp
View command line instructions

Manual merge helper

Use this merge commit message when completing the merge manually.

Checkout

From your project repository, check out a new branch and test the changes.
git fetch -u 'perf/vk-shared-pipeline-layouts:dyst-perf/vk-shared-pipeline-layouts'
git switch 'dyst-perf/vk-shared-pipeline-layouts'
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
4 participants
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!800
No description provided.