Game: Frame-time spikes while walking in circles near Kakariko bridge in Hyrule Field #1696
Labels
No labels
area/accessibility
area/audio
area/input
area/install
area/menu
area/renderer
area/saves
area/tooling
area/upscaling
beanstalk: approved
beanstalk: failed
beanstalk: needs approval
beanstalk: running
beanstalk: tested
ci: approved
good first issue
gpu/amd
gpu/igpu
gpu/intel
gpu/nvidia
help wanted
kind
bug
kind
chore
kind
docs
kind
feature
kind
performance
kind
question
needs measurements
platform/linux
platform/macos
platform/steam-deck
platform/windows
preview
priority
critical
priority
high
priority
low
regression
risk
high
risk
low
risk
medium
status
blocked
status
confirmed
status
duplicate
status
invalid
status
needs-info
status
wontfix
ci: approved
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
MagicBeanDev/Longshot#1696
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Reported on Discord
Longshot version and channel
: v3.0.0-rc.8 (source 5a1a671d), release candidate. : version not provided; they report the issue is still present, but substantially improved, on nightly 486.
What happened and steps
Just outside Kakariko Village, after the small wooden bridge toward Hyrule Field, walking in circles without changing the camera caused severe frame-time spikes and stutters. The reporter says doing the same just before the bridge, closer to Kakariko, with the camera facing the same way does not cause the spikes. This was observed in the first 10 seconds of the Bug report video. A second player reports similar spikes/stutter while attacking.
Where in the game
Hyrule Field, immediately past the small wooden bridge outside Kakariko Village. The new report identifies the map as HyruleField_Adult (adult). The second player's precise location is not provided.
How often
reports the repeatable comparison: spikes at the field-side location, not at the Kakariko-side location. reports the attacking-triggered issue still occurs slightly on nightly 486, but is overall much better than before.
Graphics card and driver version
: AMD Radeon RX 7800 XT (RADV NAVI32), Mesa RADV 26.2.6-arch3.1. : NVIDIA RTX 3070 Ti; driver version not provided.
CPU, RAM and operating system
: AMD Ryzen 7 5800X3D, 16 GB DDR4 3200 MHz, CachyOS (Linux 7.3.4-1-cachyos), KDE Wayland. : Linux, Ryzen 5 5600, 32 GB RAM.
Graphics settings
: 2560x1440 output and render resolution, game's own upscaler, V-Sync Off, frame generation Off, frame-rate cap Uncapped. Other settings are in the attached report. 's settings are not provided.
Performance measurements
's current-run log confirms profiling switches were enabled: HOS_FPS_LOG=1, HOS_THREAD_TIMES=5, RX_VK_PASS_TIMING=1, RX_VK_SLOW_LOG=50. It contains GPU pass timing samples but no actual thread-time samples, so profiling evidence is incomplete. has not provided a profiling log. See https://longshot.magicbean.dev/help/performance/ for capture steps and requested measurements.
Logs and screenshot
Discord attachments:
2026-11-18_22-43-26.zipandlongshot_log.txtfrom the earlier run;2026-11-18_23-07-04.zipcontains the laterreport.txt, current-runlongshot_log.txt, andpicture.png; a separate currentlongshot_log.txtwas also attached. The later report identifies v3.0.0-rc.8, HyruleField_Adult (adult), and native 2560x1440 with V-Sync off and no frame generation.Nightly 486 update
reports that this problem is still slightly reproducible on nightly 486, but is overall much better. The exact trigger/location and profiling measurements for that nightly have not been provided.
Still needed
: a complete performance profile captured while reproducing the spikes, with actual HOS_THREAD_TIMES and RX_VK_PASS_TIMING measurements visible in the log. : Longshot version and driver version, exact location/steps for the attacking trigger, and a profiling capture with thread-time and GPU pass timing samples. See https://longshot.magicbean.dev/help/performance/ for the profiling steps.
Files from the Discord report:
Updated with more from Discord: Added the new current-run report and logs, confirmed profiling switches are enabled and the map is identified, but noted that actual profiling samples are missing.
Files:
From Discord:
A second player reports the same stutter/frame-time spikes, specifically when attacking. Their setup: Linux, NVIDIA RTX 3070 Ti, Ryzen 5 5600, 32 GB RAM. They have not provided a Longshot version, driver version, steps/location beyond attacking, or a profiling log; please capture a performance profile as described at https://longshot.magicbean.dev/help/performance/.
From Discord:
A second player reports similar frame-time spikes/stuttering when attacking, on Linux with an NVIDIA RTX 3070 Ti, Ryzen 5 5600, and 32 GB RAM. Their Longshot version, driver, exact location/steps, and a profiling capture are not yet available. Please collect a profile with actual
HOS_THREAD_TIMESandRX_VK_PASS_TIMINGsamples while reproducing this trigger: https://longshot.magicbean.dev/help/performance/.Updated with more from Discord: Added 's report that nightly 486 substantially improves the issue, though some stutter remains.
From Discord:
From (2026-11-19): The issue still occurs on nightly 486 (commit
09a20cd6), in the same manner as originally described. They attached a video showing the performance overlay during the Hyrule Field run, including readings of 48–49 FPS and a drop to 10 FPS while approaching/engaging the enemies; the video also shows V-Sync off. Attachment: Discord message 1572942287401648128,2026-11-19_08-46-57_720p.mp4.