Game: Repeatable frame spikes when enemies die in Hyrule Field (RTX 5080, Linux) #1213

Open
opened 2026-11-16 19:21:59 +00:00 by BeanIssues · 3 comments
Collaborator

Reported on Discord

Longshot version and channel

nightly-363-0cab3548, nightly (from longshot_log.txt).

What happened and steps

The reporter loaded a save in Hyrule Field, killed several enemies, changed to night, then killed more enemies (skeletons). Each time an enemy died, repeatable frame spikes occurred. Hyrule Field was otherwise smooth.

Where in the game

Hyrule Field, while killing enemies, including skeletons at night.

How often

Reported as repeatable, occurring as soon as an enemy dies.

Graphics card and driver version

NVIDIA GeForce RTX 5080. The reporter gives driver as “620”; exact driver version unknown. The new log identifies Vulkan driver ID 0x9b118180 but does not clarify the user-facing driver version.

CPU, RAM and operating system

AMD Ryzen 7 9800X3D, Linux. The log reports 31,186 MB system memory; OS distribution/version and exact RAM amount remain unknown.

Logs and performance capture

Two supplied longshot_log.txt files. The latest reports 3840×2160 output. It contains general frame/replay timing statistics, but lacks HOS_THREAD_TIMES and RX_VK_PASS_TIMING measurements required by the performance-profile procedure, so it does not isolate the reported spikes. FPS samples: 61 total, minimum 43, median 60, maximum 120; no sample is far below the median.

Still needed

A performance profile captured during the enemy-death spikes, with HOS_FPS_LOG=1, HOS_THREAD_TIMES=5, RX_VK_PASS_TIMING=1 and RX_VK_SLOW_LOG=50 configured as described at https://longshot.magicbean.dev/help/performance/; exact NVIDIA driver version; OS distribution/version; RAM amount; graphics settings (preset, render resolution, upscaler, frame generation, V-Sync); and confirmation of spike magnitude/frequency. The reporter should capture the same Hyrule Field enemy-death scenario, then send the resulting Bug report folder and longshot_log.txt.

Possibly related but not confirmed to be the same symptom: #1111 tracks broader stuttering across the map, worst in Hyrule Field and Death Mountain on an RTX 4090. This report is specifically repeatable spikes when an enemy dies on Linux with an RTX 5080. PR #363 documents another Hyrule Field slowdown and profiling A/B plan, but it is documentation only and does not establish a fix for this enemy-death symptom.

Reported on [Discord](https://discord.com/channels/1569849200047427645/1570984952651055114/1571966153050558534) ### Longshot version and channel nightly-363-0cab3548, nightly (from `longshot_log.txt`). ### What happened and steps The reporter loaded a save in Hyrule Field, killed several enemies, changed to night, then killed more enemies (skeletons). Each time an enemy died, repeatable frame spikes occurred. Hyrule Field was otherwise smooth. ### Where in the game Hyrule Field, while killing enemies, including skeletons at night. ### How often Reported as repeatable, occurring as soon as an enemy dies. ### Graphics card and driver version NVIDIA GeForce RTX 5080. The reporter gives driver as “620”; exact driver version unknown. The new log identifies Vulkan driver ID 0x9b118180 but does not clarify the user-facing driver version. ### CPU, RAM and operating system AMD Ryzen 7 9800X3D, Linux. The log reports 31,186 MB system memory; OS distribution/version and exact RAM amount remain unknown. ### Logs and performance capture Two supplied `longshot_log.txt` files. The latest reports 3840×2160 output. It contains general frame/replay timing statistics, but lacks `HOS_THREAD_TIMES` and `RX_VK_PASS_TIMING` measurements required by the performance-profile procedure, so it does not isolate the reported spikes. FPS samples: 61 total, minimum 43, median 60, maximum 120; no sample is far below the median. ### Still needed A performance profile captured during the enemy-death spikes, with `HOS_FPS_LOG=1`, `HOS_THREAD_TIMES=5`, `RX_VK_PASS_TIMING=1` and `RX_VK_SLOW_LOG=50` configured as described at https://longshot.magicbean.dev/help/performance/; exact NVIDIA driver version; OS distribution/version; RAM amount; graphics settings (preset, render resolution, upscaler, frame generation, V-Sync); and confirmation of spike magnitude/frequency. The reporter should capture the same Hyrule Field enemy-death scenario, then send the resulting Bug report folder and `longshot_log.txt`. Possibly related but not confirmed to be the same symptom: #1111 tracks broader stuttering across the map, worst in Hyrule Field and Death Mountain on an RTX 4090. This report is specifically repeatable spikes when an enemy dies on Linux with an RTX 5080. PR #363 documents another Hyrule Field slowdown and profiling A/B plan, but it is documentation only and does not establish a fix for this enemy-death symptom.
Author
Collaborator

Files from the Discord report:

Files from the Discord report: - [longshot_log.txt](https://git.magicbean.dev/attachments/7f14eaa7-eaed-445a-80a5-d0dca6f5a5a3) - [image.png](https://git.magicbean.dev/attachments/aa0bd239-923e-4cec-9a4d-a3dfcebefc87)
Author
Collaborator

Updated with more from Discord: Added the second log review and clarified that it lacks required profile measurements; noted related Hyrule Field profiling documentation is not a fix for this symptom.

Files:

Updated with more from [Discord](https://discord.com/channels/1569849200047427645/1571966153050558534/1571968852324122634): Added the second log review and clarified that it lacks required profile measurements; noted related Hyrule Field profiling documentation is not a fix for this symptom. Files: - [longshot_log.txt](https://git.magicbean.dev/attachments/74650104-f7fc-444b-97ec-add2b5920e29)
Author
Collaborator

From Discord:

The reporter supplied another longshot_log.txt (nightly-363-0cab3548). It still does not contain HOS_THREAD_TIMES or RX_VK_PASS_TIMING profile measurements, so it does not provide the requested performance profile. It reports 3840×2160 output and frame-rate samples of min 44, median 60, max 122; these general samples do not isolate enemy-death spikes. The log identifies RTX 5080 and Ryzen 7 9800X3D; system memory is reported as 31,186 MB. The reporter asks whether the capture is correct; it is not sufficient to diagnose the specific spikes.

From [Discord](https://discord.com/channels/1569849200047427645/1570984952651055114/1571966153050558534): The reporter supplied another `longshot_log.txt` (nightly-363-0cab3548). It still does not contain `HOS_THREAD_TIMES` or `RX_VK_PASS_TIMING` profile measurements, so it does not provide the requested performance profile. It reports 3840×2160 output and frame-rate samples of min 44, median 60, max 122; these general samples do not isolate enemy-death spikes. The log identifies RTX 5080 and Ryzen 7 9800X3D; system memory is reported as 31,186 MB. The reporter asks whether the capture is correct; it is not sufficient to diagnose the specific spikes.
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#1213
No description provided.