8000 Crash: the GPU stopped making progress (3084 ms) · Issue #131 · Wooloomooloo2/bloodborne_mac_port · GitHub
[go: up one dir, main page]

Skip to content

Crash: the GPU stopped making progress (3084 ms) #131

Open

Description

@khhan0703

game-2026-11-04_23-43-50.log
game.log

What were you doing when it happened?

Diagnostics: please drag crash-2026-11-04_23-43-50.zip (the launcher showed it in Finder) onto this issue: it has the full session log and macOS's crash reports.

Bloodborne for Mac 0.5.12: crash report

  • Mac: Mac14,2 (Apple M2), 16 GB, macOS 27.2 (Build 26B123)
  • Settings: output 1280x720, upscaler off, cap 30 FPS, fixed timing, windowed
  • Session: started 2026-11-04 23:43:50 09:00, ran 76 min 29 s
  • Exit status: 1
  • Cause: the GPU stopped making progress (3084 ms) (at 01:00:20.119)
  • Log: game-2026-11-04_23-43-50.log, 7898 lines: 2 critical, 202 errors, 5716 warnings
Warnings and errors before the cause (5 most recent)
00:04:34.199 WARN   GPU [Render] <Warning> mtl_draw_core.mm:323 ReportSkips: Metal draws skipped in the last 5 s: empty scissor 6503
00:04:39.200 WARN   GPU [Render] <Warning> mtl_draw_core.mm:323 ReportSkips: Metal draws skipped in the last 5 s: empty scissor 7301
00:04:44.201 WARN   GPU [Render] <Warning> mtl_draw_core.mm:323 ReportSkips: Metal draws skipped in the last 5 s: empty scissor 584
00:18:33.463 WARN   GPU [Render.Vulkan] <Warning> buffer_cache.cpp:1046 SynchronizeMemoryFromImage: Unhandled metadata type CMask  (×5705)
00:22:50.379 WARN   GPU [Render.Vulkan] <Warning> image_view.cpp:44 LogOnce: image view format cannot view the texture in Metal (block or depth reinterpretation): using the image's format
game-2026-11-04_23-43-50.log (13 lines around the cause)
01:00:04.575 INFO   Runtime: gu
A059
est thread 'CSClothThread3' is host thread 1024132, handle 0x8105ce610 (affinity 0xf, prio 700)
01:00:04.575 INFO   Runtime: guest thread 'CSClothThread4' created (stack=65536, prio=700)
01:00:04.575 INFO   Runtime: guest thread 'CSClothThread4' affinity 0x30
01:00:04.575 INFO   Runtime: guest thread 'CSClothThread4' is host thread 1024133, handle 0x8105ce410 (affinity 0x30, prio 700)
01:00:04.575 INFO   Runtime: guest thread 'CSClothThread5' created (stack=65536, prio=700)
01:00:04.575 INFO   Runtime: guest thread 'CSClothThread5' affinity 0x30
01:00:04.575 INFO   Runtime: guest thread 'CSClothThread5' is host thread 1024134, handle 0x8105ce210 (affinity 0x30, prio 700)
01:00:04.873 INFO   Runtime: open(/app0/dvdroot_ps4/map/m24/m24_9999.tpf.dcx) -> not found
01:00:20.119 CRIT   Crash: the GPU stopped making progress (3084 ms)
01:00:20.119 ERROR  STOP: GPU watchdog: tick 277260 committed, 277257 complete, no progress for 3084 ms; exiting so the GPU does not stall WindowServer (BB_GPU_WATCHDOG_MS; GPU event 277257, handlers 277257)
01:00:20.305 INFO   GPU process: host lost (host process exited with status 75) while rendering; exiting
01:00:20.305 CRIT   Crash: the GPU process stopped unexpectedly (host process exited with status 75)
01:00:20.506 ERROR  launcher: the game stopped with status 1 after 76 min 29 s

Activity

  1. evill33t commented on Nov 4, 2026

    @evill33t

    The evidence supports a GPU queue stall; the underlying cause remains unresolved.

    • On 0.5.12, after 76 minutes, tick 277260 was committed while both the GPU event and completion handlers remained at 277257 for 3,084 ms. Three command buffers were outstanding.
    • The watchdog fix from #107 is active. Delayed CPU completion handlers alone cannot explain this report because the GPU’s own signal also stopped.
    • The 100 invalid vertex-format pipeline errors occurred during startup. The 5,705 CMask warnings ended roughly 42 minutes before the crash. Neither establishes the cause.
    • The frame heap filled about 12 minutes before failure, but allocations successfully used the spare pool. No spare-pool exhaustion was logged.
    • Both attachments are identical copies of one session.

    The report lacks the requested crash ZIP and any description of the location/action. Those, plus a save if reproducible, are the next useful evidence. The existing logs do not identify the stalled command or justify a specific fix.

  2. khhan0703 commented on Nov 4, 2026

    @khhan0703
    Author

    @evill33t
    oh, found a crash report (~/Library/Logs/Bloodborne/Reports)
    crash-2026-11-04_23-43-50.zip

  3. Wooloomooloo2 commented on Nov 5, 2026

    @Wooloomooloo2
    Owner

    Thanks @khhan0703 for the zip. As @evill33t noted, the frame heap filled about 12 minutes before the stall. That matches the long-session problem in #127: the game's per-frame memory slowly fills with occlusion queries that never finish, and things go wrong once it's full. Your stall may be a different way of hitting the same problem, so I'm keeping this open next to #127. Until it's fixed, restarting the game about every 90 minutes should avoid it.

  4. Wooloomooloo2 commented on Nov 20, 2026

    @Wooloomooloo2
    Owner

    Checking in on this one. Quite a lot has changed since this was reported, up to 0.5.17 (GPU robustness fixes, Restart, focus, controllers, and more).

    If it still happens on 0.5.17, please reply with what you see and attach the log from that session (~/Library/Logs/Bloodborne/, the game-<date and time>.log from when it happened). If we don't hear back within a week, I'll close this. You can always reopen it or file a new one.

    Written with Claude Code.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions

      0