docs(worklist): the igvk64.dll crash reproduced on a UHD 770, re-scored 0.11 -> 0.28 #2790

Closed

github-sync wants to merge 1 commit from github/pr/2790 into main

github-sync commented

2026-11-12 13:38:19 +00:00

Owner

@dankman42 opened this pull request on GitHub · 2026-11-12 13:21 UTC · original
Merged on GitHub in 1574a546a5

The worklist's Intel igvk64.dll crash row had one report and no reproduction (confidence 0.4). It now reproduces with a log on an Intel UHD 770, so it is re-scored as the workflow asks.

What changed: one row in docs/worklist.md, moved to its new place in the order:

confidence 0.4 -> 1.0, so the score goes 8 x 0.035 x 0.4 = 0.11 -> 0.28;
the reproduction and the candidate fix are added to it: main 51da0464 on a UHD 770 (Intel 101.7088, Windows 11) crashes in 4 of 4 runs at igvk64.dll+0x14d102 (0xc0000005) during first visits while pipelines are built, and #2456 ran 4 of 4 without the crash.

The details are in #2456.

Tests: docs only; no check reads this file. Nothing on screen.

🤖 Generated with Claude Code

> [@dankman42](https://github.com/dankman42) opened this pull request on GitHub · 2026-11-12 13:21 UTC · [original](https://github.com/NotRexed/Hookshot/pull/2790) > **Merged on GitHub** in https://git.magicbean.dev/MagicBeanDev/Hookshot/commit/1574a546a5657535c5635793882bff56f0403f34 The worklist's Intel `igvk64.dll` crash row had one report and no reproduction (confidence 0.4). It now reproduces with a log on an Intel UHD 770, so it is re-scored as the workflow asks. **What changed:** one row in `docs/worklist.md`, moved to its new place in the order: - confidence 0.4 -> 1.0, so the score goes 8 x 0.035 x 0.4 = 0.11 -> 0.28; - the reproduction and the candidate fix are added to it: main `51da0464` on a UHD 770 (Intel 101.7088, Windows 11) crashes in 4 of 4 runs at `igvk64.dll`+0x14d102 (0xc0000005) during first visits while pipelines are built, and #2456 ran 4 of 4 without the crash. The details are in https://git.magicbean.dev/MagicBeanDev/Hookshot/pulls/2456. **Tests:** docs only; no check reads this file. Nothing on screen. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

github-sync added 1 commit

2026-11-12 13:38:19 +00:00

docs(worklist): the igvk64.dll crash reproduced on a UHD 770, re-scored 0.11 -> 0.28 d83eda89ad

Main 51da0464 crashed in 4 of 4 runs on an Intel UHD 770 (driver
101.7088, Windows 11) at igvk64.dll+0x14d102 during first visits; #2456
ran 4 of 4 without it. Confidence 0.4 -> 1.0 (reproduced with a log).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

github-sync commented

2026-11-12 14:48:27 +00:00

Author

Owner

@Joey-Magnus approved on GitHub · 2026-11-12 14:35 UTC · original

I checked the row against the linked test report on #2456 (comment 6824750417). The UHD 770 runs on 51da0464 crashed 4 of 4 times in igvk64.dll at +0x14d102 (0xc0000005), and the four #2456 runs exited cleanly. The arithmetic is right: 8 x 0.035 x 1.0 = 0.28. The row also sits in the right place, between the 0.29 RX 580 row and the 0.22 RTX 3070 row.

One thing for whoever triages it next. The original report is a crash right after the owl intro on an i3-1115G4 (Tiger Lake). Your repro crashes while loading into Kokiri Forest after a teleport, on a desktop UHD 770 used as a secondary device (RX_VK_DEVICE=1). Same module and exception, but the fault offset from the original logs isn't compared. It may be the same bug as the 0.2 "Intel GPU: black screen or crash after the intro" row, which also lists #2456. If a later report shows they're one bug, those two rows should be merged rather than both kept at full weight. As a re-score backed by a reproduction, this is what docs/workflow.md asks for.

> [@Joey-Magnus](https://github.com/Joey-Magnus) approved on GitHub · 2026-11-12 14:35 UTC · [original](https://github.com/NotRexed/Hookshot/pull/2790#pullrequestreview-5876257989) I checked the row against the linked test report on #2456 (comment 6824750417). The UHD 770 runs on 51da0464 crashed 4 of 4 times in igvk64.dll at +0x14d102 (0xc0000005), and the four #2456 runs exited cleanly. The arithmetic is right: 8 x 0.035 x 1.0 = 0.28. The row also sits in the right place, between the 0.29 RX 580 row and the 0.22 RTX 3070 row. One thing for whoever triages it next. The original report is a crash right after the owl intro on an i3-1115G4 (Tiger Lake). Your repro crashes while loading into Kokiri Forest after a teleport, on a desktop UHD 770 used as a secondary device (RX_VK_DEVICE=1). Same module and exception, but the fault offset from the original logs isn't compared. It may be the same bug as the 0.2 "Intel GPU: black screen or crash after the intro" row, which also lists #2456. If a later report shows they're one bug, those two rows should be merged rather than both kept at full weight. As a re-score backed by a reproduction, this is what docs/workflow.md asks for.

github-sync referenced this pull request

2026-11-12 16:54:36 +00:00

fix(video): keep a slow movie's pictures and survive Intel's compiler #2849

github-sync referenced this pull request from a commit

2026-11-12 17:57:17 +00:00

docs(worklist): the igvk64.dll crash reproduced on a UHD 770, re-scored 0.11 -> 0.28 (#2790)

github-sync closed this pull request

2026-11-12 17:59:19 +00:00

Commenting is not possible because the repository is archived.

No reviewers