Crash after cutscene on RX 580 8GB #192

Closed
opened 2026-11-07 15:08:53 +00:00 by ImagineThisGuy2 · 11 comments
ImagineThisGuy2 commented 2026-11-07 15:08:53 +00:00 (Migrated from github.com)

After pressing new game the cutscene plays though after it the game crashes

here is the log:

hookshot_log.txt

After pressing new game the cutscene plays though after it the game crashes here is the log: [hookshot_log.txt](https://github.com/user-attachments/files/34522136/hookshot_log.txt)
ImagineThisGuy2 commented 2026-11-08 00:24:31 +00:00 (Migrated from github.com)

i probably should have added that this is an rx 580 8gb

i probably should have added that this is an rx 580 8gb
NotRexed commented 2026-11-08 00:41:10 +00:00 (Migrated from github.com)

#355 adds RX 570 4 GB logs (Windows 10) with the crash entering the 3D world, also at 50% resolution.

#355 adds RX 570 4 GB logs (Windows 10) with the crash entering the 3D world, also at 50% resolution.
nicholasgower commented 2026-11-08 09:04:24 +00:00 (Migrated from github.com)

The crash in the attached log is the shader recompiler, not a vendor-specific GPU path: a mesh program fails to translate because LoadGlobalS16 was not implemented, and the game then dereferences a null resource (read 0x100000000150, guest 0x150). The owner confirmed in #351 that the same LoadGlobalS16 failure appears on NVIDIA too, so this is not AMD-only.

The load half is already fixed in main by c3459dc. I opened #536 to complete the same global narrow-memory family for stores: WriteGlobalU8/S8/U16/S16 still threw NotImplementedException, so a mesh program using STG.U8/S8/U16/S16 would fail the same way. The new fixture reproduces that failure on the pre-fix code (WriteGlobalS16 is not implemented) and passes with the change, including spirv-val on every emitted module.

The RX 580/570 also lacks mesh shaders, which is a separate rendering limitation (507b95b now warns about it). The later report of VRAM exhaustion after ~10 seconds is the separate #105.

The crash in the attached log is the shader recompiler, not a vendor-specific GPU path: a mesh program fails to translate because `LoadGlobalS16` was not implemented, and the game then dereferences a null resource (`read 0x100000000150, guest 0x150`). The owner confirmed in #351 that the same `LoadGlobalS16` failure appears on NVIDIA too, so this is not AMD-only. The load half is already fixed in main by c3459dc. I opened #536 to complete the same global narrow-memory family for stores: `WriteGlobalU8/S8/U16/S16` still threw `NotImplementedException`, so a mesh program using `STG.U8/S8/U16/S16` would fail the same way. The new fixture reproduces that failure on the pre-fix code (`WriteGlobalS16 is not implemented`) and passes with the change, including `spirv-val` on every emitted module. The RX 580/570 also lacks mesh shaders, which is a separate rendering limitation (507b95b now warns about it). The later report of VRAM exhaustion after ~10 seconds is the separate #105.
nicholasgower commented 2026-11-08 09:27:42 +00:00 (Migrated from github.com)

Both halves of the narrow-global family are now covered: the load half is in main (c3459dc), and the store half (WriteGlobalU8/S8/U16/S16) is PR #536, which also extends the real-IR fixture and runs spirv-val on every module. The crash in the attached log (LoadGlobalS16, then a null resource at guest 0x150) needs a build that contains #536.

The RX 580 also lacks mesh shaders, so 507b95b will warn about that; it is a separate rendering limitation, not this crash. Please retest when a release has #536 and attach hookshot_log.txt if it still fails.

Both halves of the narrow-global family are now covered: the load half is in `main` (`c3459dc`), and the store half (`WriteGlobalU8/S8/U16/S16`) is PR #536, which also extends the real-IR fixture and runs `spirv-val` on every module. The crash in the attached log (`LoadGlobalS16`, then a null resource at guest `0x150`) needs a build that contains #536. The RX 580 also lacks mesh shaders, so `507b95b` will warn about that; it is a separate rendering limitation, not this crash. Please retest when a release has #536 and attach `hookshot_log.txt` if it still fails.
NotRexed commented 2026-11-08 10:39:51 +00:00 (Migrated from github.com)

PR #536 (8- and 16-bit global stores) is in release https://github.com/NotRexed/Hookshot/releases/tag/low-vram. Could you check whether this is fixed?

PR #536 (8- and 16-bit global stores) is in release https://github.com/NotRexed/Hookshot/releases/tag/low-vram. Could you check whether this is fixed?
ImagineThisGuy2 commented 2026-11-08 13:35:04 +00:00 (Migrated from github.com)

The game now reached the Deku Tree cutscene but crashed during it. I also noticed it is using 95% of my 24GB of RAM.

hookshot_log.txt

The game now reached the Deku Tree cutscene but crashed during it. I also noticed it is using 95% of my 24GB of RAM. [hookshot_log.txt](https://github.com/user-attachments/files/34544249/hookshot_log.txt)
ImagineThisGuy2 commented 2026-11-08 18:18:41 +00:00 (Migrated from github.com)

Newest build gets even further in the cutscene now crashes at the point where Navi flies under Mido

hookshot_log.txt

Newest build gets even further in the cutscene now crashes at the point where Navi flies under Mido [hookshot_log.txt](https://github.com/user-attachments/files/34550825/hookshot_log.txt)
Alaikin commented 2026-11-08 19:16:44 +00:00 (Migrated from github.com)
Image
<img width="1327" height="786" alt="Image" src="https://github.com/user-attachments/assets/9548f1c9-8155-4bad-9871-df7bc6986058" />
Alaikin commented 2026-11-08 19:17:12 +00:00 (Migrated from github.com)
[hookshot_log.txt](https://github.com/user-attachments/files/34553069/hookshot_log.txt) [hookshot_log.1.txt](https://github.com/user-attachments/files/34553070/hookshot_log.1.txt)
eric49a-byte commented 2026-11-08 20:28:12 +00:00 (Migrated from github.com)

i'm on amd 6950xt build 2.0.3 using bazzite linux, crash during navi cutscene in various places, see attached logs

hookshot_log.txt
hookshot_log.1.txt

i'm on amd 6950xt build 2.0.3 using bazzite linux, crash during navi cutscene in various places, see attached logs [hookshot_log.txt](https://github.com/user-attachments/files/34554301/hookshot_log.txt) [hookshot_log.1.txt](https://github.com/user-attachments/files/34554302/hookshot_log.1.txt)
NotRexed commented 2026-11-09 05:45:14 +00:00 (Migrated from github.com)

Tracked in #1001, follow there. That issue collects every report of GPUs without mesh shaders, and this one is linked from it with its GPU, version and log details. v2.0.6 brings back the fallback for GPUs without mesh shaders (PR #779). It is untested on these cards, so a v2.0.6 result from you would help. The RX 6950 XT (Bazzite) comment here is a different problem; please post that one in #1000. If it still happens on v2.0.6 (https://github.com/NotRexed/Hookshot/releases/tag/v2.0.6; Update next to Play in the launcher), please add a comment there with your hookshot_log.txt.

Tracked in #1001, follow there. That issue collects every report of GPUs without mesh shaders, and this one is linked from it with its GPU, version and log details. v2.0.6 brings back the fallback for GPUs without mesh shaders (PR #779). It is untested on these cards, so a v2.0.6 result from you would help. The RX 6950 XT (Bazzite) comment here is a different problem; please post that one in #1000. If it still happens on v2.0.6 (https://github.com/NotRexed/Hookshot/releases/tag/v2.0.6; Update next to Play in the launcher), please add a comment there with your `hookshot_log.txt`.
Commenting is not possible because the repository is archived.
No milestone
No assignees
1 participant
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/Hookshot#192
No description provided.