Axloss

Not sure how much I should trust the callstack in the crash, but it is seriosly wrong.

Note that a material and mission is not related at all, so feels like either I get bogus data or that your binary is broken.

i looked a bit deeper in the crashdump and found out that it’s likely stalling on creating ray tracing objects (regardless if you have it on or not).

To force it off you can try in game files open win32_settings.ini and set ray_tracing, to false.
Meaning it should say
ray_tracing = false
instead of
ray_tracing = true

stuck in amd driver. Try reinstall or update them. If that doesn’t help, this might help https://www.amd.com/en/support/kb/faq/gpu-601

could you post 2-3 more guids and we can see if we see anything common?

I can see that you are stuck in (physics I think) and then in the video player. The video player might help with a clean install of driver is something would be wrong on that side.

| |dxilconv.dll!hlsl::DxbcConverter::DxbcConverter(void)|Unknown|Non-user code. Symbols loaded.|
|---|---|---|---|
| |dxilconv.dll!hlsl::DxbcConverter::Alloc<>(struct IMalloc *)|Unknown|Non-user code. Symbols loaded.|
| |dxilconv.dll!CreateDxbcConverter(struct _GUID const &,void * *)|Unknown|Non-user code. Symbols loaded.|
| |dxilconv.dll!DxcCreateInstance(struct _GUID const &,struct _GUID const &,void * *)|Unknown|Non-user code. Symbols loaded.|
| |D3D12Core.dll!CShaderBytecode::FinalConstruct(struct CShaderBytecode::TConstructorArgs const &)|Unknown|Non-user code. Symbols loaded.|
| |D3D12Core.dll!CLayeredObject::CreateInstance(struct CShaderBytecode::TConstructorArgs const &,void *,void *,struct _GUID const &,void * *)|Unknown|Non-user code. Symbols loaded.|
| |D3D12Core.dll!CDevice::CreateLayeredChild(unsigned int,void const *,unsigned __int64,struct ID3D12LayeredUseCounted *,struct _GUID const &,void * *)|Unknown|Non-user code. Symbols loaded.|
| |D3D1...

To build the PSO’s we need to know both the mesh data and the shaders, and in our engine thisis decoupled and thus needs to be built in runtime. So in some sense we can control this by coupling meshes and shaders but this reduce the flexibility when building games.

In darktide we have done several improvements to fix this. We built alot of the pso’s directly in the launcher by figuring out what meshes and shaders we use. (note im not a expert in this field).

I can check with my colleague how hard it would be to move some of this changes.

all but one of the hangs is in a dll called tier0_s64.dll which is managed by steam. The last one was in a steam api call but from my perspective it doesn’t look like it’s deadlock in that case, just really slow (15s update, which we cutoff as a deadlock).

Sadly we can’t investigate further than this. Try update your steam client. In the next patch we plan to improve deadlock reporting to find info from our exe if possible. But it’s not 100% we will get any more info with that update.

We also have better tools to debug hangs in the gameplay code with the same update or before that. Since the last hang indicated that it could be something else.

let me know if you still have issues with discord overlay disabled.

you crash inside DiscordHook64.dll

Sadly we can’t investigate it further. So my best tip is to sidable to discord overlay if possible.

im in discussion with Intel to see if they have some insight in this.

with this severe issues I would try re-install the game, if that doesn’t help we need to figure out if we can get Windows to make a crashdump for you. But if the game doesn’t even launch before this happens it’s not unlikely that we wont be able to debug this issue.

[Process Exit] Exit code: 0xC000001D

STATUS_ILLEGAL_INSTRUCTION

sadly we can’t see more than you are stuck in the intel gpu driver.

also Steam Gameoverlay is in the callstack of the crash.

crashes caused by it, are usually fixed by disabling game overlay in steam. We can’t sadly not do anything here if that’s the case.

Since I see it is crashing after printing
21:53:17.147 [d3d12 pipeline state] creating pipeline state for #ID[94b534d7]

I would delete the these files

in %APPDATA%/Fatshark/Darktide
shader_cache.hans
shader_library.pso_lib
state_stream_library.pso_lib

But from the info I can see in our crash is that it could be related to your drivers so my suggestion is to see if a clean re-install of latest driver helps.

You can delete those files again after updating the driver, since updating the driver will make them invalid and we need to re-create the data in there.

Keep us updated if this helps or not.

In the logs I saw a crash in it was GPU hangs, which is very hard to debug.

But some had logs about pso creation so might be worth delete the local files for that and see if that helps. Let us know in that case.

in %APPDATA%/Fatshark/Darktide
shader_cache.hans
shader_library.pso_lib
state_stream_library.pso_lib

I’ve seen more users get stuck very early in the setup phase of the game.

We have added so we can detect such deadlocks, but not sure if this will be added to the patch we will release very soon or the one after that.

This will only help us to detect the issue since we then can trigger a crashdump during this phase which we then can analyze.

We are aware of different issues.

I added (today, meaning it will come in a future patch) so you can override the mesh streamer to disable it from user settings (currently you can only do it by modifying the steam/msstore data in settings*.ini files).

We will take a larger peek at this system in the near future to mitigate some of the issues.

I’m trying to follow up users with this kind of error that’s re-occurring.

just to be clear from the info I’ve been reading. Is the stock performance core multiplier 55x?

also from scimming through the logs it sounds like this is the issue. Like too high report rate, perhaps the windows queue quits very busy. However I will forward this to our QA so they can try reproduce the behaviour inhouse.

This also includes any kind of program that hooks into our app, like overlays.