[UE 5.8.2] MetaHuman Markerless Motion Capture Stage 4 causes GPU reset and black screens on RTX 5070

Development Character & Animation
,
1

Hello, I am experiencing a repeatable GPU/display crash when using the experimental MetaHuman Animator Markerless Motion Capture plugin in Unreal Engine 5.8.2.

System specifications:

  • Windows 11
  • Unreal Engine 5.8.2, build 56702186
  • MetaHuman Animator Markerless Motion Capture plugin 1.0.0 Beta
  • NVIDIA RTX 5070, 12 GB VRAM
  • AMD Ryzen 9 9950X3D
  • 32 GB system RAM
  • NVIDIA Studio Driver 616.56
  • DirectX 12

Source video:

  • H.264 MP4
  • 1080 × 1920
  • 30 FPS
  • Approximately 7 seconds

Problem:

Processing completes Stages 1–3 normally. As soon as Stage 4 begins, all connected monitors lose signal and become completely black. The computer appears to remain powered on and working, but the display never recovers. I eventually have to force the computer to shut down.

This has happened repeatedly, including with a very short video.

The Unreal log reports ONNX Runtime/DirectML errors during Stage 4, followed by a GPU device reset/removal:

DXGI_ERROR_DEVICE_RESET
DXGI_ERROR_DEVICE_REMOVED
HRESULT: 887A0005
onnxruntime.dll
DmlExecutionProvider
ExecutionProvider.cpp(952)

This does not appear to be a normal VRAM out-of-memory error. At the time of the failure, approximately 6.9 GB of VRAM was being used, with an approximately 11.2 GB budget. The Unreal crash information also reported bIsOOM=0.

Troubleshooting already performed:

  • Clean installation of NVIDIA Studio Driver 616.56
  • The problem also occurred with NVIDIA driver 595.76
  • Restarted Windows after driver installation
  • Removed custom TdrDelay and TdrDdiDelay registry values
  • Disabled ray tracing
  • Turned off the 4K monitor and tested using only a Full HD monitor
  • Tested body tracking with face tracking disabled
  • Tested only 30 frames
  • Disabled the preview pass
  • Used a fresh temporary performance asset
  • Confirmed DirectX 12 is enabled
  • Confirmed the problem still happens with the short 7-second clip

The 30-frame body-only test produced the same DirectML error at Stage 4 and reset the display driver.

Comparison:

The markerless body-capture workflow completed successfully on another computer using:

  • Unreal Engine 5.8.0
  • RTX 5090
  • Ryzen processor
  • 128 GB RAM
    (later i tried to install 5.8.0 to main PC rtx 5070 but UE doesnt let to install it as i had latest 5.8.2)

Therefore, I suspect a compatibility issue involving one of the following:

  • MetaHuman Animator Markerless Motion Capture plugin and UE 5.8.2
  • ONNX Runtime/DirectML and the RTX 5070
  • A regression between UE 5.8.0 and UE 5.8.2
  • Different memory requirements or plugin builds

Has anyone reproduced this on an RTX 5070 with UE 5.8.2?

Is there a known issue or workaround for the Stage 4 DirectML error 887A0005? Is it possible to test the UE 5.8.0 version of the plugin without compiling the complete engine from source?

I can provide the complete Unreal log if required.

2

Hi @Hasan_Kuchkarov

Thanks for the information that is complete!

I have seen DXGI_ERROR_DEVICE_REMOVED and DXGI_ERROR_DEVICE_RESET happen on occasion when a GPU is getting too busy with DirectML or ONNX workloads during compute passes.
This appears to happen sometimes during stage 4 in the MetaHuman Animator with some projects, even when VRAM is not full. It could be that either the GPU is getting timed out by the OS or there is an issue with the DirectML runtime that triggered a GPU reset inside UE 5.8.2

A few other things we could try to narrow it down perhaps

  • TdrDelay Registry Tweak
    You deleted your TdrDelay custom values, but increasing TdrDelay and TdrDdiDelay (as you can do in KEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\GraphicsDrivers) can provide Windows with the breathing room it needs to handle intensive DirectML passes without assuming a crash has occurred and resetting the display driver.

  • DirectX 11 / Default RHI Test
    While DirectML usually relies on DX12, I’d be curious to know if testing with different Default RHI or Command Line Arguments (-dx11 or -d3d12) changes how DirectML initializes on your RTX 5070.

  • NVIDIA Control Panel Settings
    Under 3D Settings for Unreal Engine, try changing Power Management Mode to Prefer Maximum Performance, and see if disabling Shader Cache or clearing the DirectX Shader Cache resolves the issue.

To help narrow this down further

  • Could you test if disabling GPU Power Management in NVIDIA Control Panel affects Stage 4 behavior?

  • Does the crash happen when launching Unreal Engine with -log parameter, and if so, Are there any telltale DirectML execution logs preceding the driver reset?

Here are a couple of useful resources regarding GPU crash troubleshooting and TDR registry fixes in Unreal Engine:

How to Fix a GPU Driver Crash

Dealing with a GPU Crash

Video:

Resolving “GPU Crashed or D3D Device Removed” Errors in Unreal Engine 5

Let me know how it goes or if adjusting TdrDelay changes anything!