microsoft / microsoft/onnxruntime
[Feature Request] NVIDIA GB10 (Blackwell)
- Dominant language
- C++
- Stars
- 21.9k
- Forks
- 4.2k
- Avg merge
- 4d 11h
- Merged PRs (30d)
- 184
Description
### Describe the feature request
When i build for , i have error:
` 2026-03-03 19:21:46.614105057 [E:onnxruntime:, sequential_executor.cc:572 ExecuteKernel] Non-zero status code returned while running Relu node. Name:'Relu_2' Status Message: CUDA error cudaErrorNoKernelImageForDevice:no kernel image is available for execution on the device
2026-03-03 19:21:46,614 - services.websocket_service - ERROR - Lỗi xử lý frame: [ONNXRuntimeError] : 1 : FAIL : Non-zero status code returned while running Relu node. Name:'Relu_2' Status Message: CUDA error cudaErrorNoKernelImageForDevice:no kernel image is available for execution on the device`
Pls support me.
### Describe scenario use case
I build it for face detect.
Contributor guide
Research direction
The issue names no source files, tests, or entry points. Start by reproducing the face-detection build on NVIDIA GB10 and investigate ONNX Runtime's CUDA architecture support for Blackwell; done means the Relu execution succeeds without cudaErrorNoKernelImageForDevice.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp, machine-learning
- Domain
- computer-vision, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100