yongwww commented on issue #18126: URL: https://github.com/apache/tvm/issues/18126#issuecomment-3053023955
Thanks, @gugudeshubao for raising this! TVM does support various GPU architectures such as **Hopper**, **Blackwell**, **Orin**, and **Thor**. Just a quick note, the `SM-90`target you mentioned is part of the **Hopper** family, not Blackwell/Thor. To target specific architectures in your TVM build, you can set `CMAKE_CUDA_ARCHITECTURES` in your build config, you can check your GPU arch via `nvidia-smi --query-gpu=compute_cap --format=csv`. ```cmake set(CMAKE_CUDA_ARCHITECTURES 90a) # Hopper ``` If you’re still running into issues, feel free to share the specific error messages, please. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
