Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

aarch64 Linux GPU build - CUDA version 10? #1555

Open
WilliamVenner opened this issue Nov 20, 2024 · 1 comment
Open

aarch64 Linux GPU build - CUDA version 10? #1555

WilliamVenner opened this issue Nov 20, 2024 · 1 comment

Comments

@WilliamVenner
Copy link

WilliamVenner commented Nov 20, 2024

Hi, I noticed #1500, thank you for this contribution!

I noticed that the CUDA version required for this build is CUDA 10.x, which I struggle to install specifically on Amazon Linux 2023 (for use with AWS Graviton ARM64 processors) as NVIDIA seemingly doesn't provide prebuilt installations of CUDA 10.x for ARM64. Would it be possible to support CUDA 11.x as well in order to extend support to this platform?

Perhaps also along with this the ONNX Runtime version could be upgraded to 1.17.1 like most of the other targets.

I was unable to pinpoint exactly where CUDA is installed in the GitHub actions workflow. I also could not find where you built ONNX Runtime in https://github.com/csukuangfj/onnxruntime-libs unfortunately.

@csukuangfj
Copy link
Collaborator

Yes, it's definitely possible.

The current pre-built onnxruntime lib is provided by a community user, whose board supports only cuda 10.2

If you can build onnxruntime with cuda support for linux arm64 and share the lib with us, then we can update sherpa-onnx to support it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants