Skip to content

Conversation

@t-vi
Copy link
Collaborator

@t-vi t-vi commented Nov 14, 2019

This adds the HIP_VERSION cmake variable as hip_version.
This should help detecting ROCm, e.g. in #22091.

To parallel CUDA, hip_version is a string.
An alternative variant might be to split by '.' and only take the first two parts.
The method suffers a bit from ROCm not being as monolithic as CUDA.

@t-vi
Copy link
Collaborator Author

t-vi commented Nov 14, 2019

@ezyang @iotamudelta this would seem like a plausible alternative using HIP_COMP in #22091 in all the ifs.

Copy link
Contributor

@iotamudelta iotamudelta left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this is a very good start.

@bddppq bddppq added the module: rocm AMD GPU support for Pytorch label Nov 14, 2019
@bddppq bddppq changed the title Add version.hip_version from cmake Add torch.version.hip from cmake Nov 15, 2019
@bddppq
Copy link
Contributor

bddppq commented Nov 15, 2019

Thanks!

Copy link
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@bddppq has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

@facebook-github-bot
Copy link
Contributor

@bddppq merged this pull request in 7889e1e.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Merged module: rocm AMD GPU support for Pytorch

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants