Skip to content

[CI] Fix AMDGPU arch flag for Codeplay runner #16053

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

Merged
merged 1 commit into from
Nov 12, 2024
Merged

Conversation

sarnex
Copy link
Contributor

@sarnex sarnex commented Nov 12, 2024

New HIP runner is up and running but it's a different card so we need a different arch flag.

Not the cleanest solution but this isn't really going to scale.

Confirmed working here.

@sarnex
Copy link
Contributor Author

sarnex commented Nov 12, 2024

Cancelled the CI here to not waste resources

@sarnex sarnex marked this pull request as ready for review November 12, 2024 16:54
@sarnex sarnex requested a review from a team as a code owner November 12, 2024 16:54
@sarnex sarnex requested a review from uditagarwal97 November 12, 2024 16:54
@@ -284,7 +284,11 @@ jobs:
echo "opts=$CMAKE_EXTRA_ARGS" >> $GITHUB_OUTPUT
else
if [ "${{ contains(inputs.target_devices, 'ext_oneapi_hip') }}" == "true" ]; then
echo 'opts=-DHIP_PLATFORM="AMD" -DAMD_ARCH="gfx1031"' >> $GITHUB_OUTPUT
if [ "${{ runner.name }}" == "cp-amd-runner" ]; then
Copy link
Contributor

Choose a reason for hiding this comment

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

I'm a bit hesitant about hard coding the runner name here. Just like we have extra_lit_opts, can we have another matrix input, something like extra_lit_build_opts, for passing runner specific build options? Or perhaps we can reuse inputs.extra_cmake_args?

Copy link
Contributor Author

@sarnex sarnex Nov 12, 2024

Choose a reason for hiding this comment

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

we could, but wouldn't we have to have the same hardcoded runner name check at the place where we decide what to pass for extra_lit_opts? There are two runners and they need different values for this variable. I think the real solution is running sycl-ls and grepping the output for the architecture, but I didn't want to overengineer it if the solution in this PR is acceptable.

Copy link
Contributor

Choose a reason for hiding this comment

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

True, running sycl-ls and grepping the architecture type would be an ideal solution. I don't have strong objection to the changes in this PR, so I'll approve, but can you also create a Github issue to "use sycl-ls for determining the runner architecture type"? I can implement that once I get some free cycles.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

sure, i can own it actually, seems at least somewhat interesting.

#16057

@sarnex sarnex merged commit f9c4aaf into intel:sycl Nov 12, 2024
3 of 13 checks passed
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

Successfully merging this pull request may close these issues.

2 participants