Skip to content

[BENCHMARK] Fix CUTLASS benchmark when running on BMG #4254

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

Closed
jle-quel opened this issue May 21, 2025 · 0 comments · Fixed by #4274
Closed

[BENCHMARK] Fix CUTLASS benchmark when running on BMG #4254

jle-quel opened this issue May 21, 2025 · 0 comments · Fixed by #4274
Assignees

Comments

@jle-quel
Copy link
Contributor

jle-quel commented May 21, 2025

Describe the bug

Following the integration of CUTLASS as a new provider for the GEMM benchmark, it has been observed that it does not work when running on BMG.

https://github.com/intel/intel-xpu-backend-for-triton/actions/runs/15121382503/job/42504328424

Environment details

Triton: 1d99b7a

@jle-quel jle-quel self-assigned this May 21, 2025
@jle-quel jle-quel changed the title [BENCHMARK] Fix BMG benchmark when running with CUTLASS [BENCHMARK] Fix CUTLASS benchmark when running on BMG May 21, 2025
@jle-quel jle-quel linked a pull request May 21, 2025 that will close this issue
@jle-quel jle-quel linked a pull request May 22, 2025 that will close this issue
whitneywhtsang pushed a commit that referenced this issue May 23, 2025
This PR resolves the issue
[4254](#4254)
by specifying multiple `sycl-target` to create a fatbinary and be
supported on PVC and BMG.

BMG benchmark CI:
https://github.com/intel/intel-xpu-backend-for-triton/actions/runs/15200477831

---------

Signed-off-by: Jefferson Le Quellec <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

1 participant