Skip to content

Conversation

@PaliC
Copy link
Contributor

@PaliC PaliC commented May 5, 2025

Right now we're running into the problem where the cli is timing out before the timeout which is defined in https://github.com/gpu-mode/reference-kernels.

Reference Kernels should be the source of truth so on the cli we just need the timeout to be long enough to accommodate for everything + exist. I think extending things for 1 hour should accomplish this.

Testing:
./target/release/popcorn-cli submit --gpu MI300 --leaderboard amd-mixture-of-experts --mode leaderboard submission.py timed out before the change but not after the change. The submission is the one here https://github.com/gpu-mode/reference-kernels/blob/main/problems/amd/moe/submission.py

@msaroufim msaroufim self-requested a review May 6, 2025 15:52
@msaroufim msaroufim merged commit 488050e into gpu-mode:main May 6, 2025
4 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