[release-test] GPU memory numbers are always 0 on A100 runner
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1k
- Forks
- 346
- PR merge metrics
- No merged PRs in 30d
Description
The command nvidia-smi pmon -s m -c 1 -o T does not give GPU memory numbers correctly.
https://github.com/pytorch/benchmark/blob/main/userbenchmark/release-test/monitor_proc.sh#L24
Error workflow:
https://github.com/pytorch/benchmark/actions/runs/12878326305/job/35904096937
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start at userbenchmark/release-test/monitor_proc.sh line 24 and inspect how nvidia-smi pmon -s m -c 1 -o T is used. Reproduce or review the failure in the linked release-test workflow for the A100 runner. Done means the release-test monitoring output reports nonzero, correct GPU memory numbers.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- shell
- Domain
- testing
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 52/100