Add a script to gather runner info when uploading benchmark results #6425
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎ 1 Skipped Deployment
|
facebook-github-bot
added
the
CLA Signed
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
|
This depends on #6429 |
yangw-dev
approved these changes
Mar 17, 2025
| device_type = torch.cuda.get_device_name() | ||
|
|
||
| except ImportError: | ||
| pass |
There was a problem hiding this comment.
logging the error info to help debugging
| runner_info["type"] = device_type | ||
| runner_info["gpu_count"] = torch.cuda.device_count() | ||
| runner_info["avail_gpu_mem_in_gb"] = int( | ||
| torch.cuda.get_device_properties(0).total_memory |
There was a problem hiding this comment.
is each device has same memory?
There was a problem hiding this comment.
Yup, that's the regular setup
huydhn
added a commit
that referenced
this pull request
Mar 17, 2025
While working on #6425, I discover several bugs in the upload scripts: * If there is an invalid JSON file in the directory, the script returns instead of continue, skipping all records after. Covered by https://github.com/pytorch/test-infra/blob/main/.github/scripts/benchmark-results-dir-for-testing/v3/mock.json * The script didn't handle correctly JSONEachRow format with only one record. Covered by a new test JSON from https://github.com/pytorch/test-infra/pull/6425/files#diff-bff954994eb33173b7119ff8d280f3367117b2daa9b8c54888be5f48f183a280 * The script didn't handle correctly JSONEachRow format mix with list of records. Covered by https://github.com/pytorch/test-infra/blob/main/.github/scripts/benchmark-results-dir-for-testing/v3/json-each-row.json#L3 ### Testing https://github.com/pytorch/test-infra/actions/runs/13909203687/job/38919334944#step:5:125 looks correct now
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Implement the logic to gather runner info for GPU. I adopt this logic from https://github.com/pytorch/pytorch-integration-testing/blob/master/vllm-benchmarks/upload_benchmark_results.py#L102
This also cleans up v2 logic which is not used anymore.
cc @yangw-dev Please let me know if you have a better approach in mind from the utilization monitoring project. Essentially, I want to get the device name, i.e. CUDA, ROCm, and the device type, i.e. H100, MI300X, so that they can be displayed on the dashboard. Before this change, these fields are set by the caller, now they can be set automatically by the GHA.