Skip to content

Adding proper profiler - #10

Merged
Sam-v6 merged 3 commits into
mainfrom
feature/add-profiler
Dec 8, 2025
Merged

Adding proper profiler#10
Sam-v6 merged 3 commits into
mainfrom
feature/add-profiler

Conversation

@Sam-v6

@Sam-v6 Sam-v6 commented Dec 7, 2025 •
edited
Loading

Copy link
Copy Markdown
Owner

What does this PR do?

Added PyTorch profiler where traces can be loaded into Perfetto UI (web browser)

What are some specific details on the changes?

  • Added PyTorch profiler which will for both model training and inference:
    • Summarize the top CPU and GPU kernel operations based on util time
    • Output a traces_training.json or traces_inference.json which can be loaded into https://ui.perfetto.dev/ and viewed as a timeline
  • README updated with instructions and examples

How has this been tested?

This is tied to --profile which is ran in CI

Important Notes

None

Checklist

  • Tests have been added or updated to cover the changes.
  • Documentation has been updated (if necessary)

Sam-v6 added 3 commits December 7, 2025 16:48
@Sam-v6
Profiler working for training, tied to --profile
dc65631
@Sam-v6
Updating readme with profiler instructions and timeline example
bdc6ee5
@Sam-v6
Adding:
3003c31 
-Profiler on inference
-Updates on readme for profiler, showing cpu/gpu ops and timeline
-Minor tweaks
@Sam-v6
Sam-v6 merged commit 6be6c9a into main Dec 8, 2025
1 check passed
@Sam-v6
Sam-v6 deleted the feature/add-profiler branch December 8, 2025 22:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@Sam-v6