-
Notifications
You must be signed in to change notification settings - Fork 236
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
Memory profiler for cuda #1996
base: profiler_combined_new
Are you sure you want to change the base?
Memory profiler for cuda #1996
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/1996
Note: Links to docs will display an error until the docs builds have been completed. ❌ 2 New FailuresAs of commit 2c62286 with merge base 70fc520 ( NEW FAILURES - The following jobs have failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Pull Request Overview
This PR introduces memory profiling support for CUDA in the microbenchmarks, integrating both model profiling and memory profiling into the benchmarking workflow. Key changes include the addition of utility functions for uploading trace files and generating URLs for Perfetto UI, modifications in the benchmark configuration to enable profiling, and updates to the benchmark runner to execute the new profiling functionalities.
Reviewed Changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.
File | Description |
---|---|
benchmarks/microbenchmarks/utils.py | Added functions to upload trace files, generate model and memory profiles, and generate Perfetto URLs. |
benchmarks/microbenchmarks/test/benchmark_config.yml | Enabled profiler and memory profile flags for one benchmark configuration. |
benchmarks/microbenchmarks/benchmark_runner.py | Updated error handling and conditional CSV generation based on collected results. |
benchmarks/microbenchmarks/benchmark_inference.py | Integrated calls to the newly added profiling functions with proper error logging. |
DEFAULT_TTL_SEC = 28 * 24 * 60 * 60 | ||
file_name = os.path.basename(local_path) | ||
manifold_path = os.path.join( | ||
MANIFOLD_FOLDER, f"{os.getlogin()}_{str(uuid.uuid4())}_{file_name}" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Using os.getlogin() can raise an OSError in non-interactive or service environments; consider using getpass.getuser() for improved robustness.
MANIFOLD_FOLDER, f"{os.getlogin()}_{str(uuid.uuid4())}_{file_name}" | |
MANIFOLD_FOLDER, f"{getpass.getuser()}_{str(uuid.uuid4())}_{file_name}" |
Copilot is powered by AI, so mistakes are possible. Review output carefully before use.
No description provided.