Add profiler and Perfetto UI link with comprehensive tests

[jainapurva]

This PR adds profiler support and Perfetto UI link with comprehensive tests.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/1995

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 3 New Failures

As of commit 8ef520b with merge base 70fc520 ():

NEW FAILURES - The following jobs have failed:

• .github/workflows/float8nocompile_test.yaml (gh)
• PR Label Check / Check PR Labels (gh)
  Process completed with exit code 1.
• Run TorchAO Experimental Tests / test-mps-ops (macos-m1-stable) (gh)
  Process completed with exit code 1.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[Copilot]

Pull Request Overview

This PR adds profiler support and a Perfetto UI link to the benchmark utilities along with comprehensive tests to validate the functionality. Key changes include:

• New utility functions for uploading trace files, generating Perfetto UI URLs, and generating both model and memory profiling data.
• Enhancements to the BenchmarkConfig and benchmark runner to support profiler and memory profiling, with corresponding tests.
• Updates to the configuration YAML to enable profiling for selected models.

Reviewed Changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
benchmarks/microbenchmarks/utils.py	Added functions upload_trace_file, print_perfetto_ui_url, generate_model_profile, and generate_memory_profile; updated BenchmarkConfig to include profiler settings.
benchmarks/microbenchmarks/test/test_benchmark_profiler.py	Added comprehensive tests for profiler functionality on both CPU and CUDA devices.
benchmarks/microbenchmarks/test/benchmark_config.yml	Updated YAML to enable profiler and memory profiling for selected models.
benchmarks/microbenchmarks/benchmark_runner.py	Improved error handling in benchmark execution and CSV/result generation.
benchmarks/microbenchmarks/benchmark_inference.py	Wrapped profiler and memory profiling execution in try/except blocks and refined model compilation formatting.

Comments suppressed due to low confidence (1)

benchmarks/microbenchmarks/utils.py:243

• Usage of the private API torch.cuda.memory._record_memory_history may lead to breakage in future PyTorch releases; consider switching to a supported public API if one is available.

torch.cuda.memory._record_memory_history(True, trace_alloc_max_entries=250000, trace_alloc_record_context=True)

[Copilot]

os.getlogin() may raise an OSError in non-interactive environments; consider using getpass.getuser() for improved reliability.

Suggested change

	MANIFOLD_FOLDER, f"{os.getlogin()}_{str(uuid.uuid4())}_{file_name}"
	MANIFOLD_FOLDER, f"{getpass.getuser()}_{str(uuid.uuid4())}_{file_name}"