facebookresearch/HolisticTraceAnalysis

Support the Equivalent of profiler.key_averages(group_by_stack_n=n)

Open

#294 opened on Sep 3, 2025

View on GitHubContributor guide
 (1 comment) (0 reactions) (0 assignees)Python (94 forks)auto 404
help wantedneeds triagequestion

Repository metrics

Stars
 (535 stars)
PR merge metrics
 (PR metrics pending)

Description

What is your question?

The pytorch profiler can aggregate runtimes by the operation and its traceback via key_averages(group_by_stack_n=n): https://docs.pytorch.org/tutorials/beginner/profiler.html#print-profiler-result

Identifying the top operators / kernels and their source user Python code is helpful to find the main bottlenecks and perform targeted optimizations. Does HTA support something equivalent, or could you shed some light on implementing one? Thanks!

I also found get_aten_op_kernels_and_delay(), but seems like it traces down from aten ops only instead of source user training code.

Code

No response

What have you tried?

No response

Environment

No response

Contributor guide

Research direction
Investigate how PyTorch profiler's key averages(group by stack n) works and determine if HTA can leverage existing trace data to provide similar stack trace aggregation. Explore the feasibility of adding a method to aggregate CUDA kernel and ATen operator execution times grouped by user stack frames.
Tech stack
pythonpytorch
Domain
backendperformance
Issue type
Feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
PythonPyTorch
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40