pytorch / pytorch/benchmark

We should support outputting quantile results

Open
#2,510 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
1k
Forks
346
PR merge metrics
No merged PRs in 30d

Description

Since one operator can have multiple inputs, the average results can be misleading. When the optimized operator is faster than the baseline operator on average, but some quantiles are slower, we might overlook performance inefficiencies. We should report the quantile results for each operator.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

The issue names no files, tests, or entry points. Start by tracing where benchmark results are aggregated and displayed, then determine how quantiles should be calculated per operator; done means reports show quantile results alongside averages for each operator.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
performance
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
42/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.