Skip to content

Conversation

mdboom
Copy link
Contributor

@mdboom mdboom commented Oct 2, 2024

This is mainly aimed at making sure our benchmarks catch whatever caused the GC regression in CPython. I measured about a 10% speedup on this when reverting the GC changes (and about 50% with debug builds, but performance of debug builds is not guaranteed anyway).

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This file can be deleted, it isn't used in the build.

@mdboom
Copy link
Contributor Author

mdboom commented Oct 2, 2024

Thanks for the feedback, @AA-Turner. I think I've addressed them all (and fixed the tests).

- Remove config for builders we don't use (doctest, gettext, html, latex, epub, coverage, linkcheck)
- Remove config for nitpicky mode (not enabled in the benchmark)
- Remove irrelevant setup (e.g. doctest)
- Remove setup for `notfound.extension`
- Remove non-determinism with usage of `time.strftime()`
@AA-Turner
Copy link
Member

I've created a series of commits to potentially reduce the size of the benchmark -- and opened mdboom#1 for easy viewing of them.

You may not want to include some (or any) of them, but the idea was to remove irrelevant things from the benchmark data (& also hopefully take preventative measures against so that the pyperformance repo doesn't grow in size too much!)

A

Reduce the size of the sphinx benchmark
Copy link
Member

@AA-Turner AA-Turner left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks!

A

@mdboom mdboom merged commit f473300 into python:main Oct 8, 2024
11 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants