Skip to content

Peakmem and other benchmarks always reporting worse perfomance #1891

Open

Description

@dylannelson

Version

v2026.10.1

How did you install UXarray?

Source

What happened?

We noticed lately that nearly 15 tests always show that they are performing worse, making the benchmark tests really hard to read through and find out what actually got worse:
They appear in #1710 and a few others, which can be a bit confusing when looking at benchmarking PR's like the one below:

Another PR on benchmarking
image

The work I'm doing with grid cleaning
image

#1710
image

Since ~15 tests are giving us likely wrong results, it seems like we need to fix them or remove them if they can't be fixed so we can go back to having clear results.

What did you expect to happen?

Clear results in the "got worse" tab on an empty PR

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    benchmarkingRelated to benchmarks, memory usage, and/or time profilingbugSomething isn't working

    Type

    No type

    Fields

    Priority

    None yet

    Projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions