Skip to content

Benchmarks

Every pull request to devel and main benches each operator marked @pytest.mark.bench; publish-results.yml merges the results into the ci branch, and the chart below is rebuilt from that history after every run.

Each operator gets one chart per architecture and test suite, one line per parametrization, with a metric selector (ci/scripts/build_pages.py).

Opens in a new tab

The chart page is a self-contained static build, refreshed independently of this site.

Open the benchmark history