Omnibenchmark, a continuous benchmarking tool.
With poetry or pip. With micromamba if handling software with conda. See the tutorial.
See our tags.
Please note the omni-schema dependency. Benchmark YAML schemas are updated by:
- Update src/omni_schema/schema/omni_schema.yaml manually
- `make all`
- Consider `make deploy`
And that omni-schema versions are tagged and pinned at the pytoml level (example), so omni-schema changes must precede omnibenchmark changes.
Omnibenchmark docs are served at https://omnibenchmark.org as generated on Renku's GitLab with a review/staging -> production flow.
Documentation includes a CLI reference. This reference is generated/automated via mkdocs-click and extracts the current CLI commands from omnibenchmark's main head. Hence, changes to omnibenchmark must be merged to main and precede changes to omnibenchmark's documentation.
Omnibenchmark incorporates great FOSS components, including but not limited to: Snakemake, easybuild, apptainer, lmod, LinkML, git. Thank you!