It would be nice to have a documented set of steps for contributors to follow when benchmarking any performance improvement code changes that they make as part of their contribution. They post these on PRs to facilitate better visualization for reviewers, and can serve as a reference for future.