Closed oojo12 closed 1 year ago
We shouldn't be running Criterion benchmarks at all in CI, since the results are highly unreliable. It also makes the CI pipeline slower for no reason. If we transition away from Criterion we'll update the CI when that happens.
Currently, we have a workflow to run benchmarks specifically for Iai benches which we are transitioning away from in favor of Criterion. Should we make the benches yaml generic and have it run all benchmarks instead with
cargo bench --all
? This way if we transition away from Criterion, we won't have to update the bench.yaml?