We're not running the benchmark suite in CI, as that builds with stack build --fast which is like 15-20x slower than optimized builds at running Unison code, so I think it'd be pretty misleading, while also adding more time to every build.
But it would be great to have benchmarks run automatically on every release.
@ceedubs brought up that we can't necessarily guarantee that the Github Actions runner is going to run with consistent CPU/memory, enough that the benchmark runs would be comparable.
We're not running the benchmark suite in CI, as that builds with
stack build --fast
which is like 15-20x slower than optimized builds at running Unison code, so I think it'd be pretty misleading, while also adding more time to every build.But it would be great to have benchmarks run automatically on every release.