Regular running benchmarks for all the cops?

[Aqualon]

Hi, sorry if this was discussed before but with all the ongoing improvements in terms of performance with YJIT and other JITs I wondered, if all the cops still make sense in the latest ruby versions and if there should be ongoing benchmarks run to verify that what was suggested for an older version still applies?

I saw some issues that talked about changes needed in newer ruby version for certain cops, but that seems to be based on single observations and not a consistent approach to keep rubocop-performance useful.

[Earlopain]

I openend #467 a while ago but there was no feedback on it.