When using a tool like jmh or kotlinx-benchmark to...
# random
y
When using a tool like jmh or kotlinx-benchmark to benchmark a PR change. Is it sufficient to run the benchmark pre-PR and then post-PR on the same machine? or must the benchmark also be run at the same time (by editing the PR so that it can co-exist with the pre-existing code). I'm testing a significant change (moving from a sealed class hierarchy to value classes) and so it's a bit difficult to have both cases co-exist.
­čĹÇ 1
e
forget it, pre and post-PR is the only way, imho
y
Yeah I ended up doing that and just trying to ensure that the conditions for each benchmark are the same (e.g. laptop position, turning off any heating, no apps open, etc)