Open maschad opened 11 months ago
As discussed in pl-strflt.notion.site/js-libp2p-perf-benchmarking-f7c42f96cbd44799a321347f6c15c8d4?pvs=4
In order to detect performance regressions at the component level, it may be useful to run component level benchmarks per PR. We've seen scenarios where mplex is more performant than yamux, despite mplex being deprecated. We could leverage the existing perf implementation and a Github action to create a comment such as this one https://github.com/ChainSafe/lodestar/pull/6004#issuecomment-1737257011
Add similar benchmarking that lodestar has but at the component level.
As discussed in pl-strflt.notion.site/js-libp2p-perf-benchmarking-f7c42f96cbd44799a321347f6c15c8d4?pvs=4
In order to detect performance regressions at the component level, it may be useful to run component level benchmarks per PR. We've seen scenarios where mplex is more performant than yamux, despite mplex being deprecated. We could leverage the existing perf implementation and a Github action to create a comment such as this one https://github.com/ChainSafe/lodestar/pull/6004#issuecomment-1737257011