Stream: t-release/triage

Topic: perf regression from rollup


view this post on Zulip Manish Goregaokar (Sep 12 2021 at 18:52):

I'm unsure what caused this perf regression https://github.com/rust-lang/rust/pull/88881#issuecomment-917682616

view this post on Zulip Manish Goregaokar (Sep 12 2021 at 18:52):

and unsure what I should do here

view this post on Zulip Manish Goregaokar (Sep 12 2021 at 18:52):

i think perhaps open revert PRs for some of the potential culprits and do perf runs?

view this post on Zulip Manish Goregaokar (Sep 12 2021 at 18:52):

none of the PRs seem likely

view this post on Zulip The 8472 (Sep 12 2021 at 19:07):

bisecting reverts might be faster than individual ones

view this post on Zulip Manish Goregaokar (Sep 12 2021 at 19:16):

would someone be able to take this on? idk if i'll have the time today

view this post on Zulip Manish Goregaokar (Sep 12 2021 at 19:17):

@The 8472 there aren't that many PRs in there that seem likely imo

view this post on Zulip Noah Lev (Sep 12 2021 at 20:07):

(Might be worth asking in #t-compiler/performance since they are the ones who maintain rustc-perf.)


Last updated: Jan 26 2022 at 14:20 UTC