@WG-prioritization/alerts issue #76247 has been requested for prioritization.
This perf. loss is different from what reported in (for example) in https://github.com/rust-lang/rust/issues/75992, right? Should we invoke a bisection on this issue?
yes, this is about runtime and the other is about compile time
I think we're going to find it was because of the llvm 10 bump, but a bisection wouldn't hurt
ok thanks. Adding bisect label
also this is a regression from 1.44 -> 1.45 and the otherwas 1.45 -> 1.46
@Joshua Nelson yeah pretty sure this is llvm, was on 1.44 to 1.45 when LLVM upgrade to 10 was done?
I saw that 11 doesn't fix the situation
we should mainly ping llvm crew
https://github.com/rust-lang/rust/pull/67759/commits/9f128235b49199f766f40df08c8a7eb25e143ae9 says the first tag it's in is 1.45 so yeah
if we were to prioritize this, I'd say
what others think?
but on the other hand there's no much we can do about this :)
so it's meh I guess
well, we can do something, the llvm crew can patch LLVM :P
or we could push a bit more there
maybe this is worth a nomination?
I think nomination and P-medium maybe?
the code still works, it's just slower
well yeah but hmm
let's wait for another person that suggest a priority to untie this :)
I d go p-high as well because the drop seems huge, even though there’s not much we can do
Issue #76247's prioritization request has been removed.