[PATCH] D138602: [WIP] Alwaysinliner time explosion with new pass manager
Davide Italiano via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Wed Dec 7 12:01:53 PST 2022
davide added a comment.
In D138602#3979174 <https://reviews.llvm.org/D138602#3979174>, @aeubanks wrote:
> sorry, this fell between the cracks
>
> are the alwaysinline attributes added purely for optimization purposes (counterexample, I've seen cases for testing where they are to make sure the number of frames is consistent)? if so I'd try removing literally every alwaysinline attribute and seeing what performance difference that makes
They're pulled from different libraries, so it's a mix of both. My main observation is that it was extremely hard even for compiler engineers to figure out the composition of all these attributes.
The only observable effect was something not terminating (after 5 days compiling ; )
> and I assume this is -O2/3?
This is `-Os`
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D138602/new/
https://reviews.llvm.org/D138602
More information about the llvm-commits
mailing list