This is an archive of the discontinued LLVM Phabricator instance.

[X86] Fix perf bug in `permilps` -> `shufd` in X86FixupInstTuning.
ClosedPublic

Authored by goldstein.w.n on Apr 6 2023, 11:12 AM.

Details

Summary

We shouldn't do the transformation if we either have bypass delay OR
the new opcode has worse performance. Previous code was incorrectly
using AND.

Diff Detail

Event Timeline

goldstein.w.n created this revision.Apr 6 2023, 11:12 AM
Herald added a project: Restricted Project. · View Herald TranscriptApr 6 2023, 11:12 AM
goldstein.w.n requested review of this revision.Apr 6 2023, 11:12 AM
Herald added a project: Restricted Project. · View Herald TranscriptApr 6 2023, 11:12 AM
RKSimon accepted this revision.Apr 6 2023, 1:24 PM

LGTM

This revision is now accepted and ready to land.Apr 6 2023, 1:24 PM