This is an archive of the discontinued LLVM Phabricator instance.

[mlir][vector] Add missing canonicalisation for vector.multi_reduction
ClosedPublic

Authored by awarzynski on Aug 11 2023, 8:59 AM.

Details

Summary

Make sure that when canonicalising masked vector.multi_reduction and
creating arith.select to replace the mask, scalability of the mask is
preserved.

Diff Detail

Event Timeline

awarzynski created this revision.Aug 11 2023, 8:59 AM
Herald added a project: Restricted Project. · View Herald Transcript
awarzynski requested review of this revision.Aug 11 2023, 8:59 AM
dcaballe accepted this revision.Aug 11 2023, 9:19 AM

Thanks! Yeah, we'll probably hit a ton of similar issues, unfortunately...

This revision is now accepted and ready to land.Aug 11 2023, 9:19 AM