This is an archive of the discontinued LLVM Phabricator instance.

[ARM] Better AND's for MVE compares
ClosedPublic

Authored by dmgreen on Jul 21 2019, 10:10 AM.

Details

Summary

Add a number of folds to convert and(vcmp, vcmp) into a single VPT block, where the second vcmp becomes predicated on the first.

The VCMP; VPST; VCMP will then be converted to VPT; VCMP in the VPTBlockPass.

Diff Detail

Repository
rL LLVM

Event Timeline

dmgreen created this revision.Jul 21 2019, 10:10 AM
SjoerdMeijer accepted this revision.Jul 22 2019, 12:50 AM

Looks like a good optimisation to me.

This revision is now accepted and ready to land.Jul 22 2019, 12:50 AM
This revision was automatically updated to reflect the committed changes.