This is an archive of the discontinued LLVM Phabricator instance.

[PowerPC][Altivec] Emit correct builtin for single precision vec_all_ne
ClosedPublic

Authored by nemanjai on Jul 1 2019, 10:45 AM.

Details

Summary

We currently emit a double precision comparison instruction for this, whereas we need to emit the single precision version.

Diff Detail

Event Timeline

nemanjai created this revision.Jul 1 2019, 10:45 AM
Herald added a project: Restricted Project. · View Herald TranscriptJul 1 2019, 10:45 AM
jsji accepted this revision.Jul 1 2019, 11:41 AM

LGTM.
BTW, looks like we are missing test cases all vector float vec_all* and vec_any* and also non-vsx path?

This revision is now accepted and ready to land.Jul 1 2019, 11:41 AM

LGTM.
BTW, looks like we are missing test cases all vector float vec_all* and vec_any* and also non-vsx path?

The non-VSX path is tested in test/CodeGen/builtins-ppc-altivec.c
I will add the VSX tests in this commit.

This revision was automatically updated to reflect the committed changes.