This is an archive of the discontinued LLVM Phabricator instance.

[AVX512] add float/double abs intrinsics
ClosedPublic

Authored by AsafBadouh on Jul 4 2016, 6:58 AM.

Diff Detail

Repository
rL LLVM

Event Timeline

AsafBadouh updated this revision to Diff 62679.Jul 4 2016, 6:58 AM
AsafBadouh retitled this revision from to [AVX512] add float/double abs intrinsics.
AsafBadouh updated this object.
AsafBadouh set the repository for this revision to rL LLVM.
AsafBadouh added a subscriber: cfe-commits.
delena accepted this revision.Jul 5 2016, 2:34 AM
delena edited edge metadata.
delena added inline comments.
../tunkClang/tools/clang/test/CodeGen/avx512f-builtins.c
1413–1414 ↗(On Diff #62679)

The right thing to do here:

CHECK: %[[MASK:.*]] = bitcast
CHECK: %[[AND_RES:.*]] = and <16 x i32>
// CHECK: select <16 x i1> %[[MASK]], <16 x i32>%[[AND_RES]]

Up to you.

This revision is now accepted and ready to land.Jul 5 2016, 2:34 AM
AsafBadouh added inline comments.Jul 5 2016, 4:36 AM
../tunkClang/tools/clang/test/CodeGen/avx512f-builtins.c
1413–1414 ↗(On Diff #62679)

will do,
thanks for the review!

This revision was automatically updated to reflect the committed changes.