]> granicus.if.org Git - clang/commit
[X86] As mentioned in post-commit feedback in D47174, move the 128 bit f16c intrinsic...
authorCraig Topper <craig.topper@intel.com>
Tue, 22 May 2018 22:19:19 +0000 (22:19 +0000)
committerCraig Topper <craig.topper@intel.com>
Tue, 22 May 2018 22:19:19 +0000 (22:19 +0000)
commitab89fc45f05fb49868d8cadf344a0e877481b0b8
treee2743be87301c7f0d5761b78520d7aa62d101083
parent9f33b6a567dfc557286f7ddb2dffb18da6cb8538
[X86] As mentioned in post-commit feedback in D47174, move the 128 bit f16c intrinsics into f16cintrin.h and remove __emmintrin_f16c.h

These were included in emmintrin.h to match Intel Intrinsics Guide documentation. But this is because icc is capable of emulating them on targets that don't support F16C using library calls. Clang/LLVM doesn't have this emulation support. So it makes more sense to include them in immintrin.h instead.

I've left a comment behind to hopefully deter someone from trying to move them again in the future.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@333033 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Headers/CMakeLists.txt
lib/Headers/__emmintrin_f16c.h [deleted file]
lib/Headers/emmintrin.h
lib/Headers/f16cintrin.h
lib/Headers/module.modulemap