projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: c60e180) | patch
[X86] Implement _builtin_ia32_vfmaddss and _builtin_ia32_vfmaddsd with native IR...
authorCraig Topper <craig.topper@intel.com>
Fri, 6 Jul 2018 07:14:47 +0000 (07:14 +0000)
committerCraig Topper <craig.topper@intel.com>
Fri, 6 Jul 2018 07:14:47 +0000 (07:14 +0000)
commitbe4c2933a2c370d292122fc2187d75436aeaea83
treede0c24b661056a87eeb5ecd1dd2317468679d400tree | snapshot
parentc60e1807b3f099d67a3ed7ceb77dd4257c6d61b6commit | diff
[X86] Implement _builtin_ia32_vfmaddss and _builtin_ia32_vfmaddsd with native IR using llvm.fma intrinsic.

This generates some extra zeroing currently, but we should be able to quickly address that with some isel patterns.

llvm-svn: 336417
clang/lib/CodeGen/CGBuiltin.cpp diff | blob | history
clang/test/CodeGen/fma4-builtins.c diff | blob | history
Domain: System / Toolchain;