diff options
| author | Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com> | 2017-08-30 03:03:38 +0000 |
|---|---|---|
| committer | Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com> | 2017-08-30 03:03:38 +0000 |
| commit | 06cab79e5000c2859f0deff758d7526b7f4aff68 (patch) | |
| tree | 1f428056a221ddffc218ef9bb06b49bad5506cac /clang/lib/Serialization/ModuleManager.cpp | |
| parent | 5a2898ae0329419fd3d9a9dfa98834692581f609 (diff) | |
| download | llvm-06cab79e5000c2859f0deff758d7526b7f4aff68.zip llvm-06cab79e5000c2859f0deff758d7526b7f4aff68.tar.gz llvm-06cab79e5000c2859f0deff758d7526b7f4aff68.tar.bz2 | |
[AMDGPU] Use v_max_f* for fcanonicalize
If denorms are not flushed we can use max instead of multiplication
by 1. For double that is simply faster, while for float and half
it is shorter, because mul uses constant bus and VOP3.
Differential Revision: https://reviews.llvm.org/D36856
llvm-svn: 312095
Diffstat (limited to 'clang/lib/Serialization/ModuleManager.cpp')
0 files changed, 0 insertions, 0 deletions