diff options
| author | Jingyue Wu <jingyue@google.com> | 2015-07-10 04:31:56 +0000 |
|---|---|---|
| committer | Jingyue Wu <jingyue@google.com> | 2015-07-10 04:31:56 +0000 |
| commit | ad85c8c2048708be55a9e5c82d734031bb75632a (patch) | |
| tree | 833cc443fa49b0a8339ce1e4bc8feaeceaa6addc /llvm/lib/Transforms/Utils/BasicBlockUtils.cpp | |
| parent | c851ccc3bd38b0887615f9b39950d07ffc66da0a (diff) | |
| download | llvm-ad85c8c2048708be55a9e5c82d734031bb75632a.zip llvm-ad85c8c2048708be55a9e5c82d734031bb75632a.tar.gz llvm-ad85c8c2048708be55a9e5c82d734031bb75632a.tar.bz2 | |
[NVPTX] declare no vector registers
Summary:
Without this patch, LoopVectorizer in certain cases (see loop-vectorize.ll)
produces code with complex control flow which hurts later optimizations. Since
NVPTX doesn't have vector registers in LLVM's sense
(NVPTXTTI::getRegisterBitWidth(true) == 32), we for now declare no vector
registers to effectively disable loop vectorization.
Reviewers: jholewinski
Subscribers: jingyue, llvm-commits, jholewinski
Differential Revision: http://reviews.llvm.org/D11089
llvm-svn: 241884
Diffstat (limited to 'llvm/lib/Transforms/Utils/BasicBlockUtils.cpp')
0 files changed, 0 insertions, 0 deletions