projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 1bdadf3) | patch
LoopVectorizer: Enable unrolling of conditional stores and the load/store
authorArnold Schwaighofer <aschwaighofer@apple.com>
Sun, 2 Feb 2014 03:12:34 +0000 (03:12 +0000)
committerArnold Schwaighofer <aschwaighofer@apple.com>
Sun, 2 Feb 2014 03:12:34 +0000 (03:12 +0000)
commita16c1b55e2aac49c7336f3f54b50bbe85335712e
tree2cc6f130fcbdbcba7aa30ad45315a1e519c954a7tree | snapshot
parent1bdadf38ef3273b40c2f7f253157a09a975f08a3commit | diff
LoopVectorizer: Enable unrolling of conditional stores and the load/store
unrolling heuristic per default

Benchmarking on x86_64 (thanks Chandler!) and ARM has shown those options speed
up some benchmarks while not causing any interesting regressions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@200621 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Transforms/Vectorize/LoopVectorize.cpp diff | blob | history
test/Transforms/LoopVectorize/ARM/arm-unroll.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors