projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 310fa65) | patch
BBVectorize: Don't vectorize vector-manipulation chains
authorHal Finkel <hfinkel@anl.gov>
Tue, 13 Nov 2012 03:12:40 +0000 (03:12 +0000)
committerHal Finkel <hfinkel@anl.gov>
Tue, 13 Nov 2012 03:12:40 +0000 (03:12 +0000)
commit4387b8c95971a512e07bfda30dea6459e8419e8f
tree856f3aca8f216e27a9753901c637a80ff02f4e30tree | snapshot
parent310fa65ab9acb96319d0fcb779cdd35350f5d00fcommit | diff
BBVectorize: Don't vectorize vector-manipulation chains

Don't choose a vectorization plan containing only shuffles and
vector inserts/extracts. Due to inperfections in the cost model,
these can lead to infinite recusion.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167811 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Transforms/Vectorize/BBVectorize.cpp diff | blob | history
test/Transforms/BBVectorize/X86/sh-rec3.ll [new file with mode: 0644] blob
C/C++ LLVM-based compilers that forbids OOTA behaviors