SLPVectorizer: start constructing chains at stores that are not power of two.
authorNadav Rotem <nrotem@apple.com>
Wed, 19 Jun 2013 15:57:29 +0000 (15:57 +0000)
committerNadav Rotem <nrotem@apple.com>
Wed, 19 Jun 2013 15:57:29 +0000 (15:57 +0000)
commit7d180ac7b6135bd378739dc2f760a5b31cd12cae
treea35b4fce4382d7f03231c3a7492bdc1effa00908
parent7391366cdf55ca5f5faf97a9ac2b01da47e2749c
SLPVectorizer: start constructing chains at stores that are not power of two.
The type <3 x i8> is a common in graphics and we want to be able to vectorize it.

This changes accelerates bullet by 12% and 471_omnetpp by 5%.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@184317 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Transforms/Vectorize/VecUtils.cpp
test/Transforms/SLPVectorizer/X86/odd_store.ll [new file with mode: 0644]