projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: ed3caf9) | patch
Do some peephole optimizations to remove pointless VMOVs from Neon to integer
authorCameron Zwarich <zwarich@apple.com>
Sat, 2 Apr 2011 02:40:43 +0000 (02:40 +0000)
committerCameron Zwarich <zwarich@apple.com>
Sat, 2 Apr 2011 02:40:43 +0000 (02:40 +0000)
commit4071a711126a2a75585a32b96bb5d15ea267a915
tree1e693a94ba4f9c310640ecc3540de493f3545a59tree | snapshot
parented3caf90866e183380a06c0ae49101204a9f3c28commit | diff
Do some peephole optimizations to remove pointless VMOVs from Neon to integer
registers that arise from argument shuffling with the soft float ABI. These
instructions are particularly slow on Cortex A8. This fixes one half of
<rdar://problem/8674845>.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128759 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/ARM/ARMISelLowering.cpp diff | blob | history
test/CodeGen/ARM/fp-arg-shuffle.ll [new file with mode: 0644] blob
C/C++ LLVM-based compilers that forbids OOTA behaviors