diff options
| author | Nadav Rotem <nrotem@apple.com> | 2013-01-18 23:10:30 +0000 |
|---|---|---|
| committer | Nadav Rotem <nrotem@apple.com> | 2013-01-18 23:10:30 +0000 |
| commit | 48177ac90fb940833b9deea1a6716092348cfe82 (patch) | |
| tree | 5252f0617e256f0dd1f8b26082f05a088d7232b9 /test/CodeGen/X86/v8i1-masks.ll | |
| parent | 7336f7febb5170b374a4cbffee273ad82ff8a1a3 (diff) | |
| download | external_llvm-48177ac90fb940833b9deea1a6716092348cfe82.zip external_llvm-48177ac90fb940833b9deea1a6716092348cfe82.tar.gz external_llvm-48177ac90fb940833b9deea1a6716092348cfe82.tar.bz2 | |
On Sandybridge loading unaligned 256bits using two XMM loads (vmovups and vinsertf128) is faster than using a single vmovups instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172868 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen/X86/v8i1-masks.ll')
| -rw-r--r-- | test/CodeGen/X86/v8i1-masks.ll | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/test/CodeGen/X86/v8i1-masks.ll b/test/CodeGen/X86/v8i1-masks.ll index abb4b39..ea231af 100644 --- a/test/CodeGen/X86/v8i1-masks.ll +++ b/test/CodeGen/X86/v8i1-masks.ll @@ -1,7 +1,7 @@ ; RUN: llc -march=x86-64 -mtriple=x86_64-apple-darwin -mcpu=corei7-avx -o - < %s | FileCheck %s ;CHECK: and_masks -;CHECK: vmovups +;CHECK: vmovaps ;CHECK: vcmpltp ;CHECK: vcmpltp ;CHECK: vandps |