diff options
| author | Chris Lattner <sabre@nondot.org> | 2009-07-08 00:31:33 +0000 |
|---|---|---|
| committer | Chris Lattner <sabre@nondot.org> | 2009-07-08 00:31:33 +0000 |
| commit | 2b7a271c713ff1db83990f691126bc33d6c59b52 (patch) | |
| tree | 3ca1809851786fbd67edae8a390aa2045516e1ad /test/CodeGen/X86/vec_compare.ll | |
| parent | 8e5a084d47e872a4b4228272d0246facb839e9fb (diff) | |
| download | llvm-2b7a271c713ff1db83990f691126bc33d6c59b52.tar.gz llvm-2b7a271c713ff1db83990f691126bc33d6c59b52.tar.bz2 llvm-2b7a271c713ff1db83990f691126bc33d6c59b52.tar.xz | |
dag combine sext(setcc) -> vsetcc before legalize. To make this safe,
VSETCC must define all bits, which is different than it was documented
to before. Since all targets that implement VSETCC already have this
behavior, and we don't optimize based on this, just change the
documentation. We now get nice code for vec_compare.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74978 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen/X86/vec_compare.ll')
| -rw-r--r-- | test/CodeGen/X86/vec_compare.ll | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/test/CodeGen/X86/vec_compare.ll b/test/CodeGen/X86/vec_compare.ll index 7a8dfd68cf..74e34d0202 100644 --- a/test/CodeGen/X86/vec_compare.ll +++ b/test/CodeGen/X86/vec_compare.ll @@ -1,4 +1,4 @@ -; RUN: llvm-as < %s | llc -march=x86 -mcpu=yonah | grep pcmpgtd +; RUN: llvm-as < %s | llc -march=x86 -mcpu=yonah | grep pcmpgtd | count 2 define <4 x i32> @test(<4 x i32> %A, <4 x i32> %B) nounwind { %C = vicmp sgt <4 x i32> %A, %B |