diff options
| author | Paul Redmond <paul.redmond@intel.com> | 2013-08-19 20:01:35 +0000 |
|---|---|---|
| committer | Paul Redmond <paul.redmond@intel.com> | 2013-08-19 20:01:35 +0000 |
| commit | 62f840f46a0b714d75beb42743a2d8080c9f265f (patch) | |
| tree | 07853b1f9a9f24a9ff7be46f8a7b52fdd6831362 /llvm/test/CodeGen/X86/vsplit-and.ll | |
| parent | f2bef0b15dc7caf0864c906c707cb64449dd012b (diff) | |
| download | bcm5719-llvm-62f840f46a0b714d75beb42743a2d8080c9f265f.tar.gz bcm5719-llvm-62f840f46a0b714d75beb42743a2d8080c9f265f.zip | |
Improve the widening of integral binary vector operations
- split WidenVecRes_Binary into WidenVecRes_Binary and WidenVecRes_BinaryCanTrap
- WidenVecRes_BinaryCanTrap preserves the original behaviour for operations
that can trap
- WidenVecRes_Binary simply widens the operation and improves codegen for
3-element vectors by allowing widening and promotion on x86 (matches the
behaviour of unary and ternary operation widening)
- use WidenVecRes_Binary for operations on integers.
Reviewed by: nrotem
llvm-svn: 188699
Diffstat (limited to 'llvm/test/CodeGen/X86/vsplit-and.ll')
| -rw-r--r-- | llvm/test/CodeGen/X86/vsplit-and.ll | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/test/CodeGen/X86/vsplit-and.ll b/llvm/test/CodeGen/X86/vsplit-and.ll index 3b7fdff84e3..c16b2949358 100644 --- a/llvm/test/CodeGen/X86/vsplit-and.ll +++ b/llvm/test/CodeGen/X86/vsplit-and.ll @@ -14,7 +14,7 @@ define void @t0(<2 x i64>* %dst, <2 x i64> %src1, <2 x i64> %src2) nounwind read define void @t2(<3 x i64>* %dst, <3 x i64> %src1, <3 x i64> %src2) nounwind readonly { ; CHECK: t2 -; CHECK-NOT: pand +; CHECK: pand ; CHECK: ret %cmp1 = icmp ne <3 x i64> %src1, zeroinitializer %cmp2 = icmp ne <3 x i64> %src2, zeroinitializer |
