diff options
| author | Nadav Rotem <nadav.rotem@intel.com> | 2012-04-11 06:40:27 +0000 |
|---|---|---|
| committer | Nadav Rotem <nadav.rotem@intel.com> | 2012-04-11 06:40:27 +0000 |
| commit | e611378a6e45fcb4a039d8c0089cd8fed2d311dc (patch) | |
| tree | d9891754db505d67c3d03c330ba7a2b1b7bfc666 /lib/Target/X86/X86ISelLowering.h | |
| parent | bee78fe5fcd8464f58bc729dede1a87d763ac3ae (diff) | |
| download | external_llvm-e611378a6e45fcb4a039d8c0089cd8fed2d311dc.zip external_llvm-e611378a6e45fcb4a039d8c0089cd8fed2d311dc.tar.gz external_llvm-e611378a6e45fcb4a039d8c0089cd8fed2d311dc.tar.bz2 | |
Reapply 154396 after fixing a test.
Original message:
Modify the code that lowers shuffles to blends from using blendvXX to vblendXX.
blendV uses a register for the selection while Vblend uses an immediate.
On sandybridge they still have the same latency and execute on the same execution ports.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154483 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'lib/Target/X86/X86ISelLowering.h')
| -rw-r--r-- | lib/Target/X86/X86ISelLowering.h | 7 |
1 files changed, 6 insertions, 1 deletions
diff --git a/lib/Target/X86/X86ISelLowering.h b/lib/Target/X86/X86ISelLowering.h index ca8efe6..4e00733 100644 --- a/lib/Target/X86/X86ISelLowering.h +++ b/lib/Target/X86/X86ISelLowering.h @@ -175,9 +175,14 @@ namespace llvm { /// PSIGN - Copy integer sign. PSIGN, - /// BLEND family of opcodes + /// BLENDV - Blend where the selector is an XMM. BLENDV, + /// BLENDxx - Blend where the selector is an immediate. + BLENDPW, + BLENDPS, + BLENDPD, + /// HADD - Integer horizontal add. HADD, |