projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: d50f33d) | patch
canonicalize a v2f64 blendi of 2 registers
authorSanjay Patel <spatel@rotateright.com>
Fri, 20 Feb 2015 16:55:27 +0000 (16:55 +0000)
committerSanjay Patel <spatel@rotateright.com>
Fri, 20 Feb 2015 16:55:27 +0000 (16:55 +0000)
commit1b53f74c4cd9f7028292cce860f56066b961991e
tree0d4b103a6b32b015a9bb31bba7a6a43dbbf57a9ctree | snapshot
parentd50f33d9966c380299779be48550a3ef9788af67commit | diff
canonicalize a v2f64 blendi of 2 registers

This canonicalization step saves us 3 pattern matching possibilities * 4 math ops
for scalar FP math that uses xmm regs. The backend can re-commute the operands
post-instruction-selection if that makes register allocation better.

The tests in llvm/test/CodeGen/X86/sse-scalar-fp-arith.ll cover this scenario already,
so there are no new tests with this patch.

Differential Revision: http://reviews.llvm.org/D7777

llvm-svn: 230024
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/lib/Target/X86/X86InstrSSE.td diff | blob | history
Domain: System / Toolchain;