When we use the BLEND instruction that uses the MSB as a mask, we can remove
[oota-llvm.git] / test / CodeGen / X86 / avx2-arith.ll
2011-11-09 Craig TopperAdd AVX2 instruction lowering for add, sub, and mul.