Files
clang-p2996/llvm/test/CodeGen/X86/pr35272.ll
Craig Topper 247016a735 [X86] Use vptestm/vptestnm for comparisons with zero to avoid creating a zero vector.
We can use the same input for both operands to get a free compare with zero.

We already use this trick in a couple places where we explicitly create PTESTM with the same input twice. This generalizes it.

I'm hoping to remove the ISD opcodes and move this to isel patterns like we do for scalar cmp/test.

llvm-svn: 323605
2018-01-27 20:19:09 +00:00

14 lines
508 B
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown -mcpu=skx | FileCheck %s
define <2 x i48> @PR35272(<2 x i64> %a0, <2 x i48> %a1, <2 x i48> %a2) {
; CHECK-LABEL: PR35272:
; CHECK: # %bb.0:
; CHECK-NEXT: vptestnmq %xmm0, %xmm0, %k1
; CHECK-NEXT: vpblendmq %xmm1, %xmm2, %xmm0 {%k1}
; CHECK-NEXT: retq
%1 = icmp eq <2 x i64> %a0, zeroinitializer
%2 = select <2 x i1> %1, <2 x i48> %a1, <2 x i48> %a2
ret <2 x i48> %2
}