switch (MI.getOpcode()) {
case X86::INSERTPSrr:
case X86::VINSERTPSrr:
+ case X86::VINSERTPSZrr:
// Attempt to convert the load of inserted vector into a fold load
// of a single float.
if (OpNum == 2) {
int PtrOffset = SrcIdx * 4;
unsigned NewImm = (DstIdx << 4) | ZMask;
unsigned NewOpCode =
- (MI.getOpcode() == X86::VINSERTPSrr ? X86::VINSERTPSrm
- : X86::INSERTPSrm);
+ (MI.getOpcode() == X86::VINSERTPSZrr) ? X86::VINSERTPSZrm :
+ (MI.getOpcode() == X86::VINSERTPSrr) ? X86::VINSERTPSrm :
+ X86::INSERTPSrm;
MachineInstr *NewMI =
FuseInst(MF, NewOpCode, OpNum, MOs, InsertPt, MI, *this, PtrOffset);
NewMI->getOperand(NewMI->getNumOperands() - 1).setImm(NewImm);
break;
case X86::MOVHLPSrr:
case X86::VMOVHLPSrr:
+ case X86::VMOVHLPSZrr:
// Move the upper 64-bits of the second operand to the lower 64-bits.
// To fold the load, adjust the pointer to the upper and use (V)MOVLPS.
// TODO: In most cases AVX doesn't have a 8-byte alignment requirement.
unsigned RCSize = getRegClass(MI.getDesc(), OpNum, &RI, MF)->getSize();
if (Size <= RCSize && 8 <= Align) {
unsigned NewOpCode =
- (MI.getOpcode() == X86::VMOVHLPSrr ? X86::VMOVLPSrm
- : X86::MOVLPSrm);
+ (MI.getOpcode() == X86::VMOVHLPSZrr) ? X86::VMOVLPSZ128rm :
+ (MI.getOpcode() == X86::VMOVHLPSrr) ? X86::VMOVLPSrm :
+ X86::MOVLPSrm;
MachineInstr *NewMI =
FuseInst(MF, NewOpCode, OpNum, MOs, InsertPt, MI, *this, 8);
return NewMI;
}
declare <4 x float> @llvm.x86.sse.div.ss(<4 x float>, <4 x float>) nounwind readnone
+define <4 x float> @stack_fold_insertps(<4 x float> %a0, <4 x float> %a1) {
+ ;CHECK-LABEL: stack_fold_insertps
+ ;CHECK: vinsertps $17, {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+ ;CHECK-NEXT: {{.*#+}} xmm0 = zero,mem[0],xmm0[2,3]
+ %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
+ %2 = call <4 x float> @llvm.x86.sse41.insertps(<4 x float> %a0, <4 x float> %a1, i8 209)
+ ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.sse41.insertps(<4 x float>, <4 x float>, i8) nounwind readnone
+
define double @stack_fold_mulsd(double %a0, double %a1) {
;CHECK-LABEL: stack_fold_mulsd
;CHECK: vmulsd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 8-byte Folded Reload