[SelectionDAG] Convert to or mask if all insertions are -1

We did this for 0 and and, but we can do this with or and -1.

Author: AZero13

llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

@@ -22974,7 +22974,6 @@ SDValue DAGCombiner::visitINSERT_VECTOR_ELT(SDNode *N) {
       }
 
       // If all insertions are zero value, try to convert to AND mask.
-      // TODO: Do this for -1 with OR mask?
       if (!LegalOperations && llvm::isNullConstant(InVal) &&
           all_of(Ops, [InVal](SDValue Op) { return !Op || Op == InVal; }) &&
           count_if(Ops, [InVal](SDValue Op) { return Op == InVal; }) >= 2) {
@@ -22987,6 +22986,19 @@ SDValue DAGCombiner::visitINSERT_VECTOR_ELT(SDNode *N) {
                            DAG.getBuildVector(VT, DL, Mask));
       }
 
+      // If all insertions are -1, try to convert to OR mask.
+      if (!LegalOperations && llvm::isAllOnesConstant(InVal) &&
+          all_of(Ops, [InVal](SDValue Op) { return !Op || Op == InVal; }) &&
+          count_if(Ops, [InVal](SDValue Op) { return Op == InVal; }) >= 2) {
+        SDValue Zero = DAG.getConstant(0, DL, MaxEltVT);
+        SDValue AllOnes = DAG.getAllOnesConstant(DL, MaxEltVT);
+        SmallVector<SDValue, 8> Mask(NumElts);
+        for (unsigned I = 0; I != NumElts; ++I)
+          Mask[I] = Ops[I] ? AllOnes : Zero;
+        return DAG.getNode(ISD::OR, DL, VT, CurVec,
+                           DAG.getBuildVector(VT, DL, Mask));
+      }
+
       // Failed to find a match in the chain - bail.
       break;
     }

llvm/test/CodeGen/AArch64/vecreduce-and-legalization.ll

@@ -101,19 +101,13 @@ define i8 @test_v3i8(<3 x i8> %a) nounwind {
 define i8 @test_v9i8(<9 x i8> %a) nounwind {
 ; CHECK-LABEL: test_v9i8:
 ; CHECK:       // %bb.0:
-; CHECK-NEXT:    mov v1.16b, v0.16b
-; CHECK-NEXT:    mov w8, #-1 // =0xffffffff
-; CHECK-NEXT:    mov v1.b[9], w8
-; CHECK-NEXT:    mov v1.b[10], w8
-; CHECK-NEXT:    mov v1.b[11], w8
-; CHECK-NEXT:    mov v1.b[12], w8
-; CHECK-NEXT:    mov v1.b[13], w8
-; CHECK-NEXT:    mov v1.b[14], w8
-; CHECK-NEXT:    mov v1.b[15], w8
+; CHECK-NEXT:    movi v1.2d, #0xffffff00ffffff00
+; CHECK-NEXT:    fmov x8, d0
+; CHECK-NEXT:    orr v1.16b, v0.16b, v1.16b
 ; CHECK-NEXT:    ext v1.16b, v1.16b, v1.16b, #8
 ; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
-; CHECK-NEXT:    fmov x8, d0
-; CHECK-NEXT:    and x8, x8, x8, lsr #32
+; CHECK-NEXT:    fmov x9, d0
+; CHECK-NEXT:    and x8, x9, x8, lsr #32
 ; CHECK-NEXT:    and x8, x8, x8, lsr #16
 ; CHECK-NEXT:    lsr x9, x8, #8
 ; CHECK-NEXT:    and w0, w8, w9

llvm/test/CodeGen/X86/avx-cvt-3.ll

@@ -48,17 +48,13 @@ define <8 x float> @sitofp_shuffle_zero_v8i32(<8 x i32> %a0) {
 define <8 x float> @sitofp_insert_allbits_v8i32(<8 x i32> %a0) {
 ; X86-LABEL: sitofp_insert_allbits_v8i32:
 ; X86:       # %bb.0:
-; X86-NEXT:    vxorps %xmm1, %xmm1, %xmm1
-; X86-NEXT:    vcmptrueps %ymm1, %ymm1, %ymm1
-; X86-NEXT:    vblendps {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2],ymm0[3],ymm1[4,5],ymm0[6,7]
+; X86-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0
 ; X86-NEXT:    vcvtdq2ps %ymm0, %ymm0
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: sitofp_insert_allbits_v8i32:
 ; X64:       # %bb.0:
-; X64-NEXT:    vxorps %xmm1, %xmm1, %xmm1
-; X64-NEXT:    vcmptrueps %ymm1, %ymm1, %ymm1
-; X64-NEXT:    vblendps {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2],ymm0[3],ymm1[4,5],ymm0[6,7]
+; X64-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
 ; X64-NEXT:    vcvtdq2ps %ymm0, %ymm0
 ; X64-NEXT:    retq
   %1 = insertelement <8 x i32> %a0, i32 -1, i32 0

llvm/test/CodeGen/X86/insertelement-ones.ll

@@ -150,59 +150,32 @@ define <4 x i32> @insert_v4i32_01x3(<4 x i32> %a) {
 define <8 x i32> @insert_v8i32_x12345x7(<8 x i32> %a) {
 ; SSE2-LABEL: insert_v8i32_x12345x7:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    movss {{.*#+}} xmm2 = [NaN,0.0E+0,0.0E+0,0.0E+0]
-; SSE2-NEXT:    movss {{.*#+}} xmm0 = xmm2[0],xmm0[1,2,3]
-; SSE2-NEXT:    movl $-1, %eax
-; SSE2-NEXT:    movd %eax, %xmm2
-; SSE2-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,0],xmm1[3,0]
-; SSE2-NEXT:    shufps {{.*#+}} xmm1 = xmm1[0,1],xmm2[0,2]
+; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSE2-NEXT:    retq
 ;
 ; SSE3-LABEL: insert_v8i32_x12345x7:
 ; SSE3:       # %bb.0:
-; SSE3-NEXT:    movss {{.*#+}} xmm2 = [NaN,0.0E+0,0.0E+0,0.0E+0]
-; SSE3-NEXT:    movss {{.*#+}} xmm0 = xmm2[0],xmm0[1,2,3]
-; SSE3-NEXT:    movl $-1, %eax
-; SSE3-NEXT:    movd %eax, %xmm2
-; SSE3-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,0],xmm1[3,0]
-; SSE3-NEXT:    shufps {{.*#+}} xmm1 = xmm1[0,1],xmm2[0,2]
+; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSE3-NEXT:    retq
 ;
 ; SSSE3-LABEL: insert_v8i32_x12345x7:
 ; SSSE3:       # %bb.0:
-; SSSE3-NEXT:    movss {{.*#+}} xmm2 = [NaN,0.0E+0,0.0E+0,0.0E+0]
-; SSSE3-NEXT:    movss {{.*#+}} xmm0 = xmm2[0],xmm0[1,2,3]
-; SSSE3-NEXT:    movl $-1, %eax
-; SSSE3-NEXT:    movd %eax, %xmm2
-; SSSE3-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,0],xmm1[3,0]
-; SSSE3-NEXT:    shufps {{.*#+}} xmm1 = xmm1[0,1],xmm2[0,2]
+; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: insert_v8i32_x12345x7:
 ; SSE41:       # %bb.0:
-; SSE41-NEXT:    pcmpeqd %xmm2, %xmm2
-; SSE41-NEXT:    pblendw {{.*#+}} xmm0 = xmm2[0,1],xmm0[2,3,4,5,6,7]
-; SSE41-NEXT:    pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5],xmm1[6,7]
+; SSE41-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSE41-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSE41-NEXT:    retq
 ;
-; AVX1-LABEL: insert_v8i32_x12345x7:
-; AVX1:       # %bb.0:
-; AVX1-NEXT:    vxorps %xmm1, %xmm1, %xmm1
-; AVX1-NEXT:    vcmptrueps %ymm1, %ymm1, %ymm1
-; AVX1-NEXT:    vblendps {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3,4,5],ymm1[6],ymm0[7]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: insert_v8i32_x12345x7:
-; AVX2:       # %bb.0:
-; AVX2-NEXT:    vpcmpeqd %ymm1, %ymm1, %ymm1
-; AVX2-NEXT:    vpblendd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3,4,5],ymm1[6],ymm0[7]
-; AVX2-NEXT:    retq
-;
-; AVX512-LABEL: insert_v8i32_x12345x7:
-; AVX512:       # %bb.0:
-; AVX512-NEXT:    vpcmpeqd %ymm1, %ymm1, %ymm1
-; AVX512-NEXT:    vpblendd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3,4,5],ymm1[6],ymm0[7]
-; AVX512-NEXT:    retq
+; AVX-LABEL: insert_v8i32_x12345x7:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
+; AVX-NEXT:    retq
   %1 = insertelement <8 x i32> %a, i32 -1, i32 0
   %2 = insertelement <8 x i32> %1, i32 -1, i32 6
   ret <8 x i32> %2
@@ -211,35 +184,27 @@ define <8 x i32> @insert_v8i32_x12345x7(<8 x i32> %a) {
 define <8 x i16> @insert_v8i16_x12345x7(<8 x i16> %a) {
 ; SSE2-LABEL: insert_v8i16_x12345x7:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    movl $65535, %eax # imm = 0xFFFF
-; SSE2-NEXT:    pinsrw $0, %eax, %xmm0
-; SSE2-NEXT:    pinsrw $6, %eax, %xmm0
+; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; SSE2-NEXT:    retq
 ;
 ; SSE3-LABEL: insert_v8i16_x12345x7:
 ; SSE3:       # %bb.0:
-; SSE3-NEXT:    movl $65535, %eax # imm = 0xFFFF
-; SSE3-NEXT:    pinsrw $0, %eax, %xmm0
-; SSE3-NEXT:    pinsrw $6, %eax, %xmm0
+; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; SSE3-NEXT:    retq
 ;
 ; SSSE3-LABEL: insert_v8i16_x12345x7:
 ; SSSE3:       # %bb.0:
-; SSSE3-NEXT:    movl $65535, %eax # imm = 0xFFFF
-; SSSE3-NEXT:    pinsrw $0, %eax, %xmm0
-; SSSE3-NEXT:    pinsrw $6, %eax, %xmm0
+; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: insert_v8i16_x12345x7:
 ; SSE41:       # %bb.0:
-; SSE41-NEXT:    pcmpeqd %xmm1, %xmm1
-; SSE41-NEXT:    pblendw {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3,4,5],xmm1[6],xmm0[7]
+; SSE41-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; SSE41-NEXT:    retq
 ;
 ; AVX-LABEL: insert_v8i16_x12345x7:
 ; AVX:       # %bb.0:
-; AVX-NEXT:    vpcmpeqd %xmm1, %xmm1, %xmm1
-; AVX-NEXT:    vpblendw {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3,4,5],xmm1[6],xmm0[7]
+; AVX-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
 ; AVX-NEXT:    retq
   %1 = insertelement <8 x i16> %a, i16 -1, i32 0
   %2 = insertelement <8 x i16> %1, i16 -1, i32 6
@@ -249,60 +214,32 @@ define <8 x i16> @insert_v8i16_x12345x7(<8 x i16> %a) {
 define <16 x i16> @insert_v16i16_x12345x789ABCDEx(<16 x i16> %a) {
 ; SSE2-LABEL: insert_v16i16_x12345x789ABCDEx:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    movl $65535, %eax # imm = 0xFFFF
-; SSE2-NEXT:    pinsrw $0, %eax, %xmm0
-; SSE2-NEXT:    pinsrw $6, %eax, %xmm0
-; SSE2-NEXT:    pinsrw $7, %eax, %xmm1
+; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSE2-NEXT:    retq
 ;
 ; SSE3-LABEL: insert_v16i16_x12345x789ABCDEx:
 ; SSE3:       # %bb.0:
-; SSE3-NEXT:    movl $65535, %eax # imm = 0xFFFF
-; SSE3-NEXT:    pinsrw $0, %eax, %xmm0
-; SSE3-NEXT:    pinsrw $6, %eax, %xmm0
-; SSE3-NEXT:    pinsrw $7, %eax, %xmm1
+; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSE3-NEXT:    retq
 ;
 ; SSSE3-LABEL: insert_v16i16_x12345x789ABCDEx:
 ; SSSE3:       # %bb.0:
-; SSSE3-NEXT:    movl $65535, %eax # imm = 0xFFFF
-; SSSE3-NEXT:    pinsrw $0, %eax, %xmm0
-; SSSE3-NEXT:    pinsrw $6, %eax, %xmm0
-; SSSE3-NEXT:    pinsrw $7, %eax, %xmm1
+; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: insert_v16i16_x12345x789ABCDEx:
 ; SSE41:       # %bb.0:
-; SSE41-NEXT:    pcmpeqd %xmm2, %xmm2
-; SSE41-NEXT:    pblendw {{.*#+}} xmm0 = xmm2[0],xmm0[1,2,3,4,5],xmm2[6],xmm0[7]
-; SSE41-NEXT:    pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,5,6],xmm2[7]
+; SSE41-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSE41-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSE41-NEXT:    retq
 ;
-; AVX1-LABEL: insert_v16i16_x12345x789ABCDEx:
-; AVX1:       # %bb.0:
-; AVX1-NEXT:    vmovss {{.*#+}} xmm1 = [65535,0,0,0]
-; AVX1-NEXT:    vorps %ymm1, %ymm0, %ymm0
-; AVX1-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
-; AVX1-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: insert_v16i16_x12345x789ABCDEx:
-; AVX2:       # %bb.0:
-; AVX2-NEXT:    vpcmpeqd %xmm1, %xmm1, %xmm1
-; AVX2-NEXT:    vpblendw {{.*#+}} xmm1 = xmm1[0],xmm0[1,2,3,4,5],xmm1[6],xmm0[7]
-; AVX2-NEXT:    vpcmpeqd %ymm2, %ymm2, %ymm2
-; AVX2-NEXT:    vpblendw {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5,6],ymm2[7],ymm0[8,9,10,11,12,13,14],ymm2[15]
-; AVX2-NEXT:    vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
-; AVX2-NEXT:    retq
-;
-; AVX512-LABEL: insert_v16i16_x12345x789ABCDEx:
-; AVX512:       # %bb.0:
-; AVX512-NEXT:    vpcmpeqd %xmm1, %xmm1, %xmm1
-; AVX512-NEXT:    vpblendw {{.*#+}} xmm1 = xmm1[0],xmm0[1,2,3,4,5],xmm1[6],xmm0[7]
-; AVX512-NEXT:    vpcmpeqd %ymm2, %ymm2, %ymm2
-; AVX512-NEXT:    vpblendw {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5,6],ymm2[7],ymm0[8,9,10,11,12,13,14],ymm2[15]
-; AVX512-NEXT:    vpblendd {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
-; AVX512-NEXT:    retq
+; AVX-LABEL: insert_v16i16_x12345x789ABCDEx:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
+; AVX-NEXT:    retq
   %1 = insertelement <16 x i16> %a, i16 -1, i32 0
   %2 = insertelement <16 x i16> %1, i16 -1, i32 6
   %3 = insertelement <16 x i16> %2, i16 -1, i32 15
@@ -313,33 +250,26 @@ define <16 x i8> @insert_v16i8_x123456789ABCDEx(<16 x i8> %a) {
 ; SSE2-LABEL: insert_v16i8_x123456789ABCDEx:
 ; SSE2:       # %bb.0:
 ; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
-; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; SSE2-NEXT:    retq
 ;
 ; SSE3-LABEL: insert_v16i8_x123456789ABCDEx:
 ; SSE3:       # %bb.0:
 ; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
-; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; SSE3-NEXT:    retq
 ;
 ; SSSE3-LABEL: insert_v16i8_x123456789ABCDEx:
 ; SSSE3:       # %bb.0:
 ; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
-; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: insert_v16i8_x123456789ABCDEx:
 ; SSE41:       # %bb.0:
-; SSE41-NEXT:    movl $255, %eax
-; SSE41-NEXT:    pinsrb $0, %eax, %xmm0
-; SSE41-NEXT:    pinsrb $15, %eax, %xmm0
+; SSE41-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; SSE41-NEXT:    retq
 ;
 ; AVX-LABEL: insert_v16i8_x123456789ABCDEx:
 ; AVX:       # %bb.0:
-; AVX-NEXT:    movl $255, %eax
-; AVX-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $15, %eax, %xmm0, %xmm0
+; AVX-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
 ; AVX-NEXT:    retq
   %1 = insertelement <16 x i8> %a, i8 -1, i32 0
   %2 = insertelement <16 x i8> %1, i8 -1, i32 15
@@ -350,69 +280,31 @@ define <32 x i8> @insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx(<32 x i8> %a) {
 ; SSE2-LABEL: insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx:
 ; SSE2:       # %bb.0:
 ; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
-; SSE2-NEXT:    movaps {{.*#+}} xmm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255]
-; SSE2-NEXT:    orps %xmm2, %xmm0
 ; SSE2-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
-; SSE2-NEXT:    orps %xmm2, %xmm1
 ; SSE2-NEXT:    retq
 ;
 ; SSE3-LABEL: insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx:
 ; SSE3:       # %bb.0:
 ; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
-; SSE3-NEXT:    movaps {{.*#+}} xmm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255]
-; SSE3-NEXT:    orps %xmm2, %xmm0
 ; SSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
-; SSE3-NEXT:    orps %xmm2, %xmm1
 ; SSE3-NEXT:    retq
 ;
 ; SSSE3-LABEL: insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx:
 ; SSSE3:       # %bb.0:
 ; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
-; SSSE3-NEXT:    movaps {{.*#+}} xmm2 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255]
-; SSSE3-NEXT:    orps %xmm2, %xmm0
 ; SSSE3-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
-; SSSE3-NEXT:    orps %xmm2, %xmm1
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx:
 ; SSE41:       # %bb.0:
-; SSE41-NEXT:    movl $255, %eax
-; SSE41-NEXT:    pinsrb $0, %eax, %xmm0
-; SSE41-NEXT:    pinsrb $15, %eax, %xmm0
-; SSE41-NEXT:    pinsrb $14, %eax, %xmm1
-; SSE41-NEXT:    pinsrb $15, %eax, %xmm1
+; SSE41-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; SSE41-NEXT:    orps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; SSE41-NEXT:    retq
 ;
-; AVX1-LABEL: insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx:
-; AVX1:       # %bb.0:
-; AVX1-NEXT:    vmovss {{.*#+}} xmm1 = [255,0,0,0]
-; AVX1-NEXT:    vorps %ymm1, %ymm0, %ymm0
-; AVX1-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
-; AVX1-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
-; AVX1-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx:
-; AVX2:       # %bb.0:
-; AVX2-NEXT:    movl $255, %eax
-; AVX2-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm1
-; AVX2-NEXT:    vpinsrb $15, %eax, %xmm1, %xmm1
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm0
-; AVX2-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0
-; AVX2-NEXT:    vpinsrb $15, %eax, %xmm0, %xmm0
-; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
-; AVX2-NEXT:    retq
-;
-; AVX512-LABEL: insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx:
-; AVX512:       # %bb.0:
-; AVX512-NEXT:    movl $255, %eax
-; AVX512-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm1
-; AVX512-NEXT:    vpinsrb $15, %eax, %xmm1, %xmm1
-; AVX512-NEXT:    vextracti128 $1, %ymm0, %xmm0
-; AVX512-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0
-; AVX512-NEXT:    vpinsrb $15, %eax, %xmm0, %xmm0
-; AVX512-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
-; AVX512-NEXT:    retq
+; AVX-LABEL: insert_v32i8_x123456789ABCDEzGHIJKLMNOPQRSTxx:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
+; AVX-NEXT:    retq
   %1 = insertelement <32 x i8> %a, i8 -1, i32 0
   %2 = insertelement <32 x i8> %1, i8 -1, i32 15
   %3 = insertelement <32 x i8> %2, i8 -1, i32 30