[llvm] r347300 - [X86][SSE] Lower immediately to PACKUS instead of VECTOR_SHUFFLE.

Simon Pilgrim via llvm-commits llvm-commits at lists.llvm.org
Tue Nov 20 03:46:37 PST 2018


Author: rksimon
Date: Tue Nov 20 03:46:37 2018
New Revision: 347300

URL: http://llvm.org/viewvc/llvm-project?rev=347300&view=rev
Log:
[X86][SSE] Lower immediately to PACKUS instead of VECTOR_SHUFFLE.

As discussed on rL347240, this avoids some regressions on D54679 and also helps some combines to kick in a bit earlier.

Modified:
    llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
    llvm/trunk/test/CodeGen/X86/vector-mul.ll
    llvm/trunk/test/CodeGen/X86/vector-reduce-mul.ll

Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=347300&r1=347299&r2=347300&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Tue Nov 20 03:46:37 2018
@@ -23413,7 +23413,7 @@ static SDValue LowerMUL(SDValue Op, cons
 
     MVT ExVT = MVT::getVectorVT(MVT::i16, NumElts / 2);
 
-    // Extract the lo parts to any extend to i16
+    // Extract the lo parts to any extend to i16.
     // We're going to mask off the low byte of each result element of the
     // pmullw, so it doesn't matter what's in the high byte of each 16-bit
     // element.
@@ -23423,7 +23423,7 @@ static SDValue LowerMUL(SDValue Op, cons
     ALo = DAG.getBitcast(ExVT, ALo);
     BLo = DAG.getBitcast(ExVT, BLo);
 
-    // Extract the hi parts to any extend to i16
+    // Extract the hi parts to any extend to i16.
     // We're going to mask off the low byte of each result element of the
     // pmullw, so it doesn't matter what's in the high byte of each 16-bit
     // element.
@@ -23432,20 +23432,12 @@ static SDValue LowerMUL(SDValue Op, cons
     AHi = DAG.getBitcast(ExVT, AHi);
     BHi = DAG.getBitcast(ExVT, BHi);
 
-    // Multiply, mask the lower 8bits of the lo/hi results and pack
+    // Multiply, mask the lower 8bits of the lo/hi results and pack.
     SDValue RLo = DAG.getNode(ISD::MUL, dl, ExVT, ALo, BLo);
     SDValue RHi = DAG.getNode(ISD::MUL, dl, ExVT, AHi, BHi);
     RLo = DAG.getNode(ISD::AND, dl, ExVT, RLo, DAG.getConstant(255, dl, ExVT));
     RHi = DAG.getNode(ISD::AND, dl, ExVT, RHi, DAG.getConstant(255, dl, ExVT));
-    RLo = DAG.getBitcast(VT, RLo);
-    RHi = DAG.getBitcast(VT, RHi);
-
-    // For each 128-bit lane, we need to take the 8 even elements from RLo then
-    // the 8 even elements from RHi.
-    SmallVector<int, 64> PackMask;
-    createPackShuffleMask(VT, PackMask, /*Unary*/false);
-
-    return DAG.getVectorShuffle(VT, dl, RLo, RHi, PackMask);
+    return DAG.getNode(X86ISD::PACKUS, dl, VT, RLo, RHi);
   }
 
   // Lower v4i32 mul as 2x shuffle, 2x pmuludq, 2x shuffle.

Modified: llvm/trunk/test/CodeGen/X86/vector-mul.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/vector-mul.ll?rev=347300&r1=347299&r2=347300&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/vector-mul.ll (original)
+++ llvm/trunk/test/CodeGen/X86/vector-mul.ll Tue Nov 20 03:46:37 2018
@@ -564,12 +564,9 @@ define <16 x i8> @mul_v16i8_2_3_9_17_33_
 ; X64-XOP-NEXT:    vpunpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
 ; X64-XOP-NEXT:    vpunpckhbw {{.*#+}} xmm2 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
 ; X64-XOP-NEXT:    vpmullw %xmm1, %xmm2, %xmm1
-; X64-XOP-NEXT:    vmovdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
-; X64-XOP-NEXT:    vpand %xmm2, %xmm1, %xmm1
 ; X64-XOP-NEXT:    vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
 ; X64-XOP-NEXT:    vpmullw {{.*}}(%rip), %xmm0, %xmm0
-; X64-XOP-NEXT:    vpand %xmm2, %xmm0, %xmm0
-; X64-XOP-NEXT:    vpackuswb %xmm1, %xmm0, %xmm0
+; X64-XOP-NEXT:    vpperm {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14],xmm1[0,2,4,6,8,10,12,14]
 ; X64-XOP-NEXT:    retq
 ;
 ; X64-AVX2-LABEL: mul_v16i8_2_3_9_17_33_65_129_2_3_9_17_33_65_129_2_3:
@@ -1146,12 +1143,9 @@ define <16 x i8> @mul_v16i8_0_1_3_7_15_3
 ; X64-XOP-NEXT:    vpunpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
 ; X64-XOP-NEXT:    vpunpckhbw {{.*#+}} xmm2 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
 ; X64-XOP-NEXT:    vpmullw %xmm1, %xmm2, %xmm1
-; X64-XOP-NEXT:    vmovdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
-; X64-XOP-NEXT:    vpand %xmm2, %xmm1, %xmm1
 ; X64-XOP-NEXT:    vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
 ; X64-XOP-NEXT:    vpmullw {{.*}}(%rip), %xmm0, %xmm0
-; X64-XOP-NEXT:    vpand %xmm2, %xmm0, %xmm0
-; X64-XOP-NEXT:    vpackuswb %xmm1, %xmm0, %xmm0
+; X64-XOP-NEXT:    vpperm {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14],xmm1[0,2,4,6,8,10,12,14]
 ; X64-XOP-NEXT:    retq
 ;
 ; X64-AVX2-LABEL: mul_v16i8_0_1_3_7_15_31_63_127_0_1_3_7_15_31_63_127:

Modified: llvm/trunk/test/CodeGen/X86/vector-reduce-mul.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/vector-reduce-mul.ll?rev=347300&r1=347299&r2=347300&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/vector-reduce-mul.ll (original)
+++ llvm/trunk/test/CodeGen/X86/vector-reduce-mul.ll Tue Nov 20 03:46:37 2018
@@ -1811,12 +1811,10 @@ define i8 @test_v32i8(<32 x i8> %a0) {
 ; AVX1-NEXT:    vpmullw %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vpand %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpackuswb %xmm3, %xmm0, %xmm0
-; AVX1-NEXT:    vpsrlw $8, %xmm0, %xmm2
+; AVX1-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX1-NEXT:    vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; AVX1-NEXT:    vpmovzxbw {{.*#+}} xmm2 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
-; AVX1-NEXT:    vpmullw %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpand %xmm1, %xmm0, %xmm0
-; AVX1-NEXT:    vpackuswb %xmm0, %xmm0, %xmm0
+; AVX1-NEXT:    vpmovzxbw {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
+; AVX1-NEXT:    vpmullw %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX1-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX1-NEXT:    vzeroupper
@@ -1824,40 +1822,38 @@ define i8 @test_v32i8(<32 x i8> %a0) {
 ;
 ; AVX2-LABEL: test_v32i8:
 ; AVX2:       # %bb.0:
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm2[8],ymm0[8],ymm2[9],ymm0[9],ymm2[10],ymm0[10],ymm2[11],ymm0[11],ymm2[12],ymm0[12],ymm2[13],ymm0[13],ymm2[14],ymm0[14],ymm2[15],ymm0[15],ymm2[24],ymm0[24],ymm2[25],ymm0[25],ymm2[26],ymm0[26],ymm2[27],ymm0[27],ymm2[28],ymm0[28],ymm2[29],ymm0[29],ymm2[30],ymm0[30],ymm2[31],ymm0[31]
-; AVX2-NEXT:    vpmullw %ymm3, %ymm1, %ymm3
-; AVX2-NEXT:    vmovdqa {{.*#+}} ymm1 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX2-NEXT:    vpand %ymm1, %ymm3, %ymm3
+; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
+; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm1[8],ymm0[8],ymm1[9],ymm0[9],ymm1[10],ymm0[10],ymm1[11],ymm0[11],ymm1[12],ymm0[12],ymm1[13],ymm0[13],ymm1[14],ymm0[14],ymm1[15],ymm0[15],ymm1[24],ymm0[24],ymm1[25],ymm0[25],ymm1[26],ymm0[26],ymm1[27],ymm0[27],ymm1[28],ymm0[28],ymm1[29],ymm0[29],ymm1[30],ymm0[30],ymm1[31],ymm0[31]
+; AVX2-NEXT:    vpmullw %ymm3, %ymm2, %ymm2
+; AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; AVX2-NEXT:    vpand %ymm3, %ymm2, %ymm2
 ; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm2[0],ymm0[0],ymm2[1],ymm0[1],ymm2[2],ymm0[2],ymm2[3],ymm0[3],ymm2[4],ymm0[4],ymm2[5],ymm0[5],ymm2[6],ymm0[6],ymm2[7],ymm0[7],ymm2[16],ymm0[16],ymm2[17],ymm0[17],ymm2[18],ymm0[18],ymm2[19],ymm0[19],ymm2[20],ymm0[20],ymm2[21],ymm0[21],ymm2[22],ymm0[22],ymm2[23],ymm0[23]
-; AVX2-NEXT:    vpmullw %ymm2, %ymm0, %ymm0
-; AVX2-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpackuswb %ymm3, %ymm0, %ymm0
-; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX2-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX2-NEXT:    vpmullw %ymm0, %ymm2, %ymm0
-; AVX2-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0],ymm0[0],ymm1[1],ymm0[1],ymm1[2],ymm0[2],ymm1[3],ymm0[3],ymm1[4],ymm0[4],ymm1[5],ymm0[5],ymm1[6],ymm0[6],ymm1[7],ymm0[7],ymm1[16],ymm0[16],ymm1[17],ymm0[17],ymm1[18],ymm0[18],ymm1[19],ymm0[19],ymm1[20],ymm0[20],ymm1[21],ymm0[21],ymm1[22],ymm0[22],ymm1[23],ymm0[23]
+; AVX2-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpand %ymm3, %ymm0, %ymm0
 ; AVX2-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
-; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm3 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX2-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX2-NEXT:    vpmullw %ymm0, %ymm1, %ymm0
+; AVX2-NEXT:    vpand %ymm3, %ymm0, %ymm0
+; AVX2-NEXT:    vpxor %xmm1, %xmm1, %xmm1
+; AVX2-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX2-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX2-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
-; AVX2-NEXT:    vpmullw %ymm0, %ymm3, %ymm0
-; AVX2-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
-; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm3 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX2-NEXT:    vpmullw %ymm0, %ymm2, %ymm0
+; AVX2-NEXT:    vpand %ymm3, %ymm0, %ymm0
+; AVX2-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX2-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX2-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[1,1,2,3]
-; AVX2-NEXT:    vpmullw %ymm0, %ymm3, %ymm0
-; AVX2-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
-; AVX2-NEXT:    vpsrlw $8, %xmm0, %xmm2
+; AVX2-NEXT:    vpmullw %ymm0, %ymm2, %ymm0
+; AVX2-NEXT:    vpand %ymm3, %ymm0, %ymm0
+; AVX2-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX2-NEXT:    vpmullw %ymm2, %ymm0, %ymm0
-; AVX2-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
+; AVX2-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX2-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX2-NEXT:    vzeroupper
@@ -1921,40 +1917,38 @@ define i8 @test_v32i8(<32 x i8> %a0) {
 ;
 ; AVX512DQ-LABEL: test_v32i8:
 ; AVX512DQ:       # %bb.0:
-; AVX512DQ-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX512DQ-NEXT:    vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512DQ-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm2[8],ymm0[8],ymm2[9],ymm0[9],ymm2[10],ymm0[10],ymm2[11],ymm0[11],ymm2[12],ymm0[12],ymm2[13],ymm0[13],ymm2[14],ymm0[14],ymm2[15],ymm0[15],ymm2[24],ymm0[24],ymm2[25],ymm0[25],ymm2[26],ymm0[26],ymm2[27],ymm0[27],ymm2[28],ymm0[28],ymm2[29],ymm0[29],ymm2[30],ymm0[30],ymm2[31],ymm0[31]
-; AVX512DQ-NEXT:    vpmullw %ymm3, %ymm1, %ymm3
-; AVX512DQ-NEXT:    vmovdqa {{.*#+}} ymm1 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX512DQ-NEXT:    vpand %ymm1, %ymm3, %ymm3
+; AVX512DQ-NEXT:    vextracti128 $1, %ymm0, %xmm1
+; AVX512DQ-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512DQ-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm1[8],ymm0[8],ymm1[9],ymm0[9],ymm1[10],ymm0[10],ymm1[11],ymm0[11],ymm1[12],ymm0[12],ymm1[13],ymm0[13],ymm1[14],ymm0[14],ymm1[15],ymm0[15],ymm1[24],ymm0[24],ymm1[25],ymm0[25],ymm1[26],ymm0[26],ymm1[27],ymm0[27],ymm1[28],ymm0[28],ymm1[29],ymm0[29],ymm1[30],ymm0[30],ymm1[31],ymm0[31]
+; AVX512DQ-NEXT:    vpmullw %ymm3, %ymm2, %ymm2
+; AVX512DQ-NEXT:    vmovdqa {{.*#+}} ymm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; AVX512DQ-NEXT:    vpand %ymm3, %ymm2, %ymm2
 ; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm2[0],ymm0[0],ymm2[1],ymm0[1],ymm2[2],ymm0[2],ymm2[3],ymm0[3],ymm2[4],ymm0[4],ymm2[5],ymm0[5],ymm2[6],ymm0[6],ymm2[7],ymm0[7],ymm2[16],ymm0[16],ymm2[17],ymm0[17],ymm2[18],ymm0[18],ymm2[19],ymm0[19],ymm2[20],ymm0[20],ymm2[21],ymm0[21],ymm2[22],ymm0[22],ymm2[23],ymm0[23]
-; AVX512DQ-NEXT:    vpmullw %ymm2, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpackuswb %ymm3, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512DQ-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX512DQ-NEXT:    vpmullw %ymm0, %ymm2, %ymm0
-; AVX512DQ-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0],ymm0[0],ymm1[1],ymm0[1],ymm1[2],ymm0[2],ymm1[3],ymm0[3],ymm1[4],ymm0[4],ymm1[5],ymm0[5],ymm1[6],ymm0[6],ymm1[7],ymm0[7],ymm1[16],ymm0[16],ymm1[17],ymm0[17],ymm1[18],ymm0[18],ymm1[19],ymm0[19],ymm1[20],ymm0[20],ymm1[21],ymm0[21],ymm1[22],ymm0[22],ymm1[23],ymm0[23]
+; AVX512DQ-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
+; AVX512DQ-NEXT:    vpand %ymm3, %ymm0, %ymm0
 ; AVX512DQ-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm3 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512DQ-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX512DQ-NEXT:    vpmullw %ymm0, %ymm1, %ymm0
+; AVX512DQ-NEXT:    vpand %ymm3, %ymm0, %ymm0
+; AVX512DQ-NEXT:    vpxor %xmm1, %xmm1, %xmm1
+; AVX512DQ-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
+; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX512DQ-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
-; AVX512DQ-NEXT:    vpmullw %ymm0, %ymm3, %ymm0
-; AVX512DQ-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm3 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512DQ-NEXT:    vpmullw %ymm0, %ymm2, %ymm0
+; AVX512DQ-NEXT:    vpand %ymm3, %ymm0, %ymm0
+; AVX512DQ-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
+; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX512DQ-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[1,1,2,3]
-; AVX512DQ-NEXT:    vpmullw %ymm0, %ymm3, %ymm0
-; AVX512DQ-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpsrlw $8, %xmm0, %xmm2
+; AVX512DQ-NEXT:    vpmullw %ymm0, %ymm2, %ymm0
+; AVX512DQ-NEXT:    vpand %ymm3, %ymm0, %ymm0
+; AVX512DQ-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
+; AVX512DQ-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512DQ-NEXT:    vpmullw %ymm2, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpand %ymm1, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
+; AVX512DQ-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
 ; AVX512DQ-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512DQ-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512DQ-NEXT:    vzeroupper
@@ -1994,8 +1988,6 @@ define i8 @test_v32i8(<32 x i8> %a0) {
 ; AVX512DQVL-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm3, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512DQVL-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512DQVL-NEXT:    vzeroupper
@@ -2190,8 +2182,6 @@ define i8 @test_v64i8(<64 x i8> %a0) {
 ; AVX1-NEXT:    vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
 ; AVX1-NEXT:    vpmovzxbw {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
 ; AVX1-NEXT:    vpmullw %xmm1, %xmm0, %xmm0
-; AVX1-NEXT:    vpand %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpackuswb %xmm0, %xmm0, %xmm0
 ; AVX1-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX1-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX1-NEXT:    vzeroupper
@@ -2240,8 +2230,6 @@ define i8 @test_v64i8(<64 x i8> %a0) {
 ; AVX2-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX2-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpand %ymm2, %ymm0, %ymm0
-; AVX2-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
 ; AVX2-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX2-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX2-NEXT:    vzeroupper
@@ -2291,8 +2279,6 @@ define i8 @test_v64i8(<64 x i8> %a0) {
 ; AVX512BW-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BW-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpackuswb %zmm0, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512BW-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512BW-NEXT:    vzeroupper
@@ -2342,8 +2328,6 @@ define i8 @test_v64i8(<64 x i8> %a0) {
 ; AVX512BWVL-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BWVL-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpackuswb %zmm0, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512BWVL-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512BWVL-NEXT:    vzeroupper
@@ -2392,8 +2376,6 @@ define i8 @test_v64i8(<64 x i8> %a0) {
 ; AVX512DQ-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQ-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpand %ymm2, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
 ; AVX512DQ-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512DQ-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512DQ-NEXT:    vzeroupper
@@ -2442,8 +2424,6 @@ define i8 @test_v64i8(<64 x i8> %a0) {
 ; AVX512DQVL-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm3, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512DQVL-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512DQVL-NEXT:    vzeroupper
@@ -2756,8 +2736,6 @@ define i8 @test_v128i8(<128 x i8> %a0) {
 ; AVX1-NEXT:    vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
 ; AVX1-NEXT:    vpmovzxbw {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
 ; AVX1-NEXT:    vpmullw %xmm1, %xmm0, %xmm0
-; AVX1-NEXT:    vpand %xmm4, %xmm0, %xmm0
-; AVX1-NEXT:    vpackuswb %xmm0, %xmm0, %xmm0
 ; AVX1-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX1-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX1-NEXT:    vzeroupper
@@ -2824,8 +2802,6 @@ define i8 @test_v128i8(<128 x i8> %a0) {
 ; AVX2-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX2-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpand %ymm4, %ymm0, %ymm0
-; AVX2-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
 ; AVX2-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX2-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX2-NEXT:    vzeroupper
@@ -2835,57 +2811,55 @@ define i8 @test_v128i8(<128 x i8> %a0) {
 ; AVX512BW:       # %bb.0:
 ; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm2 = zmm1[8],zmm0[8],zmm1[9],zmm0[9],zmm1[10],zmm0[10],zmm1[11],zmm0[11],zmm1[12],zmm0[12],zmm1[13],zmm0[13],zmm1[14],zmm0[14],zmm1[15],zmm0[15],zmm1[24],zmm0[24],zmm1[25],zmm0[25],zmm1[26],zmm0[26],zmm1[27],zmm0[27],zmm1[28],zmm0[28],zmm1[29],zmm0[29],zmm1[30],zmm0[30],zmm1[31],zmm0[31],zmm1[40],zmm0[40],zmm1[41],zmm0[41],zmm1[42],zmm0[42],zmm1[43],zmm0[43],zmm1[44],zmm0[44],zmm1[45],zmm0[45],zmm1[46],zmm0[46],zmm1[47],zmm0[47],zmm1[56],zmm0[56],zmm1[57],zmm0[57],zmm1[58],zmm0[58],zmm1[59],zmm0[59],zmm1[60],zmm0[60],zmm1[61],zmm0[61],zmm1[62],zmm0[62],zmm1[63],zmm0[63]
 ; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm3 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
-; AVX512BW-NEXT:    vpmullw %zmm2, %zmm3, %zmm3
-; AVX512BW-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm3, %zmm3
+; AVX512BW-NEXT:    vpmullw %zmm2, %zmm3, %zmm2
+; AVX512BW-NEXT:    vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm2, %zmm2
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm1[0],zmm0[0],zmm1[1],zmm0[1],zmm1[2],zmm0[2],zmm1[3],zmm0[3],zmm1[4],zmm0[4],zmm1[5],zmm0[5],zmm1[6],zmm0[6],zmm1[7],zmm0[7],zmm1[16],zmm0[16],zmm1[17],zmm0[17],zmm1[18],zmm0[18],zmm1[19],zmm0[19],zmm1[20],zmm0[20],zmm1[21],zmm0[21],zmm1[22],zmm0[22],zmm1[23],zmm0[23],zmm1[32],zmm0[32],zmm1[33],zmm0[33],zmm1[34],zmm0[34],zmm1[35],zmm0[35],zmm1[36],zmm0[36],zmm1[37],zmm0[37],zmm1[38],zmm0[38],zmm1[39],zmm0[39],zmm1[48],zmm0[48],zmm1[49],zmm0[49],zmm1[50],zmm0[50],zmm1[51],zmm0[51],zmm1[52],zmm0[52],zmm1[53],zmm0[53],zmm1[54],zmm0[54],zmm1[55],zmm0[55]
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BW-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpackuswb %zmm3, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpackuswb %zmm2, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vextracti64x4 $1, %zmm0, %ymm1
-; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm3 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
 ; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm4 = zmm1[8],zmm0[8],zmm1[9],zmm0[9],zmm1[10],zmm0[10],zmm1[11],zmm0[11],zmm1[12],zmm0[12],zmm1[13],zmm0[13],zmm1[14],zmm0[14],zmm1[15],zmm0[15],zmm1[24],zmm0[24],zmm1[25],zmm0[25],zmm1[26],zmm0[26],zmm1[27],zmm0[27],zmm1[28],zmm0[28],zmm1[29],zmm0[29],zmm1[30],zmm0[30],zmm1[31],zmm0[31],zmm1[40],zmm0[40],zmm1[41],zmm0[41],zmm1[42],zmm0[42],zmm1[43],zmm0[43],zmm1[44],zmm0[44],zmm1[45],zmm0[45],zmm1[46],zmm0[46],zmm1[47],zmm0[47],zmm1[56],zmm0[56],zmm1[57],zmm0[57],zmm1[58],zmm0[58],zmm1[59],zmm0[59],zmm1[60],zmm0[60],zmm1[61],zmm0[61],zmm1[62],zmm0[62],zmm1[63],zmm0[63]
-; AVX512BW-NEXT:    vpmullw %zmm4, %zmm3, %zmm3
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm3, %zmm3
+; AVX512BW-NEXT:    vpmullw %zmm4, %zmm2, %zmm2
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm2, %zmm2
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm1[0],zmm0[0],zmm1[1],zmm0[1],zmm1[2],zmm0[2],zmm1[3],zmm0[3],zmm1[4],zmm0[4],zmm1[5],zmm0[5],zmm1[6],zmm0[6],zmm1[7],zmm0[7],zmm1[16],zmm0[16],zmm1[17],zmm0[17],zmm1[18],zmm0[18],zmm1[19],zmm0[19],zmm1[20],zmm0[20],zmm1[21],zmm0[21],zmm1[22],zmm0[22],zmm1[23],zmm0[23],zmm1[32],zmm0[32],zmm1[33],zmm0[33],zmm1[34],zmm0[34],zmm1[35],zmm0[35],zmm1[36],zmm0[36],zmm1[37],zmm0[37],zmm1[38],zmm0[38],zmm1[39],zmm0[39],zmm1[48],zmm0[48],zmm1[49],zmm0[49],zmm1[50],zmm0[50],zmm1[51],zmm0[51],zmm1[52],zmm0[52],zmm1[53],zmm0[53],zmm1[54],zmm0[54],zmm1[55],zmm0[55]
 ; AVX512BW-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpackuswb %zmm3, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpackuswb %zmm2, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vextracti128 $1, %ymm0, %xmm1
-; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm3 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
 ; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm4 = zmm1[8],zmm0[8],zmm1[9],zmm0[9],zmm1[10],zmm0[10],zmm1[11],zmm0[11],zmm1[12],zmm0[12],zmm1[13],zmm0[13],zmm1[14],zmm0[14],zmm1[15],zmm0[15],zmm1[24],zmm0[24],zmm1[25],zmm0[25],zmm1[26],zmm0[26],zmm1[27],zmm0[27],zmm1[28],zmm0[28],zmm1[29],zmm0[29],zmm1[30],zmm0[30],zmm1[31],zmm0[31],zmm1[40],zmm0[40],zmm1[41],zmm0[41],zmm1[42],zmm0[42],zmm1[43],zmm0[43],zmm1[44],zmm0[44],zmm1[45],zmm0[45],zmm1[46],zmm0[46],zmm1[47],zmm0[47],zmm1[56],zmm0[56],zmm1[57],zmm0[57],zmm1[58],zmm0[58],zmm1[59],zmm0[59],zmm1[60],zmm0[60],zmm1[61],zmm0[61],zmm1[62],zmm0[62],zmm1[63],zmm0[63]
-; AVX512BW-NEXT:    vpmullw %zmm4, %zmm3, %zmm3
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm3, %zmm3
+; AVX512BW-NEXT:    vpmullw %zmm4, %zmm2, %zmm2
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm2, %zmm2
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm1[0],zmm0[0],zmm1[1],zmm0[1],zmm1[2],zmm0[2],zmm1[3],zmm0[3],zmm1[4],zmm0[4],zmm1[5],zmm0[5],zmm1[6],zmm0[6],zmm1[7],zmm0[7],zmm1[16],zmm0[16],zmm1[17],zmm0[17],zmm1[18],zmm0[18],zmm1[19],zmm0[19],zmm1[20],zmm0[20],zmm1[21],zmm0[21],zmm1[22],zmm0[22],zmm1[23],zmm0[23],zmm1[32],zmm0[32],zmm1[33],zmm0[33],zmm1[34],zmm0[34],zmm1[35],zmm0[35],zmm1[36],zmm0[36],zmm1[37],zmm0[37],zmm1[38],zmm0[38],zmm1[39],zmm0[39],zmm1[48],zmm0[48],zmm1[49],zmm0[49],zmm1[50],zmm0[50],zmm1[51],zmm0[51],zmm1[52],zmm0[52],zmm1[53],zmm0[53],zmm1[54],zmm0[54],zmm1[55],zmm0[55]
 ; AVX512BW-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpackuswb %zmm3, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpackuswb %zmm2, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
 ; AVX512BW-NEXT:    vpmullw %zmm0, %zmm1, %zmm0
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpxor %xmm1, %xmm1, %xmm1
 ; AVX512BW-NEXT:    vpackuswb %zmm1, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
+; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm2 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX512BW-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
-; AVX512BW-NEXT:    vpmullw %zmm0, %zmm3, %zmm0
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpmullw %zmm0, %zmm2, %zmm0
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpackuswb %zmm1, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
+; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm2 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX512BW-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[1,1,2,3]
-; AVX512BW-NEXT:    vpmullw %zmm0, %zmm3, %zmm0
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm0, %zmm0
+; AVX512BW-NEXT:    vpmullw %zmm0, %zmm2, %zmm0
+; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpackuswb %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BW-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpandq %zmm2, %zmm0, %zmm0
-; AVX512BW-NEXT:    vpackuswb %zmm0, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512BW-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512BW-NEXT:    vzeroupper
@@ -2895,57 +2869,55 @@ define i8 @test_v128i8(<128 x i8> %a0) {
 ; AVX512BWVL:       # %bb.0:
 ; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} zmm2 = zmm1[8],zmm0[8],zmm1[9],zmm0[9],zmm1[10],zmm0[10],zmm1[11],zmm0[11],zmm1[12],zmm0[12],zmm1[13],zmm0[13],zmm1[14],zmm0[14],zmm1[15],zmm0[15],zmm1[24],zmm0[24],zmm1[25],zmm0[25],zmm1[26],zmm0[26],zmm1[27],zmm0[27],zmm1[28],zmm0[28],zmm1[29],zmm0[29],zmm1[30],zmm0[30],zmm1[31],zmm0[31],zmm1[40],zmm0[40],zmm1[41],zmm0[41],zmm1[42],zmm0[42],zmm1[43],zmm0[43],zmm1[44],zmm0[44],zmm1[45],zmm0[45],zmm1[46],zmm0[46],zmm1[47],zmm0[47],zmm1[56],zmm0[56],zmm1[57],zmm0[57],zmm1[58],zmm0[58],zmm1[59],zmm0[59],zmm1[60],zmm0[60],zmm1[61],zmm0[61],zmm1[62],zmm0[62],zmm1[63],zmm0[63]
 ; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} zmm3 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
-; AVX512BWVL-NEXT:    vpmullw %zmm2, %zmm3, %zmm3
-; AVX512BWVL-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm3, %zmm3
+; AVX512BWVL-NEXT:    vpmullw %zmm2, %zmm3, %zmm2
+; AVX512BWVL-NEXT:    vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm2, %zmm2
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm1[0],zmm0[0],zmm1[1],zmm0[1],zmm1[2],zmm0[2],zmm1[3],zmm0[3],zmm1[4],zmm0[4],zmm1[5],zmm0[5],zmm1[6],zmm0[6],zmm1[7],zmm0[7],zmm1[16],zmm0[16],zmm1[17],zmm0[17],zmm1[18],zmm0[18],zmm1[19],zmm0[19],zmm1[20],zmm0[20],zmm1[21],zmm0[21],zmm1[22],zmm0[22],zmm1[23],zmm0[23],zmm1[32],zmm0[32],zmm1[33],zmm0[33],zmm1[34],zmm0[34],zmm1[35],zmm0[35],zmm1[36],zmm0[36],zmm1[37],zmm0[37],zmm1[38],zmm0[38],zmm1[39],zmm0[39],zmm1[48],zmm0[48],zmm1[49],zmm0[49],zmm1[50],zmm0[50],zmm1[51],zmm0[51],zmm1[52],zmm0[52],zmm1[53],zmm0[53],zmm1[54],zmm0[54],zmm1[55],zmm0[55]
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BWVL-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpackuswb %zmm3, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpackuswb %zmm2, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vextracti64x4 $1, %zmm0, %ymm1
-; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} zmm3 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
 ; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} zmm4 = zmm1[8],zmm0[8],zmm1[9],zmm0[9],zmm1[10],zmm0[10],zmm1[11],zmm0[11],zmm1[12],zmm0[12],zmm1[13],zmm0[13],zmm1[14],zmm0[14],zmm1[15],zmm0[15],zmm1[24],zmm0[24],zmm1[25],zmm0[25],zmm1[26],zmm0[26],zmm1[27],zmm0[27],zmm1[28],zmm0[28],zmm1[29],zmm0[29],zmm1[30],zmm0[30],zmm1[31],zmm0[31],zmm1[40],zmm0[40],zmm1[41],zmm0[41],zmm1[42],zmm0[42],zmm1[43],zmm0[43],zmm1[44],zmm0[44],zmm1[45],zmm0[45],zmm1[46],zmm0[46],zmm1[47],zmm0[47],zmm1[56],zmm0[56],zmm1[57],zmm0[57],zmm1[58],zmm0[58],zmm1[59],zmm0[59],zmm1[60],zmm0[60],zmm1[61],zmm0[61],zmm1[62],zmm0[62],zmm1[63],zmm0[63]
-; AVX512BWVL-NEXT:    vpmullw %zmm4, %zmm3, %zmm3
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm3, %zmm3
+; AVX512BWVL-NEXT:    vpmullw %zmm4, %zmm2, %zmm2
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm2, %zmm2
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm1[0],zmm0[0],zmm1[1],zmm0[1],zmm1[2],zmm0[2],zmm1[3],zmm0[3],zmm1[4],zmm0[4],zmm1[5],zmm0[5],zmm1[6],zmm0[6],zmm1[7],zmm0[7],zmm1[16],zmm0[16],zmm1[17],zmm0[17],zmm1[18],zmm0[18],zmm1[19],zmm0[19],zmm1[20],zmm0[20],zmm1[21],zmm0[21],zmm1[22],zmm0[22],zmm1[23],zmm0[23],zmm1[32],zmm0[32],zmm1[33],zmm0[33],zmm1[34],zmm0[34],zmm1[35],zmm0[35],zmm1[36],zmm0[36],zmm1[37],zmm0[37],zmm1[38],zmm0[38],zmm1[39],zmm0[39],zmm1[48],zmm0[48],zmm1[49],zmm0[49],zmm1[50],zmm0[50],zmm1[51],zmm0[51],zmm1[52],zmm0[52],zmm1[53],zmm0[53],zmm1[54],zmm0[54],zmm1[55],zmm0[55]
 ; AVX512BWVL-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpackuswb %zmm3, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpackuswb %zmm2, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vextracti128 $1, %ymm0, %xmm1
-; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} zmm3 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
 ; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} zmm4 = zmm1[8],zmm0[8],zmm1[9],zmm0[9],zmm1[10],zmm0[10],zmm1[11],zmm0[11],zmm1[12],zmm0[12],zmm1[13],zmm0[13],zmm1[14],zmm0[14],zmm1[15],zmm0[15],zmm1[24],zmm0[24],zmm1[25],zmm0[25],zmm1[26],zmm0[26],zmm1[27],zmm0[27],zmm1[28],zmm0[28],zmm1[29],zmm0[29],zmm1[30],zmm0[30],zmm1[31],zmm0[31],zmm1[40],zmm0[40],zmm1[41],zmm0[41],zmm1[42],zmm0[42],zmm1[43],zmm0[43],zmm1[44],zmm0[44],zmm1[45],zmm0[45],zmm1[46],zmm0[46],zmm1[47],zmm0[47],zmm1[56],zmm0[56],zmm1[57],zmm0[57],zmm1[58],zmm0[58],zmm1[59],zmm0[59],zmm1[60],zmm0[60],zmm1[61],zmm0[61],zmm1[62],zmm0[62],zmm1[63],zmm0[63]
-; AVX512BWVL-NEXT:    vpmullw %zmm4, %zmm3, %zmm3
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm3, %zmm3
+; AVX512BWVL-NEXT:    vpmullw %zmm4, %zmm2, %zmm2
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm2, %zmm2
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm1[0],zmm0[0],zmm1[1],zmm0[1],zmm1[2],zmm0[2],zmm1[3],zmm0[3],zmm1[4],zmm0[4],zmm1[5],zmm0[5],zmm1[6],zmm0[6],zmm1[7],zmm0[7],zmm1[16],zmm0[16],zmm1[17],zmm0[17],zmm1[18],zmm0[18],zmm1[19],zmm0[19],zmm1[20],zmm0[20],zmm1[21],zmm0[21],zmm1[22],zmm0[22],zmm1[23],zmm0[23],zmm1[32],zmm0[32],zmm1[33],zmm0[33],zmm1[34],zmm0[34],zmm1[35],zmm0[35],zmm1[36],zmm0[36],zmm1[37],zmm0[37],zmm1[38],zmm0[38],zmm1[39],zmm0[39],zmm1[48],zmm0[48],zmm1[49],zmm0[49],zmm1[50],zmm0[50],zmm1[51],zmm0[51],zmm1[52],zmm0[52],zmm1[53],zmm0[53],zmm1[54],zmm0[54],zmm1[55],zmm0[55]
 ; AVX512BWVL-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpackuswb %zmm3, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpackuswb %zmm2, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BWVL-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
 ; AVX512BWVL-NEXT:    vpmullw %zmm0, %zmm1, %zmm0
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vpxor %xmm1, %xmm1, %xmm1
 ; AVX512BWVL-NEXT:    vpackuswb %zmm1, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
+; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm2 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX512BWVL-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
-; AVX512BWVL-NEXT:    vpmullw %zmm0, %zmm3, %zmm0
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmullw %zmm0, %zmm2, %zmm0
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vpackuswb %zmm1, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
+; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm2 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX512BWVL-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[1,1,2,3]
-; AVX512BWVL-NEXT:    vpmullw %zmm0, %zmm3, %zmm0
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmullw %zmm0, %zmm2, %zmm0
+; AVX512BWVL-NEXT:    vpandq %zmm3, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vpackuswb %zmm1, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512BWVL-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
 ; AVX512BWVL-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpandq %zmm2, %zmm0, %zmm0
-; AVX512BWVL-NEXT:    vpackuswb %zmm0, %zmm0, %zmm0
 ; AVX512BWVL-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512BWVL-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512BWVL-NEXT:    vzeroupper
@@ -3012,8 +2984,6 @@ define i8 @test_v128i8(<128 x i8> %a0) {
 ; AVX512DQ-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512DQ-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQ-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpand %ymm4, %ymm0, %ymm0
-; AVX512DQ-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
 ; AVX512DQ-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512DQ-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512DQ-NEXT:    vzeroupper
@@ -3023,65 +2993,63 @@ define i8 @test_v128i8(<128 x i8> %a0) {
 ; AVX512DQVL:       # %bb.0:
 ; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm4 = ymm2[8],ymm0[8],ymm2[9],ymm0[9],ymm2[10],ymm0[10],ymm2[11],ymm0[11],ymm2[12],ymm0[12],ymm2[13],ymm0[13],ymm2[14],ymm0[14],ymm2[15],ymm0[15],ymm2[24],ymm0[24],ymm2[25],ymm0[25],ymm2[26],ymm0[26],ymm2[27],ymm0[27],ymm2[28],ymm0[28],ymm2[29],ymm0[29],ymm2[30],ymm0[30],ymm2[31],ymm0[31]
 ; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm5 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512DQVL-NEXT:    vpmullw %ymm4, %ymm5, %ymm5
-; AVX512DQVL-NEXT:    vmovdqa {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm5, %ymm5
+; AVX512DQVL-NEXT:    vpmullw %ymm4, %ymm5, %ymm4
+; AVX512DQVL-NEXT:    vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm4, %ymm4
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm2[0],ymm0[0],ymm2[1],ymm0[1],ymm2[2],ymm0[2],ymm2[3],ymm0[3],ymm2[4],ymm0[4],ymm2[5],ymm0[5],ymm2[6],ymm0[6],ymm2[7],ymm0[7],ymm2[16],ymm0[16],ymm2[17],ymm0[17],ymm2[18],ymm0[18],ymm2[19],ymm0[19],ymm2[20],ymm0[20],ymm2[21],ymm0[21],ymm2[22],ymm0[22],ymm2[23],ymm0[23]
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpmullw %ymm2, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpackuswb %ymm5, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpackuswb %ymm4, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm3[8],ymm0[8],ymm3[9],ymm0[9],ymm3[10],ymm0[10],ymm3[11],ymm0[11],ymm3[12],ymm0[12],ymm3[13],ymm0[13],ymm3[14],ymm0[14],ymm3[15],ymm0[15],ymm3[24],ymm0[24],ymm3[25],ymm0[25],ymm3[26],ymm0[26],ymm3[27],ymm0[27],ymm3[28],ymm0[28],ymm3[29],ymm0[29],ymm3[30],ymm0[30],ymm3[31],ymm0[31]
-; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm5 = ymm1[8],ymm0[8],ymm1[9],ymm0[9],ymm1[10],ymm0[10],ymm1[11],ymm0[11],ymm1[12],ymm0[12],ymm1[13],ymm0[13],ymm1[14],ymm0[14],ymm1[15],ymm0[15],ymm1[24],ymm0[24],ymm1[25],ymm0[25],ymm1[26],ymm0[26],ymm1[27],ymm0[27],ymm1[28],ymm0[28],ymm1[29],ymm0[29],ymm1[30],ymm0[30],ymm1[31],ymm0[31]
-; AVX512DQVL-NEXT:    vpmullw %ymm2, %ymm5, %ymm2
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm4 = ymm1[8],ymm0[8],ymm1[9],ymm0[9],ymm1[10],ymm0[10],ymm1[11],ymm0[11],ymm1[12],ymm0[12],ymm1[13],ymm0[13],ymm1[14],ymm0[14],ymm1[15],ymm0[15],ymm1[24],ymm0[24],ymm1[25],ymm0[25],ymm1[26],ymm0[26],ymm1[27],ymm0[27],ymm1[28],ymm0[28],ymm1[29],ymm0[29],ymm1[30],ymm0[30],ymm1[31],ymm0[31]
+; AVX512DQVL-NEXT:    vpmullw %ymm2, %ymm4, %ymm2
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm2, %ymm2
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm3 = ymm3[0],ymm0[0],ymm3[1],ymm0[1],ymm3[2],ymm0[2],ymm3[3],ymm0[3],ymm3[4],ymm0[4],ymm3[5],ymm0[5],ymm3[6],ymm0[6],ymm3[7],ymm0[7],ymm3[16],ymm0[16],ymm3[17],ymm0[17],ymm3[18],ymm0[18],ymm3[19],ymm0[19],ymm3[20],ymm0[20],ymm3[21],ymm0[21],ymm3[22],ymm0[22],ymm3[23],ymm0[23]
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0],ymm0[0],ymm1[1],ymm0[1],ymm1[2],ymm0[2],ymm1[3],ymm0[3],ymm1[4],ymm0[4],ymm1[5],ymm0[5],ymm1[6],ymm0[6],ymm1[7],ymm0[7],ymm1[16],ymm0[16],ymm1[17],ymm0[17],ymm1[18],ymm0[18],ymm1[19],ymm0[19],ymm1[20],ymm0[20],ymm1[21],ymm0[21],ymm1[22],ymm0[22],ymm1[23],ymm0[23]
 ; AVX512DQVL-NEXT:    vpmullw %ymm3, %ymm1, %ymm1
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm1, %ymm1
 ; AVX512DQVL-NEXT:    vpackuswb %ymm2, %ymm1, %ymm1
 ; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm1[8],ymm0[8],ymm1[9],ymm0[9],ymm1[10],ymm0[10],ymm1[11],ymm0[11],ymm1[12],ymm0[12],ymm1[13],ymm0[13],ymm1[14],ymm0[14],ymm1[15],ymm0[15],ymm1[24],ymm0[24],ymm1[25],ymm0[25],ymm1[26],ymm0[26],ymm1[27],ymm0[27],ymm1[28],ymm0[28],ymm1[29],ymm0[29],ymm1[30],ymm0[30],ymm1[31],ymm0[31]
 ; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
 ; AVX512DQVL-NEXT:    vpmullw %ymm2, %ymm3, %ymm2
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm2, %ymm2
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0],ymm0[0],ymm1[1],ymm0[1],ymm1[2],ymm0[2],ymm1[3],ymm0[3],ymm1[4],ymm0[4],ymm1[5],ymm0[5],ymm1[6],ymm0[6],ymm1[7],ymm0[7],ymm1[16],ymm0[16],ymm1[17],ymm0[17],ymm1[18],ymm0[18],ymm1[19],ymm0[19],ymm1[20],ymm0[20],ymm1[21],ymm0[21],ymm1[22],ymm0[22],ymm1[23],ymm0[23]
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vextracti128 $1, %ymm0, %xmm1
 ; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
 ; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm1[8],ymm0[8],ymm1[9],ymm0[9],ymm1[10],ymm0[10],ymm1[11],ymm0[11],ymm1[12],ymm0[12],ymm1[13],ymm0[13],ymm1[14],ymm0[14],ymm1[15],ymm0[15],ymm1[24],ymm0[24],ymm1[25],ymm0[25],ymm1[26],ymm0[26],ymm1[27],ymm0[27],ymm1[28],ymm0[28],ymm1[29],ymm0[29],ymm1[30],ymm0[30],ymm1[31],ymm0[31]
 ; AVX512DQVL-NEXT:    vpmullw %ymm3, %ymm2, %ymm2
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm2, %ymm2
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0],ymm0[0],ymm1[1],ymm0[1],ymm1[2],ymm0[2],ymm1[3],ymm0[3],ymm1[4],ymm0[4],ymm1[5],ymm0[5],ymm1[6],ymm0[6],ymm1[7],ymm0[7],ymm1[16],ymm0[16],ymm1[17],ymm0[17],ymm1[18],ymm0[18],ymm1[19],ymm0[19],ymm1[20],ymm0[20],ymm1[21],ymm0[21],ymm1[22],ymm0[22],ymm1[23],ymm0[23]
 ; AVX512DQVL-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
 ; AVX512DQVL-NEXT:    vpmullw %ymm0, %ymm1, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpxor %xmm1, %xmm1, %xmm1
 ; AVX512DQVL-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX512DQVL-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
 ; AVX512DQVL-NEXT:    vpmullw %ymm0, %ymm2, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; AVX512DQVL-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[1,1,2,3]
 ; AVX512DQVL-NEXT:    vpmullw %ymm0, %ymm2, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpand %ymm5, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpsrlw $8, %xmm0, %xmm1
 ; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
 ; AVX512DQVL-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpand %ymm4, %ymm0, %ymm0
-; AVX512DQVL-NEXT:    vpackuswb %ymm0, %ymm0, %ymm0
 ; AVX512DQVL-NEXT:    vpextrb $0, %xmm0, %eax
 ; AVX512DQVL-NEXT:    # kill: def $al killed $al killed $eax
 ; AVX512DQVL-NEXT:    vzeroupper




More information about the llvm-commits mailing list