[AVX512] Fix insertelement i1 lowering.

[igorb]

Summary

[AVX512] Fix insertelement i1 lowering.
1 . Use shuffle to insert element i1 into vector. The previous implementation was incorrect ( dest_bit OR src_bit , it doesn't clear the bit if src_bit=0 )

2. Improve shuffle i1 vector, use CVT2MASK if supported instead TRUNCATE.

[delena]

clear -> clean

lib/Target/X86/X86ISelLowering.cpp

12651

You can do this if IdxVal !=0 and Vec is defined.
Otherwise it can be done cheaper with shifts and "or"s

[igorb]

Thanks for review,
fixed according to comments.

[delena]

lib/Target/X86/X86ISelLowering.cpp

12650–12681

Clean the last bit in the source vector.

12651

Insertion of one bit into first or last position can be done with two SHIFTs + OR.

12654

lsb -> the first

12655

Line alignment.

12662

MaskVec[i] = (i == IdxVal) ? 0 : i+NumElements;

or swap parameters in getVectorShuffle()

12663

Move the bit to the last position inside the vector.

[delena]

Comments about shuffle are wrong, taking them back.
You can commit after fixing all comments inside the code.

lib/Target/X86/X86ISelLowering.cpp

if (!Subtarget.useSoftFloat() && Subtarget.hasVLX()) {
     setOperationAction(ISD::TRUNCATE,           MVT::v2i1, Custom);
     setOperationAction(ISD::TRUNCATE,           MVT::v4i1, Custom);
     setOperationAction(ISD::SETCC,              MVT::v4i1, Custom);
     setOperationAction(ISD::SETCC,              MVT::v2i1, Custom);
     setOperationAction(ISD::CONCAT_VECTORS,     MVT::v4i1, Custom);
     setOperationAction(ISD::CONCAT_VECTORS,     MVT::v8i1, Custom);
     setOperationAction(ISD::INSERT_SUBVECTOR,   MVT::v8i1, Custom);
     setOperationAction(ISD::INSERT_SUBVECTOR,   MVT::v4i1, Custom);
+    setOperationAction(ISD::INSERT_VECTOR_ELT,  MVT::v8i1, Custom);
+    setOperationAction(ISD::INSERT_VECTOR_ELT,  MVT::v4i1, Custom);
     setOperationAction(ISD::SELECT,             MVT::v4i1, Custom);
     setOperationAction(ISD::SELECT,             MVT::v2i1, Custom);
     setOperationAction(ISD::BUILD_VECTOR,       MVT::v4i1, Custom);
     setOperationAction(ISD::BUILD_VECTOR,       MVT::v2i1, Custom);
     setOperationAction(ISD::VECTOR_SHUFFLE,     MVT::v2i1, Custom);
     setOperationAction(ISD::VECTOR_SHUFFLE,     MVT::v4i1, Custom);
     setOperationAction(ISD::VSELECT,            MVT::v2i1, Expand);
     setOperationAction(ISD::VSELECT,            MVT::v4i1, Expand);
static SDValue lower1BitVectorShuffle(const SDLoc &DL, ArrayRef<int> Mask,
   if (V2.isUndef())
     V2 = DAG.getUNDEF(ExtVT);
   else if (ISD::isBuildVectorAllZeros(V2.getNode()))
     V2 = getZeroVector(ExtVT, Subtarget, DAG, DL);
   else if (ISD::isBuildVectorAllOnes(V2.getNode()))
     V2 = getOnesVector(ExtVT, Subtarget, DAG, DL);
   else
     V2 = DAG.getNode(ISD::SIGN_EXTEND, DL, ExtVT, V2);
-  return DAG.getNode(ISD::TRUNCATE, DL, VT,
-                     DAG.getVectorShuffle(ExtVT, DL, V1, V2, Mask));
+
+  SDValue Shuffle = DAG.getVectorShuffle(ExtVT, DL, V1, V2, Mask);
+  // i1 was sign extended we can use X86ISD::CVT2MASK.
+  int NumElems = VT.getVectorNumElements();
+  if ((Subtarget.hasBWI() && (NumElems >= 32)) ||
+      (Subtarget.hasDQI() && (NumElems < 32)))
+    return DAG.getNode(X86ISD::CVT2MASK, DL, VT, Shuffle);
+
+  return DAG.getNode(ISD::TRUNCATE, DL, VT, Shuffle);
 }
 
 /// Helper function that returns true if the shuffle mask should be
 /// commuted to improve canonicalization.
 static bool canonicalizeShuffleMaskWithCommute(ArrayRef<int> Mask) {
   int NumElements = Mask.size();
 
   int NumV1Elements = 0, NumV2Elements = 0, NumSentinelElements = 0;
if (!isa<ConstantSDNode>(Idx)) {
     SDValue ExtOp = DAG.getNode(ISD::INSERT_VECTOR_ELT, dl, ExtVecVT,
       DAG.getNode(ISD::ZERO_EXTEND, dl, ExtVecVT, Vec),
       DAG.getNode(ISD::ZERO_EXTEND, dl, ExtEltVT, Elt), Idx);
     return DAG.getNode(ISD::TRUNCATE, dl, VecVT, ExtOp);
   }
 
   unsigned IdxVal = cast<ConstantSDNode>(Idx)->getZExtValue();
   SDValue EltInVec = DAG.getNode(ISD::SCALAR_TO_VECTOR, dl, VecVT, Elt);
+  unsigned NumElems = VecVT.getVectorNumElements();
+
+  if(Vec.isUndef()) {
     if (IdxVal)
       EltInVec = DAG.getNode(X86ISD::VSHLI, dl, VecVT, EltInVec,
                              DAG.getConstant(IdxVal, dl, MVT::i8));
-  if (Vec.isUndef())
     return EltInVec;
+  }
+
+  //insert into first/last elm
+  if (IdxVal == 0 || IdxVal == NumElems -1) {
+    unsigned MaxSift = VecVT.getVectorNumElements() - 1;
+    //set bit at index , clean all other bits
+    EltInVec = DAG.getNode(X86ISD::VSHLI, dl, VecVT, EltInVec,
+                           DAG.getConstant(MaxSift , dl, MVT::i8));
+    EltInVec = DAG.getNode(X86ISD::VSRLI, dl, VecVT, EltInVec,
+                           DAG.getConstant(MaxSift - IdxVal, dl, MVT::i8));
+    //clean bit in vector
+    if (IdxVal == 0) {
+      Vec = DAG.getNode(X86ISD::VSRLI, dl, VecVT, Vec,
+                             DAG.getConstant(1 , dl, MVT::i8));
+      Vec = DAG.getNode(X86ISD::VSHLI, dl, VecVT, Vec,
+                             DAG.getConstant(1, dl, MVT::i8));
+    } else {
+      Vec = DAG.getNode(X86ISD::VSHLI, dl, VecVT, Vec,
+                             DAG.getConstant(1, dl, MVT::i8));
+      Vec = DAG.getNode(X86ISD::VSRLI, dl, VecVT, Vec,
+                             DAG.getConstant(1 , dl, MVT::i8));
+    }
+
     return DAG.getNode(ISD::OR, dl, VecVT, Vec, EltInVec);
   }
 
+  // Use shuffle to insert element.
+  SmallVector<int, 64> MaskVec(NumElems);
+  for (unsigned i = 0; i != NumElems; ++i)
+    MaskVec[i] = (i == IdxVal) ? NumElems : i;
+
+  return DAG.getVectorShuffle(VecVT, dl, Vec, EltInVec, MaskVec);
+}
+
 SDValue X86TargetLowering::LowerINSERT_VECTOR_ELT(SDValue Op,
                                                   SelectionDAG &DAG) const {
   MVT VT = Op.getSimpleValueType();
   MVT EltVT = VT.getVectorElementType();
   unsigned NumElts = VT.getVectorNumElements();
 
   if (EltVT == MVT::i1)
     return InsertBitToMaskVector(Op, DAG);

lib/Target/X86/X86InstrAVX512.td

 }
 
 defm KSET0 : avx512_mask_setop_w<immAllZerosV>;
 defm KSET1 : avx512_mask_setop_w<immAllOnesV>;
 
 // With AVX-512 only, 8-bit mask is promoted to 16-bit mask.
 let Predicates = [HasAVX512] in {
   def : Pat<(v8i1 immAllZerosV), (COPY_TO_REGCLASS (KSET0W), VK8)>;
+  def : Pat<(v4i1 immAllZerosV), (COPY_TO_REGCLASS (KSET0W), VK4)>;
+  def : Pat<(v2i1 immAllZerosV), (COPY_TO_REGCLASS (KSET0W), VK2)>;
   def : Pat<(v8i1 immAllOnesV),  (COPY_TO_REGCLASS (KSET1W), VK8)>;
   def : Pat<(v4i1 immAllOnesV),  (COPY_TO_REGCLASS (KSET1W), VK4)>;
   def : Pat<(v2i1 immAllOnesV),  (COPY_TO_REGCLASS (KSET1W), VK2)>;
   def : Pat<(i1 0), (COPY_TO_REGCLASS (KSET0W), VK1)>;
   def : Pat<(i1 1), (COPY_TO_REGCLASS (KSHIFTRWri (KSET1W), (i8 15)), VK1)>;
   def : Pat<(i1 -1), (COPY_TO_REGCLASS (KSHIFTRWri (KSET1W), (i8 15)), VK1)>;
 }
 
def : Pat<(v4i1 (extract_subvector (v8i1 VK8:$src), (iPTR 4))),
                    VK4))>;
 def : Pat<(v8i1 (extract_subvector (v16i1 VK16:$src), (iPTR 8))),
           (v8i1 (COPY_TO_REGCLASS (KSHIFTRWri VK16:$src, (i8 8)), VK8))>;
 def : Pat<(v16i1 (extract_subvector (v32i1 VK32:$src), (iPTR 16))),
           (v16i1 (COPY_TO_REGCLASS (KSHIFTRDri VK32:$src, (i8 16)), VK16))>;
 def : Pat<(v32i1 (extract_subvector (v64i1 VK64:$src), (iPTR 32))),
           (v32i1 (COPY_TO_REGCLASS (KSHIFTRQri VK64:$src, (i8 32)), VK32))>;
 
-def : Pat<(v8i1 (X86vshli VK8:$src, (i8 imm:$imm))),
-          (v8i1 (COPY_TO_REGCLASS
-                 (KSHIFTLWri (COPY_TO_REGCLASS VK8:$src, VK16),
-                  (I8Imm $imm)), VK8))>, Requires<[HasAVX512, NoDQI]>;
 
-def : Pat<(v4i1 (X86vshli VK4:$src, (i8 imm:$imm))),
-          (v4i1 (COPY_TO_REGCLASS
-                 (KSHIFTLWri (COPY_TO_REGCLASS VK4:$src, VK16),
-                  (I8Imm $imm)), VK4))>, Requires<[HasAVX512]>;
+// Patterns for kmask shift
+multiclass mask_shift_lowering<RegisterClass RC, ValueType VT> {
+  def : Pat<(VT (X86vshli RC:$src, (i8 imm:$imm))),
+            (VT (COPY_TO_REGCLASS                
+                   (KSHIFTLWri (COPY_TO_REGCLASS RC:$src, VK16),
+                               (I8Imm $imm)), 
+                   RC))>;
+  def : Pat<(VT (X86vsrli RC:$src, (i8 imm:$imm))),
+            (VT (COPY_TO_REGCLASS                
+                   (KSHIFTRWri (COPY_TO_REGCLASS RC:$src, VK16),
+                               (I8Imm $imm)), 
+                   RC))>;
+}
+
+defm : mask_shift_lowering<VK8, v8i1>, Requires<[HasAVX512, NoDQI]>;
+defm : mask_shift_lowering<VK4, v4i1>, Requires<[HasAVX512]>;
+defm : mask_shift_lowering<VK2, v2i1>, Requires<[HasAVX512]>;
 //===----------------------------------------------------------------------===//
 // AVX-512 - Aligned and unaligned load and store
 //
 
 
 multiclass avx512_load<bits<8> opc, string OpcodeStr, X86VectorVTInfo _,
                          PatFrag ld_frag, PatFrag mload,
                          bit IsReMaterializable = 1,

test/CodeGen/X86/avx512-ext.ll

 
 define i16 @trunc_i32_to_i1(i32 %a) {
 ; ALL-LABEL: trunc_i32_to_i1:
 ; ALL:       ## BB#0:
 ; ALL-NEXT:    andl $1, %edi
 ; ALL-NEXT:    kmovw %edi, %k0
 ; ALL-NEXT:    movw $-4, %ax
 ; ALL-NEXT:    kmovw %eax, %k1
+; ALL-NEXT:    kshiftrw $1, %k1, %k1
+; ALL-NEXT:    kshiftlw $1, %k1, %k1
+; ALL-NEXT:    kshiftlw $15, %k0, %k0
+; ALL-NEXT:    kshiftrw $15, %k0, %k0
 ; ALL-NEXT:    korw %k0, %k1, %k0
 ; ALL-NEXT:    kmovw %k0, %eax
 ; ALL-NEXT:    retq
   %a_i = trunc i32 %a to i1
   %maskv = insertelement <16 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, i1 %a_i, i32 0
   %res = bitcast <16 x i1> %maskv to i16
   ret i16 %res
 }

test/CodeGen/X86/avx512-insert-extract.ll

 }
 
 define i16 @test13(i32 %a, i32 %b) {
 ; KNL-LABEL: test13:
 ; KNL:       ## BB#0:
 ; KNL-NEXT:    cmpl %esi, %edi
 ; KNL-NEXT:    setb %al
 ; KNL-NEXT:    kmovw %eax, %k0
+; KNL-NEXT:    kshiftlw $15, %k0, %k0
+; KNL-NEXT:    kshiftrw $15, %k0, %k0
 ; KNL-NEXT:    movw $-4, %ax
 ; KNL-NEXT:    kmovw %eax, %k1
+; KNL-NEXT:    kshiftrw $1, %k1, %k1
+; KNL-NEXT:    kshiftlw $1, %k1, %k1
 ; KNL-NEXT:    korw %k0, %k1, %k0
 ; KNL-NEXT:    kmovw %k0, %eax
 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: test13:
 ; SKX:       ## BB#0:
 ; SKX-NEXT:    cmpl %esi, %edi
 ; SKX-NEXT:    setb %al
 ; SKX-NEXT:    kmovw %eax, %k0
+; SKX-NEXT:    kshiftlw $15, %k0, %k0
+; SKX-NEXT:    kshiftrw $15, %k0, %k0
 ; SKX-NEXT:    movw $-4, %ax
 ; SKX-NEXT:    kmovw %eax, %k1
+; SKX-NEXT:    kshiftrw $1, %k1, %k1
+; SKX-NEXT:    kshiftlw $1, %k1, %k1
 ; SKX-NEXT:    korw %k0, %k1, %k0
 ; SKX-NEXT:    kmovw %k0, %eax
 ; SKX-NEXT:    retq
   %cmp_res = icmp ult i32 %a, %b
   %maskv = insertelement <16 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, i1 %cmp_res, i32 0
   %res = bitcast <16 x i1> %maskv to i16
   ret i16 %res
 }
; SKX-NEXT:    retq
   ret i16 %x2
 }
 
 define i16 @test16(i1 *%addr, i16 %a) {
 ; KNL-LABEL: test16:
 ; KNL:       ## BB#0:
 ; KNL-NEXT:    movzbl (%rdi), %eax
 ; KNL-NEXT:    andl $1, %eax
-; KNL-NEXT:    kmovw %eax, %k0
-; KNL-NEXT:    kmovw %esi, %k1
-; KNL-NEXT:    kshiftlw $10, %k0, %k0
-; KNL-NEXT:    korw %k0, %k1, %k0
+; KNL-NEXT:    kmovw %eax, %k1
+; KNL-NEXT:    kmovw %esi, %k2
+; KNL-NEXT:    vpternlogd $255, %zmm0, %zmm0, %zmm0
+; KNL-NEXT:    vmovdqa32 %zmm0, %zmm1 {%k2} {z}
+; KNL-NEXT:    vmovdqa32 %zmm0, %zmm0 {%k1} {z}
+; KNL-NEXT:    vmovdqa32 {{.*#+}} zmm2 = [0,1,2,3,4,5,6,7,8,9,16,11,12,13,14,15]
+; KNL-NEXT:    vpermt2d %zmm0, %zmm2, %zmm1
+; KNL-NEXT:    vpslld $31, %zmm1, %zmm0
+; KNL-NEXT:    vptestmd %zmm0, %zmm0, %k0
 ; KNL-NEXT:    kmovw %k0, %eax
 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: test16:
 ; SKX:       ## BB#0:
 ; SKX-NEXT:    movzbl (%rdi), %eax
 ; SKX-NEXT:    andl $1, %eax
 ; SKX-NEXT:    kmovd %eax, %k0
 ; SKX-NEXT:    kmovw %esi, %k1
-; SKX-NEXT:    kshiftlw $10, %k0, %k0
-; SKX-NEXT:    korw %k0, %k1, %k0
+; SKX-NEXT:    vpmovm2d %k1, %zmm0
+; SKX-NEXT:    vpmovm2d %k0, %zmm1
+; SKX-NEXT:    vmovdqa32 {{.*#+}} zmm2 = [0,1,2,3,4,5,6,7,8,9,16,11,12,13,14,15]
+; SKX-NEXT:    vpermt2d %zmm1, %zmm2, %zmm0
+; SKX-NEXT:    vpmovd2m %zmm0, %k0
 ; SKX-NEXT:    kmovw %k0, %eax
 ; SKX-NEXT:    retq
   %x = load i1 , i1 * %addr, align 128
   %a1 = bitcast i16 %a to <16 x i1>
   %x1 = insertelement <16 x i1> %a1, i1 %x, i32 10
   %x2 = bitcast <16 x i1>%x1 to i16
   ret i16 %x2
 }
 
 define i8 @test17(i1 *%addr, i8 %a) {
 ; KNL-LABEL: test17:
 ; KNL:       ## BB#0:
 ; KNL-NEXT:    movzbl (%rdi), %eax
 ; KNL-NEXT:    andl $1, %eax
-; KNL-NEXT:    kmovw %eax, %k0
-; KNL-NEXT:    kmovw %esi, %k1
-; KNL-NEXT:    kshiftlw $4, %k0, %k0
-; KNL-NEXT:    korw %k0, %k1, %k0
+; KNL-NEXT:    kmovw %eax, %k1
+; KNL-NEXT:    kmovw %esi, %k2
+; KNL-NEXT:    vpternlogd $255, %zmm0, %zmm0, %zmm0
+; KNL-NEXT:    vmovdqa64 %zmm0, %zmm1 {%k2} {z}
+; KNL-NEXT:    vmovdqa64 %zmm0, %zmm0 {%k1} {z}
+; KNL-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [0,1,2,3,8,5,6,7]
+; KNL-NEXT:    vpermt2q %zmm0, %zmm2, %zmm1
+; KNL-NEXT:    vpsllq $63, %zmm1, %zmm0
+; KNL-NEXT:    vptestmq %zmm0, %zmm0, %k0
 ; KNL-NEXT:    kmovw %k0, %eax
 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: test17:
 ; SKX:       ## BB#0:
 ; SKX-NEXT:    movzbl (%rdi), %eax
 ; SKX-NEXT:    andl $1, %eax
 ; SKX-NEXT:    kmovd %eax, %k0
 ; SKX-NEXT:    kmovb %esi, %k1
-; SKX-NEXT:    kshiftlb $4, %k0, %k0
-; SKX-NEXT:    korb %k0, %k1, %k0
+; SKX-NEXT:    vpmovm2q %k1, %zmm0
+; SKX-NEXT:    vpmovm2q %k0, %zmm1
+; SKX-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [0,1,2,3,8,5,6,7]
+; SKX-NEXT:    vpermt2q %zmm1, %zmm2, %zmm0
+; SKX-NEXT:    vpmovq2m %zmm0, %k0
 ; SKX-NEXT:    kmovb %k0, %eax
 ; SKX-NEXT:    retq
   %x = load i1 , i1 * %addr, align 128
   %a1 = bitcast i8 %a to <8 x i1>
   %x1 = insertelement <8 x i1> %a1, i1 %x, i32 4
   %x2 = bitcast <8 x i1>%x1 to i8
   ret i8 %x2
 }

 ; SKX:       ## BB#0:
 ; SKX-NEXT:    vextracti32x4 $1, %ymm0, %xmm1
 ; SKX-NEXT:    vpinsrb $4, %edi, %xmm1, %xmm1
 ; SKX-NEXT:    vinserti32x4 $1, %xmm1, %ymm0, %ymm0
 ; SKX-NEXT:    retq
   %r = insertelement <32 x i8> %x, i8 %y, i32 20
   ret <32 x i8> %r
 }
+
+define i32 @test_insertelement_v32i1(i32 %a, i32 %b, <32 x i32> %x , <32 x i32> %y) {
+; SKX-LABEL: test_insertelement_v32i1:
+; SKX:       ## BB#0:
+; SKX-NEXT:    cmpl %esi, %edi
+; SKX-NEXT:    setb %al
+; SKX-NEXT:    kmovw %eax, %k0
+; SKX-NEXT:    vpcmpltud %zmm2, %zmm0, %k1
+; SKX-NEXT:    vpcmpltud %zmm3, %zmm1, %k2
+; SKX-NEXT:    kunpckwd %k1, %k2, %k1
+; SKX-NEXT:    vpmovm2w %k1, %zmm0
+; SKX-NEXT:    vpmovm2w %k0, %zmm1
+; SKX-NEXT:    vmovdqu16 {{.*#+}} zmm2 = [0,1,2,3,32,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31]
+; SKX-NEXT:    vpermt2w %zmm1, %zmm2, %zmm0
+; SKX-NEXT:    vpmovw2m %zmm0, %k0
+; SKX-NEXT:    kmovd %k0, %eax
+; SKX-NEXT:    retq
+  %cmp_res_i1 = icmp ult i32 %a, %b
+  %cmp_cmp_vec = icmp ult <32 x i32> %x, %y
+  %maskv = insertelement <32 x i1> %cmp_cmp_vec, i1 %cmp_res_i1, i32 4
+  %res = bitcast <32 x i1> %maskv to i32
+  ret i32 %res
+}
+
+define i8 @test_iinsertelement_v4i1(i32 %a, i32 %b, <4 x i32> %x , <4 x i32> %y) {
+; SKX-LABEL: test_iinsertelement_v4i1:
+; SKX:       ## BB#0:
+; SKX-NEXT:    cmpl %esi, %edi
+; SKX-NEXT:    setb %al
+; SKX-NEXT:    kmovw %eax, %k0
+; SKX-NEXT:    vpcmpltud %xmm1, %xmm0, %k1
+; SKX-NEXT:    vpmovm2d %k1, %xmm0
+; SKX-NEXT:    vpmovm2d %k0, %xmm1
+; SKX-NEXT:    vpbroadcastq %xmm1, %xmm1
+; SKX-NEXT:    vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2],xmm0[3]
+; SKX-NEXT:    vpmovd2m %xmm0, %k0
+; SKX-NEXT:    kmovb %k0, %eax
+; SKX-NEXT:    retq
+  %cmp_res_i1 = icmp ult i32 %a, %b
+  %cmp_cmp_vec = icmp ult <4 x i32> %x, %y
+  %maskv = insertelement <4 x i1> %cmp_cmp_vec, i1 %cmp_res_i1, i32 2
+  %res0 = shufflevector <4 x i1> %maskv, <4 x i1> undef , <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 4, i32 4, i32 4>
+  %res = bitcast <8 x i1> %res0 to i8
+  ret i8 %res
+}

test/CodeGen/X86/avx512-mask-op.ll

 ; KNL-NEXT:    popq %rbp
 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: test16:
 ; SKX:       ## BB#0:
 ; SKX-NEXT:    kmovq %rdi, %k0
 ; SKX-NEXT:    kxnorw %k0, %k0, %k1
 ; SKX-NEXT:    kshiftrw $15, %k1, %k1
-; SKX-NEXT:    kshiftlq $5, %k1, %k1
-; SKX-NEXT:    korq %k1, %k0, %k0
+; SKX-NEXT:    vpmovm2b %k1, %zmm0
+; SKX-NEXT:    vpsllq $40, %xmm0, %xmm0
+; SKX-NEXT:    vpmovm2b %k0, %zmm1
+; SKX-NEXT:    vmovdqu8 {{.*#+}} ymm2 = [255,255,255,255,255,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; SKX-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
+; SKX-NEXT:    vextracti64x4 $1, %zmm1, %ymm1
+; SKX-NEXT:    vinserti64x4 $1, %ymm1, %zmm0, %zmm0
+; SKX-NEXT:    vpmovb2m %zmm0, %k0
 ; SKX-NEXT:    vpmovm2b %k0, %zmm0
 ; SKX-NEXT:    retq
   %a = bitcast i64 %x to <64 x i1>
   %b = insertelement <64 x i1>%a, i1 true, i32 5
   %c = sext <64 x i1>%b to <64 x i8>
   ret <64 x i8>%c
 }
 

 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: test17:
 ; SKX:       ## BB#0:
 ; SKX-NEXT:    kmovq %rdi, %k0
 ; SKX-NEXT:    cmpl %edx, %esi
 ; SKX-NEXT:    setg %al
 ; SKX-NEXT:    kmovw %eax, %k1
-; SKX-NEXT:    kshiftlq $5, %k1, %k1
-; SKX-NEXT:    korq %k1, %k0, %k0
+; SKX-NEXT:    vpmovm2b %k1, %zmm0
+; SKX-NEXT:    vpsllq $40, %xmm0, %xmm0
+; SKX-NEXT:    vpmovm2b %k0, %zmm1
+; SKX-NEXT:    vmovdqu8 {{.*#+}} ymm2 = [255,255,255,255,255,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; SKX-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
+; SKX-NEXT:    vextracti64x4 $1, %zmm1, %ymm1
+; SKX-NEXT:    vinserti64x4 $1, %ymm1, %zmm0, %zmm0
+; SKX-NEXT:    vpmovb2m %zmm0, %k0
 ; SKX-NEXT:    vpmovm2b %k0, %zmm0
 ; SKX-NEXT:    retq
   %a = bitcast i64 %x to <64 x i1>
   %b = icmp sgt i32 %y, %z
   %c = insertelement <64 x i1>%a, i1 %b, i32 5
   %d = sext <64 x i1>%c to <64 x i8>
   ret <64 x i8>%d
 }
 
 define <8 x i1> @test18(i8 %a, i16 %y) {
 ; KNL-LABEL: test18:
 ; KNL:       ## BB#0:
-; KNL-NEXT:    kmovw %edi, %k0
-; KNL-NEXT:    kmovw %esi, %k1
-; KNL-NEXT:    kshiftlw $7, %k1, %k2
+; KNL-NEXT:    kmovw %edi, %k1
+; KNL-NEXT:    kmovw %esi, %k2
+; KNL-NEXT:    kshiftlw $7, %k2, %k0
+; KNL-NEXT:    kshiftrw $15, %k0, %k0
+; KNL-NEXT:    kshiftlw $6, %k2, %k2
 ; KNL-NEXT:    kshiftrw $15, %k2, %k2
-; KNL-NEXT:    kshiftlw $6, %k1, %k1
-; KNL-NEXT:    kshiftrw $15, %k1, %k1
-; KNL-NEXT:    kshiftlw $6, %k1, %k1
-; KNL-NEXT:    korw %k1, %k0, %k0
-; KNL-NEXT:    kshiftlw $7, %k2, %k1
-; KNL-NEXT:    korw %k1, %k0, %k1
 ; KNL-NEXT:    vpternlogd $255, %zmm0, %zmm0, %zmm0
+; KNL-NEXT:    vmovdqa64 %zmm0, %zmm1 {%k1} {z}
+; KNL-NEXT:    vmovdqa64 %zmm0, %zmm2 {%k2} {z}
+; KNL-NEXT:    vmovdqa64 {{.*#+}} zmm3 = [0,1,2,3,4,5,8,7]
+; KNL-NEXT:    vpermt2q %zmm2, %zmm3, %zmm1
+; KNL-NEXT:    vpsllq $63, %zmm1, %zmm1
+; KNL-NEXT:    vptestmq %zmm1, %zmm1, %k1
+; KNL-NEXT:    kshiftlw $1, %k1, %k1
+; KNL-NEXT:    kshiftrw $1, %k1, %k1
+; KNL-NEXT:    kshiftlw $7, %k0, %k0
+; KNL-NEXT:    kshiftrw $0, %k0, %k0
+; KNL-NEXT:    korw %k0, %k1, %k1
 ; KNL-NEXT:    vmovdqa64 %zmm0, %zmm0 {%k1} {z}
 ; KNL-NEXT:    vpmovqw %zmm0, %xmm0
 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: test18:
 ; SKX:       ## BB#0:
 ; SKX-NEXT:    kmovb %edi, %k0
 ; SKX-NEXT:    kmovw %esi, %k1
-; SKX-NEXT:    kshiftlw $6, %k1, %k2
+; SKX-NEXT:    kshiftlw $7, %k1, %k2
 ; SKX-NEXT:    kshiftrw $15, %k2, %k2
-; SKX-NEXT:    kshiftlw $7, %k1, %k1
+; SKX-NEXT:    kshiftlw $6, %k1, %k1
 ; SKX-NEXT:    kshiftrw $15, %k1, %k1
-; SKX-NEXT:    kshiftlb $7, %k1, %k1
-; SKX-NEXT:    kshiftlb $6, %k2, %k2
-; SKX-NEXT:    korb %k2, %k0, %k0
+; SKX-NEXT:    vpmovm2q %k0, %zmm0
+; SKX-NEXT:    vpmovm2q %k1, %zmm1
+; SKX-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [0,1,2,3,4,5,8,7]
+; SKX-NEXT:    vpermt2q %zmm1, %zmm2, %zmm0
+; SKX-NEXT:    vpmovq2m %zmm0, %k0
+; SKX-NEXT:    kshiftlb $1, %k0, %k0
+; SKX-NEXT:    kshiftrb $1, %k0, %k0
+; SKX-NEXT:    kshiftlb $7, %k2, %k1
+; SKX-NEXT:    kshiftrb $0, %k1, %k1
 ; SKX-NEXT:    korb %k1, %k0, %k0
 ; SKX-NEXT:    vpmovm2w %k0, %xmm0
 ; SKX-NEXT:    retq
   %b = bitcast i8 %a to <8 x i1>
   %b1 = bitcast i16 %y to <16 x i1>
   %el1 = extractelement <16 x i1>%b1, i32 8
   %el2 = extractelement <16 x i1>%b1, i32 9
   %c = insertelement <8 x i1>%b, i1 %el1, i32 7

test/CodeGen/X86/avx512-skx-insert-subvec.ll

 define <8 x i1> @test2(<2 x i1> %a) {
 ; CHECK-LABEL: test2:
 ; CHECK:       # BB#0:
 ; CHECK-NEXT:    vpsllq $63, %xmm0, %xmm0
 ; CHECK-NEXT:    vptestmq %xmm0, %xmm0, %k0
 ; CHECK-NEXT:    vpmovm2q %k0, %zmm0
 ; CHECK-NEXT:    vpxord %zmm1, %zmm1, %zmm1
 ; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm0 = zmm1[0,1,0,1],zmm0[0,1,0,1]
-; CHECK-NEXT:    vpsllq $63, %zmm0, %zmm0
-; CHECK-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; CHECK-NEXT:    vpmovq2m %zmm0, %k0
 ; CHECK-NEXT:    vpmovm2w %k0, %xmm0
 ; CHECK-NEXT:    retq
   %res = shufflevector <2 x i1> %a, <2 x i1> zeroinitializer, <8 x i32> <i32 3, i32 3, i32 undef, i32 undef, i32 0, i32 1, i32 undef, i32 undef>
   ret <8 x i1> %res
 }
 
 define <8 x i1> @test3(<4 x i1> %a) {
 ; CHECK-LABEL: test3:

test/CodeGen/X86/vector-shuffle-v1.ll

 ; AVX512F-NEXT:    retq
 ;
 ; VL_BW_DQ-LABEL: shuf2i1_1_0:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    vpsllq $63, %xmm0, %xmm0
 ; VL_BW_DQ-NEXT:    vptestmq %xmm0, %xmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %xmm0
 ; VL_BW_DQ-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
-; VL_BW_DQ-NEXT:    vpsllq $63, %xmm0, %xmm0
-; VL_BW_DQ-NEXT:    vptestmq %xmm0, %xmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %xmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %xmm0
 ; VL_BW_DQ-NEXT:    retq
   %b = shufflevector <2 x i1> %a, <2 x i1> undef, <2 x i32> <i32 1, i32 0>
   ret <2 x i1> %b
 }
 
 define <2 x i1> @shuf2i1_1_2(<2 x i1> %a) {
 ; AVX512F-LABEL: shuf2i1_1_2:
 ; AVX512F:       # BB#0:
 ; AVX512F-NEXT:    movl $1, %eax
 ; AVX512F-NEXT:    vmovq %rax, %xmm1
 ; AVX512F-NEXT:    vpalignr {{.*#+}} xmm0 = xmm0[8,9,10,11,12,13,14,15],xmm1[0,1,2,3,4,5,6,7]
 ; AVX512F-NEXT:    retq
 ;
 ; VL_BW_DQ-LABEL: shuf2i1_1_2:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    vpsllq $63, %xmm0, %xmm0
 ; VL_BW_DQ-NEXT:    vptestmq %xmm0, %xmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %xmm0
 ; VL_BW_DQ-NEXT:    movb $1, %al
 ; VL_BW_DQ-NEXT:    kmovb %eax, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %xmm1
 ; VL_BW_DQ-NEXT:    vpalignr {{.*#+}} xmm0 = xmm0[8,9,10,11,12,13,14,15],xmm1[0,1,2,3,4,5,6,7]
-; VL_BW_DQ-NEXT:    vpsllq $63, %xmm0, %xmm0
-; VL_BW_DQ-NEXT:    vptestmq %xmm0, %xmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %xmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %xmm0
 ; VL_BW_DQ-NEXT:    retq
   %b = shufflevector <2 x i1> %a, <2 x i1> <i1 1, i1 0>, <2 x i32> <i32 1, i32 2>
   ret <2 x i1> %b
 }
 
 
 define <4 x i1> @shuf4i1_3_2_10(<4 x i1> %a) {
 ; AVX512F-LABEL: shuf4i1_3_2_10:
 ; AVX512F:       # BB#0:
 ; AVX512F-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[3,2,1,0]
 ; AVX512F-NEXT:    retq
 ;
 ; VL_BW_DQ-LABEL: shuf4i1_3_2_10:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    vpslld $31, %xmm0, %xmm0
 ; VL_BW_DQ-NEXT:    vptestmd %xmm0, %xmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2d %k0, %xmm0
 ; VL_BW_DQ-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; VL_BW_DQ-NEXT:    vpslld $31, %xmm0, %xmm0
-; VL_BW_DQ-NEXT:    vptestmd %xmm0, %xmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovd2m %xmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2d %k0, %xmm0
 ; VL_BW_DQ-NEXT:    retq
   %b = shufflevector <4 x i1> %a, <4 x i1> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
   ret <4 x i1> %b
 }
 
 define <8 x i1> @shuf8i1_3_6_1_0_3_7_7_0(<8 x i64> %a, <8 x i64> %b, <8 x i64> %a1, <8 x i64> %b1) {
 ; AVX512F-LABEL: shuf8i1_3_6_1_0_3_7_7_0:

 ; AVX512F-NEXT:    retq
 ;
 ; VL_BW_DQ-LABEL: shuf8i1_3_6_1_0_3_7_7_0:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    vpcmpeqq %zmm2, %zmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [3,6,1,0,3,7,7,0]
 ; VL_BW_DQ-NEXT:    vpermq %zmm0, %zmm1, %zmm0
-; VL_BW_DQ-NEXT:    vpsllq $63, %zmm0, %zmm0
-; VL_BW_DQ-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2w %k0, %xmm0
 ; VL_BW_DQ-NEXT:    retq
   %a2 = icmp eq <8 x i64> %a, %a1
   %b2 = icmp eq <8 x i64> %b, %b1
   %c = shufflevector <8 x i1> %a2, <8 x i1> %b2, <8 x i32> <i32 3, i32 6, i32 1, i32 0, i32 3, i32 7, i32 7, i32 0>
   ret <8 x i1> %c
 }
 

 ; VL_BW_DQ-LABEL: shuf16i1_3_6_22_12_3_7_7_0_3_6_1_13_3_21_7_0:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    vpcmpeqd %zmm2, %zmm0, %k0
 ; VL_BW_DQ-NEXT:    vpcmpeqd %zmm3, %zmm1, %k1
 ; VL_BW_DQ-NEXT:    vpmovm2d %k1, %zmm0
 ; VL_BW_DQ-NEXT:    vpmovm2d %k0, %zmm1
 ; VL_BW_DQ-NEXT:    vmovdqa32 {{.*#+}} zmm2 = [3,6,22,12,3,7,7,0,3,6,1,13,3,21,7,0]
 ; VL_BW_DQ-NEXT:    vpermt2d %zmm0, %zmm2, %zmm1
-; VL_BW_DQ-NEXT:    vpslld $31, %zmm1, %zmm0
-; VL_BW_DQ-NEXT:    vptestmd %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovd2m %zmm1, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2b %k0, %xmm0
 ; VL_BW_DQ-NEXT:    retq
   %a2 = icmp eq <16 x i32> %a, %a1
   %b2 = icmp eq <16 x i32> %b, %b1
   %c = shufflevector <16 x i1> %a2, <16 x i1> %b2, <16 x i32> <i32 3, i32 6, i32 22, i32 12, i32 3, i32 7, i32 7, i32 0, i32 3, i32 6, i32 1, i32 13, i32 3, i32 21, i32 7, i32 0>
   ret <16 x i1> %c
 }
 

 ;
 ; VL_BW_DQ-LABEL: shuf32i1_3_6_22_12_3_7_7_0_3_6_1_13_3_21_7_0_3_6_22_12_3_7_7_0_3_6_1_13_3_21_7_0:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    vpsllw $7, %ymm0, %ymm0
 ; VL_BW_DQ-NEXT:    vpmovb2m %ymm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2w %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vmovdqu16 {{.*#+}} zmm1 = [3,6,22,12,3,7,7,0,3,6,1,13,3,21,7,0,3,6,22,12,3,7,7,0,3,6,1,13,3,21,7,0]
 ; VL_BW_DQ-NEXT:    vpermw %zmm0, %zmm1, %zmm0
-; VL_BW_DQ-NEXT:    vpsllw $15, %zmm0, %zmm0
 ; VL_BW_DQ-NEXT:    vpmovw2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2b %k0, %ymm0
 ; VL_BW_DQ-NEXT:    retq
   %b = shufflevector <32 x i1> %a, <32 x i1> undef, <32 x i32> <i32 3, i32 6, i32 22, i32 12, i32 3, i32 7, i32 7, i32 0, i32 3, i32 6, i32 1, i32 13, i32 3, i32 21, i32 7, i32 0, i32 3, i32 6, i32 22, i32 12, i32 3, i32 7, i32 7, i32 0, i32 3, i32 6, i32 1, i32 13, i32 3, i32 21, i32 7, i32 0>
   ret <32 x i1> %b
 }
 
 define <8 x i1> @shuf8i1_u_2_u_u_2_u_2_u(i8 %a) {

 ; AVX512F-NEXT:    retq
 ;
 ; VL_BW_DQ-LABEL: shuf8i1_u_2_u_u_2_u_2_u:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    kmovb %edi, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vextracti64x2 $1, %zmm0, %xmm0
 ; VL_BW_DQ-NEXT:    vpbroadcastq %xmm0, %zmm0
-; VL_BW_DQ-NEXT:    vpsllq $63, %zmm0, %zmm0
-; VL_BW_DQ-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2w %k0, %xmm0
 ; VL_BW_DQ-NEXT:    retq
   %b = bitcast i8 %a to <8 x i1>
   %c = shufflevector < 8 x i1> %b, <8 x i1>undef, <8 x i32> <i32 undef, i32 2, i32 undef, i32 undef, i32 2, i32 undef, i32 2, i32 undef>
   ret <8 x i1> %c
 }
 
 define i8 @shuf8i1_10_2_9_u_3_u_2_u(i8 %a) {

 ;
 ; VL_BW_DQ-LABEL: shuf8i1_10_2_9_u_3_u_2_u:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    kmovb %edi, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vpxord %zmm1, %zmm1, %zmm1
 ; VL_BW_DQ-NEXT:    vmovdqa64 {{.*#+}} zmm2 = <8,2,10,u,3,u,2,u>
 ; VL_BW_DQ-NEXT:    vpermt2q %zmm1, %zmm2, %zmm0
-; VL_BW_DQ-NEXT:    vpsllq $63, %zmm0, %zmm0
-; VL_BW_DQ-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    kmovb %k0, %eax
 ; VL_BW_DQ-NEXT:    retq
   %b = bitcast i8 %a to <8 x i1>
   %c = shufflevector < 8 x i1> %b, <8 x i1> zeroinitializer, <8 x i32> <i32 10, i32 2, i32 9, i32 undef, i32 3, i32 undef, i32 2, i32 undef>
   %d = bitcast <8 x i1> %c to i8
   ret i8 %d
 }
 

 ; AVX512F-NEXT:    kmovw %k0, %eax
 ; AVX512F-NEXT:    retq
 ;
 ; VL_BW_DQ-LABEL: shuf8i1_0_1_4_5_u_u_u_u:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    kmovb %edi, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,4,5,0,1,0,1]
-; VL_BW_DQ-NEXT:    vpsllq $63, %zmm0, %zmm0
-; VL_BW_DQ-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    kmovb %k0, %eax
 ; VL_BW_DQ-NEXT:    retq
   %b = bitcast i8 %a to <8 x i1>
   %c = shufflevector < 8 x i1> %b, <8 x i1> undef, <8 x i32> <i32 0, i32 1, i32 4, i32 5, i32 undef, i32 undef, i32 undef, i32 undef>
   %d = bitcast <8 x i1> %c to i8
   ret i8 %d
 }
 

 ;
 ; VL_BW_DQ-LABEL: shuf8i1_9_6_1_0_3_7_7_0:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    kmovb %edi, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vpxord %zmm1, %zmm1, %zmm1
 ; VL_BW_DQ-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [8,6,1,0,3,7,7,0]
 ; VL_BW_DQ-NEXT:    vpermt2q %zmm1, %zmm2, %zmm0
-; VL_BW_DQ-NEXT:    vpsllq $63, %zmm0, %zmm0
-; VL_BW_DQ-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    kmovb %k0, %eax
 ; VL_BW_DQ-NEXT:    retq
   %b = bitcast i8 %a to <8 x i1>
   %c = shufflevector <8 x i1> %b, <8 x i1> zeroinitializer, <8 x i32> <i32 9, i32 6, i32 1, i32 0, i32 3, i32 7, i32 7, i32 0>
   %d = bitcast <8 x i1>%c to i8
   ret i8 %d
 }
 

 ;
 ; VL_BW_DQ-LABEL: shuf8i1_9_6_1_10_3_7_7_0:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    kmovb %edi, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [9,1,2,10,4,5,6,7]
 ; VL_BW_DQ-NEXT:    vpxord %zmm2, %zmm2, %zmm2
 ; VL_BW_DQ-NEXT:    vpermt2q %zmm0, %zmm1, %zmm2
-; VL_BW_DQ-NEXT:    vpsllq $63, %zmm2, %zmm0
-; VL_BW_DQ-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %zmm2, %k0
 ; VL_BW_DQ-NEXT:    kmovb %k0, %eax
 ; VL_BW_DQ-NEXT:    retq
   %b = bitcast i8 %a to <8 x i1>
   %c = shufflevector <8 x i1> zeroinitializer, <8 x i1> %b, <8 x i32> <i32 9, i32 6, i32 1, i32 10, i32 3, i32 7, i32 7, i32 0>
   %d = bitcast <8 x i1>%c to i8
   ret i8 %d
 }
 

 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    kmovb %edi, %k0
 ; VL_BW_DQ-NEXT:    movb $51, %al
 ; VL_BW_DQ-NEXT:    kmovb %eax, %k1
 ; VL_BW_DQ-NEXT:    vpmovm2q %k1, %zmm0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %zmm1
 ; VL_BW_DQ-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [9,6,1,0,3,7,7,1]
 ; VL_BW_DQ-NEXT:    vpermt2q %zmm1, %zmm2, %zmm0
-; VL_BW_DQ-NEXT:    vpsllq $63, %zmm0, %zmm0
-; VL_BW_DQ-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    kmovb %k0, %eax
 ; VL_BW_DQ-NEXT:    retq
   %b = bitcast i8 %a to <8 x i1>
   %c = shufflevector <8 x i1> <i1 1, i1 1, i1 0, i1 0, i1 1, i1 1, i1 0, i1 0>, <8 x i1> %b, <8 x i32> <i32 9, i32 6, i32 1, i32 0, i32 3, i32 7, i32 7, i32 1>
   %c1 = bitcast <8 x i1>%c to i8
   ret i8 %c1
 }
 

 ; VL_BW_DQ-LABEL: shuf8i1_9_6_1_10_3_7_7_0_all_ones:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    vpsllw $15, %xmm0, %xmm0
 ; VL_BW_DQ-NEXT:    vpmovw2m %xmm0, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2q %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [9,1,2,3,4,5,6,7]
 ; VL_BW_DQ-NEXT:    vpternlogd $255, %zmm2, %zmm2, %zmm2
 ; VL_BW_DQ-NEXT:    vpermt2q %zmm0, %zmm1, %zmm2
-; VL_BW_DQ-NEXT:    vpsllq $63, %zmm2, %zmm0
-; VL_BW_DQ-NEXT:    vptestmq %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovq2m %zmm2, %k0
 ; VL_BW_DQ-NEXT:    kmovb %k0, %eax
 ; VL_BW_DQ-NEXT:    retq
   %c = shufflevector <8 x i1> <i1 1, i1 1, i1 1, i1 1, i1 1, i1 1, i1 1, i1 1>, <8 x i1> %a, <8 x i32> <i32 9, i32 6, i32 1, i32 0, i32 3, i32 7, i32 7, i32 0>
   %c1 = bitcast <8 x i1>%c to i8
   ret i8 %c1
 }
 
 

 ; AVX512F-NEXT:    kmovw %k0, %eax
 ; AVX512F-NEXT:    retq
 ;
 ; VL_BW_DQ-LABEL: shuf16i1_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0_0:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    kmovw %edi, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2d %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vpbroadcastd %xmm0, %zmm0
-; VL_BW_DQ-NEXT:    vpslld $31, %zmm0, %zmm0
-; VL_BW_DQ-NEXT:    vptestmd %zmm0, %zmm0, %k0
+; VL_BW_DQ-NEXT:    vpmovd2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    kmovw %k0, %eax
 ; VL_BW_DQ-NEXT:    retq
   %b = bitcast i16 %a to <16 x i1>
   %c = shufflevector < 16 x i1> %b, <16 x i1> undef, <16 x i32> zeroinitializer
   %d = bitcast <16 x i1> %c to i16
   ret i16 %d
 }
 

 ; AVX512F-NEXT:    popq %rbp
 ; AVX512F-NEXT:    retq
 ;
 ; VL_BW_DQ-LABEL: shuf64i1_zero:
 ; VL_BW_DQ:       # BB#0:
 ; VL_BW_DQ-NEXT:    kmovq %rdi, %k0
 ; VL_BW_DQ-NEXT:    vpmovm2b %k0, %zmm0
 ; VL_BW_DQ-NEXT:    vpbroadcastb %xmm0, %zmm0
-; VL_BW_DQ-NEXT:    vpsllw $7, %zmm0, %zmm0
 ; VL_BW_DQ-NEXT:    vpmovb2m %zmm0, %k0
 ; VL_BW_DQ-NEXT:    kmovq %k0, %rax
 ; VL_BW_DQ-NEXT:    retq
   %b = bitcast i64 %a to <64 x i1>
   %c = shufflevector < 64 x i1> %b, <64 x i1> undef, <64 x i32> zeroinitializer
   %d = bitcast <64 x i1> %c to i64
   ret i64 %d
 }