[AVX512] Remove and autoupgrade many of the broadcast intrinsics

[craig.topper]

Summary

This autoupgrades most of the broadcast intrinsics. They've been unused in clang for some time.

This leaves the 32x2 intrinsics because they are still used in clang.

[RKSimon]

LGTM

Do you think we should make SUBV_BROADCAST a memory intrinsic? Maybe BROADCAST as well, although that has reg implementations and splitting it into two may be awkward.

[craig.topper]

I think I remember having that thought before. It's been a while. I actually found most of this patch in a "git stash" from quite a while ago.

llvm/trunk/include/llvm/IR/IntrinsicsX86.td

def int_x86_avx512_mask_broadcasti32x2_256 :
           Intrinsic<[llvm_v8i32_ty],
                     [llvm_v4i32_ty, llvm_v8i32_ty, llvm_i8_ty], [IntrNoMem]>;
 
   def int_x86_avx512_mask_broadcasti32x2_512 :
           GCCBuiltin<"__builtin_ia32_broadcasti32x2_512_mask">,
           Intrinsic<[llvm_v16i32_ty],
                     [llvm_v4i32_ty, llvm_v16i32_ty, llvm_i16_ty], [IntrNoMem]>;
 
-  def int_x86_avx512_mask_broadcastf32x4_256 :
-            GCCBuiltin<"__builtin_ia32_broadcastf32x4_256_mask">,
-            Intrinsic<[llvm_v8f32_ty],
-                    [llvm_v4f32_ty, llvm_v8f32_ty, llvm_i8_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcastf32x4_512 :
-          GCCBuiltin<"__builtin_ia32_broadcastf32x4_512">,
-          Intrinsic<[llvm_v16f32_ty],
-                    [llvm_v4f32_ty, llvm_v16f32_ty, llvm_i16_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcastf32x8_512 :
-          GCCBuiltin<"__builtin_ia32_broadcastf32x8_512_mask">,
-          Intrinsic<[llvm_v16f32_ty],
-                    [llvm_v8f32_ty, llvm_v16f32_ty, llvm_i16_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcastf64x2_256 :
-          GCCBuiltin<"__builtin_ia32_broadcastf64x2_256_mask">,
-          Intrinsic<[llvm_v4f64_ty],
-                    [llvm_v2f64_ty, llvm_v4f64_ty, llvm_i8_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcastf64x2_512 :
-          GCCBuiltin<"__builtin_ia32_broadcastf64x2_512_mask">,
-          Intrinsic<[llvm_v8f64_ty],
-                    [llvm_v2f64_ty, llvm_v8f64_ty, llvm_i8_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcastf64x4_512 :
-          GCCBuiltin<"__builtin_ia32_broadcastf64x4_512">,
-          Intrinsic<[llvm_v8f64_ty],
-                    [llvm_v4f64_ty, llvm_v8f64_ty, llvm_i8_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcasti32x4_256 :
-          GCCBuiltin<"__builtin_ia32_broadcasti32x4_256_mask">,
-          Intrinsic<[llvm_v8i32_ty],
-                    [llvm_v4i32_ty, llvm_v8i32_ty, llvm_i8_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcasti32x4_512 :
-          GCCBuiltin<"__builtin_ia32_broadcasti32x4_512">,
-          Intrinsic<[llvm_v16i32_ty],
-                    [llvm_v4i32_ty, llvm_v16i32_ty, llvm_i16_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcasti32x8_512 :
-          GCCBuiltin<"__builtin_ia32_broadcasti32x8_512_mask">,
-          Intrinsic<[llvm_v16i32_ty],
-                    [llvm_v8i32_ty, llvm_v16i32_ty, llvm_i16_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcasti64x2_256 :
-          GCCBuiltin<"__builtin_ia32_broadcasti64x2_256_mask">,
-          Intrinsic<[llvm_v4i64_ty],
-                    [llvm_v2i64_ty, llvm_v4i64_ty, llvm_i8_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcasti64x2_512 :
-          GCCBuiltin<"__builtin_ia32_broadcasti64x2_512_mask">,
-          Intrinsic<[llvm_v8i64_ty],
-                    [llvm_v2i64_ty, llvm_v8i64_ty, llvm_i8_ty], [IntrNoMem]>;
-
-  def int_x86_avx512_mask_broadcasti64x4_512 :
-          GCCBuiltin<"__builtin_ia32_broadcasti64x4_512">,
-          Intrinsic<[llvm_v8i64_ty],
-                    [llvm_v4i64_ty, llvm_v8i64_ty, llvm_i8_ty], [IntrNoMem]>;
-
    def int_x86_avx512_broadcastmw_512 :
           GCCBuiltin<"__builtin_ia32_broadcastmw512">,
           Intrinsic<[llvm_v16i32_ty], [llvm_i16_ty], [IntrNoMem]>;
    def int_x86_avx512_broadcastmw_256 :
           GCCBuiltin<"__builtin_ia32_broadcastmw256">,
           Intrinsic<[llvm_v8i32_ty], [llvm_i16_ty], [IntrNoMem]>;
    def int_x86_avx512_broadcastmw_128 :
           GCCBuiltin<"__builtin_ia32_broadcastmw128">,

llvm/trunk/lib/IR/AutoUpgrade.cpp

if (Name.startswith("sse2.pcmpeq.") || // Added in 3.1
       Name.startswith("avx512.psrl.dq") || // Added in 3.9
       Name == "sse41.pblendw" || // Added in 3.7
       Name.startswith("sse41.blendp") || // Added in 3.7
       Name.startswith("avx.blend.p") || // Added in 3.7
       Name == "avx2.pblendw" || // Added in 3.7
       Name.startswith("avx2.pblendd.") || // Added in 3.7
       Name.startswith("avx.vbroadcastf128") || // Added in 4.0
       Name == "avx2.vbroadcasti128" || // Added in 3.7
+      Name.startswith("avx512.mask.broadcastf32x4.") || // Added in 6.0
+      Name.startswith("avx512.mask.broadcastf64x2.") || // Added in 6.0
+      Name.startswith("avx512.mask.broadcasti32x4.") || // Added in 6.0
+      Name.startswith("avx512.mask.broadcasti64x2.") || // Added in 6.0
+      Name == "avx512.mask.broadcastf32x8.512" || // Added in 6.0
+      Name == "avx512.mask.broadcasti32x8.512" || // Added in 6.0
+      Name == "avx512.mask.broadcastf64x4.512" || // Added in 6.0
+      Name == "avx512.mask.broadcasti64x4.512" || // Added in 6.0
       Name == "xop.vpcmov" || // Added in 3.8
       Name == "xop.vpcmov.256" || // Added in 5.0
       Name.startswith("avx512.mask.move.s") || // Added in 4.0
       Name.startswith("avx512.cvtmask2") || // Added in 5.0
       (Name.startswith("xop.vpcom") && // Added in 3.2
        F->arg_size() == 2))
     return true;
 
if (IsX86 && (Name.startswith("sse2.pcmp") ||
                                             PointerType::getUnqual(VT));
       Value *Load = Builder.CreateAlignedLoad(Op, 1);
       if (NumSrcElts == 2)
         Rep = Builder.CreateShuffleVector(Load, UndefValue::get(Load->getType()),
                                           { 0, 1, 0, 1 });
       else
         Rep = Builder.CreateShuffleVector(Load, UndefValue::get(Load->getType()),
                                           { 0, 1, 2, 3, 0, 1, 2, 3 });
+    } else if (IsX86 && (Name.startswith("avx512.mask.broadcastf") ||
+                         Name.startswith("avx512.mask.broadcasti"))) {
+      unsigned NumSrcElts =
+                        CI->getArgOperand(0)->getType()->getVectorNumElements();
+      unsigned NumDstElts = CI->getType()->getVectorNumElements();
+
+      SmallVector<uint32_t, 8> ShuffleMask(NumDstElts);
+      for (unsigned i = 0; i != NumDstElts; ++i)
+        ShuffleMask[i] = i % NumSrcElts;
+
+      Rep = Builder.CreateShuffleVector(CI->getArgOperand(0),
+                                        CI->getArgOperand(0),
+                                        ShuffleMask);
+      Rep = EmitX86Select(Builder, CI->getArgOperand(2), Rep,
+                          CI->getArgOperand(1));
     } else if (IsX86 && (Name.startswith("avx2.pbroadcast") ||
                          Name.startswith("avx2.vbroadcast") ||
                          Name.startswith("avx512.pbroadcast") ||
                          Name.startswith("avx512.mask.broadcast.s"))) {
       // Replace vp?broadcasts with a vector shuffle.
       Value *Op = CI->getArgOperand(0);
       unsigned NumElts = CI->getType()->getVectorNumElements();
       Type *MaskTy = VectorType::get(Type::getInt32Ty(C), NumElts);

llvm/trunk/lib/Target/X86/X86ISelLowering.cpp

case CONVERT_TO_MASK: {
 
       SDValue CvtMask = DAG.getNode(IntrData->Opc0, dl, MaskVT,
                                     Op.getOperand(1));
       SDValue Res = DAG.getNode(ISD::INSERT_SUBVECTOR, dl, BitcastVT,
                                 DAG.getUNDEF(BitcastVT), CvtMask,
                                 DAG.getIntPtrConstant(0, dl));
       return DAG.getBitcast(Op.getValueType(), Res);
     }
-    case BRCST_SUBVEC_TO_VEC: {
-      SDValue Src = Op.getOperand(1);
-      SDValue Passthru = Op.getOperand(2);
-      SDValue Mask = Op.getOperand(3);
-      EVT resVT = Passthru.getValueType();
-      SDValue subVec = DAG.getNode(ISD::INSERT_SUBVECTOR, dl, resVT,
-                                       DAG.getUNDEF(resVT), Src,
-                                       DAG.getIntPtrConstant(0, dl));
-      SDValue immVal;
-      if (Src.getSimpleValueType().is256BitVector() && resVT.is512BitVector())
-        immVal = DAG.getConstant(0x44, dl, MVT::i8);
-      else
-        immVal = DAG.getConstant(0, dl, MVT::i8);
-      return getVectorMaskingNode(DAG.getNode(IntrData->Opc0, dl, VT,
-                                              subVec, subVec, immVal),
-                                  Mask, Passthru, Subtarget, DAG);
-    }
     case BRCST32x2_TO_VEC: {
       SDValue Src = Op.getOperand(1);
       SDValue PassThru = Op.getOperand(2);
       SDValue Mask = Op.getOperand(3);
 
       assert((VT.getScalarType() == MVT::i32 ||
               VT.getScalarType() == MVT::f32) && "Unexpected type!");
       //bitcast Src to packed 64

llvm/trunk/lib/Target/X86/X86IntrinsicsInfo.h

enum IntrinsicType : uint16_t {
   CVTPD2PS, CVTPD2PS_MASK,
   INTR_TYPE_1OP_MASK, INTR_TYPE_1OP_MASK_RM,
   INTR_TYPE_2OP_MASK, INTR_TYPE_2OP_MASK_RM, INTR_TYPE_2OP_IMM8_MASK,
   INTR_TYPE_3OP_MASK, INTR_TYPE_3OP_MASK_RM, INTR_TYPE_3OP_IMM8_MASK,
   FMA_OP_MASK, FMA_OP_MASKZ, FMA_OP_MASK3,
   FMA_OP_SCALAR_MASK, FMA_OP_SCALAR_MASKZ, FMA_OP_SCALAR_MASK3,
   VPERM_2OP_MASK, VPERM_3OP_MASK, VPERM_3OP_MASKZ, INTR_TYPE_SCALAR_MASK,
   INTR_TYPE_SCALAR_MASK_RM, INTR_TYPE_3OP_SCALAR_MASK_RM,
-  COMPRESS_EXPAND_IN_REG, COMPRESS_TO_MEM, BRCST_SUBVEC_TO_VEC, BRCST32x2_TO_VEC,
+  COMPRESS_EXPAND_IN_REG, COMPRESS_TO_MEM, BRCST32x2_TO_VEC,
   TRUNCATE_TO_MEM_VI8, TRUNCATE_TO_MEM_VI16, TRUNCATE_TO_MEM_VI32,
   EXPAND_FROM_MEM,
   TERLOG_OP_MASK, TERLOG_OP_MASKZ, BROADCASTM, KUNPCK, FIXUPIMM, FIXUPIMM_MASKZ, FIXUPIMMS,
   FIXUPIMMS_MASKZ, CONVERT_TO_MASK, GATHER_AVX2, MASK_BINOP,
 };
 
 struct IntrinsicData {
 
static const IntrinsicData  IntrinsicsWithoutChain[] = {
   X86_INTRINSIC_DATA(avx512_mask_add_sd_round, INTR_TYPE_SCALAR_MASK_RM,
                      X86ISD::FADDS_RND, 0),
   X86_INTRINSIC_DATA(avx512_mask_add_ss_round, INTR_TYPE_SCALAR_MASK_RM,
                      X86ISD::FADDS_RND, 0),
   X86_INTRINSIC_DATA(avx512_mask_broadcastf32x2_256, BRCST32x2_TO_VEC,
                      X86ISD::VBROADCAST, 0),
   X86_INTRINSIC_DATA(avx512_mask_broadcastf32x2_512, BRCST32x2_TO_VEC,
                      X86ISD::VBROADCAST, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcastf32x4_256, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcastf32x4_512, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcastf32x8_512, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcastf64x2_256, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcastf64x2_512, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcastf64x4_512, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
   X86_INTRINSIC_DATA(avx512_mask_broadcasti32x2_128, BRCST32x2_TO_VEC,
                      X86ISD::VBROADCAST, 0),
   X86_INTRINSIC_DATA(avx512_mask_broadcasti32x2_256, BRCST32x2_TO_VEC,
                      X86ISD::VBROADCAST, 0),
   X86_INTRINSIC_DATA(avx512_mask_broadcasti32x2_512, BRCST32x2_TO_VEC,
                      X86ISD::VBROADCAST, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcasti32x4_256, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcasti32x4_512, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcasti32x8_512, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcasti64x2_256, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcasti64x2_512, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
-  X86_INTRINSIC_DATA(avx512_mask_broadcasti64x4_512, BRCST_SUBVEC_TO_VEC,
-                     X86ISD::SHUF128, 0),
   X86_INTRINSIC_DATA(avx512_mask_cmp_pd_128, CMP_MASK_CC, X86ISD::CMPM, 0),
   X86_INTRINSIC_DATA(avx512_mask_cmp_pd_256, CMP_MASK_CC, X86ISD::CMPM, 0),
   X86_INTRINSIC_DATA(avx512_mask_cmp_pd_512, CMP_MASK_CC, X86ISD::CMPM,
                      X86ISD::CMPM_RND),
   X86_INTRINSIC_DATA(avx512_mask_cmp_ps_128, CMP_MASK_CC, X86ISD::CMPM, 0),
   X86_INTRINSIC_DATA(avx512_mask_cmp_ps_256, CMP_MASK_CC, X86ISD::CMPM, 0),
   X86_INTRINSIC_DATA(avx512_mask_cmp_ps_512, CMP_MASK_CC, X86ISD::CMPM,
                      X86ISD::CMPM_RND),

llvm/trunk/test/CodeGen/X86/avx512-intrinsics-upgrade.ll

; CHECK-NEXT:    retq
   %res6 = call i8 @llvm.x86.avx512.mask.ucmp.q.512(<8 x i64> %a0, <8 x i64> %a1, i32 6, i8 %mask)
   %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
   %res7 = call i8 @llvm.x86.avx512.mask.ucmp.q.512(<8 x i64> %a0, <8 x i64> %a1, i32 7, i8 %mask)
   %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
   ret <8 x i8> %vec7
 }
 
 declare i8 @llvm.x86.avx512.mask.ucmp.q.512(<8 x i64>, <8 x i64>, i32, i8) nounwind readnone
+
+declare <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float>, <16 x float>, i16)
+
+define <16 x float>@test_int_x86_avx512_mask_broadcastf32x4_512(<4 x float> %x0, <16 x float> %x2, i16 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
+; CHECK-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; CHECK-NEXT:    vinsertf64x4 $1, %ymm0, %zmm0, %zmm0
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vmovaps %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vmovaps %zmm0, %zmm2 {%k1} {z}
+; CHECK-NEXT:    vaddps %zmm1, %zmm0, %zmm0
+; CHECK-NEXT:    vaddps %zmm0, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+
+  %res1 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float> %x0, <16 x float> %x2, i16 -1)
+  %res2 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float> %x0, <16 x float> %x2, i16 %mask)
+  %res3 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float> %x0, <16 x float> zeroinitializer, i16 %mask)
+  %res4 = fadd <16 x float> %res1, %res2
+  %res5 = fadd <16 x float> %res3, %res4
+  ret <16 x float> %res5
+}
+
+define <16 x float>@test_int_x86_avx512_mask_broadcastf32x4_512_load(<4 x float>* %x0ptr, <16 x float> %x2, i16 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_512_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vbroadcastf32x4 {{.*#+}} zmm0 {%k1} = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
+; CHECK-NEXT:    retq
+  %x0 = load <4 x float>, <4 x float>* %x0ptr
+  %res = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float> %x0, <16 x float> %x2, i16 %mask)
+  ret <16 x float> %res
+}
+
+declare <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double>, <8 x double>, i8)
+
+define <8 x double>@test_int_x86_avx512_mask_broadcastf64x4_512(<4 x double> %x0, <8 x double> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x4_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %YMM0<def> %YMM0<kill> %ZMM0<def>
+; CHECK-NEXT:    vinsertf64x4 $1, %ymm0, %zmm0, %zmm2
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vinsertf64x4 $1, %ymm0, %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vaddpd %zmm1, %zmm2, %zmm1
+; CHECK-NEXT:    vinsertf64x4 $1, %ymm0, %zmm0, %zmm0 {%k1} {z}
+; CHECK-NEXT:    vaddpd %zmm1, %zmm0, %zmm0
+; CHECK-NEXT:    retq
+
+  %res1 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double> %x0, <8 x double> %x2, i8 -1)
+  %res2 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double> %x0, <8 x double> %x2, i8 %mask)
+  %res3 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double> %x0, <8 x double> zeroinitializer, i8 %mask)
+  %res4 = fadd <8 x double> %res1, %res2
+  %res5 = fadd <8 x double> %res3, %res4
+  ret <8 x double> %res5
+}
+
+define <8 x double>@test_int_x86_avx512_mask_broadcastf64x4_512_load(<4 x double>* %x0ptr, <8 x double> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x4_512_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vbroadcastf64x4 {{.*#+}} zmm0 {%k1} = mem[0,1,2,3,0,1,2,3]
+; CHECK-NEXT:    retq
+
+  %x0 = load <4 x double>, <4 x double>* %x0ptr
+  %res = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double> %x0, <8 x double> %x2, i8 %mask)
+  ret <8 x double> %res
+}
+
+declare <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32>, <16 x i32>, i16)
+
+define <16 x i32>@test_int_x86_avx512_mask_broadcasti32x4_512(<4 x i32> %x0, <16 x i32> %x2, i16 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
+; CHECK-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; CHECK-NEXT:    vinserti64x4 $1, %ymm0, %zmm0, %zmm0
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vmovdqa32 %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vmovdqa32 %zmm0, %zmm2 {%k1} {z}
+; CHECK-NEXT:    vpaddd %zmm1, %zmm0, %zmm0
+; CHECK-NEXT:    vpaddd %zmm0, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+
+  %res1 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32> %x0, <16 x i32> %x2, i16 -1)
+  %res2 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32> %x0, <16 x i32> %x2, i16 %mask)
+  %res3 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32> %x0, <16 x i32> zeroinitializer, i16 %mask)
+  %res4 = add <16 x i32> %res1, %res2
+  %res5 = add <16 x i32> %res3, %res4
+  ret <16 x i32> %res5
+}
+
+define <16 x i32>@test_int_x86_avx512_mask_broadcasti32x4_512_load(<4 x i32>* %x0ptr, <16 x i32> %x2, i16 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_512_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vbroadcasti32x4 {{.*#+}} zmm0 {%k1} = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
+; CHECK-NEXT:    retq
+
+  %x0 = load <4 x i32>, <4 x i32>* %x0ptr
+  %res = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32> %x0, <16 x i32> %x2, i16 %mask)
+  ret <16 x i32> %res
+}
+
+declare <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64>, <8 x i64>, i8)
+
+define <8 x i64>@test_int_x86_avx512_mask_broadcasti64x4_512(<4 x i64> %x0, <8 x i64> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x4_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %YMM0<def> %YMM0<kill> %ZMM0<def>
+; CHECK-NEXT:    vinserti64x4 $1, %ymm0, %zmm0, %zmm2
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vinserti64x4 $1, %ymm0, %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vpaddq %zmm1, %zmm2, %zmm1
+; CHECK-NEXT:    vinserti64x4 $1, %ymm0, %zmm0, %zmm0 {%k1} {z}
+; CHECK-NEXT:    vpaddq %zmm1, %zmm0, %zmm0
+; CHECK-NEXT:    retq
+
+  %res1 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64> %x0, <8 x i64> %x2, i8 -1)
+  %res2 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64> %x0, <8 x i64> %x2, i8 %mask)
+  %res3 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64> %x0, <8 x i64> zeroinitializer, i8 %mask)
+  %res4 = add <8 x i64> %res1, %res2
+  %res5 = add <8 x i64> %res3, %res4
+  ret <8 x i64> %res5
+}
+
+define <8 x i64>@test_int_x86_avx512_mask_broadcasti64x4_512_load(<4 x i64>* %x0ptr, <8 x i64> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x4_512_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vbroadcasti64x4 {{.*#+}} zmm0 {%k1} = mem[0,1,2,3,0,1,2,3]
+; CHECK-NEXT:    retq
+
+  %x0 = load <4 x i64>, <4 x i64>* %x0ptr
+  %res = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64> %x0, <8 x i64> %x2, i8 %mask)
+  ret <8 x i64> %res
+}

llvm/trunk/test/CodeGen/X86/avx512-intrinsics.ll

 ; CHECK-NEXT:    kmovw %k0, %eax
 ; CHECK-NEXT:    retq
   %res = call i32 @llvm.x86.avx512.vcomi.ss(<4 x float> %a0, <4 x float> %a1, i32 9, i32 4)
   ret i32 %res
 }
 
 declare i32 @llvm.x86.avx512.vcomi.ss(<4 x float>, <4 x float>, i32, i32)
 
-declare <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float>, <16 x float>, i16)
-
-define <16 x float>@test_int_x86_avx512_mask_broadcastf32x4_512(<4 x float> %x0, <16 x float> %x2, i16 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_512:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %ZMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vshuff32x4 {{.*#+}} zmm2 {%k1} {z} = zmm0[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshuff32x4 {{.*#+}} zmm1 {%k1} = zmm0[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshuff32x4 {{.*#+}} zmm0 = zmm0[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vaddps %zmm1, %zmm0, %zmm0
-; CHECK-NEXT:    vaddps %zmm0, %zmm2, %zmm0
-; CHECK-NEXT:    retq
-
-  %res1 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float> %x0, <16 x float> %x2, i16 -1)
-  %res2 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float> %x0, <16 x float> %x2, i16 %mask)
-  %res3 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float> %x0, <16 x float> zeroinitializer, i16 %mask)
-  %res4 = fadd <16 x float> %res1, %res2
-  %res5 = fadd <16 x float> %res3, %res4
-  ret <16 x float> %res5
-}
-
-define <16 x float>@test_int_x86_avx512_mask_broadcastf32x4_512_load(<4 x float>* %x0ptr, <16 x float> %x2, i16 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_512_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1
-; CHECK-NEXT:    vmovaps (%rdi), %xmm1
-; CHECK-NEXT:    vshuff32x4 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    retq
-  %x0 = load <4 x float>, <4 x float>* %x0ptr
-  %res = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x4.512(<4 x float> %x0, <16 x float> %x2, i16 %mask)
-  ret <16 x float> %res
-}
-
-declare <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double>, <8 x double>, i8)
-
-define <8 x double>@test_int_x86_avx512_mask_broadcastf64x4_512(<4 x double> %x0, <8 x double> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x4_512:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %YMM0<def> %YMM0<kill> %ZMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vshuff64x2 {{.*#+}} zmm2 {%k1} {z} = zmm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshuff64x2 {{.*#+}} zmm1 {%k1} = zmm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshuff64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vaddpd %zmm1, %zmm0, %zmm0
-; CHECK-NEXT:    vaddpd %zmm0, %zmm2, %zmm0
-; CHECK-NEXT:    retq
-
-  %res1 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double> %x0, <8 x double> %x2, i8 -1)
-  %res2 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double> %x0, <8 x double> %x2, i8 %mask)
-  %res3 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double> %x0, <8 x double> zeroinitializer, i8 %mask)
-  %res4 = fadd <8 x double> %res1, %res2
-  %res5 = fadd <8 x double> %res3, %res4
-  ret <8 x double> %res5
-}
-
-define <8 x double>@test_int_x86_avx512_mask_broadcastf64x4_512_load(<4 x double>* %x0ptr, <8 x double> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x4_512_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1
-; CHECK-NEXT:    vmovapd (%rdi), %ymm1
-; CHECK-NEXT:    vshuff64x2 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    retq
-
-  %x0 = load <4 x double>, <4 x double>* %x0ptr
-  %res = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x4.512(<4 x double> %x0, <8 x double> %x2, i8 %mask)
-  ret <8 x double> %res
-}
-
-declare <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32>, <16 x i32>, i16)
-
-define <16 x i32>@test_int_x86_avx512_mask_broadcasti32x4_512(<4 x i32> %x0, <16 x i32> %x2, i16 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_512:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %ZMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vshufi32x4 {{.*#+}} zmm2 {%k1} {z} = zmm0[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshufi32x4 {{.*#+}} zmm1 {%k1} = zmm0[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshufi32x4 {{.*#+}} zmm0 = zmm0[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vpaddd %zmm1, %zmm0, %zmm0
-; CHECK-NEXT:    vpaddd %zmm0, %zmm2, %zmm0
-; CHECK-NEXT:    retq
-
-  %res1 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32> %x0, <16 x i32> %x2, i16 -1)
-  %res2 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32> %x0, <16 x i32> %x2, i16 %mask)
-  %res3 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32> %x0, <16 x i32> zeroinitializer, i16 %mask)
-  %res4 = add <16 x i32> %res1, %res2
-  %res5 = add <16 x i32> %res3, %res4
-  ret <16 x i32> %res5
-}
-
-define <16 x i32>@test_int_x86_avx512_mask_broadcasti32x4_512_load(<4 x i32>* %x0ptr, <16 x i32> %x2, i16 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_512_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1
-; CHECK-NEXT:    vmovdqa (%rdi), %xmm1
-; CHECK-NEXT:    vshufi32x4 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    retq
-
-  %x0 = load <4 x i32>, <4 x i32>* %x0ptr
-  %res = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x4.512(<4 x i32> %x0, <16 x i32> %x2, i16 %mask)
-  ret <16 x i32> %res
-}
-
-declare <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64>, <8 x i64>, i8)
-
-define <8 x i64>@test_int_x86_avx512_mask_broadcasti64x4_512(<4 x i64> %x0, <8 x i64> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x4_512:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %YMM0<def> %YMM0<kill> %ZMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm2 {%k1} {z} = zmm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm1 {%k1} = zmm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vpaddq %zmm1, %zmm0, %zmm0
-; CHECK-NEXT:    vpaddq %zmm0, %zmm2, %zmm0
-; CHECK-NEXT:    retq
-
-  %res1 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64> %x0, <8 x i64> %x2, i8 -1)
-  %res2 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64> %x0, <8 x i64> %x2, i8 %mask)
-  %res3 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64> %x0, <8 x i64> zeroinitializer, i8 %mask)
-  %res4 = add <8 x i64> %res1, %res2
-  %res5 = add <8 x i64> %res3, %res4
-  ret <8 x i64> %res5
-}
-
-define <8 x i64>@test_int_x86_avx512_mask_broadcasti64x4_512_load(<4 x i64>* %x0ptr, <8 x i64> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x4_512_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1
-; CHECK-NEXT:    vmovdqa (%rdi), %ymm1
-; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    retq
-
-  %x0 = load <4 x i64>, <4 x i64>* %x0ptr
-  %res = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x4.512(<4 x i64> %x0, <8 x i64> %x2, i8 %mask)
-  ret <8 x i64> %res
-}
-
 declare <16 x i32> @llvm.x86.avx512.mask.prorv.d.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
 
 define <16 x i32>@test_int_x86_avx512_mask_prorv_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x3) {
 ; CHECK-LABEL: test_int_x86_avx512_mask_prorv_d_512:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1
 ; CHECK-NEXT:    vprorvd %zmm1, %zmm0, %zmm3
 ; CHECK-NEXT:    vprorvd %zmm1, %zmm0, %zmm2 {%k1}

llvm/trunk/test/CodeGen/X86/avx512dq-intrinsics-upgrade.ll

 ; CHECK-LABEL: test_int_x86_avx512_cvtmask2q_512:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k0
 ; CHECK-NEXT:    vpmovm2q %k0, %zmm0
 ; CHECK-NEXT:    retq
   %res = call <8 x i64> @llvm.x86.avx512.cvtmask2q.512(i8 %x0)
   ret <8 x i64> %res
 }
+
+declare <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float>, <16 x float>, i16)
+
+define <16 x float>@test_int_x86_avx512_mask_broadcastf32x8_512(<8 x float> %x0, <16 x float> %x2, i16 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x8_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %YMM0<def> %YMM0<kill> %ZMM0<def>
+; CHECK-NEXT:    vinsertf32x8 $1, %ymm0, %zmm0, %zmm2
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vinsertf32x8 $1, %ymm0, %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vaddps %zmm1, %zmm2, %zmm1
+; CHECK-NEXT:    vinsertf32x8 $1, %ymm0, %zmm0, %zmm0 {%k1} {z}
+; CHECK-NEXT:    vaddps %zmm1, %zmm0, %zmm0
+; CHECK-NEXT:    retq
+
+  %res1 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float> %x0, <16 x float> %x2, i16 -1)
+  %res2 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float> %x0, <16 x float> %x2, i16 %mask)
+  %res3 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float> %x0, <16 x float> zeroinitializer, i16 %mask)
+  %res4 = fadd <16 x float> %res1, %res2
+  %res5 = fadd <16 x float> %res3, %res4
+  ret <16 x float> %res5
+}
+
+define <16 x float>@test_int_x86_avx512_mask_broadcastf32x8_512_load(<8 x float>* %x0ptr, <16 x float> %x2, i16 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x8_512_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vbroadcastf32x8 {{.*#+}} zmm0 {%k1} = mem[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
+; CHECK-NEXT:    retq
+
+  %x0 = load <8 x float>, <8 x float>* %x0ptr
+  %res = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float> %x0, <16 x float> %x2, i16 %mask)
+  ret <16 x float> %res
+}
+
+declare <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double>, <8 x double>, i8)
+
+define <8 x double>@test_int_x86_avx512_mask_broadcastf64x2_512(<2 x double> %x0, <8 x double> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
+; CHECK-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; CHECK-NEXT:    vinsertf64x4 $1, %ymm0, %zmm0, %zmm2
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vinsertf64x4 $1, %ymm0, %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vaddpd %zmm1, %zmm2, %zmm1
+; CHECK-NEXT:    vinsertf64x4 $1, %ymm0, %zmm0, %zmm0 {%k1} {z}
+; CHECK-NEXT:    vaddpd %zmm1, %zmm0, %zmm0
+; CHECK-NEXT:    retq
+
+  %res1 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double> %x0, <8 x double> %x2, i8 -1)
+  %res2 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double> %x0, <8 x double> %x2, i8 %mask)
+  %res3 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double> %x0, <8 x double> zeroinitializer, i8 %mask)
+  %res4 = fadd <8 x double> %res1, %res2
+  %res5 = fadd <8 x double> %res3, %res4
+  ret <8 x double> %res5
+}
+
+define <8 x double>@test_int_x86_avx512_mask_broadcastf64x2_512_load(<2 x double>* %x0ptr, <8 x double> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_512_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vbroadcastf64x2 {{.*#+}} zmm0 {%k1} = mem[0,1,0,1,0,1,0,1]
+; CHECK-NEXT:    retq
+
+  %x0 = load <2 x double>, <2 x double>* %x0ptr
+  %res = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double> %x0, <8 x double> %x2, i8 %mask)
+  ret <8 x double> %res
+}
+
+declare <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32>, <16 x i32>, i16)
+
+define <16 x i32>@test_int_x86_avx512_mask_broadcasti32x8_512(<8 x i32> %x0, <16 x i32> %x2, i16 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x8_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %YMM0<def> %YMM0<kill> %ZMM0<def>
+; CHECK-NEXT:    vinserti32x8 $1, %ymm0, %zmm0, %zmm2
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vinserti32x8 $1, %ymm0, %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vpaddd %zmm1, %zmm2, %zmm1
+; CHECK-NEXT:    vinserti32x8 $1, %ymm0, %zmm0, %zmm0 {%k1} {z}
+; CHECK-NEXT:    vpaddd %zmm1, %zmm0, %zmm0
+; CHECK-NEXT:    retq
+
+  %res1 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32> %x0, <16 x i32> %x2, i16 -1)
+  %res2 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32> %x0, <16 x i32> %x2, i16 %mask)
+  %res3 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32> %x0, <16 x i32> zeroinitializer, i16 %mask)
+  %res4 = add <16 x i32> %res1, %res2
+  %res5 = add <16 x i32> %res3, %res4
+  ret <16 x i32> %res5
+}
+
+define <16 x i32>@test_int_x86_avx512_mask_broadcasti32x8_512_load(<8 x i32>* %x0ptr, <16 x i32> %x2, i16 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x8_512_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vbroadcasti32x8 {{.*#+}} zmm0 {%k1} = mem[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
+; CHECK-NEXT:    retq
+
+  %x0 = load <8 x i32>, <8 x i32>* %x0ptr
+  %res = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32> %x0, <16 x i32> %x2, i16 %mask)
+  ret <16 x i32> %res
+}
+
+declare <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64>, <8 x i64>, i8)
+
+define <8 x i64>@test_int_x86_avx512_mask_broadcasti64x2_512(<2 x i64> %x0, <8 x i64> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_512:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
+; CHECK-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm0
+; CHECK-NEXT:    vinserti64x4 $1, %ymm0, %zmm0, %zmm2
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vinserti64x4 $1, %ymm0, %zmm0, %zmm1 {%k1}
+; CHECK-NEXT:    vpaddq %zmm1, %zmm2, %zmm1
+; CHECK-NEXT:    vinserti64x4 $1, %ymm0, %zmm0, %zmm0 {%k1} {z}
+; CHECK-NEXT:    vpaddq %zmm1, %zmm0, %zmm0
+; CHECK-NEXT:    retq
+
+  %res1 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64> %x0, <8 x i64> %x2, i8 -1)
+  %res2 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64> %x0, <8 x i64> %x2, i8 %mask)
+  %res3 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64> %x0, <8 x i64> zeroinitializer, i8 %mask)
+  %res4 = add <8 x i64> %res1, %res2
+  %res5 = add <8 x i64> %res3, %res4
+  ret <8 x i64> %res5
+}
+
+define <8 x i64>@test_int_x86_avx512_mask_broadcasti64x2_512_load(<2 x i64>* %x0ptr, <8 x i64> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_512_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vbroadcasti64x2 {{.*#+}} zmm0 {%k1} = mem[0,1,0,1,0,1,0,1]
+; CHECK-NEXT:    retq
+
+  %x0 = load <2 x i64>, <2 x i64>* %x0ptr
+  %res = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64> %x0, <8 x i64> %x2, i8 %mask)
+  ret <8 x i64> %res
+}

llvm/trunk/test/CodeGen/X86/avx512dq-intrinsics.ll

 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    vpmovq2m %zmm0, %k0
 ; CHECK-NEXT:    kmovw %k0, %eax
 ; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
 ; CHECK-NEXT:    retq
   %res = call i8 @llvm.x86.avx512.cvtq2mask.512(<8 x i64> %x0)
   ret i8 %res
 }
-
-declare <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float>, <16 x float>, i16)
-
-define <16 x float>@test_int_x86_avx512_mask_broadcastf32x8_512(<8 x float> %x0, <16 x float> %x2, i16 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x8_512:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %YMM0<def> %YMM0<kill> %ZMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vshuff32x4 {{.*#+}} zmm2 {%k1} {z} = zmm0[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
-; CHECK-NEXT:    vshuff32x4 {{.*#+}} zmm1 {%k1} = zmm0[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
-; CHECK-NEXT:    vshuff32x4 {{.*#+}} zmm0 = zmm0[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
-; CHECK-NEXT:    vaddps %zmm1, %zmm0, %zmm0
-; CHECK-NEXT:    vaddps %zmm0, %zmm2, %zmm0
-; CHECK-NEXT:    retq
-
-  %res1 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float> %x0, <16 x float> %x2, i16 -1)
-  %res2 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float> %x0, <16 x float> %x2, i16 %mask)
-  %res3 = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float> %x0, <16 x float> zeroinitializer, i16 %mask)
-  %res4 = fadd <16 x float> %res1, %res2
-  %res5 = fadd <16 x float> %res3, %res4
-  ret <16 x float> %res5
-}
-
-define <16 x float>@test_int_x86_avx512_mask_broadcastf32x8_512_load(<8 x float>* %x0ptr, <16 x float> %x2, i16 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x8_512_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1
-; CHECK-NEXT:    vmovaps (%rdi), %ymm1
-; CHECK-NEXT:    vshuff32x4 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
-; CHECK-NEXT:    retq
-
-  %x0 = load <8 x float>, <8 x float>* %x0ptr
-  %res = call <16 x float> @llvm.x86.avx512.mask.broadcastf32x8.512(<8 x float> %x0, <16 x float> %x2, i16 %mask)
-  ret <16 x float> %res
-}
-
-declare <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double>, <8 x double>, i8)
-
-define <8 x double>@test_int_x86_avx512_mask_broadcastf64x2_512(<2 x double> %x0, <8 x double> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_512:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %ZMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vshuff64x2 {{.*#+}} zmm2 {%k1} {z} = zmm0[0,1,0,1,0,1,0,1]
-; CHECK-NEXT:    vshuff64x2 {{.*#+}} zmm1 {%k1} = zmm0[0,1,0,1,0,1,0,1]
-; CHECK-NEXT:    vshuff64x2 {{.*#+}} zmm0 = zmm0[0,1,0,1,0,1,0,1]
-; CHECK-NEXT:    vaddpd %zmm1, %zmm0, %zmm0
-; CHECK-NEXT:    vaddpd %zmm0, %zmm2, %zmm0
-; CHECK-NEXT:    retq
-
-  %res1 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double> %x0, <8 x double> %x2, i8 -1)
-  %res2 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double> %x0, <8 x double> %x2, i8 %mask)
-  %res3 = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double> %x0, <8 x double> zeroinitializer, i8 %mask)
-  %res4 = fadd <8 x double> %res1, %res2
-  %res5 = fadd <8 x double> %res3, %res4
-  ret <8 x double> %res5
-}
-
-define <8 x double>@test_int_x86_avx512_mask_broadcastf64x2_512_load(<2 x double>* %x0ptr, <8 x double> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_512_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1
-; CHECK-NEXT:    vmovapd (%rdi), %xmm1
-; CHECK-NEXT:    vshuff64x2 {{.*#+}} zmm0 {%k1} = zmm1[0,1,0,1,0,1,0,1]
-; CHECK-NEXT:    retq
-
-  %x0 = load <2 x double>, <2 x double>* %x0ptr
-  %res = call <8 x double> @llvm.x86.avx512.mask.broadcastf64x2.512(<2 x double> %x0, <8 x double> %x2, i8 %mask)
-  ret <8 x double> %res
-}
-
-declare <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32>, <16 x i32>, i16)
-
-define <16 x i32>@test_int_x86_avx512_mask_broadcasti32x8_512(<8 x i32> %x0, <16 x i32> %x2, i16 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x8_512:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %YMM0<def> %YMM0<kill> %ZMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vshufi32x4 {{.*#+}} zmm2 {%k1} {z} = zmm0[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
-; CHECK-NEXT:    vshufi32x4 {{.*#+}} zmm1 {%k1} = zmm0[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
-; CHECK-NEXT:    vshufi32x4 {{.*#+}} zmm0 = zmm0[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
-; CHECK-NEXT:    vpaddd %zmm1, %zmm0, %zmm0
-; CHECK-NEXT:    vpaddd %zmm0, %zmm2, %zmm0
-; CHECK-NEXT:    retq
-
-  %res1 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32> %x0, <16 x i32> %x2, i16 -1)
-  %res2 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32> %x0, <16 x i32> %x2, i16 %mask)
-  %res3 = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32> %x0, <16 x i32> zeroinitializer, i16 %mask)
-  %res4 = add <16 x i32> %res1, %res2
-  %res5 = add <16 x i32> %res3, %res4
-  ret <16 x i32> %res5
-}
-
-define <16 x i32>@test_int_x86_avx512_mask_broadcasti32x8_512_load(<8 x i32>* %x0ptr, <16 x i32> %x2, i16 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x8_512_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1
-; CHECK-NEXT:    vmovdqa (%rdi), %ymm1
-; CHECK-NEXT:    vshufi32x4 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7]
-; CHECK-NEXT:    retq
-
-  %x0 = load <8 x i32>, <8 x i32>* %x0ptr
-  %res = call <16 x i32> @llvm.x86.avx512.mask.broadcasti32x8.512(<8 x i32> %x0, <16 x i32> %x2, i16 %mask)
-  ret <16 x i32> %res
-}
-
-declare <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64>, <8 x i64>, i8)
-
-define <8 x i64>@test_int_x86_avx512_mask_broadcasti64x2_512(<2 x i64> %x0, <8 x i64> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_512:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %ZMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm2 {%k1} {z} = zmm0[0,1,0,1,0,1,0,1]
-; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm1 {%k1} = zmm0[0,1,0,1,0,1,0,1]
-; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,0,1,0,1,0,1]
-; CHECK-NEXT:    vpaddq %zmm1, %zmm0, %zmm0
-; CHECK-NEXT:    vpaddq %zmm0, %zmm2, %zmm0
-; CHECK-NEXT:    retq
-
-  %res1 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64> %x0, <8 x i64> %x2, i8 -1)
-  %res2 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64> %x0, <8 x i64> %x2, i8 %mask)
-  %res3 = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64> %x0, <8 x i64> zeroinitializer, i8 %mask)
-  %res4 = add <8 x i64> %res1, %res2
-  %res5 = add <8 x i64> %res3, %res4
-  ret <8 x i64> %res5
-}
-
-define <8 x i64>@test_int_x86_avx512_mask_broadcasti64x2_512_load(<2 x i64>* %x0ptr, <8 x i64> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_512_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1
-; CHECK-NEXT:    vmovdqa (%rdi), %xmm1
-; CHECK-NEXT:    vshufi64x2 {{.*#+}} zmm0 {%k1} = zmm1[0,1,0,1,0,1,0,1]
-; CHECK-NEXT:    retq
-
-  %x0 = load <2 x i64>, <2 x i64>* %x0ptr
-  %res = call <8 x i64> @llvm.x86.avx512.mask.broadcasti64x2.512(<2 x i64> %x0, <8 x i64> %x2, i8 %mask)
-  ret <8 x i64> %res
-}

llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics-upgrade.ll

 ; CHECK-LABEL: test_int_x86_avx512_cvtmask2q_256:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k0 ## encoding: [0xc5,0xf8,0x92,0xc7]
 ; CHECK-NEXT:    vpmovm2q %k0, %ymm0 ## encoding: [0x62,0xf2,0xfe,0x28,0x38,0xc0]
 ; CHECK-NEXT:    retq ## encoding: [0xc3]
   %res = call <4 x i64> @llvm.x86.avx512.cvtmask2q.256(i8 %x0)
   ret <4 x i64> %res
 }
+
+declare <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double>, <4 x double>, i8)
+
+define <4 x double>@test_int_x86_avx512_mask_broadcastf64x2_256(<2 x double> %x0, <4 x double> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
+; CHECK-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm2 ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x18,0xd0,0x01]
+; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT:    vinsertf64x2 $1, %xmm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x18,0xc8,0x01]
+; CHECK-NEXT:    vaddpd %ymm1, %ymm2, %ymm1 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0x58,0xc9]
+; CHECK-NEXT:    vinsertf64x2 $1, %xmm0, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x18,0xc0,0x01]
+; CHECK-NEXT:    vaddpd %ymm1, %ymm0, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xfd,0x58,0xc1]
+; CHECK-NEXT:    retq ## encoding: [0xc3]
+
+  %res1 = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> %x2, i8 -1)
+  %res2 = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> %x2, i8 %mask)
+  %res3 = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> zeroinitializer, i8 %mask)
+  %res4 = fadd <4 x double> %res1, %res2
+  %res5 = fadd <4 x double> %res3, %res4
+  ret <4 x double> %res5
+}
+
+define <4 x double>@test_int_x86_avx512_mask_broadcastf64x2_256_load(<2 x double>* %x0ptr, <4 x double> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_256_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT:    vbroadcastf64x2 (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x1a,0x07]
+; CHECK-NEXT:    ## ymm0 {%k1} = mem[0,1,0,1]
+; CHECK-NEXT:    retq ## encoding: [0xc3]
+
+  %x0 = load <2 x double>, <2 x double>* %x0ptr
+  %res = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> %x2, i8 %mask)
+  ret <4 x double> %res
+}
+
+declare <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64>, <4 x i64>, i8)
+
+define <4 x i64>@test_int_x86_avx512_mask_broadcasti64x2_256(<2 x i64> %x0, <4 x i64> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
+; CHECK-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm2 ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x38,0xd0,0x01]
+; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT:    vinserti64x2 $1, %xmm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x38,0xc8,0x01]
+; CHECK-NEXT:    vpaddq %ymm1, %ymm2, %ymm1 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc9]
+; CHECK-NEXT:    vinserti64x2 $1, %xmm0, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x38,0xc0,0x01]
+; CHECK-NEXT:    vpaddq %ymm1, %ymm0, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xfd,0xd4,0xc1]
+; CHECK-NEXT:    retq ## encoding: [0xc3]
+
+  %res1 = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> %x2, i8 -1)
+  %res2 = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> %x2, i8 %mask)
+  %res3 = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> zeroinitializer, i8 %mask)
+  %res4 = add <4 x i64> %res1, %res2
+  %res5 = add <4 x i64> %res3, %res4
+  ret <4 x i64> %res5
+}
+
+define <4 x i64>@test_int_x86_avx512_mask_broadcasti64x2_256_load(<2 x i64>* %x0ptr, <4 x i64> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_256_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT:    vbroadcasti64x2 (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x5a,0x07]
+; CHECK-NEXT:    ## ymm0 {%k1} = mem[0,1,0,1]
+; CHECK-NEXT:    retq ## encoding: [0xc3]
+
+  %x0 = load <2 x i64>, <2 x i64>* %x0ptr
+  %res = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> %x2, i8 %mask)
+  ret <4 x i64> %res
+}

llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics.ll

 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    vpmovq2m %ymm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x28,0x39,0xc0]
 ; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
 ; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
 ; CHECK-NEXT:    retq ## encoding: [0xc3]
     %res = call i8 @llvm.x86.avx512.cvtq2mask.256(<4 x i64> %x0)
     ret i8 %res
 }
-
-declare <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double>, <4 x double>, i8)
-
-define <4 x double>@test_int_x86_avx512_mask_broadcastf64x2_256(<2 x double> %x0, <4 x double> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_256:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vshuff64x2 $0, %ymm0, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x23,0xd0,0x00]
-; CHECK-NEXT:    ## ymm2 {%k1} {z} = ymm0[0,1,0,1]
-; CHECK-NEXT:    vshuff64x2 $0, %ymm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x23,0xc8,0x00]
-; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,1,0,1]
-; CHECK-NEXT:    vshuff64x2 $0, %ymm0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x23,0xc0,0x00]
-; CHECK-NEXT:    ## ymm0 = ymm0[0,1,0,1]
-; CHECK-NEXT:    vaddpd %ymm1, %ymm0, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xfd,0x58,0xc1]
-; CHECK-NEXT:    vaddpd %ymm0, %ymm2, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0x58,0xc0]
-; CHECK-NEXT:    retq ## encoding: [0xc3]
-
-  %res1 = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> %x2, i8 -1)
-  %res2 = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> %x2, i8 %mask)
-  %res3 = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> zeroinitializer, i8 %mask)
-  %res4 = fadd <4 x double> %res1, %res2
-  %res5 = fadd <4 x double> %res3, %res4
-  ret <4 x double> %res5
-}
-
-define <4 x double>@test_int_x86_avx512_mask_broadcastf64x2_256_load(<2 x double>* %x0ptr, <4 x double> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_256_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT:    vmovapd (%rdi), %xmm1 ## EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0x0f]
-; CHECK-NEXT:    vshuff64x2 $0, %ymm1, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf3,0xf5,0x29,0x23,0xc1,0x00]
-; CHECK-NEXT:    ## ymm0 {%k1} = ymm1[0,1,0,1]
-; CHECK-NEXT:    retq ## encoding: [0xc3]
-
-  %x0 = load <2 x double>, <2 x double>* %x0ptr
-  %res = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> %x2, i8 %mask)
-  ret <4 x double> %res
-}
-
-declare <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64>, <4 x i64>, i8)
-
-define <4 x i64>@test_int_x86_avx512_mask_broadcasti64x2_256(<2 x i64> %x0, <4 x i64> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_256:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vshufi64x2 $0, %ymm0, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x43,0xd0,0x00]
-; CHECK-NEXT:    ## ymm2 {%k1} {z} = ymm0[0,1,0,1]
-; CHECK-NEXT:    vshufi64x2 $0, %ymm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x43,0xc8,0x00]
-; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,1,0,1]
-; CHECK-NEXT:    vshufi64x2 $0, %ymm0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x43,0xc0,0x00]
-; CHECK-NEXT:    ## ymm0 = ymm0[0,1,0,1]
-; CHECK-NEXT:    vpaddq %ymm1, %ymm0, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xfd,0xd4,0xc1]
-; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
-; CHECK-NEXT:    retq ## encoding: [0xc3]
-
-  %res1 = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> %x2, i8 -1)
-  %res2 = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> %x2, i8 %mask)
-  %res3 = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> zeroinitializer, i8 %mask)
-  %res4 = add <4 x i64> %res1, %res2
-  %res5 = add <4 x i64> %res3, %res4
-  ret <4 x i64> %res5
-}
-
-define <4 x i64>@test_int_x86_avx512_mask_broadcasti64x2_256_load(<2 x i64>* %x0ptr, <4 x i64> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_256_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT:    vmovdqa (%rdi), %xmm1 ## EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0x0f]
-; CHECK-NEXT:    vshufi64x2 $0, %ymm1, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf3,0xf5,0x29,0x43,0xc1,0x00]
-; CHECK-NEXT:    ## ymm0 {%k1} = ymm1[0,1,0,1]
-; CHECK-NEXT:    retq ## encoding: [0xc3]
-
-  %x0 = load <2 x i64>, <2 x i64>* %x0ptr
-  %res = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> %x2, i8 %mask)
-  ret <4 x i64> %res
-}

llvm/trunk/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll

; CHECK-NEXT:    retq ## encoding: [0xc3]
   %res6 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 6, i8 %mask)
   %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
   %res7 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 7, i8 %mask)
   %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
   ret <8 x i8> %vec7
 }
 
 declare i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64>, <2 x i64>, i32, i8) nounwind readnone
+
+declare <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float>, <8 x float>, i8)
+
+define <8 x float>@test_int_x86_avx512_mask_broadcastf32x4_256(<4 x float> %x0, <8 x float> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
+; CHECK-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm2 ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x18,0xd0,0x01]
+; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT:    vinsertf32x4 $1, %xmm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x18,0xc8,0x01]
+; CHECK-NEXT:    vaddps %ymm1, %ymm2, %ymm1 ## EVEX TO VEX Compression encoding: [0xc5,0xec,0x58,0xc9]
+; CHECK-NEXT:    vinsertf32x4 $1, %xmm0, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x18,0xc0,0x01]
+; CHECK-NEXT:    vaddps %ymm1, %ymm0, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xfc,0x58,0xc1]
+; CHECK-NEXT:    retq ## encoding: [0xc3]
+  %res1 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 -1)
+  %res2 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 %mask)
+  %res3 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> zeroinitializer, i8 %mask)
+  %res4 = fadd <8 x float> %res1, %res2
+  %res5 = fadd <8 x float> %res3, %res4
+  ret <8 x float> %res5
+}
+
+define <8 x float>@test_int_x86_avx512_mask_broadcastf32x4_256_load(<4 x float>* %x0ptr, <8 x float> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_256_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT:    vbroadcastf32x4 (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x1a,0x07]
+; CHECK-NEXT:    ## ymm0 {%k1} = mem[0,1,2,3,0,1,2,3]
+; CHECK-NEXT:    retq ## encoding: [0xc3]
+  %x0 = load <4 x float>, <4 x float>* %x0ptr
+  %res = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 %mask)
+  ret <8 x float> %res
+}
+
+declare <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32>, <8 x i32>, i8)
+
+define <8 x i32>@test_int_x86_avx512_mask_broadcasti32x4_256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_256:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
+; CHECK-NEXT:    vinserti128 $1, %xmm0, %ymm0, %ymm2 ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x38,0xd0,0x01]
+; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT:    vinserti32x4 $1, %xmm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x38,0xc8,0x01]
+; CHECK-NEXT:    vpaddd %ymm1, %ymm2, %ymm1 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc9]
+; CHECK-NEXT:    vinserti32x4 $1, %xmm0, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x38,0xc0,0x01]
+; CHECK-NEXT:    vpaddd %ymm1, %ymm0, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfe,0xc1]
+; CHECK-NEXT:    retq ## encoding: [0xc3]
+  %res1 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 -1)
+  %res2 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask)
+  %res3 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> zeroinitializer, i8 %mask)
+  %res4 = add <8 x i32> %res1, %res2
+  %res5 = add <8 x i32> %res3, %res4
+  ret <8 x i32> %res5
+}
+
+define <8 x i32>@test_int_x86_avx512_mask_broadcasti32x4_256_load(<4 x i32>* %x0ptr, <8 x i32> %x2, i8 %mask) {
+; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_256_load:
+; CHECK:       ## BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT:    vbroadcasti32x4 (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x5a,0x07]
+; CHECK-NEXT:    ## ymm0 {%k1} = mem[0,1,2,3,0,1,2,3]
+; CHECK-NEXT:    retq ## encoding: [0xc3]
+  %x0 = load <4 x i32>, <4 x i32>* %x0ptr
+  %res = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask)
+  ret <8 x i32> %res
+}

llvm/trunk/test/CodeGen/X86/avx512vl-intrinsics.ll

 ; CHECK-NEXT:    retq ## encoding: [0xc3]
   %res = call <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double> %a0, <2 x double> %a1, i8 %mask)
   ret <2 x double> %res
 }
 
 declare <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone
 declare <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double>, <2 x double>, i8) nounwind readnone
 
-declare <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float>, <8 x float>, i8)
-
-define <8 x float>@test_int_x86_avx512_mask_broadcastf32x4_256(<4 x float> %x0, <8 x float> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_256:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vshuff32x4 $0, %ymm0, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x23,0xd0,0x00]
-; CHECK-NEXT:    ## ymm2 {%k1} {z} = ymm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshuff32x4 $0, %ymm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x23,0xc8,0x00]
-; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshuff32x4 $0, %ymm0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x23,0xc0,0x00]
-; CHECK-NEXT:    ## ymm0 = ymm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vaddps %ymm1, %ymm0, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xfc,0x58,0xc1]
-; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xec,0x58,0xc0]
-; CHECK-NEXT:    retq ## encoding: [0xc3]
-  %res1 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 -1)
-  %res2 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 %mask)
-  %res3 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> zeroinitializer, i8 %mask)
-  %res4 = fadd <8 x float> %res1, %res2
-  %res5 = fadd <8 x float> %res3, %res4
-  ret <8 x float> %res5
-}
-
-define <8 x float>@test_int_x86_avx512_mask_broadcastf32x4_256_load(<4 x float>* %x0ptr, <8 x float> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_256_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT:    vmovaps (%rdi), %xmm1 ## EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0x0f]
-; CHECK-NEXT:    vshuff32x4 $0, %ymm1, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf3,0x75,0x29,0x23,0xc1,0x00]
-; CHECK-NEXT:    ## ymm0 {%k1} = ymm1[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    retq ## encoding: [0xc3]
-  %x0 = load <4 x float>, <4 x float>* %x0ptr
-  %res = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 %mask)
-  ret <8 x float> %res
-}
-
-declare <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32>, <8 x i32>, i8)
-
-define <8 x i32>@test_int_x86_avx512_mask_broadcasti32x4_256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_256:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
-; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vshufi32x4 $0, %ymm0, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x43,0xd0,0x00]
-; CHECK-NEXT:    ## ymm2 {%k1} {z} = ymm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshufi32x4 $0, %ymm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x43,0xc8,0x00]
-; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vshufi32x4 $0, %ymm0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x43,0xc0,0x00]
-; CHECK-NEXT:    ## ymm0 = ymm0[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    vpaddd %ymm1, %ymm0, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfe,0xc1]
-; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc0]
-; CHECK-NEXT:    retq ## encoding: [0xc3]
-  %res1 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 -1)
-  %res2 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask)
-  %res3 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> zeroinitializer, i8 %mask)
-  %res4 = add <8 x i32> %res1, %res2
-  %res5 = add <8 x i32> %res3, %res4
-  ret <8 x i32> %res5
-}
-
-define <8 x i32>@test_int_x86_avx512_mask_broadcasti32x4_256_load(<4 x i32>* %x0ptr, <8 x i32> %x2, i8 %mask) {
-; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_256_load:
-; CHECK:       ## BB#0:
-; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
-; CHECK-NEXT:    vmovdqa (%rdi), %xmm1 ## EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0x0f]
-; CHECK-NEXT:    vshufi32x4 $0, %ymm1, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf3,0x75,0x29,0x43,0xc1,0x00]
-; CHECK-NEXT:    ## ymm0 {%k1} = ymm1[0,1,2,3,0,1,2,3]
-; CHECK-NEXT:    retq ## encoding: [0xc3]
-  %x0 = load <4 x i32>, <4 x i32>* %x0ptr
-  %res = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask)
-  ret <8 x i32> %res
-}
-
 declare <4 x i32> @llvm.x86.avx512.mask.prorv.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
 
 define <4 x i32>@test_int_x86_avx512_mask_prorv_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
 ; CHECK-LABEL: test_int_x86_avx512_mask_prorv_d_128:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
 ; CHECK-NEXT:    vprorvd %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf2,0x7d,0x08,0x14,0xd9]
 ; CHECK-NEXT:    vprorvd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x14,0xd1]