Switch AArch64 to use reduction intrinsics

[aemerson]

Summary

This patch enables the use of the experimental reduction intrinsics from D30086 for AArch64.

The existing idiom recognition for the shufflevector sequence is removed and tests updated to use the new form.

[rengolin]

As soon as the dependency patch is approved, this one looks ok.

Did you submit the patch that will make the MC layer recognise these patterns into the actual instructions?

lib/Target/AArch64/AArch64TargetTransformInfo.cpp

690 ↗

(On Diff #95824)

For now, shouldn't this be == 128?

[aemerson]

No changes are needed on the MC side. The same target-specific reduction DAG nodes (e.g. AArch64ISD::UADDV) should be created and from then on everything should work as before.

lib/Target/AArch64/AArch64TargetTransformInfo.cpp

690 ↗

(On Diff #95824)

It's needed because vectors can be oversize, e.g. v8i32. We still want to use reductions for that as legalisation will split the vector later.

697 ↗

(On Diff #95824)

FYI, I haven't yet updated this version of the patch, the current generic reduction patch combines the two hooks into one so this function will merge into the one above when I do.

[aemerson]

New patch, rebased on latest ToT and using the different API implemented in the previous patch in D30086.

Marking this as accepted based on Renato's earlier approval, I'm just updating this review for future reference.

llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.h

private:
   SDValue LowerFP_EXTEND(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerFP_ROUND(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerFP_TO_INT(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerINT_TO_FP(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerVectorAND(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerVectorOR(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerCONCAT_VECTORS(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerFSINCOS(SDValue Op, SelectionDAG &DAG) const;
+  SDValue LowerVECREDUCE(SDValue Op, SelectionDAG &DAG) const;
 
   SDValue BuildSDIVPow2(SDNode *N, const APInt &Divisor, SelectionDAG &DAG,
                         std::vector<SDNode *> *Created) const override;
   SDValue getSqrtEstimate(SDValue Operand, SelectionDAG &DAG, int Enabled,
                           int &ExtraSteps, bool &UseOneConst,
                           bool Reciprocal) const override;
   SDValue getRecipEstimate(SDValue Operand, SelectionDAG &DAG, int Enabled,
                            int &ExtraSteps) const override;

llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.cpp

AArch64TargetLowering::AArch64TargetLowering(const TargetMachine &TM,
   setTargetDAGCombine(ISD::MUL);
 
   setTargetDAGCombine(ISD::SELECT);
   setTargetDAGCombine(ISD::VSELECT);
 
   setTargetDAGCombine(ISD::INTRINSIC_VOID);
   setTargetDAGCombine(ISD::INTRINSIC_W_CHAIN);
   setTargetDAGCombine(ISD::INSERT_VECTOR_ELT);
-  setTargetDAGCombine(ISD::EXTRACT_VECTOR_ELT);
 
   MaxStoresPerMemset = MaxStoresPerMemsetOptSize = 8;
   MaxStoresPerMemcpy = MaxStoresPerMemcpyOptSize = 4;
   MaxStoresPerMemmove = MaxStoresPerMemmoveOptSize = 4;
 
   setStackPointerRegisterToSaveRestore(AArch64::SP);
 
   setSchedulingPreference(Sched::Hybrid);
if (Subtarget->hasNEON()) {
 
     // AArch64 doesn't have MUL.2d:
     setOperationAction(ISD::MUL, MVT::v2i64, Expand);
     // Custom handling for some quad-vector types to detect MULL.
     setOperationAction(ISD::MUL, MVT::v8i16, Custom);
     setOperationAction(ISD::MUL, MVT::v4i32, Custom);
     setOperationAction(ISD::MUL, MVT::v2i64, Custom);
 
+    // Vector reductions
+    for (MVT VT : MVT::integer_valuetypes()) {
+      setOperationAction(ISD::VECREDUCE_ADD, VT, Custom);
+      setOperationAction(ISD::VECREDUCE_SMAX, VT, Custom);
+      setOperationAction(ISD::VECREDUCE_SMIN, VT, Custom);
+      setOperationAction(ISD::VECREDUCE_UMAX, VT, Custom);
+      setOperationAction(ISD::VECREDUCE_UMIN, VT, Custom);
+    }
+    for (MVT VT : MVT::fp_valuetypes()) {
+      setOperationAction(ISD::VECREDUCE_FMAX, VT, Custom);
+      setOperationAction(ISD::VECREDUCE_FMIN, VT, Custom);
+    }
+
     setOperationAction(ISD::ANY_EXTEND, MVT::v4i32, Legal);
     setTruncStoreAction(MVT::v2i32, MVT::v2i16, Expand);
     // Likewise, narrowing and extending vector loads/stores aren't handled
     // directly.
     for (MVT VT : MVT::vector_valuetypes()) {
       setOperationAction(ISD::SIGN_EXTEND_INREG, VT, Expand);
 
       setOperationAction(ISD::MULHS, VT, Expand);
SDValue AArch64TargetLowering::LowerOperation(SDValue Op,
   case ISD::FP_TO_UINT:
     return LowerFP_TO_INT(Op, DAG);
   case ISD::FSINCOS:
     return LowerFSINCOS(Op, DAG);
   case ISD::MUL:
     return LowerMUL(Op, DAG);
   case ISD::INTRINSIC_WO_CHAIN:
     return LowerINTRINSIC_WO_CHAIN(Op, DAG);
+  case ISD::VECREDUCE_ADD:
+  case ISD::VECREDUCE_SMAX:
+  case ISD::VECREDUCE_SMIN:
+  case ISD::VECREDUCE_UMAX:
+  case ISD::VECREDUCE_UMIN:
+  case ISD::VECREDUCE_FMAX:
+  case ISD::VECREDUCE_FMIN:
+    return LowerVECREDUCE(Op, DAG);
   }
 }
 
 //===----------------------------------------------------------------------===//
 //                      Calling Convention Implementation
 //===----------------------------------------------------------------------===//
 
 #include "AArch64GenCallingConv.inc"
SDValue AArch64TargetLowering::LowerVSETCC(SDValue Op,
   Cmp = DAG.getSExtOrTrunc(Cmp, dl, Op.getValueType());
 
   if (ShouldInvert)
     return Cmp = DAG.getNOT(dl, Cmp, Cmp.getValueType());
 
   return Cmp;
 }
 
+static SDValue getReductionSDNode(unsigned Op, SDLoc DL, SDValue ScalarOp,
+                                  SelectionDAG &DAG) {
+  SDValue VecOp = ScalarOp.getOperand(0);
+  auto Rdx = DAG.getNode(Op, DL, VecOp.getSimpleValueType(), VecOp);
+  return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ScalarOp.getValueType(), Rdx,
+                     DAG.getConstant(0, DL, MVT::i64));
+}
+
+SDValue AArch64TargetLowering::LowerVECREDUCE(SDValue Op,
+                                              SelectionDAG &DAG) const {
+  SDLoc dl(Op);
+  switch (Op.getOpcode()) {
+  case ISD::VECREDUCE_ADD:
+    return getReductionSDNode(AArch64ISD::UADDV, dl, Op, DAG);
+  case ISD::VECREDUCE_SMAX:
+    return getReductionSDNode(AArch64ISD::SMAXV, dl, Op, DAG);
+  case ISD::VECREDUCE_SMIN:
+    return getReductionSDNode(AArch64ISD::SMINV, dl, Op, DAG);
+  case ISD::VECREDUCE_UMAX:
+    return getReductionSDNode(AArch64ISD::UMAXV, dl, Op, DAG);
+  case ISD::VECREDUCE_UMIN:
+    return getReductionSDNode(AArch64ISD::UMINV, dl, Op, DAG);
+  case ISD::VECREDUCE_FMAX: {
+    assert(Op->getFlags().hasNoNaNs() && "fmax vector reduction needs NoNaN flag");
+    return DAG.getNode(
+        ISD::INTRINSIC_WO_CHAIN, dl, Op.getValueType(),
+        DAG.getConstant(Intrinsic::aarch64_neon_fmaxnmv, dl, MVT::i32),
+        Op.getOperand(0));
+  }
+  case ISD::VECREDUCE_FMIN: {
+    assert(Op->getFlags().hasNoNaNs() && "fmin vector reduction needs NoNaN flag");
+    return DAG.getNode(
+        ISD::INTRINSIC_WO_CHAIN, dl, Op.getValueType(),
+        DAG.getConstant(Intrinsic::aarch64_neon_fminnmv, dl, MVT::i32),
+        Op.getOperand(0));
+  }
+  default:
+    llvm_unreachable("Unhandled reduction");
+  }
+}
+
 /// getTgtMemIntrinsic - Represent NEON load and store intrinsics as
 /// MemIntrinsicNodes.  The associated MachineMemOperands record the alignment
 /// specified in the intrinsic calls.
 bool AArch64TargetLowering::getTgtMemIntrinsic(IntrinsicInfo &Info,
                                                const CallInst &I,
                                                unsigned Intrinsic) const {
   auto &DL = I.getModule()->getDataLayout();
   switch (Intrinsic) {
static SDValue performSTORECombine(SDNode *N,
 
   if (Subtarget->supportsAddressTopByteIgnored() &&
       performTBISimplification(N->getOperand(2), DCI, DAG))
     return SDValue(N, 0);
 
   return SDValue();
 }
 
-/// This function handles the log2-shuffle pattern produced by the
-/// LoopVectorizer for the across vector reduction. It consists of
-/// log2(NumVectorElements) steps and, in each step, 2^(s) elements
-/// are reduced, where s is an induction variable from 0 to
-/// log2(NumVectorElements).
-static SDValue tryMatchAcrossLaneShuffleForReduction(SDNode *N, SDValue OpV,
-                                                     unsigned Op,
-                                                     SelectionDAG &DAG) {
-  EVT VTy = OpV->getOperand(0).getValueType();
-  if (!VTy.isVector())
-    return SDValue();
-
-  int NumVecElts = VTy.getVectorNumElements();
-  if (Op == ISD::FMAXNUM || Op == ISD::FMINNUM) {
-    if (NumVecElts != 4)
-      return SDValue();
-  } else {
-    if (NumVecElts != 4 && NumVecElts != 8 && NumVecElts != 16)
-      return SDValue();
-  }
-
-  int NumExpectedSteps = APInt(8, NumVecElts).logBase2();
-  SDValue PreOp = OpV;
-  // Iterate over each step of the across vector reduction.
-  for (int CurStep = 0; CurStep != NumExpectedSteps; ++CurStep) {
-    SDValue CurOp = PreOp.getOperand(0);
-    SDValue Shuffle = PreOp.getOperand(1);
-    if (Shuffle.getOpcode() != ISD::VECTOR_SHUFFLE) {
-      // Try to swap the 1st and 2nd operand as add and min/max instructions
-      // are commutative.
-      CurOp = PreOp.getOperand(1);
-      Shuffle = PreOp.getOperand(0);
-      if (Shuffle.getOpcode() != ISD::VECTOR_SHUFFLE)
-        return SDValue();
-    }
-
-    // Check if the input vector is fed by the operator we want to handle,
-    // except the last step; the very first input vector is not necessarily
-    // the same operator we are handling.
-    if (CurOp.getOpcode() != Op && (CurStep != (NumExpectedSteps - 1)))
-      return SDValue();
-
-    // Check if it forms one step of the across vector reduction.
-    // E.g.,
-    //   %cur = add %1, %0
-    //   %shuffle = vector_shuffle %cur, <2, 3, u, u>
-    //   %pre = add %cur, %shuffle
-    if (Shuffle.getOperand(0) != CurOp)
-      return SDValue();
-
-    int NumMaskElts = 1 << CurStep;
-    ArrayRef<int> Mask = cast<ShuffleVectorSDNode>(Shuffle)->getMask();
-    // Check mask values in each step.
-    // We expect the shuffle mask in each step follows a specific pattern
-    // denoted here by the <M, U> form, where M is a sequence of integers
-    // starting from NumMaskElts, increasing by 1, and the number integers
-    // in M should be NumMaskElts. U is a sequence of UNDEFs and the number
-    // of undef in U should be NumVecElts - NumMaskElts.
-    // E.g., for <8 x i16>, mask values in each step should be :
-    //   step 0 : <1,u,u,u,u,u,u,u>
-    //   step 1 : <2,3,u,u,u,u,u,u>
-    //   step 2 : <4,5,6,7,u,u,u,u>
-    for (int i = 0; i < NumVecElts; ++i)
-      if ((i < NumMaskElts && Mask[i] != (NumMaskElts + i)) ||
-          (i >= NumMaskElts && !(Mask[i] < 0)))
-        return SDValue();
-
-    PreOp = CurOp;
-  }
-  unsigned Opcode;
-  bool IsIntrinsic = false;
-
-  switch (Op) {
-  default:
-    llvm_unreachable("Unexpected operator for across vector reduction");
-  case ISD::ADD:
-    Opcode = AArch64ISD::UADDV;
-    break;
-  case ISD::SMAX:
-    Opcode = AArch64ISD::SMAXV;
-    break;
-  case ISD::UMAX:
-    Opcode = AArch64ISD::UMAXV;
-    break;
-  case ISD::SMIN:
-    Opcode = AArch64ISD::SMINV;
-    break;
-  case ISD::UMIN:
-    Opcode = AArch64ISD::UMINV;
-    break;
-  case ISD::FMAXNUM:
-    Opcode = Intrinsic::aarch64_neon_fmaxnmv;
-    IsIntrinsic = true;
-    break;
-  case ISD::FMINNUM:
-    Opcode = Intrinsic::aarch64_neon_fminnmv;
-    IsIntrinsic = true;
-    break;
-  }
-  SDLoc DL(N);
-
-  return IsIntrinsic
-             ? DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, N->getValueType(0),
-                           DAG.getConstant(Opcode, DL, MVT::i32), PreOp)
-             : DAG.getNode(
-                   ISD::EXTRACT_VECTOR_ELT, DL, N->getValueType(0),
-                   DAG.getNode(Opcode, DL, PreOp.getSimpleValueType(), PreOp),
-                   DAG.getConstant(0, DL, MVT::i64));
-}
-
-/// Target-specific DAG combine for the across vector min/max reductions.
-/// This function specifically handles the final clean-up step of the vector
-/// min/max reductions produced by the LoopVectorizer. It is the log2-shuffle
-/// pattern, which narrows down and finds the final min/max value from all
-/// elements of the vector.
-/// For example, for a <16 x i8> vector :
-///   svn0 = vector_shuffle %0, undef<8,9,10,11,12,13,14,15,u,u,u,u,u,u,u,u>
-///   %smax0 = smax %arr, svn0
-///   %svn1 = vector_shuffle %smax0, undef<4,5,6,7,u,u,u,u,u,u,u,u,u,u,u,u>
-///   %smax1 = smax %smax0, %svn1
-///   %svn2 = vector_shuffle %smax1, undef<2,3,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
-///   %smax2 = smax %smax1, svn2
-///   %svn3 = vector_shuffle %smax2, undef<1,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
-///   %sc = setcc %smax2, %svn3, gt
-///   %n0 = extract_vector_elt %sc, #0
-///   %n1 = extract_vector_elt %smax2, #0
-///   %n2 = extract_vector_elt $smax2, #1
-///   %result = select %n0, %n1, n2
-///     becomes :
-///   %1 = smaxv %0
-///   %result = extract_vector_elt %1, 0
-static SDValue
-performAcrossLaneMinMaxReductionCombine(SDNode *N, SelectionDAG &DAG,
-                                        const AArch64Subtarget *Subtarget) {
-  if (!Subtarget->hasNEON())
-    return SDValue();
-
-  SDValue N0 = N->getOperand(0);
-  SDValue IfTrue = N->getOperand(1);
-  SDValue IfFalse = N->getOperand(2);
-
-  // Check if the SELECT merges up the final result of the min/max
-  // from a vector.
-  if (N0.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
-      IfTrue.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
-      IfFalse.getOpcode() != ISD::EXTRACT_VECTOR_ELT)
-    return SDValue();
-
-  // Expect N0 is fed by SETCC.
-  SDValue SetCC = N0.getOperand(0);
-  EVT SetCCVT = SetCC.getValueType();
-  if (SetCC.getOpcode() != ISD::SETCC || !SetCCVT.isVector() ||
-      SetCCVT.getVectorElementType() != MVT::i1)
-    return SDValue();
-
-  SDValue VectorOp = SetCC.getOperand(0);
-  unsigned Op = VectorOp->getOpcode();
-  // Check if the input vector is fed by the operator we want to handle.
-  if (Op != ISD::SMAX && Op != ISD::UMAX && Op != ISD::SMIN &&
-      Op != ISD::UMIN && Op != ISD::FMAXNUM && Op != ISD::FMINNUM)
-    return SDValue();
-
-  EVT VTy = VectorOp.getValueType();
-  if (!VTy.isVector())
-    return SDValue();
-
-  if (VTy.getSizeInBits() < 64)
-    return SDValue();
-
-  EVT EltTy = VTy.getVectorElementType();
-  if (Op == ISD::FMAXNUM || Op == ISD::FMINNUM) {
-    if (EltTy != MVT::f32)
-      return SDValue();
-  } else {
-    if (EltTy != MVT::i32 && EltTy != MVT::i16 && EltTy != MVT::i8)
-      return SDValue();
-  }
-
-  // Check if extracting from the same vector.
-  // For example,
-  //   %sc = setcc %vector, %svn1, gt
-  //   %n0 = extract_vector_elt %sc, #0
-  //   %n1 = extract_vector_elt %vector, #0
-  //   %n2 = extract_vector_elt $vector, #1
-  if (!(VectorOp == IfTrue->getOperand(0) &&
-        VectorOp == IfFalse->getOperand(0)))
-    return SDValue();
-
-  // Check if the condition code is matched with the operator type.
-  ISD::CondCode CC = cast<CondCodeSDNode>(SetCC->getOperand(2))->get();
-  if ((Op == ISD::SMAX && CC != ISD::SETGT && CC != ISD::SETGE) ||
-      (Op == ISD::UMAX && CC != ISD::SETUGT && CC != ISD::SETUGE) ||
-      (Op == ISD::SMIN && CC != ISD::SETLT && CC != ISD::SETLE) ||
-      (Op == ISD::UMIN && CC != ISD::SETULT && CC != ISD::SETULE) ||
-      (Op == ISD::FMAXNUM && CC != ISD::SETOGT && CC != ISD::SETOGE &&
-       CC != ISD::SETUGT && CC != ISD::SETUGE && CC != ISD::SETGT &&
-       CC != ISD::SETGE) ||
-      (Op == ISD::FMINNUM && CC != ISD::SETOLT && CC != ISD::SETOLE &&
-       CC != ISD::SETULT && CC != ISD::SETULE && CC != ISD::SETLT &&
-       CC != ISD::SETLE))
-    return SDValue();
-
-  // Expect to check only lane 0 from the vector SETCC.
-  if (!isNullConstant(N0.getOperand(1)))
-    return SDValue();
-
-  // Expect to extract the true value from lane 0.
-  if (!isNullConstant(IfTrue.getOperand(1)))
-    return SDValue();
-
-  // Expect to extract the false value from lane 1.
-  if (!isOneConstant(IfFalse.getOperand(1)))
-    return SDValue();
-
-  return tryMatchAcrossLaneShuffleForReduction(N, SetCC, Op, DAG);
-}
-
-/// Target-specific DAG combine for the across vector add reduction.
-/// This function specifically handles the final clean-up step of the vector
-/// add reduction produced by the LoopVectorizer. It is the log2-shuffle
-/// pattern, which adds all elements of a vector together.
-/// For example, for a <4 x i32> vector :
-///   %1 = vector_shuffle %0, <2,3,u,u>
-///   %2 = add %0, %1
-///   %3 = vector_shuffle %2, <1,u,u,u>
-///   %4 = add %2, %3
-///   %result = extract_vector_elt %4, 0
-/// becomes :
-///   %0 = uaddv %0
-///   %result = extract_vector_elt %0, 0
-static SDValue
-performAcrossLaneAddReductionCombine(SDNode *N, SelectionDAG &DAG,
-                                     const AArch64Subtarget *Subtarget) {
-  if (!Subtarget->hasNEON())
-    return SDValue();
-  SDValue N0 = N->getOperand(0);
-  SDValue N1 = N->getOperand(1);
-
-  // Check if the input vector is fed by the ADD.
-  if (N0->getOpcode() != ISD::ADD)
-    return SDValue();
-
-  // The vector extract idx must constant zero because we only expect the final
-  // result of the reduction is placed in lane 0.
-  if (!isNullConstant(N1))
-    return SDValue();
-
-  EVT VTy = N0.getValueType();
-  if (!VTy.isVector())
-    return SDValue();
-
-  EVT EltTy = VTy.getVectorElementType();
-  if (EltTy != MVT::i32 && EltTy != MVT::i16 && EltTy != MVT::i8)
-    return SDValue();
-
-  if (VTy.getSizeInBits() < 64)
-    return SDValue();
-
-  return tryMatchAcrossLaneShuffleForReduction(N, N0, ISD::ADD, DAG);
-}
 
 /// Target-specific DAG combine function for NEON load/store intrinsics
 /// to merge base address updates.
 static SDValue performNEONPostLDSTCombine(SDNode *N,
                                           TargetLowering::DAGCombinerInfo &DCI,
                                           SelectionDAG &DAG) {
   if (DCI.isBeforeLegalize() || DCI.isCalledByLegalizer())
     return SDValue();
SDValue AArch64TargetLowering::PerformDAGCombine(SDNode *N,
   case ISD::ANY_EXTEND:
   case ISD::ZERO_EXTEND:
   case ISD::SIGN_EXTEND:
     return performExtendCombine(N, DCI, DAG);
   case ISD::BITCAST:
     return performBitcastCombine(N, DCI, DAG);
   case ISD::CONCAT_VECTORS:
     return performConcatVectorsCombine(N, DCI, DAG);
-  case ISD::SELECT: {
-    SDValue RV = performSelectCombine(N, DCI);
-    if (!RV.getNode())
-      RV = performAcrossLaneMinMaxReductionCombine(N, DAG, Subtarget);
-    return RV;
-  }
+  case ISD::SELECT:
+    return performSelectCombine(N, DCI);
   case ISD::VSELECT:
     return performVSelectCombine(N, DCI.DAG);
   case ISD::LOAD:
     if (performTBISimplification(N->getOperand(1), DCI, DAG))
       return SDValue(N, 0);
     break;
   case ISD::STORE:
     return performSTORECombine(N, DCI, DAG, Subtarget);
   case AArch64ISD::BRCOND:
     return performBRCONDCombine(N, DCI, DAG);
   case AArch64ISD::TBNZ:
   case AArch64ISD::TBZ:
     return performTBZCombine(N, DCI, DAG);
   case AArch64ISD::CSEL:
     return performCONDCombine(N, DCI, DAG, 2, 3);
   case AArch64ISD::DUP:
     return performPostLD1Combine(N, DCI, false);
   case AArch64ISD::NVCAST:
     return performNVCASTCombine(N);
   case ISD::INSERT_VECTOR_ELT:
     return performPostLD1Combine(N, DCI, true);
-  case ISD::EXTRACT_VECTOR_ELT:
-    return performAcrossLaneAddReductionCombine(N, DAG, Subtarget);
   case ISD::INTRINSIC_VOID:
   case ISD::INTRINSIC_W_CHAIN:
     switch (cast<ConstantSDNode>(N->getOperand(1))->getZExtValue()) {
     case Intrinsic::aarch64_neon_ld2:
     case Intrinsic::aarch64_neon_ld3:
     case Intrinsic::aarch64_neon_ld4:
     case Intrinsic::aarch64_neon_ld1x2:
     case Intrinsic::aarch64_neon_ld1x3:

 void AArch64TargetLowering::ReplaceNodeResults(
     SDNode *N, SmallVectorImpl<SDValue> &Results, SelectionDAG &DAG) const {
   switch (N->getOpcode()) {
   default:
     llvm_unreachable("Don't know how to custom expand this");
   case ISD::BITCAST:
     ReplaceBITCASTResults(N, Results, DAG);
     return;
+  case ISD::VECREDUCE_ADD:
+  case ISD::VECREDUCE_SMAX:
+  case ISD::VECREDUCE_SMIN:
+  case ISD::VECREDUCE_UMAX:
+  case ISD::VECREDUCE_UMIN:
+    Results.push_back(LowerVECREDUCE(SDValue(N, 0), DAG));
+    return;
+
   case AArch64ISD::SADDV:
     ReplaceReductionResults(N, Results, DAG, ISD::ADD, AArch64ISD::SADDV);
     return;
   case AArch64ISD::UADDV:
     ReplaceReductionResults(N, Results, DAG, ISD::ADD, AArch64ISD::UADDV);
     return;
   case AArch64ISD::SMINV:
     ReplaceReductionResults(N, Results, DAG, ISD::SMIN, AArch64ISD::SMINV);

llvm/trunk/lib/Target/AArch64/AArch64TargetTransformInfo.h

public:
 
   unsigned getMinPrefetchStride();
 
   unsigned getMaxPrefetchIterationsAhead();
 
   bool shouldExpandReduction(const IntrinsicInst *II) const {
     return false;
   }
+
+  bool useReductionIntrinsic(unsigned Opcode, Type *Ty,
+                             TTI::ReductionFlags Flags) const;
   /// @}
 };
 
 } // end namespace llvm
 
 #endif

llvm/trunk/lib/Target/AArch64/AArch64TargetTransformInfo.cpp

 
 unsigned AArch64TTIImpl::getMinPrefetchStride() {
   return ST->getMinPrefetchStride();
 }
 
 unsigned AArch64TTIImpl::getMaxPrefetchIterationsAhead() {
   return ST->getMaxPrefetchIterationsAhead();
 }
+
+bool AArch64TTIImpl::useReductionIntrinsic(unsigned Opcode, Type *Ty,
+                                           TTI::ReductionFlags Flags) const {
+  assert(isa<VectorType>(Ty) && "Expected Ty to be a vector type");
+  switch (Opcode) {
+  case Instruction::FAdd:
+  case Instruction::FMul:
+  case Instruction::And:
+  case Instruction::Or:
+  case Instruction::Xor:
+  case Instruction::Mul:
+    return false;
+  case Instruction::Add:
+    return Ty->getScalarSizeInBits() * Ty->getVectorNumElements() >= 128;
+  case Instruction::ICmp:
+    return Ty->getScalarSizeInBits() < 64;
+  case Instruction::FCmp:
+    return Flags.NoNaN;
+  default:
+    llvm_unreachable("Unhandled reduction opcode");
+  }
+  return false;
+}

llvm/trunk/test/CodeGen/AArch64/aarch64-addv.ll

 ; RUN: llc < %s -mtriple=aarch64-eabi -aarch64-neon-syntax=generic | FileCheck %s
 
+; Function Attrs: nounwind readnone
+declare i64 @llvm.experimental.vector.reduce.add.i64.v2i64(<2 x i64>)
+declare i32 @llvm.experimental.vector.reduce.add.i32.v4i32(<4 x i32>)
+declare i16 @llvm.experimental.vector.reduce.add.i16.v8i16(<8 x i16>)
+declare i8 @llvm.experimental.vector.reduce.add.i8.v16i8(<16 x i8>)
+
 define i8 @add_B(<16 x i8>* %arr)  {
 ; CHECK-LABEL: add_B
 ; CHECK: addv {{b[0-9]+}}, {{v[0-9]+}}.16b
   %bin.rdx = load <16 x i8>, <16 x i8>* %arr
-  %rdx.shuf0 = shufflevector <16 x i8> %bin.rdx, <16 x i8> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx0 = add <16 x i8> %bin.rdx, %rdx.shuf0
-  %rdx.shuf = shufflevector <16 x i8> %bin.rdx0, <16 x i8> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef >
-  %bin.rdx11 = add <16 x i8> %bin.rdx0, %rdx.shuf
-  %rdx.shuf12 = shufflevector <16 x i8> %bin.rdx11, <16 x i8> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef>
-  %bin.rdx13 = add <16 x i8> %bin.rdx11, %rdx.shuf12
-  %rdx.shuf13 = shufflevector <16 x i8> %bin.rdx13, <16 x i8> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef>
-  %bin.rdx14 = add <16 x i8> %bin.rdx13, %rdx.shuf13
-  %r = extractelement <16 x i8> %bin.rdx14, i32 0
+  %r = call i8 @llvm.experimental.vector.reduce.add.i8.v16i8(<16 x i8> %bin.rdx)
   ret i8 %r
 }
 
 define i16 @add_H(<8 x i16>* %arr)  {
 ; CHECK-LABEL: add_H
 ; CHECK: addv {{h[0-9]+}}, {{v[0-9]+}}.8h
   %bin.rdx = load <8 x i16>, <8 x i16>* %arr
-  %rdx.shuf = shufflevector <8 x i16> %bin.rdx, <8 x i16> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef,i32 undef, i32 undef>
-  %bin.rdx11 = add <8 x i16> %bin.rdx, %rdx.shuf
-  %rdx.shuf12 = shufflevector <8 x i16> %bin.rdx11, <8 x i16> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx13 = add <8 x i16> %bin.rdx11, %rdx.shuf12
-  %rdx.shuf13 = shufflevector <8 x i16> %bin.rdx13, <8 x i16> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx14 = add <8 x i16> %bin.rdx13, %rdx.shuf13
-  %r = extractelement <8 x i16> %bin.rdx14, i32 0
+  %r = call i16 @llvm.experimental.vector.reduce.add.i16.v8i16(<8 x i16> %bin.rdx)
   ret i16 %r
 }
 
 define i32 @add_S( <4 x i32>* %arr)  {
 ; CHECK-LABEL: add_S
 ; CHECK: addv {{s[0-9]+}}, {{v[0-9]+}}.4s
   %bin.rdx = load <4 x i32>, <4 x i32>* %arr
-  %rdx.shuf = shufflevector <4 x i32> %bin.rdx, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
-  %bin.rdx11 = add <4 x i32> %bin.rdx, %rdx.shuf
-  %rdx.shuf12 = shufflevector <4 x i32> %bin.rdx11, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
-  %bin.rdx13 = add <4 x i32> %bin.rdx11, %rdx.shuf12
-  %r = extractelement <4 x i32> %bin.rdx13, i32 0
+  %r = call i32 @llvm.experimental.vector.reduce.add.i32.v4i32(<4 x i32> %bin.rdx)
   ret i32 %r
 }
 
 define i64 @add_D(<2 x i64>* %arr)  {
 ; CHECK-LABEL: add_D
 ; CHECK-NOT: addv
   %bin.rdx = load <2 x i64>, <2 x i64>* %arr
-  %rdx.shuf0 = shufflevector <2 x i64> %bin.rdx, <2 x i64> undef, <2 x i32> <i32 1, i32 undef>
-  %bin.rdx0 = add <2 x i64> %bin.rdx, %rdx.shuf0
-  %r = extractelement <2 x i64> %bin.rdx0, i32 0
+  %r = call i64 @llvm.experimental.vector.reduce.add.i64.v2i64(<2 x i64> %bin.rdx)
   ret i64 %r
 }
 
+declare i32 @llvm.experimental.vector.reduce.add.i32.v8i32(<8 x i32>)
+
 define i32 @oversized_ADDV_256(i8* noalias nocapture readonly %arg1, i8* noalias nocapture readonly %arg2) {
 ; CHECK-LABEL: oversized_ADDV_256
 ; CHECK: addv {{s[0-9]+}}, {{v[0-9]+}}.4s
 entry:
   %0 = bitcast i8* %arg1 to <8 x i8>*
   %1 = load <8 x i8>, <8 x i8>* %0, align 1
   %2 = zext <8 x i8> %1 to <8 x i32>
   %3 = bitcast i8* %arg2 to <8 x i8>*
   %4 = load <8 x i8>, <8 x i8>* %3, align 1
   %5 = zext <8 x i8> %4 to <8 x i32>
   %6 = sub nsw <8 x i32> %2, %5
   %7 = icmp slt <8 x i32> %6, zeroinitializer
   %8 = sub nsw <8 x i32> zeroinitializer, %6
   %9 = select <8 x i1> %7, <8 x i32> %8, <8 x i32> %6
-  %rdx.shuf = shufflevector <8 x i32> %9, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx = add <8 x i32> %9, %rdx.shuf
-  %rdx.shuf1 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx2 = add <8 x i32> %bin.rdx, %rdx.shuf1
-  %rdx.shuf3 = shufflevector <8 x i32> %bin.rdx2, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx4 = add <8 x i32> %bin.rdx2, %rdx.shuf3
-  %10 = extractelement <8 x i32> %bin.rdx4, i32 0
-  ret i32 %10
+  %r = call i32 @llvm.experimental.vector.reduce.add.i32.v8i32(<8 x i32> %9)
+  ret i32 %r
 }
 
+declare i32 @llvm.experimental.vector.reduce.add.i32.v16i32(<16 x i32>)
+
 define i32 @oversized_ADDV_512(<16 x i32>* %arr)  {
 ; CHECK-LABEL: oversized_ADDV_512
 ; CHECK: addv {{s[0-9]+}}, {{v[0-9]+}}.4s
   %bin.rdx = load <16 x i32>, <16 x i32>* %arr
-
-  %rdx.shuf0 = shufflevector <16 x i32> %bin.rdx, <16 x i32> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx0 = add <16 x i32> %bin.rdx, %rdx.shuf0
-
-  %rdx.shuf = shufflevector <16 x i32> %bin.rdx0, <16 x i32> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef >
-  %bin.rdx11 = add <16 x i32> %bin.rdx0, %rdx.shuf
-
-  %rdx.shuf12 = shufflevector <16 x i32> %bin.rdx11, <16 x i32> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef>
-  %bin.rdx13 = add <16 x i32> %bin.rdx11, %rdx.shuf12
-
-  %rdx.shuf13 = shufflevector <16 x i32> %bin.rdx13, <16 x i32> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef, i32 undef, i32 undef,i32 undef, i32 undef>
-  %bin.rdx14 = add <16 x i32> %bin.rdx13, %rdx.shuf13
-
-  %r = extractelement <16 x i32> %bin.rdx14, i32 0
+  %r = call i32 @llvm.experimental.vector.reduce.add.i32.v16i32(<16 x i32> %bin.rdx)
   ret i32 %r
 }

llvm/trunk/test/CodeGen/AArch64/aarch64-minmaxv.ll

 ; RUN: llc < %s -mtriple=aarch64-linux--gnu -aarch64-neon-syntax=generic | FileCheck %s
 
 target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
 
+declare i8 @llvm.experimental.vector.reduce.smax.i8.v16i8(<16 x i8>)
+declare i16 @llvm.experimental.vector.reduce.smax.i16.v8i16(<8 x i16>)
+declare i32 @llvm.experimental.vector.reduce.smax.i32.v4i32(<4 x i32>)
+declare i8 @llvm.experimental.vector.reduce.umax.i8.v16i8(<16 x i8>)
+declare i16 @llvm.experimental.vector.reduce.umax.i16.v8i16(<8 x i16>)
+declare i32 @llvm.experimental.vector.reduce.umax.i32.v4i32(<4 x i32>)
+
+declare i8 @llvm.experimental.vector.reduce.smin.i8.v16i8(<16 x i8>)
+declare i16 @llvm.experimental.vector.reduce.smin.i16.v8i16(<8 x i16>)
+declare i32 @llvm.experimental.vector.reduce.smin.i32.v4i32(<4 x i32>)
+declare i8 @llvm.experimental.vector.reduce.umin.i8.v16i8(<16 x i8>)
+declare i16 @llvm.experimental.vector.reduce.umin.i16.v8i16(<8 x i16>)
+declare i32 @llvm.experimental.vector.reduce.umin.i32.v4i32(<4 x i32>)
+
+declare float @llvm.experimental.vector.reduce.fmax.f32.v4f32(<4 x float>)
+declare float @llvm.experimental.vector.reduce.fmin.f32.v4f32(<4 x float>)
+
 ; CHECK-LABEL: smax_B
 ; CHECK: smaxv {{b[0-9]+}}, {{v[0-9]+}}.16b
 define i8 @smax_B(<16 x i8>* nocapture readonly %arr)  {
   %arr.load = load <16 x i8>, <16 x i8>* %arr
-  %rdx.shuf = shufflevector <16 x i8> %arr.load, <16 x i8> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp sgt <16 x i8> %arr.load, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i8> %arr.load, <16 x i8> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i8> %rdx.minmax.select23, <16 x i8> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp sgt <16 x i8> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i8> %rdx.minmax.select23, <16 x i8> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i8> %rdx.minmax.select26, <16 x i8> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp sgt <16 x i8> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i8> %rdx.minmax.select26, <16 x i8> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i8> %rdx.minmax.select29, <16 x i8> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp sgt <16 x i8> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i8> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i8> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i8 %rdx.minmax.select29.elt, i8 %rdx.shuf30.elt
+  %r = call i8 @llvm.experimental.vector.reduce.smax.i8.v16i8(<16 x i8> %arr.load)
   ret i8 %r
 }
 
 ; CHECK-LABEL: smax_H
 ; CHECK: smaxv {{h[0-9]+}}, {{v[0-9]+}}.8h
 define i16 @smax_H(<8 x i16>* nocapture readonly %arr) {
-  %rdx.minmax.select = load <8 x i16>, <8 x i16>* %arr
-  %rdx.shuf = shufflevector <8 x i16> %rdx.minmax.select, <8 x i16> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp23 = icmp sgt <8 x i16> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select24 = select <8 x i1> %rdx.minmax.cmp23, <8 x i16> %rdx.minmax.select, <8 x i16> %rdx.shuf
-  %rdx.shuf25 = shufflevector <8 x i16> %rdx.minmax.select24, <8 x i16> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp26 = icmp sgt <8 x i16> %rdx.minmax.select24, %rdx.shuf25
-  %rdx.minmax.select27 = select <8 x i1> %rdx.minmax.cmp26, <8 x i16> %rdx.minmax.select24, <8 x i16> %rdx.shuf25
-  %rdx.shuf28 = shufflevector <8 x i16> %rdx.minmax.select27, <8 x i16> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp29 = icmp sgt <8 x i16> %rdx.minmax.select27, %rdx.shuf28
-  %rdx.minmax.cmp29.elt = extractelement <8 x i1> %rdx.minmax.cmp29, i32 0
-  %rdx.minmax.select27.elt = extractelement <8 x i16> %rdx.minmax.select27, i32 0
-  %rdx.shuf28.elt = extractelement <8 x i16> %rdx.minmax.select27, i32 1
-  %r = select i1 %rdx.minmax.cmp29.elt, i16 %rdx.minmax.select27.elt, i16 %rdx.shuf28.elt
+  %arr.load = load <8 x i16>, <8 x i16>* %arr
+  %r = call i16 @llvm.experimental.vector.reduce.smax.i16.v8i16(<8 x i16> %arr.load)
   ret i16 %r
 }
 
 ; CHECK-LABEL: smax_S
 ; CHECK: smaxv {{s[0-9]+}}, {{v[0-9]+}}.4s
 define i32 @smax_S(<4 x i32> * nocapture readonly %arr)  {
-  %rdx.minmax.select = load <4 x i32>, <4 x i32>* %arr
-  %rdx.shuf = shufflevector <4 x i32> %rdx.minmax.select, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
-  %rdx.minmax.cmp18 = icmp sgt <4 x i32> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select19 = select <4 x i1> %rdx.minmax.cmp18, <4 x i32> %rdx.minmax.select, <4 x i32> %rdx.shuf
-  %rdx.shuf20 = shufflevector <4 x i32> %rdx.minmax.select19, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp21 = icmp sgt <4 x i32> %rdx.minmax.select19, %rdx.shuf20
-  %rdx.minmax.cmp21.elt = extractelement <4 x i1> %rdx.minmax.cmp21, i32 0
-  %rdx.minmax.select19.elt = extractelement <4 x i32> %rdx.minmax.select19, i32 0
-  %rdx.shuf20.elt = extractelement <4 x i32> %rdx.minmax.select19, i32 1
-  %r = select i1 %rdx.minmax.cmp21.elt, i32 %rdx.minmax.select19.elt, i32 %rdx.shuf20.elt
+  %arr.load = load <4 x i32>, <4 x i32>* %arr
+  %r = call i32 @llvm.experimental.vector.reduce.smax.i32.v4i32(<4 x i32> %arr.load)
   ret i32 %r
 }
 
-; CHECK-LABEL: smax_D
-; CHECK-NOT: smaxv
-define i64 @smax_D(<2 x i64>* nocapture readonly %arr) {
-  %rdx.minmax.select = load <2 x i64>, <2 x i64>* %arr
-  %rdx.shuf = shufflevector <2 x i64> %rdx.minmax.select, <2 x i64> undef, <2 x i32> <i32 1, i32 undef>
-  %rdx.minmax.cmp18 = icmp sgt <2 x i64> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.cmp18.elt = extractelement <2 x i1> %rdx.minmax.cmp18, i32 0
-  %rdx.minmax.select.elt = extractelement <2 x i64> %rdx.minmax.select, i32 0
-  %rdx.shuf.elt = extractelement <2 x i64> %rdx.minmax.select, i32 1
-  %r = select i1 %rdx.minmax.cmp18.elt, i64 %rdx.minmax.select.elt, i64 %rdx.shuf.elt
-  ret i64 %r
-}
-
-
 ; CHECK-LABEL: umax_B
 ; CHECK: umaxv {{b[0-9]+}}, {{v[0-9]+}}.16b
 define i8 @umax_B(<16 x i8>* nocapture readonly %arr)  {
-  %rdx.minmax.select = load <16 x i8>, <16 x i8>* %arr
-  %rdx.shuf = shufflevector <16 x i8> %rdx.minmax.select, <16 x i8> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp ugt <16 x i8> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i8> %rdx.minmax.select, <16 x i8> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i8> %rdx.minmax.select23, <16 x i8> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp ugt <16 x i8> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i8> %rdx.minmax.select23, <16 x i8> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i8> %rdx.minmax.select26, <16 x i8> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp ugt <16 x i8> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i8> %rdx.minmax.select26, <16 x i8> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i8> %rdx.minmax.select29, <16 x i8> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp ugt <16 x i8> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i8> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i8> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i8 %rdx.minmax.select29.elt, i8 %rdx.shuf30.elt
+  %arr.load = load <16 x i8>, <16 x i8>* %arr
+  %r = call i8 @llvm.experimental.vector.reduce.umax.i8.v16i8(<16 x i8> %arr.load)
   ret i8 %r
 }
 
 ; CHECK-LABEL: umax_H
 ; CHECK: umaxv {{h[0-9]+}}, {{v[0-9]+}}.8h
 define i16 @umax_H(<8 x i16>* nocapture readonly %arr)  {
-  %rdx.minmax.select = load <8 x i16>, <8 x i16>* %arr
-  %rdx.shuf = shufflevector <8 x i16> %rdx.minmax.select, <8 x i16> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp23 = icmp ugt <8 x i16> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select24 = select <8 x i1> %rdx.minmax.cmp23, <8 x i16> %rdx.minmax.select, <8 x i16> %rdx.shuf
-  %rdx.shuf25 = shufflevector <8 x i16> %rdx.minmax.select24, <8 x i16> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp26 = icmp ugt <8 x i16> %rdx.minmax.select24, %rdx.shuf25
-  %rdx.minmax.select27 = select <8 x i1> %rdx.minmax.cmp26, <8 x i16> %rdx.minmax.select24, <8 x i16> %rdx.shuf25
-  %rdx.shuf28 = shufflevector <8 x i16> %rdx.minmax.select27, <8 x i16> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp29 = icmp ugt <8 x i16> %rdx.minmax.select27, %rdx.shuf28
-  %rdx.minmax.cmp29.elt = extractelement <8 x i1> %rdx.minmax.cmp29, i32 0
-  %rdx.minmax.select27.elt = extractelement <8 x i16> %rdx.minmax.select27, i32 0
-  %rdx.shuf28.elt = extractelement <8 x i16> %rdx.minmax.select27, i32 1
-  %r = select i1 %rdx.minmax.cmp29.elt, i16 %rdx.minmax.select27.elt, i16 %rdx.shuf28.elt
+  %arr.load = load <8 x i16>, <8 x i16>* %arr
+  %r = call i16 @llvm.experimental.vector.reduce.umax.i16.v8i16(<8 x i16> %arr.load)
   ret i16 %r
 }
 
 ; CHECK-LABEL: umax_S
 ; CHECK: umaxv {{s[0-9]+}}, {{v[0-9]+}}.4s
 define i32 @umax_S(<4 x i32>* nocapture readonly %arr) {
-  %rdx.minmax.select  = load <4 x i32>, <4 x i32>* %arr
-  %rdx.shuf = shufflevector <4 x i32> %rdx.minmax.select, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
-  %rdx.minmax.cmp18 = icmp ugt <4 x i32> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select19 = select <4 x i1> %rdx.minmax.cmp18, <4 x i32> %rdx.minmax.select, <4 x i32> %rdx.shuf
-  %rdx.shuf20 = shufflevector <4 x i32> %rdx.minmax.select19, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp21 = icmp ugt <4 x i32> %rdx.minmax.select19, %rdx.shuf20
-  %rdx.minmax.cmp21.elt = extractelement <4 x i1> %rdx.minmax.cmp21, i32 0
-  %rdx.minmax.select19.elt = extractelement <4 x i32> %rdx.minmax.select19, i32 0
-  %rdx.shuf20.elt = extractelement <4 x i32> %rdx.minmax.select19, i32 1
-  %r = select i1 %rdx.minmax.cmp21.elt, i32 %rdx.minmax.select19.elt, i32 %rdx.shuf20.elt
+  %arr.load = load <4 x i32>, <4 x i32>* %arr
+  %r = call i32 @llvm.experimental.vector.reduce.umax.i32.v4i32(<4 x i32> %arr.load)
   ret i32 %r
 }
 
-; CHECK-LABEL: umax_D
-; CHECK-NOT: umaxv
-define i64 @umax_D(<2 x i64>* nocapture readonly %arr)  {
-  %rdx.minmax.select = load <2 x i64>, <2 x i64>* %arr
-  %rdx.shuf = shufflevector <2 x i64> %rdx.minmax.select, <2 x i64> undef, <2 x i32> <i32 1, i32 undef>
-  %rdx.minmax.cmp18 = icmp ugt <2 x i64> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.cmp18.elt = extractelement <2 x i1> %rdx.minmax.cmp18, i32 0
-  %rdx.minmax.select.elt = extractelement <2 x i64> %rdx.minmax.select, i32 0
-  %rdx.shuf.elt = extractelement <2 x i64> %rdx.minmax.select, i32 1
-  %r = select i1 %rdx.minmax.cmp18.elt, i64 %rdx.minmax.select.elt, i64 %rdx.shuf.elt
-  ret i64 %r
-}
-
-
 ; CHECK-LABEL: smin_B
 ; CHECK: sminv {{b[0-9]+}}, {{v[0-9]+}}.16b
 define i8 @smin_B(<16 x i8>* nocapture readonly %arr) {
-  %rdx.minmax.select = load <16 x i8>, <16 x i8>* %arr
-  %rdx.shuf = shufflevector <16 x i8> %rdx.minmax.select, <16 x i8> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp slt <16 x i8> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i8> %rdx.minmax.select, <16 x i8> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i8> %rdx.minmax.select23, <16 x i8> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp slt <16 x i8> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i8> %rdx.minmax.select23, <16 x i8> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i8> %rdx.minmax.select26, <16 x i8> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp slt <16 x i8> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i8> %rdx.minmax.select26, <16 x i8> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i8> %rdx.minmax.select29, <16 x i8> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp slt <16 x i8> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i8> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i8> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i8 %rdx.minmax.select29.elt, i8 %rdx.shuf30.elt
+  %arr.load = load <16 x i8>, <16 x i8>* %arr
+  %r = call i8 @llvm.experimental.vector.reduce.smin.i8.v16i8(<16 x i8> %arr.load)
   ret i8 %r
 }
 
 ; CHECK-LABEL: smin_H
 ; CHECK: sminv {{h[0-9]+}}, {{v[0-9]+}}.8h
 define i16 @smin_H(<8 x i16>* nocapture readonly %arr) {
-  %rdx.minmax.select = load <8 x i16>, <8 x i16>* %arr
-  %rdx.shuf = shufflevector <8 x i16> %rdx.minmax.select, <8 x i16> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp23 = icmp slt <8 x i16> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select24 = select <8 x i1> %rdx.minmax.cmp23, <8 x i16> %rdx.minmax.select, <8 x i16> %rdx.shuf
-  %rdx.shuf25 = shufflevector <8 x i16> %rdx.minmax.select24, <8 x i16> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp26 = icmp slt <8 x i16> %rdx.minmax.select24, %rdx.shuf25
-  %rdx.minmax.select27 = select <8 x i1> %rdx.minmax.cmp26, <8 x i16> %rdx.minmax.select24, <8 x i16> %rdx.shuf25
-  %rdx.shuf28 = shufflevector <8 x i16> %rdx.minmax.select27, <8 x i16> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp29 = icmp slt <8 x i16> %rdx.minmax.select27, %rdx.shuf28
-  %rdx.minmax.cmp29.elt = extractelement <8 x i1> %rdx.minmax.cmp29, i32 0
-  %rdx.minmax.select27.elt = extractelement <8 x i16> %rdx.minmax.select27, i32 0
-  %rdx.shuf28.elt = extractelement <8 x i16> %rdx.minmax.select27, i32 1
-  %r = select i1 %rdx.minmax.cmp29.elt, i16 %rdx.minmax.select27.elt, i16 %rdx.shuf28.elt
+  %arr.load = load <8 x i16>, <8 x i16>* %arr
+  %r = call i16 @llvm.experimental.vector.reduce.smin.i16.v8i16(<8 x i16> %arr.load)
   ret i16 %r
 }
 
 ; CHECK-LABEL: smin_S
 ; CHECK: sminv {{s[0-9]+}}, {{v[0-9]+}}.4s
 define i32 @smin_S(<4 x i32>* nocapture readonly %arr) {
-  %rdx.minmax.select = load <4 x i32>, <4 x i32>* %arr
-  %rdx.shuf = shufflevector <4 x i32> %rdx.minmax.select, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
-  %rdx.minmax.cmp18 = icmp slt <4 x i32> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select19 = select <4 x i1> %rdx.minmax.cmp18, <4 x i32> %rdx.minmax.select, <4 x i32> %rdx.shuf
-  %rdx.shuf20 = shufflevector <4 x i32> %rdx.minmax.select19, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp21 = icmp slt <4 x i32> %rdx.minmax.select19, %rdx.shuf20
-  %rdx.minmax.cmp21.elt = extractelement <4 x i1> %rdx.minmax.cmp21, i32 0
-  %rdx.minmax.select19.elt = extractelement <4 x i32> %rdx.minmax.select19, i32 0
-  %rdx.shuf20.elt = extractelement <4 x i32> %rdx.minmax.select19, i32 1
-  %r = select i1 %rdx.minmax.cmp21.elt, i32 %rdx.minmax.select19.elt, i32 %rdx.shuf20.elt
+  %arr.load = load <4 x i32>, <4 x i32>* %arr
+  %r = call i32 @llvm.experimental.vector.reduce.smin.i32.v4i32(<4 x i32> %arr.load)
   ret i32 %r
 }
 
-; CHECK-LABEL: smin_D
-; CHECK-NOT: sminv
-define i64 @smin_D(<2 x i64>* nocapture readonly %arr) {
-  %rdx.minmax.select = load <2 x i64>, <2 x i64>* %arr
-  %rdx.shuf = shufflevector <2 x i64> %rdx.minmax.select, <2 x i64> undef, <2 x i32> <i32 1, i32 undef>
-  %rdx.minmax.cmp18 = icmp slt <2 x i64> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.cmp18.elt = extractelement <2 x i1> %rdx.minmax.cmp18, i32 0
-  %rdx.minmax.select.elt = extractelement <2 x i64> %rdx.minmax.select, i32 0
-  %rdx.shuf.elt = extractelement <2 x i64> %rdx.minmax.select, i32 1
-  %r = select i1 %rdx.minmax.cmp18.elt, i64 %rdx.minmax.select.elt, i64 %rdx.shuf.elt
-  ret i64 %r
-}
-
-
 ; CHECK-LABEL: umin_B
 ; CHECK: uminv {{b[0-9]+}}, {{v[0-9]+}}.16b
 define i8 @umin_B(<16 x i8>* nocapture readonly %arr)  {
-  %rdx.minmax.select = load <16 x i8>, <16 x i8>* %arr
-  %rdx.shuf = shufflevector <16 x i8> %rdx.minmax.select, <16 x i8> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp ult <16 x i8> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i8> %rdx.minmax.select, <16 x i8> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i8> %rdx.minmax.select23, <16 x i8> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp ult <16 x i8> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i8> %rdx.minmax.select23, <16 x i8> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i8> %rdx.minmax.select26, <16 x i8> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp ult <16 x i8> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i8> %rdx.minmax.select26, <16 x i8> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i8> %rdx.minmax.select29, <16 x i8> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp ult <16 x i8> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i8> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i8> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i8 %rdx.minmax.select29.elt, i8 %rdx.shuf30.elt
+  %arr.load = load <16 x i8>, <16 x i8>* %arr
+  %r = call i8 @llvm.experimental.vector.reduce.umin.i8.v16i8(<16 x i8> %arr.load)
   ret i8 %r
 }
 
 ; CHECK-LABEL: umin_H
 ; CHECK: uminv {{h[0-9]+}}, {{v[0-9]+}}.8h
 define i16 @umin_H(<8 x i16>* nocapture readonly %arr)  {
-  %rdx.minmax.select = load <8 x i16>, <8 x i16>* %arr
-  %rdx.shuf = shufflevector <8 x i16> %rdx.minmax.select, <8 x i16> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp23 = icmp ult <8 x i16> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select24 = select <8 x i1> %rdx.minmax.cmp23, <8 x i16> %rdx.minmax.select, <8 x i16> %rdx.shuf
-  %rdx.shuf25 = shufflevector <8 x i16> %rdx.minmax.select24, <8 x i16> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp26 = icmp ult <8 x i16> %rdx.minmax.select24, %rdx.shuf25
-  %rdx.minmax.select27 = select <8 x i1> %rdx.minmax.cmp26, <8 x i16> %rdx.minmax.select24, <8 x i16> %rdx.shuf25
-  %rdx.shuf28 = shufflevector <8 x i16> %rdx.minmax.select27, <8 x i16> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp29 = icmp ult <8 x i16> %rdx.minmax.select27, %rdx.shuf28
-  %rdx.minmax.cmp29.elt = extractelement <8 x i1> %rdx.minmax.cmp29, i32 0
-  %rdx.minmax.select27.elt = extractelement <8 x i16> %rdx.minmax.select27, i32 0
-  %rdx.shuf28.elt = extractelement <8 x i16> %rdx.minmax.select27, i32 1
-  %r = select i1 %rdx.minmax.cmp29.elt, i16 %rdx.minmax.select27.elt, i16 %rdx.shuf28.elt
+  %arr.load = load <8 x i16>, <8 x i16>* %arr
+  %r = call i16 @llvm.experimental.vector.reduce.umin.i16.v8i16(<8 x i16> %arr.load)
   ret i16 %r
 }
 
 ; CHECK-LABEL: umin_S
 ; CHECK: uminv {{s[0-9]+}}, {{v[0-9]+}}.4s
 define i32 @umin_S(<4 x i32>* nocapture readonly %arr) {
-  %rdx.minmax.select = load <4 x i32>, <4 x i32>* %arr
-  %rdx.shuf = shufflevector <4 x i32> %rdx.minmax.select, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
-  %rdx.minmax.cmp18 = icmp ult <4 x i32> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select19 = select <4 x i1> %rdx.minmax.cmp18, <4 x i32> %rdx.minmax.select, <4 x i32> %rdx.shuf
-  %rdx.shuf20 = shufflevector <4 x i32> %rdx.minmax.select19, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp21 = icmp ult <4 x i32> %rdx.minmax.select19, %rdx.shuf20
-  %rdx.minmax.cmp21.elt = extractelement <4 x i1> %rdx.minmax.cmp21, i32 0
-  %rdx.minmax.select19.elt = extractelement <4 x i32> %rdx.minmax.select19, i32 0
-  %rdx.shuf20.elt = extractelement <4 x i32> %rdx.minmax.select19, i32 1
-  %r = select i1 %rdx.minmax.cmp21.elt, i32 %rdx.minmax.select19.elt, i32 %rdx.shuf20.elt
+  %arr.load = load <4 x i32>, <4 x i32>* %arr
+  %r = call i32 @llvm.experimental.vector.reduce.umin.i32.v4i32(<4 x i32> %arr.load)
   ret i32 %r
 }
 
-; CHECK-LABEL: umin_D
-; CHECK-NOT: uminv
-define i64 @umin_D(<2 x i64>* nocapture readonly %arr)  {
-  %rdx.minmax.select = load <2 x i64>, <2 x i64>* %arr
-  %rdx.shuf = shufflevector <2 x i64> %rdx.minmax.select, <2 x i64> undef, <2 x i32> <i32 1, i32 undef>
-  %rdx.minmax.cmp18 = icmp ult <2 x i64> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.cmp18.elt = extractelement <2 x i1> %rdx.minmax.cmp18, i32 0
-  %rdx.minmax.select.elt = extractelement <2 x i64> %rdx.minmax.select, i32 0
-  %rdx.shuf.elt = extractelement <2 x i64> %rdx.minmax.select, i32 1
-  %r = select i1 %rdx.minmax.cmp18.elt, i64 %rdx.minmax.select.elt, i64 %rdx.shuf.elt
-  ret i64 %r
-}
-
 ; CHECK-LABEL: fmaxnm_S
 ; CHECK: fmaxnmv
 define float @fmaxnm_S(<4 x float>* nocapture readonly %arr) {
-  %rdx.minmax.select  = load <4 x float>, <4 x float>* %arr
-  %rdx.shuf = shufflevector <4 x float> %rdx.minmax.select, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
-  %rdx.minmax.cmp = fcmp fast oge <4 x float> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select1 = select <4 x i1> %rdx.minmax.cmp, <4 x float> %rdx.minmax.select, <4 x float> %rdx.shuf
-  %rdx.shuf1 = shufflevector <4 x float> %rdx.minmax.select1, <4 x float> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp1 = fcmp fast oge <4 x float> %rdx.minmax.select1, %rdx.shuf1
-  %rdx.minmax.cmp1.elt = extractelement <4 x i1> %rdx.minmax.cmp1, i32 0
-  %rdx.minmax.select1.elt = extractelement <4 x float> %rdx.minmax.select1, i32 0
-  %rdx.shuf1.elt = extractelement <4 x float> %rdx.minmax.select1, i32 1
-  %r = select i1 %rdx.minmax.cmp1.elt, float %rdx.minmax.select1.elt, float %rdx.shuf1.elt
+  %arr.load  = load <4 x float>, <4 x float>* %arr
+  %r = call nnan float @llvm.experimental.vector.reduce.fmax.f32.v4f32(<4 x float> %arr.load)
   ret float %r
 }
 
 ; CHECK-LABEL: fminnm_S
 ; CHECK: fminnmv
 define float @fminnm_S(<4 x float>* nocapture readonly %arr) {
-  %rdx.minmax.select  = load <4 x float>, <4 x float>* %arr
-  %rdx.shuf = shufflevector <4 x float> %rdx.minmax.select, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
-  %rdx.minmax.cmp = fcmp fast ole <4 x float> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select1 = select <4 x i1> %rdx.minmax.cmp, <4 x float> %rdx.minmax.select, <4 x float> %rdx.shuf
-  %rdx.shuf1 = shufflevector <4 x float> %rdx.minmax.select1, <4 x float> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp1 = fcmp fast ole <4 x float> %rdx.minmax.select1, %rdx.shuf1
-  %rdx.minmax.cmp1.elt = extractelement <4 x i1> %rdx.minmax.cmp1, i32 0
-  %rdx.minmax.select1.elt = extractelement <4 x float> %rdx.minmax.select1, i32 0
-  %rdx.shuf1.elt = extractelement <4 x float> %rdx.minmax.select1, i32 1
-  %r = select i1 %rdx.minmax.cmp1.elt, float %rdx.minmax.select1.elt, float %rdx.shuf1.elt
+  %arr.load  = load <4 x float>, <4 x float>* %arr
+  %r = call nnan float @llvm.experimental.vector.reduce.fmin.f32.v4f32(<4 x float> %arr.load)
   ret float %r
 }
 
+declare i16 @llvm.experimental.vector.reduce.umax.i16.v16i16(<16 x i16>)
+
 define i16 @oversized_umax_256(<16 x i16>* nocapture readonly %arr)  {
 ; CHECK-LABEL: oversized_umax_256
 ; CHECK: umax [[V0:v[0-9]+]].8h, {{v[0-9]+}}.8h, {{v[0-9]+}}.8h
 ; CHECK: umaxv {{h[0-9]+}}, [[V0]]
-  %rdx.minmax.select = load <16 x i16>, <16 x i16>* %arr
-  %rdx.shuf = shufflevector <16 x i16> %rdx.minmax.select, <16 x i16> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp ugt <16 x i16> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i16> %rdx.minmax.select, <16 x i16> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i16> %rdx.minmax.select23, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp ugt <16 x i16> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i16> %rdx.minmax.select23, <16 x i16> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i16> %rdx.minmax.select26, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp ugt <16 x i16> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i16> %rdx.minmax.select26, <16 x i16> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i16> %rdx.minmax.select29, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp ugt <16 x i16> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i16> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i16> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i16 %rdx.minmax.select29.elt, i16 %rdx.shuf30.elt
+  %arr.load = load <16 x i16>, <16 x i16>* %arr
+  %r = call i16 @llvm.experimental.vector.reduce.umax.i16.v16i16(<16 x i16> %arr.load)
   ret i16 %r
 }
 
+declare i32 @llvm.experimental.vector.reduce.umax.i32.v16i32(<16 x i32>)
+
 define i32 @oversized_umax_512(<16 x i32>* nocapture readonly %arr)  {
 ; CHECK-LABEL: oversized_umax_512
 ; CHECK: umax v
 ; CHECK-NEXT: umax v
 ; CHECK-NEXT: umax [[V0:v[0-9]+]].4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
 ; CHECK-NEXT: umaxv {{s[0-9]+}}, [[V0]]
   %arr.load = load <16 x i32>, <16 x i32>* %arr
-  %rdx.shuf = shufflevector <16 x i32> %arr.load, <16 x i32> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp ugt <16 x i32> %arr.load, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i32> %arr.load, <16 x i32> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i32> %rdx.minmax.select23, <16 x i32> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp ugt <16 x i32> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i32> %rdx.minmax.select23, <16 x i32> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i32> %rdx.minmax.select26, <16 x i32> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp ugt <16 x i32> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i32> %rdx.minmax.select26, <16 x i32> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i32> %rdx.minmax.select29, <16 x i32> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp ugt <16 x i32> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i32> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i32> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i32 %rdx.minmax.select29.elt, i32 %rdx.shuf30.elt
+  %r = call i32 @llvm.experimental.vector.reduce.umax.i32.v16i32(<16 x i32> %arr.load)
   ret i32 %r
 }
 
+declare i16 @llvm.experimental.vector.reduce.umin.i16.v16i16(<16 x i16>)
+
 define i16 @oversized_umin_256(<16 x i16>* nocapture readonly %arr)  {
 ; CHECK-LABEL: oversized_umin_256
 ; CHECK: umin [[V0:v[0-9]+]].8h, {{v[0-9]+}}.8h, {{v[0-9]+}}.8h
 ; CHECK: uminv {{h[0-9]+}}, [[V0]]
-  %rdx.minmax.select = load <16 x i16>, <16 x i16>* %arr
-  %rdx.shuf = shufflevector <16 x i16> %rdx.minmax.select, <16 x i16> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp ult <16 x i16> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i16> %rdx.minmax.select, <16 x i16> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i16> %rdx.minmax.select23, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp ult <16 x i16> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i16> %rdx.minmax.select23, <16 x i16> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i16> %rdx.minmax.select26, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp ult <16 x i16> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i16> %rdx.minmax.select26, <16 x i16> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i16> %rdx.minmax.select29, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp ult <16 x i16> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i16> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i16> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i16 %rdx.minmax.select29.elt, i16 %rdx.shuf30.elt
+  %arr.load = load <16 x i16>, <16 x i16>* %arr
+  %r = call i16 @llvm.experimental.vector.reduce.umin.i16.v16i16(<16 x i16> %arr.load)
   ret i16 %r
 }
 
+declare i32 @llvm.experimental.vector.reduce.umin.i32.v16i32(<16 x i32>)
+
 define i32 @oversized_umin_512(<16 x i32>* nocapture readonly %arr)  {
 ; CHECK-LABEL: oversized_umin_512
 ; CHECK: umin v
 ; CHECK-NEXT: umin v
 ; CHECK-NEXT: umin [[V0:v[0-9]+]].4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
 ; CHECK-NEXT: uminv {{s[0-9]+}}, [[V0]]
   %arr.load = load <16 x i32>, <16 x i32>* %arr
-  %rdx.shuf = shufflevector <16 x i32> %arr.load, <16 x i32> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp ult <16 x i32> %arr.load, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i32> %arr.load, <16 x i32> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i32> %rdx.minmax.select23, <16 x i32> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp ult <16 x i32> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i32> %rdx.minmax.select23, <16 x i32> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i32> %rdx.minmax.select26, <16 x i32> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp ult <16 x i32> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i32> %rdx.minmax.select26, <16 x i32> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i32> %rdx.minmax.select29, <16 x i32> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp ult <16 x i32> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i32> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i32> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i32 %rdx.minmax.select29.elt, i32 %rdx.shuf30.elt
+  %r = call i32 @llvm.experimental.vector.reduce.umin.i32.v16i32(<16 x i32> %arr.load)
   ret i32 %r
 }
 
+declare i16 @llvm.experimental.vector.reduce.smax.i16.v16i16(<16 x i16>)
+
 define i16 @oversized_smax_256(<16 x i16>* nocapture readonly %arr)  {
 ; CHECK-LABEL: oversized_smax_256
 ; CHECK: smax [[V0:v[0-9]+]].8h, {{v[0-9]+}}.8h, {{v[0-9]+}}.8h
 ; CHECK: smaxv {{h[0-9]+}}, [[V0]]
   %arr.load = load <16 x i16>, <16 x i16>* %arr
-  %rdx.shuf = shufflevector <16 x i16> %arr.load, <16 x i16> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp sgt <16 x i16> %arr.load, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i16> %arr.load, <16 x i16> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i16> %rdx.minmax.select23, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp sgt <16 x i16> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i16> %rdx.minmax.select23, <16 x i16> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i16> %rdx.minmax.select26, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp sgt <16 x i16> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i16> %rdx.minmax.select26, <16 x i16> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i16> %rdx.minmax.select29, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp sgt <16 x i16> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i16> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i16> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i16 %rdx.minmax.select29.elt, i16 %rdx.shuf30.elt
+  %r = call i16 @llvm.experimental.vector.reduce.smax.i16.v16i16(<16 x i16> %arr.load)
   ret i16 %r
 }
 
+declare i32 @llvm.experimental.vector.reduce.smax.i32.v16i32(<16 x i32>)
+
 define i32 @oversized_smax_512(<16 x i32>* nocapture readonly %arr)  {
 ; CHECK-LABEL: oversized_smax_512
 ; CHECK: smax v
 ; CHECK-NEXT: smax v
 ; CHECK-NEXT: smax [[V0:v[0-9]+]].4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
 ; CHECK-NEXT: smaxv {{s[0-9]+}}, [[V0]]
   %arr.load = load <16 x i32>, <16 x i32>* %arr
-  %rdx.shuf = shufflevector <16 x i32> %arr.load, <16 x i32> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp sgt <16 x i32> %arr.load, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i32> %arr.load, <16 x i32> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i32> %rdx.minmax.select23, <16 x i32> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp sgt <16 x i32> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i32> %rdx.minmax.select23, <16 x i32> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i32> %rdx.minmax.select26, <16 x i32> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp sgt <16 x i32> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i32> %rdx.minmax.select26, <16 x i32> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i32> %rdx.minmax.select29, <16 x i32> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp sgt <16 x i32> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i32> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i32> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i32 %rdx.minmax.select29.elt, i32 %rdx.shuf30.elt
+  %r = call i32 @llvm.experimental.vector.reduce.smax.i32.v16i32(<16 x i32> %arr.load)
   ret i32 %r
 }
 
+declare i16 @llvm.experimental.vector.reduce.smin.i16.v16i16(<16 x i16>)
+
 define i16 @oversized_smin_256(<16 x i16>* nocapture readonly %arr)  {
 ; CHECK-LABEL: oversized_smin_256
 ; CHECK: smin [[V0:v[0-9]+]].8h, {{v[0-9]+}}.8h, {{v[0-9]+}}.8h
 ; CHECK: sminv {{h[0-9]+}}, [[V0]]
-  %rdx.minmax.select = load <16 x i16>, <16 x i16>* %arr
-  %rdx.shuf = shufflevector <16 x i16> %rdx.minmax.select, <16 x i16> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp slt <16 x i16> %rdx.minmax.select, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i16> %rdx.minmax.select, <16 x i16> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i16> %rdx.minmax.select23, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp slt <16 x i16> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i16> %rdx.minmax.select23, <16 x i16> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i16> %rdx.minmax.select26, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp slt <16 x i16> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i16> %rdx.minmax.select26, <16 x i16> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i16> %rdx.minmax.select29, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp slt <16 x i16> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i16> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i16> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i16 %rdx.minmax.select29.elt, i16 %rdx.shuf30.elt
+  %arr.load = load <16 x i16>, <16 x i16>* %arr
+  %r = call i16 @llvm.experimental.vector.reduce.smin.i16.v16i16(<16 x i16> %arr.load)
   ret i16 %r
 }
 
+declare i32 @llvm.experimental.vector.reduce.smin.i32.v16i32(<16 x i32>)
+
 define i32 @oversized_smin_512(<16 x i32>* nocapture readonly %arr)  {
 ; CHECK-LABEL: oversized_smin_512
 ; CHECK: smin v
 ; CHECK-NEXT: smin v
 ; CHECK-NEXT: smin [[V0:v[0-9]+]].4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
 ; CHECK-NEXT: sminv {{s[0-9]+}}, [[V0]]
   %arr.load = load <16 x i32>, <16 x i32>* %arr
-  %rdx.shuf = shufflevector <16 x i32> %arr.load, <16 x i32> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp22 = icmp slt <16 x i32> %arr.load, %rdx.shuf
-  %rdx.minmax.select23 = select <16 x i1> %rdx.minmax.cmp22, <16 x i32> %arr.load, <16 x i32> %rdx.shuf
-  %rdx.shuf24 = shufflevector <16 x i32> %rdx.minmax.select23, <16 x i32> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp25 = icmp slt <16 x i32> %rdx.minmax.select23, %rdx.shuf24
-  %rdx.minmax.select26 = select <16 x i1> %rdx.minmax.cmp25, <16 x i32> %rdx.minmax.select23, <16 x i32> %rdx.shuf24
-  %rdx.shuf27 = shufflevector <16 x i32> %rdx.minmax.select26, <16 x i32> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp28 = icmp slt <16 x i32> %rdx.minmax.select26, %rdx.shuf27
-  %rdx.minmax.select29 = select <16 x i1> %rdx.minmax.cmp28, <16 x i32> %rdx.minmax.select26, <16 x i32> %rdx.shuf27
-  %rdx.shuf30 = shufflevector <16 x i32> %rdx.minmax.select29, <16 x i32> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %rdx.minmax.cmp31 = icmp slt <16 x i32> %rdx.minmax.select29, %rdx.shuf30
-  %rdx.minmax.cmp31.elt = extractelement <16 x i1> %rdx.minmax.cmp31, i32 0
-  %rdx.minmax.select29.elt = extractelement <16 x i32> %rdx.minmax.select29, i32 0
-  %rdx.shuf30.elt = extractelement <16 x i32> %rdx.minmax.select29, i32 1
-  %r = select i1 %rdx.minmax.cmp31.elt, i32 %rdx.minmax.select29.elt, i32 %rdx.shuf30.elt
+  %r = call i32 @llvm.experimental.vector.reduce.smin.i32.v16i32(<16 x i32> %arr.load)
   ret i32 %r
 }

llvm/trunk/test/CodeGen/AArch64/arm64-vabs.ll

;CHECK: uabdl2.2d
   %load2 = load <4 x i32>, <4 x i32>* %B
   %tmp1 = shufflevector <4 x i32> %load1, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
   %tmp2 = shufflevector <4 x i32> %load2, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
   %tmp3 = call <2 x i32> @llvm.aarch64.neon.uabd.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
   %tmp4 = zext <2 x i32> %tmp3 to <2 x i64>
   ret <2 x i64> %tmp4
 }
 
-define i16 @uabdl8h_log2_shuffle(<16 x i8>* %a, <16 x i8>* %b) {
-; CHECK-LABEL: uabdl8h_log2_shuffle
+declare i16 @llvm.experimental.vector.reduce.add.i16.v16i16(<16 x i16>)
+
+define i16 @uabdl8h_rdx(<16 x i8>* %a, <16 x i8>* %b) {
+; CHECK-LABEL: uabdl8h_rdx
 ; CHECK: uabdl2.8h
 ; CHECK: uabdl.8h
   %aload = load <16 x i8>, <16 x i8>* %a, align 1
   %bload = load <16 x i8>, <16 x i8>* %b, align 1
   %aext = zext <16 x i8> %aload to <16 x i16>
   %bext = zext <16 x i8> %bload to <16 x i16>
   %abdiff = sub nsw <16 x i16> %aext, %bext
   %abcmp = icmp slt <16 x i16> %abdiff, zeroinitializer
   %ababs = sub nsw <16 x i16> zeroinitializer, %abdiff
   %absel = select <16 x i1> %abcmp, <16 x i16> %ababs, <16 x i16> %abdiff
-  %rdx.shuf = shufflevector <16 x i16> %absel, <16 x i16> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin1.rdx = add <16 x i16> %absel, %rdx.shuf
-  %rdx.shufx = shufflevector <16 x i16> %bin1.rdx, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx = add <16 x i16> %bin1.rdx, %rdx.shufx
-  %rdx.shuf136 = shufflevector <16 x i16> %bin.rdx, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx137 = add <16 x i16> %bin.rdx, %rdx.shuf136
-  %rdx.shuf138 = shufflevector <16 x i16> %bin.rdx137, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx139 = add <16 x i16> %bin.rdx137, %rdx.shuf138
-  %reduced_v = extractelement <16 x i16> %bin.rdx139, i16 0
+  %reduced_v = call i16 @llvm.experimental.vector.reduce.add.i16.v16i16(<16 x i16> %absel)
   ret i16 %reduced_v
 }
 
-define i32 @uabdl4s_log2_shuffle(<8 x i16>* %a, <8 x i16>* %b) {
-; CHECK-LABEL: uabdl4s_log2_shuffle
+declare i32 @llvm.experimental.vector.reduce.add.i32.v8i32(<8 x i32>)
+
+define i32 @uabdl4s_rdx(<8 x i16>* %a, <8 x i16>* %b) {
+; CHECK-LABEL: uabdl4s_rdx
 ; CHECK: uabdl2.4s
 ; CHECK: uabdl.4s
   %aload = load <8 x i16>, <8 x i16>* %a, align 1
   %bload = load <8 x i16>, <8 x i16>* %b, align 1
   %aext = zext <8 x i16> %aload to <8 x i32>
   %bext = zext <8 x i16> %bload to <8 x i32>
   %abdiff = sub nsw <8 x i32> %aext, %bext
   %abcmp = icmp slt <8 x i32> %abdiff, zeroinitializer
   %ababs = sub nsw <8 x i32> zeroinitializer, %abdiff
   %absel = select <8 x i1> %abcmp, <8 x i32> %ababs, <8 x i32> %abdiff
-  %rdx.shuf = shufflevector <8 x i32> %absel, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx = add <8 x i32> %absel, %rdx.shuf
-  %rdx.shuf136 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx137 = add <8 x i32> %bin.rdx, %rdx.shuf136
-  %rdx.shuf138 = shufflevector <8 x i32> %bin.rdx137, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-  %bin.rdx139 = add <8 x i32> %bin.rdx137, %rdx.shuf138
-  %reduced_v = extractelement <8 x i32> %bin.rdx139, i32 0
+  %reduced_v = call i32 @llvm.experimental.vector.reduce.add.i32.v8i32(<8 x i32> %absel)
   ret i32 %reduced_v
 }
 
-define i64 @uabdl2d_log2_shuffle(<4 x i32>* %a, <4 x i32>* %b, i32 %h) {
-; CHECK: uabdl2d_log2_shuffle
+declare i64 @llvm.experimental.vector.reduce.add.i64.v4i64(<4 x i64>)
+
+define i64 @uabdl2d_rdx(<4 x i32>* %a, <4 x i32>* %b, i32 %h) {
+; CHECK: uabdl2d_rdx
 ; CHECK: uabdl2.2d
 ; CHECK: uabdl.2d
   %aload = load <4 x i32>, <4 x i32>* %a, align 1
   %bload = load <4 x i32>, <4 x i32>* %b, align 1
   %aext = zext <4 x i32> %aload to <4 x i64>
   %bext = zext <4 x i32> %bload to <4 x i64>
   %abdiff = sub nsw <4 x i64> %aext, %bext
   %abcmp = icmp slt <4 x i64> %abdiff, zeroinitializer
   %ababs = sub nsw <4 x i64> zeroinitializer, %abdiff
   %absel = select <4 x i1> %abcmp, <4 x i64> %ababs, <4 x i64> %abdiff
-  %rdx.shuf136 = shufflevector <4 x i64> %absel, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
-  %bin.rdx137 = add <4 x i64> %absel, %rdx.shuf136
-  %rdx.shuf138 = shufflevector <4 x i64> %bin.rdx137, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
-  %bin.rdx139 = add <4 x i64> %bin.rdx137, %rdx.shuf138
-  %reduced_v = extractelement <4 x i64> %bin.rdx139, i16 0
+  %reduced_v = call i64 @llvm.experimental.vector.reduce.add.i64.v4i64(<4 x i64> %absel)
   ret i64 %reduced_v
 }
 
 define <2 x float> @fabd_2s(<2 x float>* %A, <2 x float>* %B) nounwind {
 ;CHECK-LABEL: fabd_2s:
 ;CHECK: fabd.2s
         %tmp1 = load <2 x float>, <2 x float>* %A
         %tmp2 = load <2 x float>, <2 x float>* %B

llvm/trunk/test/Transforms/LoopVectorize/AArch64/reduction-small-size.ll

 ; CHECK: vector.body:
 ; CHECK:   phi <16 x i8>
 ; CHECK:   load <16 x i8>
 ; CHECK:   load <16 x i8>
 ; CHECK:   add <16 x i8>
 ; CHECK:   add <16 x i8>
 ;
 ; CHECK: middle.block:
-; CHECK:   shufflevector <16 x i8>
-; CHECK:   add <16 x i8>
-; CHECK:   shufflevector <16 x i8>
-; CHECK:   add <16 x i8>
-; CHECK:   shufflevector <16 x i8>
-; CHECK:   add <16 x i8>
-; CHECK:   shufflevector <16 x i8>
-; CHECK:   add <16 x i8>
-; CHECK:   [[Rdx:%[a-zA-Z0-9.]+]] = extractelement <16 x i8>
+; CHECK:   [[Rdx:%[a-zA-Z0-9.]+]] = call i8 @llvm.experimental.vector.reduce.add.i8.v16i8(<16 x i8>
 ; CHECK:   zext i8 [[Rdx]] to i32
 ;
 define i8 @reduction_i8(i8* nocapture readonly %a, i8* nocapture readonly %b, i32 %n) {
 entry:
   %cmp.12 = icmp sgt i32 %n, 0
   br i1 %cmp.12, label %for.body.preheader, label %for.cond.cleanup
 
 for.body.preheader:

 ; CHECK: vector.body:
 ; CHECK:   phi <8 x i16>
 ; CHECK:   load <8 x i16>
 ; CHECK:   load <8 x i16>
 ; CHECK:   add <8 x i16>
 ; CHECK:   add <8 x i16>
 ;
 ; CHECK: middle.block:
-; CHECK:   shufflevector <8 x i16>
-; CHECK:   add <8 x i16>
-; CHECK:   shufflevector <8 x i16>
-; CHECK:   add <8 x i16>
-; CHECK:   shufflevector <8 x i16>
-; CHECK:   add <8 x i16>
-; CHECK:   [[Rdx:%[a-zA-Z0-9.]+]] = extractelement <8 x i16>
+; CHECK:   [[Rdx:%[a-zA-Z0-9.]+]] = call i16 @llvm.experimental.vector.reduce.add.i16.v8i16(<8 x i16>
 ; CHECK:   zext i16 [[Rdx]] to i32
 ;
 define i16 @reduction_i16_1(i16* nocapture readonly %a, i16* nocapture readonly %b, i32 %n) {
 entry:
   %cmp.16 = icmp sgt i32 %n, 0
   br i1 %cmp.16, label %for.body.preheader, label %for.cond.cleanup
 
 for.body.preheader:

 ; CHECK:   [[Ld1:%[a-zA-Z0-9.]+]] = load <8 x i8>
 ; CHECK:   zext <8 x i8> [[Ld1]] to <8 x i16>
 ; CHECK:   [[Ld2:%[a-zA-Z0-9.]+]] = load <8 x i8>
 ; CHECK:   zext <8 x i8> [[Ld2]] to <8 x i16>
 ; CHECK:   add <8 x i16>
 ; CHECK:   add <8 x i16>
 ;
 ; CHECK: middle.block:
-; CHECK:   shufflevector <8 x i16>
-; CHECK:   add <8 x i16>
-; CHECK:   shufflevector <8 x i16>
-; CHECK:   add <8 x i16>
-; CHECK:   shufflevector <8 x i16>
-; CHECK:   add <8 x i16>
-; CHECK:   [[Rdx:%[a-zA-Z0-9.]+]] = extractelement <8 x i16>
+; CHECK:   [[Rdx:%[a-zA-Z0-9.]+]] = call i16 @llvm.experimental.vector.reduce.add.i16.v8i16(<8 x i16>
 ; CHECK:   zext i16 [[Rdx]] to i32
 ;
 define i16 @reduction_i16_2(i8* nocapture readonly %a, i8* nocapture readonly %b, i32 %n) {
 entry:
   %cmp.14 = icmp sgt i32 %n, 0
   br i1 %cmp.14, label %for.body.preheader, label %for.cond.cleanup
 
 for.body.preheader:

llvm/trunk/test/Transforms/SLPVectorizer/AArch64/gather-root.ll

 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; RUN: opt < %s -slp-vectorizer -S | FileCheck %s --check-prefix=DEFAULT
 ; RUN: opt < %s -slp-schedule-budget=0 -slp-min-tree-size=0 -slp-threshold=-30 -slp-vectorizer -S | FileCheck %s --check-prefix=GATHER
 ; RUN: opt < %s -slp-schedule-budget=0 -slp-threshold=-30 -slp-vectorizer -S | FileCheck %s --check-prefix=MAX-COST
 
 target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128"
 target triple = "aarch64--linux-gnu"
 
 @a = common global [80 x i8] zeroinitializer, align 16
 
 ; DEFAULT-LABEL: @PR28330(
 ; DEFAULT: %tmp17 = phi i32 [ %bin.extra, %for.body ], [ 0, %entry ]
 ; DEFAULT: %[[S0:.+]] = select <8 x i1> %1, <8 x i32> <i32 -720, i32 -720, i32 -720, i32 -720, i32 -720, i32 -720, i32 -720, i32 -720>, <8 x i32> <i32 -80, i32 -80, i32 -80, i32 -80, i32 -80, i32 -80, i32 -80, i32 -80>
-; DEFAULT: %[[R0:.+]] = shufflevector <8 x i32> %[[S0]], <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-; DEFAULT: %[[R1:.+]] = add <8 x i32> %[[S0]], %[[R0]]
-; DEFAULT: %[[R2:.+]] = shufflevector <8 x i32> %[[R1]], <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; DEFAULT: %[[R3:.+]] = add <8 x i32> %[[R1]], %[[R2]]
-; DEFAULT: %[[R4:.+]] = shufflevector <8 x i32> %[[R3]], <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; DEFAULT: %[[R5:.+]] = add <8 x i32> %[[R3]], %[[R4]]
-; DEFAULT: %[[R6:.+]] = extractelement <8 x i32> %[[R5]], i32 0
-; DEFAULT: %bin.extra = add i32 %[[R6]], %tmp17
+; DEFAULT: %[[Rdx:.+]] = call i32 @llvm.experimental.vector.reduce.add.i32.v8i32(<8 x i32> %[[S0]])
+; DEFAULT: %bin.extra = add i32 %[[Rdx]], %tmp17
 ;
 ; GATHER-LABEL: @PR28330(
 ; GATHER: %tmp17 = phi i32 [ %bin.extra, %for.body ], [ 0, %entry ]
 ; GATHER: %tmp19 = select i1 %tmp1, i32 -720, i32 -80
 ; GATHER: %tmp21 = select i1 %tmp3, i32 -720, i32 -80
 ; GATHER: %tmp23 = select i1 %tmp5, i32 -720, i32 -80
 ; GATHER: %tmp25 = select i1 %tmp7, i32 -720, i32 -80
 ; GATHER: %tmp27 = select i1 %tmp9, i32 -720, i32 -80
 ; GATHER: %tmp29 = select i1 %tmp11, i32 -720, i32 -80
 ; GATHER: %tmp31 = select i1 %tmp13, i32 -720, i32 -80
 ; GATHER: %tmp33 = select i1 %tmp15, i32 -720, i32 -80
 ; GATHER: %[[I0:.+]] = insertelement <8 x i32> undef, i32 %tmp19, i32 0
 ; GATHER: %[[I1:.+]] = insertelement <8 x i32> %[[I0]], i32 %tmp21, i32 1
 ; GATHER: %[[I2:.+]] = insertelement <8 x i32> %[[I1]], i32 %tmp23, i32 2
 ; GATHER: %[[I3:.+]] = insertelement <8 x i32> %[[I2]], i32 %tmp25, i32 3
 ; GATHER: %[[I4:.+]] = insertelement <8 x i32> %[[I3]], i32 %tmp27, i32 4
 ; GATHER: %[[I5:.+]] = insertelement <8 x i32> %[[I4]], i32 %tmp29, i32 5
 ; GATHER: %[[I6:.+]] = insertelement <8 x i32> %[[I5]], i32 %tmp31, i32 6
 ; GATHER: %[[I7:.+]] = insertelement <8 x i32> %[[I6]], i32 %tmp33, i32 7
-; GATHER: %[[R0:.+]] = shufflevector <8 x i32> %[[I7]], <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-; GATHER: %[[R1:.+]] = add <8 x i32> %[[I7]], %[[R0]]
-; GATHER: %[[R2:.+]] = shufflevector <8 x i32> %[[R1]], <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; GATHER: %[[R3:.+]] = add <8 x i32> %[[R1]], %[[R2]]
-; GATHER: %[[R4:.+]] = shufflevector <8 x i32> %[[R3]], <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; GATHER: %[[R5:.+]] = add <8 x i32> %[[R3]], %[[R4]]
-; GATHER: %[[R6:.+]] = extractelement <8 x i32> %[[R5]], i32 0
-; GATHER: %bin.extra = add i32 %[[R6]], %tmp17
+; GATHER: %[[Rdx:.+]] = call i32 @llvm.experimental.vector.reduce.add.i32.v8i32(<8 x i32> %[[I7]])
+; GATHER: %bin.extra = add i32 %[[Rdx]], %tmp17
 ;
 ; MAX-COST-LABEL: @PR28330(
 ; MAX-COST-NOT: shufflevector
 
 define void @PR28330(i32 %n) {
 entry:
   %tmp0 = load i8, i8* getelementptr inbounds ([80 x i8], [80 x i8]* @a, i64 0, i64 1), align 1
   %tmp1 = icmp eq i8 %tmp0, 0

 ; DEFAULT-NEXT:    [[TMP2:%.*]] = select <8 x i1> [[TMP1]], <8 x i32> <i32 -720, i32 -720, i32 -720, i32 -720, i32 -720, i32 -720, i32 -720, i32 -720>, <8 x i32> <i32 -80, i32 -80, i32 -80, i32 -80, i32 -80, i32 -80, i32 -80, i32 -80>
 ; DEFAULT-NEXT:    [[TMP20:%.*]] = add i32 -5, undef
 ; DEFAULT-NEXT:    [[TMP22:%.*]] = add i32 [[TMP20]], undef
 ; DEFAULT-NEXT:    [[TMP24:%.*]] = add i32 [[TMP22]], undef
 ; DEFAULT-NEXT:    [[TMP26:%.*]] = add i32 [[TMP24]], undef
 ; DEFAULT-NEXT:    [[TMP28:%.*]] = add i32 [[TMP26]], undef
 ; DEFAULT-NEXT:    [[TMP30:%.*]] = add i32 [[TMP28]], undef
 ; DEFAULT-NEXT:    [[TMP32:%.*]] = add i32 [[TMP30]], undef
-; DEFAULT-NEXT:    [[RDX_SHUF:%.*]] = shufflevector <8 x i32> [[TMP2]], <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-; DEFAULT-NEXT:    [[BIN_RDX:%.*]] = add <8 x i32> [[TMP2]], [[RDX_SHUF]]
-; DEFAULT-NEXT:    [[RDX_SHUF1:%.*]] = shufflevector <8 x i32> [[BIN_RDX]], <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; DEFAULT-NEXT:    [[BIN_RDX2:%.*]] = add <8 x i32> [[BIN_RDX]], [[RDX_SHUF1]]
-; DEFAULT-NEXT:    [[RDX_SHUF3:%.*]] = shufflevector <8 x i32> [[BIN_RDX2]], <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; DEFAULT-NEXT:    [[BIN_RDX4:%.*]] = add <8 x i32> [[BIN_RDX2]], [[RDX_SHUF3]]
-; DEFAULT-NEXT:    [[TMP3:%.*]] = extractelement <8 x i32> [[BIN_RDX4]], i32 0
-; DEFAULT-NEXT:    [[BIN_EXTRA]] = add i32 [[TMP3]], -5
+; DEFAULT-NEXT:    [[Rdx:%.*]] = call i32 @llvm.experimental.vector.reduce.add.i32.v8i32(<8 x i32> [[TMP2]])
+; DEFAULT-NEXT:    [[BIN_EXTRA]] = add i32 [[Rdx]], -5
 ; DEFAULT-NEXT:    [[TMP34:%.*]] = add i32 [[TMP32]], undef
 ; DEFAULT-NEXT:    br label [[FOR_BODY]]
 ;
 ; GATHER-LABEL: @PR32038(
 ; GATHER-NEXT:  entry:
 ; GATHER-NEXT:    [[TMP0:%.*]] = load i8, i8* getelementptr inbounds ([80 x i8], [80 x i8]* @a, i64 0, i64 1), align 1
 ; GATHER-NEXT:    [[TMP1:%.*]] = icmp eq i8 [[TMP0]], 0
 ; GATHER-NEXT:    [[TMP2:%.*]] = load i8, i8* getelementptr inbounds ([80 x i8], [80 x i8]* @a, i64 0, i64 2), align 2

 ; GATHER-NEXT:    [[TMP0:%.*]] = insertelement <8 x i32> undef, i32 [[TMP19]], i32 0
 ; GATHER-NEXT:    [[TMP1:%.*]] = insertelement <8 x i32> [[TMP0]], i32 [[TMP21]], i32 1
 ; GATHER-NEXT:    [[TMP2:%.*]] = insertelement <8 x i32> [[TMP1]], i32 [[TMP23]], i32 2
 ; GATHER-NEXT:    [[TMP3:%.*]] = insertelement <8 x i32> [[TMP2]], i32 [[TMP25]], i32 3
 ; GATHER-NEXT:    [[TMP4:%.*]] = insertelement <8 x i32> [[TMP3]], i32 [[TMP27]], i32 4
 ; GATHER-NEXT:    [[TMP5:%.*]] = insertelement <8 x i32> [[TMP4]], i32 [[TMP29]], i32 5
 ; GATHER-NEXT:    [[TMP6:%.*]] = insertelement <8 x i32> [[TMP5]], i32 [[TMP31]], i32 6
 ; GATHER-NEXT:    [[TMP7:%.*]] = insertelement <8 x i32> [[TMP6]], i32 [[TMP33]], i32 7
-; GATHER-NEXT:    [[RDX_SHUF:%.*]] = shufflevector <8 x i32> [[TMP7]], <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
-; GATHER-NEXT:    [[BIN_RDX:%.*]] = add <8 x i32> [[TMP7]], [[RDX_SHUF]]
-; GATHER-NEXT:    [[RDX_SHUF1:%.*]] = shufflevector <8 x i32> [[BIN_RDX]], <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; GATHER-NEXT:    [[BIN_RDX2:%.*]] = add <8 x i32> [[BIN_RDX]], [[RDX_SHUF1]]
-; GATHER-NEXT:    [[RDX_SHUF3:%.*]] = shufflevector <8 x i32> [[BIN_RDX2]], <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; GATHER-NEXT:    [[BIN_RDX4:%.*]] = add <8 x i32> [[BIN_RDX2]], [[RDX_SHUF3]]
-; GATHER-NEXT:    [[TMP8:%.*]] = extractelement <8 x i32> [[BIN_RDX4]], i32 0
-; GATHER-NEXT:    [[BIN_EXTRA]] = add i32 [[TMP8]], -5
+; GATHER-NEXT:    [[Rdx:%.*]] = call i32 @llvm.experimental.vector.reduce.add.i32.v8i32(<8 x i32> [[TMP7]])
+; GATHER-NEXT:    [[BIN_EXTRA]] = add i32 [[Rdx]], -5
 ; GATHER-NEXT:    [[TMP34:%.*]] = add i32 [[TMP32]], [[TMP33]]
 ; GATHER-NEXT:    br label [[FOR_BODY]]
 ;
 ; MAX-COST-LABEL: @PR32038(
 entry:
   %tmp0 = load i8, i8* getelementptr inbounds ([80 x i8], [80 x i8]* @a, i64 0, i64 1), align 1
   %tmp1 = icmp eq i8 %tmp0, 0
   %tmp2 = load i8, i8* getelementptr inbounds ([80 x i8], [80 x i8]* @a, i64 0, i64 2), align 2