[LoopVectorize] Support widening memory ops with irregular types

[LemonBoy]

Summary

Successive elements of types whose Alloc size doesn't match the type size have some padding bits in between, preventing the vectorizer from kicking in.
Let's handle this case by promoting the load/store to a type that's as big as the Alloc size and truncating/extending the values as needed.

[LemonBoy]

Rebase to main.

[lebedev.ri]

llvm/include/llvm/IR/DataLayout.h

519

Split this into a preparatory patch?

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

2812

[lebedev.ri]

llvm/include/llvm/IR/DataLayout.h

519

... and replace hasIrregularType() in LoopVectorize.cpp with this check?

[LemonBoy]

Ping! (As suggested by fhahn)

[LemonBoy]

Weekly ping.

[LemonBoy]

Ping?

[lebedev.ri]

This review seems to be stuck/dead, consider abandoning if no longer relevant.

llvm/include/llvm/IR/DataLayout.h

public:
   /// the runtime size will be a positive integer multiple of the base size.
   ///
   /// This is the amount that alloca reserves for this type. For example,
   /// returns 96 or 128 for x86_fp80, depending on alignment.
   TypeSize getTypeAllocSizeInBits(Type *Ty) const {
     return 8 * getTypeAllocSize(Ty);
   }
 
+  /// Returns true if no extra padding bits are needed between successive
+  /// elements of the specified type.
+  ///
+  /// For example, returns false for i19 that for some targets has a 32-bit
+  /// store size.
+  bool typeSizeEqualsAllocSize(Type *Ty) const {
+    return getTypeSizeInBits(Ty) == getTypeAllocSizeInBits(Ty);
+  }
+
   /// Returns the minimum ABI-required alignment for the specified type.
   /// FIXME: Deprecate this function once migration to Align is over.
   unsigned getABITypeAlignment(Type *Ty) const;
 
   /// Returns the minimum ABI-required alignment for the specified type.
   Align getABITypeAlign(Type *Ty) const;
 
   /// Helper function to return `Alignment` if it's set or the result of

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

 static Type *getMemInstValueType(Value *I) {
   assert((isa<LoadInst>(I) || isa<StoreInst>(I)) &&
          "Expected Load or Store instruction");
   if (auto *LI = dyn_cast<LoadInst>(I))
     return LI->getType();
   return cast<StoreInst>(I)->getValueOperand()->getType();
 }
 
-/// A helper function that returns true if the given type is irregular. The
-/// type is irregular if its allocated size doesn't equal the store size of an
-/// element of the corresponding vector type.
-static bool hasIrregularType(Type *Ty, const DataLayout &DL) {
-  // Determine if an array of N elements of type Ty is "bitcast compatible"
-  // with a <N x Ty> vector.
-  // This is only true if there is no padding between the array elements.
-  return DL.getTypeAllocSizeInBits(Ty) != DL.getTypeSizeInBits(Ty);
-}
-
 /// A helper function that returns the reciprocal of the block probability of
 /// predicated blocks. If we return X, we are assuming the predicated block
 /// will execute once for every X iterations of the loop header.
 ///
 /// TODO: We should use actual block probability here, if available. Currently,
 ///       we always assume predicated blocks have a 50% chance of executing.
 static unsigned getReciprocalPredBlockProb() { return 2; }
 
void InnerLoopVectorizer::vectorizeMemoryInstruction(
 
   LoopVectorizationCostModel::InstWidening Decision =
       Cost->getWideningDecision(Instr, VF);
   assert((Decision == LoopVectorizationCostModel::CM_Widen ||
           Decision == LoopVectorizationCostModel::CM_Widen_Reverse ||
           Decision == LoopVectorizationCostModel::CM_GatherScatter) &&
          "CM decision is not to widen the memory instruction");
 
+  const DataLayout &DL = Instr->getModule()->getDataLayout();
   Type *ScalarDataTy = getMemInstValueType(Instr);
+  Type *OrigScalarDataTy = nullptr;
+
+  if (!DL.typeSizeEqualsAllocSize(ScalarDataTy)) {
+    assert(ScalarDataTy->isIntegerTy() &&
+           "Irregular type widening requires integer type");
+
+    OrigScalarDataTy = ScalarDataTy;
+    ScalarDataTy = IntegerType::get(ScalarDataTy->getContext(),
+                                    DL.getTypeAllocSizeInBits(ScalarDataTy));
+  }
 
   auto *DataTy = VectorType::get(ScalarDataTy, VF);
   const Align Alignment = getLoadStoreAlignment(Instr);
 
   // Determine if the pointer operand of the access is either consecutive or
   // reverse consecutive.
   bool Reverse = (Decision == LoopVectorizationCostModel::CM_Widen_Reverse);
   bool ConsecutiveStride =
       Reverse || (Decision == LoopVectorizationCostModel::CM_Widen);
   bool CreateGatherScatter =
       (Decision == LoopVectorizationCostModel::CM_GatherScatter);
 
   // Either Ptr feeds a vector load/store, or a vector GEP should feed a vector
   // gather/scatter. Otherwise Decision should have been to Scalarize.
   assert((ConsecutiveStride || CreateGatherScatter) &&
          "The instruction should be scalarized");
   (void)ConsecutiveStride;
 
   VectorParts BlockInMaskParts(UF);
   bool isMaskRequired = BlockInMask;
   if (isMaskRequired)
     for (unsigned Part = 0; Part < UF; ++Part)
       BlockInMaskParts[Part] = State.get(BlockInMask, Part);
 
-  const auto CreateVecPtr = [&](unsigned Part, Value *Ptr) -> Value * {
+  const auto CreateVecPtr = [&](unsigned Part, Value *Ptr,
+                                Type *CastTy) -> Value * {
     // Calculate the pointer for the specific unroll-part.
     GetElementPtrInst *PartPtr = nullptr;
 
     bool InBounds = false;
     if (auto *gep = dyn_cast<GetElementPtrInst>(Ptr->stripPointerCasts()))
       InBounds = gep->isInBounds();
+
+    if (CastTy) {
+      unsigned AddressSpace = Ptr->getType()->getPointerAddressSpace();
+      Ptr = Builder.CreateBitCast(Ptr, CastTy->getPointerTo(AddressSpace));
+    }
+
     if (Reverse) {
       // If the address is consecutive but reversed, then the
       // wide store needs to start at the last vector element.
       // RunTimeVF =  VScale * VF.getKnownMinValue()
       // For fixed-width VScale is 1, then RunTimeVF = VF.getKnownMinValue()
       Value *RunTimeVF = getRuntimeVF(Builder, Builder.getInt32Ty(), VF);
       // NumElt = -Part * RunTimeVF
       Value *NumElt = Builder.CreateMul(Builder.getInt32(-Part), RunTimeVF);
void InnerLoopVectorizer::vectorizeMemoryInstruction(
 
   // Handle Stores:
   if (SI) {
     setDebugLocFromInst(Builder, SI);
 
     for (unsigned Part = 0; Part < UF; ++Part) {
       Instruction *NewSI = nullptr;
       Value *StoredVal = State.get(StoredValue, Part);
+
+      // Widen the stored value if the access type was promoted.
+      if (OrigScalarDataTy)
+        StoredVal = Builder.CreateZExt(StoredVal, DataTy);
+
       if (CreateGatherScatter) {
         Value *MaskPart = isMaskRequired ? BlockInMaskParts[Part] : nullptr;
         Value *VectorGep = State.get(Addr, Part);
+
+        if (OrigScalarDataTy) {
+          auto *PtrsTy = cast<VectorType>(VectorGep->getType());
+          auto *PointeeTy = PtrsTy->getElementType();
+
+          unsigned AddressSpace = PointeeTy->getPointerAddressSpace();
+          auto *NewPtrsTy =
+              VectorType::get(ScalarDataTy->getPointerTo(AddressSpace), VF);
+          VectorGep = Builder.CreateBitOrPointerCast(VectorGep, NewPtrsTy);
+        }
+
         NewSI = Builder.CreateMaskedScatter(StoredVal, VectorGep, Alignment,
                                             MaskPart);
       } else {
         if (Reverse) {
           // If we store to reverse consecutive memory locations, then we need
           // to reverse the order of elements in the stored value.
           StoredVal = reverseVector(StoredVal);
           // We don't want to update the value in the map as it might be used in
           // another expression. So don't call resetVectorValue(StoredVal).
         }
-        auto *VecPtr = CreateVecPtr(Part, State.get(Addr, VPIteration(0, 0)));
+        auto *VecPtr = CreateVecPtr(Part, State.get(Addr, VPIteration(0, 0)),
+                                    ScalarDataTy);
         if (isMaskRequired)
           NewSI = Builder.CreateMaskedStore(StoredVal, VecPtr, Alignment,
                                             BlockInMaskParts[Part]);
         else
           NewSI = Builder.CreateAlignedStore(StoredVal, VecPtr, Alignment);
       }
       addMetadata(NewSI, SI);
     }
     return;
   }
 
   // Handle loads.
   assert(LI && "Must have a load instruction");
   setDebugLocFromInst(Builder, LI);
   for (unsigned Part = 0; Part < UF; ++Part) {
     Value *NewLI;
     if (CreateGatherScatter) {
       Value *MaskPart = isMaskRequired ? BlockInMaskParts[Part] : nullptr;
       Value *VectorGep = State.get(Addr, Part);
+
+      if (OrigScalarDataTy) {
+        auto *PtrsTy = cast<VectorType>(VectorGep->getType());
+        auto *PointeeTy = PtrsTy->getElementType();
+
+        unsigned AddressSpace = PointeeTy->getPointerAddressSpace();
+        auto *NewPtrsTy =
+            VectorType::get(ScalarDataTy->getPointerTo(AddressSpace), VF);
+        VectorGep = Builder.CreateBitOrPointerCast(VectorGep, NewPtrsTy);
+      }
+
       NewLI = Builder.CreateMaskedGather(VectorGep, Alignment, MaskPart,
                                          nullptr, "wide.masked.gather");
       addMetadata(NewLI, LI);
     } else {
-      auto *VecPtr = CreateVecPtr(Part, State.get(Addr, VPIteration(0, 0)));
+      auto *VecPtr =
+          CreateVecPtr(Part, State.get(Addr, VPIteration(0, 0)), ScalarDataTy);
       if (isMaskRequired)
         NewLI = Builder.CreateMaskedLoad(
             VecPtr, Alignment, BlockInMaskParts[Part], PoisonValue::get(DataTy),
             "wide.masked.load");
       else
         NewLI =
             Builder.CreateAlignedLoad(DataTy, VecPtr, Alignment, "wide.load");
 
       // Add metadata to the load, but setVectorValue to the reverse shuffle.
       addMetadata(NewLI, LI);
       if (Reverse)
         NewLI = reverseVector(NewLI);
     }
 
+    // Narrow the loaded values if the access type was promoted.
+    if (OrigScalarDataTy) {
+      auto *OrigVecTy = VectorType::get(OrigScalarDataTy, VF);
+      NewLI = Builder.CreateTrunc(NewLI, OrigVecTy);
+    }
+
     State.set(Def, NewLI, Part);
   }
 }
 
 void InnerLoopVectorizer::scalarizeInstruction(Instruction *Instr, VPValue *Def,
                                                VPUser &User,
                                                const VPIteration &Instance,
                                                bool IfPredicateInstr,
assert(getWideningDecision(I, VF) == CM_Unknown &&
          "Decision should not be set yet.");
   auto *Group = getInterleavedAccessGroup(I);
   assert(Group && "Must have a group.");
 
   // If the instruction's allocated size doesn't equal it's type size, it
   // requires padding and will be scalarized.
   auto &DL = I->getModule()->getDataLayout();
   auto *ScalarTy = getMemInstValueType(I);
-  if (hasIrregularType(ScalarTy, DL))
+  if (!DL.typeSizeEqualsAllocSize(ScalarTy))
     return false;
 
   // Check if masking is required.
   // A Group may need masking for one of two reasons: it resides in a block that
   // needs predication, or it was decided to use masking to deal with gaps.
   bool PredicatedAccessRequiresMasking =
       Legal->blockNeedsPredication(I->getParent()) && Legal->isMaskRequired(I);
   bool AccessWithGapsRequiresMasking =
bool LoopVectorizationCostModel::memoryInstructionCanBeWidened(
   if (!Legal->isConsecutivePtr(Ptr))
     return false;
 
   // If the instruction is a store located in a predicated block, it will be
   // scalarized.
   if (isScalarWithPredication(I))
     return false;
 
-  // If the instruction's allocated size doesn't equal it's type size, it
-  // requires padding and will be scalarized.
   auto &DL = I->getModule()->getDataLayout();
   auto *ScalarTy = LI ? LI->getType() : SI->getValueOperand()->getType();
-  if (hasIrregularType(ScalarTy, DL))
-    return false;
+  // Handle types with padding between elements by promoting the memory op to an
+  // integer type as big as the alloc size.
+  if (!DL.typeSizeEqualsAllocSize(ScalarTy))
+    return ScalarTy->isIntegerTy();
 
   return true;
 }
 
 void LoopVectorizationCostModel::collectLoopUniforms(ElementCount VF) {
   // We should not collect Uniforms more than once per VF. Right now,
   // this function is called from collectUniformsAndScalars(), which
   // already does this check. Collecting Uniforms for VF=1 does not make any
LoopVectorizationCostModel::getMemInstScalarizationCost(Instruction *I,
   }
 
   return Cost;
 }
 
 InstructionCost
 LoopVectorizationCostModel::getConsecutiveMemOpCost(Instruction *I,
                                                     ElementCount VF) {
+  auto &DL = I->getModule()->getDataLayout();
   Type *ValTy = getMemInstValueType(I);
   auto *VectorTy = cast<VectorType>(ToVectorTy(ValTy, VF));
   Value *Ptr = getLoadStorePointerOperand(I);
   unsigned AS = getLoadStoreAddressSpace(I);
   int ConsecutiveStride = Legal->isConsecutivePtr(Ptr);
   enum TTI::TargetCostKind CostKind = TTI::TCK_RecipThroughput;
+  bool NeedsPromotion = !DL.typeSizeEqualsAllocSize(ValTy);
+  Type *PromotedTy = NeedsPromotion
+                         ? IntegerType::get(ValTy->getContext(),
+                                            DL.getTypeAllocSizeInBits(ValTy))
+                         : ValTy;
+  Type *PromotedVectorTy =
+      NeedsPromotion ? ToVectorTy(PromotedTy, VF) : VectorTy;
 
   assert((ConsecutiveStride == 1 || ConsecutiveStride == -1) &&
          "Stride should be 1 or -1 for consecutive memory access");
+
   const Align Alignment = getLoadStoreAlignment(I);
   InstructionCost Cost = 0;
   if (Legal->isMaskRequired(I))
-    Cost += TTI.getMaskedMemoryOpCost(I->getOpcode(), VectorTy, Alignment, AS,
-                                      CostKind);
+    Cost += TTI.getMaskedMemoryOpCost(I->getOpcode(), PromotedVectorTy,
+                                      Alignment, AS, CostKind);
   else
-    Cost += TTI.getMemoryOpCost(I->getOpcode(), VectorTy, Alignment, AS,
+    Cost += TTI.getMemoryOpCost(I->getOpcode(), PromotedVectorTy, Alignment, AS,
                                 CostKind, I);
 
+  if (NeedsPromotion) {
+    const TTI::CastContextHint CCH = TTI::CastContextHint::None;
+
+    if (isa<LoadInst>(I))
+      Cost += TTI.getCastInstrCost(Instruction::Trunc, VectorTy,
+                                   PromotedVectorTy, CCH, CostKind);
+    else
+      Cost += TTI.getCastInstrCost(Instruction::ZExt, PromotedVectorTy,
+                                   VectorTy, CCH, CostKind);
+  }
+
   bool Reverse = ConsecutiveStride < 0;
   if (Reverse)
     Cost +=
         TTI.getShuffleCost(TargetTransformInfo::SK_Reverse, VectorTy, None, 0);
   return Cost;
 }
 
 InstructionCost

llvm/test/Transforms/LoopVectorize/X86/load-deref-pred.ll

 ; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [4096 x i32], align 4
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast [4096 x i32]* [[ALLOCA]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP80:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP40:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = bitcast i32* [[TMP68]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, <4 x i32>* [[TMP69]], align 4
-; CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 4
-; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[TMP70]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i32>, <4 x i32>* [[TMP71]], align 4
-; CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 8
-; CHECK-NEXT:    [[TMP73:%.*]] = bitcast i32* [[TMP72]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i32>, <4 x i32>* [[TMP73]], align 4
-; CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 12
-; CHECK-NEXT:    [[TMP75:%.*]] = bitcast i32* [[TMP74]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i32>, <4 x i32>* [[TMP75]], align 4
-; CHECK-NEXT:    [[TMP76:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[WIDE_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[WIDE_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[WIDE_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[WIDE_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = bitcast i32* [[TMP28]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD7:%.*]] = load <4 x i32>, <4 x i32>* [[TMP29]], align 4
+; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 4
+; CHECK-NEXT:    [[TMP31:%.*]] = bitcast i32* [[TMP30]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD8:%.*]] = load <4 x i32>, <4 x i32>* [[TMP31]], align 4
+; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 8
+; CHECK-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD9:%.*]] = load <4 x i32>, <4 x i32>* [[TMP33]], align 4
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 12
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i32* [[TMP34]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD10:%.*]] = load <4 x i32>, <4 x i32>* [[TMP35]], align 4
+; CHECK-NEXT:    [[TMP36:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[WIDE_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[WIDE_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[WIDE_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI13:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[WIDE_LOAD10]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP40]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI12]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI13]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP84:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP84]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP4:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP44:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP44]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP4:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP81]], [[TMP80]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP82]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP85:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP41]], [[TMP40]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP42]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX15:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX14]]
+; CHECK-NEXT:    [[TMP45:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX15]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP5:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [4096 x i32]
   %base = bitcast [4096 x i32]* %alloca to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [4096 x i32], align 4
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast [4096 x i32]* [[ALLOCA]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP100:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP101:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP102:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP103:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP60:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP61:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP62:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP63:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP65:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP66:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP67:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP68:%.*]] = insertelement <4 x i32> poison, i32 [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = insertelement <4 x i32> [[TMP68]], i32 [[TMP65]], i32 1
-; CHECK-NEXT:    [[TMP70:%.*]] = insertelement <4 x i32> [[TMP69]], i32 [[TMP66]], i32 2
-; CHECK-NEXT:    [[TMP71:%.*]] = insertelement <4 x i32> [[TMP70]], i32 [[TMP67]], i32 3
-; CHECK-NEXT:    [[TMP72:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP73:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP74:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP75:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP76:%.*]] = insertelement <4 x i32> poison, i32 [[TMP72]], i32 0
-; CHECK-NEXT:    [[TMP77:%.*]] = insertelement <4 x i32> [[TMP76]], i32 [[TMP73]], i32 1
-; CHECK-NEXT:    [[TMP78:%.*]] = insertelement <4 x i32> [[TMP77]], i32 [[TMP74]], i32 2
-; CHECK-NEXT:    [[TMP79:%.*]] = insertelement <4 x i32> [[TMP78]], i32 [[TMP75]], i32 3
-; CHECK-NEXT:    [[TMP80:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP81:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP82:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP83:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP84:%.*]] = insertelement <4 x i32> poison, i32 [[TMP80]], i32 0
-; CHECK-NEXT:    [[TMP85:%.*]] = insertelement <4 x i32> [[TMP84]], i32 [[TMP81]], i32 1
-; CHECK-NEXT:    [[TMP86:%.*]] = insertelement <4 x i32> [[TMP85]], i32 [[TMP82]], i32 2
-; CHECK-NEXT:    [[TMP87:%.*]] = insertelement <4 x i32> [[TMP86]], i32 [[TMP83]], i32 3
-; CHECK-NEXT:    [[TMP88:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP89:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP90:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP91:%.*]] = load i32, i32* [[BASE]], align 4
-; CHECK-NEXT:    [[TMP92:%.*]] = insertelement <4 x i32> poison, i32 [[TMP88]], i32 0
-; CHECK-NEXT:    [[TMP93:%.*]] = insertelement <4 x i32> [[TMP92]], i32 [[TMP89]], i32 1
-; CHECK-NEXT:    [[TMP94:%.*]] = insertelement <4 x i32> [[TMP93]], i32 [[TMP90]], i32 2
-; CHECK-NEXT:    [[TMP95:%.*]] = insertelement <4 x i32> [[TMP94]], i32 [[TMP91]], i32 3
-; CHECK-NEXT:    [[TMP96:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP97:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP98:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP99:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[TMP71]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI4:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[TMP79]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI5:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[TMP87]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI6:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[TMP95]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP100]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP101]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI4]]
-; CHECK-NEXT:    [[TMP102]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI5]]
-; CHECK-NEXT:    [[TMP103]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI6]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP25:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP26:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP27:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP28:%.*]] = insertelement <4 x i32> poison, i32 [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = insertelement <4 x i32> [[TMP28]], i32 [[TMP25]], i32 1
+; CHECK-NEXT:    [[TMP30:%.*]] = insertelement <4 x i32> [[TMP29]], i32 [[TMP26]], i32 2
+; CHECK-NEXT:    [[TMP31:%.*]] = insertelement <4 x i32> [[TMP30]], i32 [[TMP27]], i32 3
+; CHECK-NEXT:    [[TMP32:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP33:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP34:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP35:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i32> poison, i32 [[TMP32]], i32 0
+; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i32> [[TMP36]], i32 [[TMP33]], i32 1
+; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i32> [[TMP37]], i32 [[TMP34]], i32 2
+; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i32> [[TMP38]], i32 [[TMP35]], i32 3
+; CHECK-NEXT:    [[TMP40:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP41:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP42:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP43:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i32> poison, i32 [[TMP40]], i32 0
+; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i32> [[TMP44]], i32 [[TMP41]], i32 1
+; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i32> [[TMP45]], i32 [[TMP42]], i32 2
+; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i32> [[TMP46]], i32 [[TMP43]], i32 3
+; CHECK-NEXT:    [[TMP48:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP49:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP50:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP51:%.*]] = load i32, i32* [[BASE]], align 4
+; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i32> poison, i32 [[TMP48]], i32 0
+; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i32> [[TMP52]], i32 [[TMP49]], i32 1
+; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i32> [[TMP53]], i32 [[TMP50]], i32 2
+; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i32> [[TMP54]], i32 [[TMP51]], i32 3
+; CHECK-NEXT:    [[TMP56:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP57:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP58:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP59:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[TMP31]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[TMP39]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[TMP47]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[TMP55]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP60]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP61]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
+; CHECK-NEXT:    [[TMP62]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
+; CHECK-NEXT:    [[TMP63]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP104:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP104]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP6:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP64:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP64]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP6:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP101]], [[TMP100]]
-; CHECK-NEXT:    [[BIN_RDX7:%.*]] = add <4 x i32> [[TMP102]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX8:%.*]] = add <4 x i32> [[TMP103]], [[BIN_RDX7]]
-; CHECK-NEXT:    [[TMP105:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX8]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP61]], [[TMP60]]
+; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP62]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP63]], [[BIN_RDX10]]
+; CHECK-NEXT:    [[TMP65:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP105]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP65]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[BASE]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP7:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP105]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP65]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [4096 x i32]
   %base = bitcast [4096 x i32]* %alloca to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [4096 x i32], align 4
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast [4096 x i32]* [[ALLOCA]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
-; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[PRED_LOAD_CONTINUE33:%.*]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP180:%.*]], [[PRED_LOAD_CONTINUE33]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP181:%.*]], [[PRED_LOAD_CONTINUE33]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP182:%.*]], [[PRED_LOAD_CONTINUE33]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP183:%.*]], [[PRED_LOAD_CONTINUE33]] ]
+; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[PRED_LOAD_CONTINUE36:%.*]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP152:%.*]], [[PRED_LOAD_CONTINUE36]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP153:%.*]], [[PRED_LOAD_CONTINUE36]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP154:%.*]], [[PRED_LOAD_CONTINUE36]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP155:%.*]], [[PRED_LOAD_CONTINUE36]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = extractelement <4 x i1> [[TMP39]], i32 0
-; CHECK-NEXT:    br i1 [[TMP64]], label [[PRED_LOAD_IF:%.*]], label [[PRED_LOAD_CONTINUE:%.*]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = extractelement <4 x i1> [[TMP11]], i32 0
+; CHECK-NEXT:    br i1 [[TMP24]], label [[PRED_LOAD_IF:%.*]], label [[PRED_LOAD_CONTINUE:%.*]]
 ; CHECK:       pred.load.if:
-; CHECK-NEXT:    [[TMP65:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i16, i16* [[TMP65]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP67:%.*]] = bitcast i16* [[TMP66]] to i32*
-; CHECK-NEXT:    [[TMP68:%.*]] = load i32, i32* [[TMP67]], align 4
-; CHECK-NEXT:    [[TMP69:%.*]] = insertelement <4 x i32> poison, i32 [[TMP68]], i32 0
+; CHECK-NEXT:    [[TMP25:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i16, i16* [[TMP25]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP27:%.*]] = bitcast i16* [[TMP26]] to i32*
+; CHECK-NEXT:    [[TMP28:%.*]] = load i32, i32* [[TMP27]], align 4
+; CHECK-NEXT:    [[TMP29:%.*]] = insertelement <4 x i32> poison, i32 [[TMP28]], i32 0
 ; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE]]
 ; CHECK:       pred.load.continue:
-; CHECK-NEXT:    [[TMP70:%.*]] = phi <4 x i32> [ poison, [[VECTOR_BODY]] ], [ [[TMP69]], [[PRED_LOAD_IF]] ]
-; CHECK-NEXT:    [[TMP71:%.*]] = extractelement <4 x i1> [[TMP39]], i32 1
-; CHECK-NEXT:    br i1 [[TMP71]], label [[PRED_LOAD_IF4:%.*]], label [[PRED_LOAD_CONTINUE5:%.*]]
-; CHECK:       pred.load.if4:
-; CHECK-NEXT:    [[TMP72:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP73:%.*]] = getelementptr inbounds i16, i16* [[TMP72]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP30:%.*]] = phi <4 x i32> [ poison, [[VECTOR_BODY]] ], [ [[TMP29]], [[PRED_LOAD_IF]] ]
+; CHECK-NEXT:    [[TMP31:%.*]] = extractelement <4 x i1> [[TMP11]], i32 1
+; CHECK-NEXT:    br i1 [[TMP31]], label [[PRED_LOAD_IF7:%.*]], label [[PRED_LOAD_CONTINUE8:%.*]]
+; CHECK:       pred.load.if7:
+; CHECK-NEXT:    [[TMP32:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP33:%.*]] = add i64 [[INDEX]], 1
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i16, i16* [[TMP32]], i64 [[TMP33]]
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i16* [[TMP34]] to i32*
+; CHECK-NEXT:    [[TMP36:%.*]] = load i32, i32* [[TMP35]], align 4
+; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i32> [[TMP30]], i32 [[TMP36]], i32 1
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE8]]
+; CHECK:       pred.load.continue8:
+; CHECK-NEXT:    [[TMP38:%.*]] = phi <4 x i32> [ [[TMP30]], [[PRED_LOAD_CONTINUE]] ], [ [[TMP37]], [[PRED_LOAD_IF7]] ]
+; CHECK-NEXT:    [[TMP39:%.*]] = extractelement <4 x i1> [[TMP11]], i32 2
+; CHECK-NEXT:    br i1 [[TMP39]], label [[PRED_LOAD_IF9:%.*]], label [[PRED_LOAD_CONTINUE10:%.*]]
+; CHECK:       pred.load.if9:
+; CHECK-NEXT:    [[TMP40:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP41:%.*]] = add i64 [[INDEX]], 2
+; CHECK-NEXT:    [[TMP42:%.*]] = getelementptr inbounds i16, i16* [[TMP40]], i64 [[TMP41]]
+; CHECK-NEXT:    [[TMP43:%.*]] = bitcast i16* [[TMP42]] to i32*
+; CHECK-NEXT:    [[TMP44:%.*]] = load i32, i32* [[TMP43]], align 4
+; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i32> [[TMP38]], i32 [[TMP44]], i32 2
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE10]]
+; CHECK:       pred.load.continue10:
+; CHECK-NEXT:    [[TMP46:%.*]] = phi <4 x i32> [ [[TMP38]], [[PRED_LOAD_CONTINUE8]] ], [ [[TMP45]], [[PRED_LOAD_IF9]] ]
+; CHECK-NEXT:    [[TMP47:%.*]] = extractelement <4 x i1> [[TMP11]], i32 3
+; CHECK-NEXT:    br i1 [[TMP47]], label [[PRED_LOAD_IF11:%.*]], label [[PRED_LOAD_CONTINUE12:%.*]]
+; CHECK:       pred.load.if11:
+; CHECK-NEXT:    [[TMP48:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP49:%.*]] = add i64 [[INDEX]], 3
+; CHECK-NEXT:    [[TMP50:%.*]] = getelementptr inbounds i16, i16* [[TMP48]], i64 [[TMP49]]
+; CHECK-NEXT:    [[TMP51:%.*]] = bitcast i16* [[TMP50]] to i32*
+; CHECK-NEXT:    [[TMP52:%.*]] = load i32, i32* [[TMP51]], align 4
+; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i32> [[TMP46]], i32 [[TMP52]], i32 3
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE12]]
+; CHECK:       pred.load.continue12:
+; CHECK-NEXT:    [[TMP54:%.*]] = phi <4 x i32> [ [[TMP46]], [[PRED_LOAD_CONTINUE10]] ], [ [[TMP53]], [[PRED_LOAD_IF11]] ]
+; CHECK-NEXT:    [[TMP55:%.*]] = extractelement <4 x i1> [[TMP15]], i32 0
+; CHECK-NEXT:    br i1 [[TMP55]], label [[PRED_LOAD_IF13:%.*]], label [[PRED_LOAD_CONTINUE14:%.*]]
+; CHECK:       pred.load.if13:
+; CHECK-NEXT:    [[TMP56:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP57:%.*]] = getelementptr inbounds i16, i16* [[TMP56]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP58:%.*]] = bitcast i16* [[TMP57]] to i32*
+; CHECK-NEXT:    [[TMP59:%.*]] = load i32, i32* [[TMP58]], align 4
+; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i32> poison, i32 [[TMP59]], i32 0
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE14]]
+; CHECK:       pred.load.continue14:
+; CHECK-NEXT:    [[TMP61:%.*]] = phi <4 x i32> [ poison, [[PRED_LOAD_CONTINUE12]] ], [ [[TMP60]], [[PRED_LOAD_IF13]] ]
+; CHECK-NEXT:    [[TMP62:%.*]] = extractelement <4 x i1> [[TMP15]], i32 1
+; CHECK-NEXT:    br i1 [[TMP62]], label [[PRED_LOAD_IF15:%.*]], label [[PRED_LOAD_CONTINUE16:%.*]]
+; CHECK:       pred.load.if15:
+; CHECK-NEXT:    [[TMP63:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP64:%.*]] = add i64 [[INDEX]], 5
+; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i16, i16* [[TMP63]], i64 [[TMP64]]
+; CHECK-NEXT:    [[TMP66:%.*]] = bitcast i16* [[TMP65]] to i32*
+; CHECK-NEXT:    [[TMP67:%.*]] = load i32, i32* [[TMP66]], align 4
+; CHECK-NEXT:    [[TMP68:%.*]] = insertelement <4 x i32> [[TMP61]], i32 [[TMP67]], i32 1
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE16]]
+; CHECK:       pred.load.continue16:
+; CHECK-NEXT:    [[TMP69:%.*]] = phi <4 x i32> [ [[TMP61]], [[PRED_LOAD_CONTINUE14]] ], [ [[TMP68]], [[PRED_LOAD_IF15]] ]
+; CHECK-NEXT:    [[TMP70:%.*]] = extractelement <4 x i1> [[TMP15]], i32 2
+; CHECK-NEXT:    br i1 [[TMP70]], label [[PRED_LOAD_IF17:%.*]], label [[PRED_LOAD_CONTINUE18:%.*]]
+; CHECK:       pred.load.if17:
+; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP72:%.*]] = add i64 [[INDEX]], 6
+; CHECK-NEXT:    [[TMP73:%.*]] = getelementptr inbounds i16, i16* [[TMP71]], i64 [[TMP72]]
 ; CHECK-NEXT:    [[TMP74:%.*]] = bitcast i16* [[TMP73]] to i32*
 ; CHECK-NEXT:    [[TMP75:%.*]] = load i32, i32* [[TMP74]], align 4
-; CHECK-NEXT:    [[TMP76:%.*]] = insertelement <4 x i32> [[TMP70]], i32 [[TMP75]], i32 1
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE5]]
-; CHECK:       pred.load.continue5:
-; CHECK-NEXT:    [[TMP77:%.*]] = phi <4 x i32> [ [[TMP70]], [[PRED_LOAD_CONTINUE]] ], [ [[TMP76]], [[PRED_LOAD_IF4]] ]
-; CHECK-NEXT:    [[TMP78:%.*]] = extractelement <4 x i1> [[TMP39]], i32 2
-; CHECK-NEXT:    br i1 [[TMP78]], label [[PRED_LOAD_IF6:%.*]], label [[PRED_LOAD_CONTINUE7:%.*]]
-; CHECK:       pred.load.if6:
+; CHECK-NEXT:    [[TMP76:%.*]] = insertelement <4 x i32> [[TMP69]], i32 [[TMP75]], i32 2
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE18]]
+; CHECK:       pred.load.continue18:
+; CHECK-NEXT:    [[TMP77:%.*]] = phi <4 x i32> [ [[TMP69]], [[PRED_LOAD_CONTINUE16]] ], [ [[TMP76]], [[PRED_LOAD_IF17]] ]
+; CHECK-NEXT:    [[TMP78:%.*]] = extractelement <4 x i1> [[TMP15]], i32 3
+; CHECK-NEXT:    br i1 [[TMP78]], label [[PRED_LOAD_IF19:%.*]], label [[PRED_LOAD_CONTINUE20:%.*]]
+; CHECK:       pred.load.if19:
 ; CHECK-NEXT:    [[TMP79:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP80:%.*]] = getelementptr inbounds i16, i16* [[TMP79]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP81:%.*]] = bitcast i16* [[TMP80]] to i32*
-; CHECK-NEXT:    [[TMP82:%.*]] = load i32, i32* [[TMP81]], align 4
-; CHECK-NEXT:    [[TMP83:%.*]] = insertelement <4 x i32> [[TMP77]], i32 [[TMP82]], i32 2
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE7]]
-; CHECK:       pred.load.continue7:
-; CHECK-NEXT:    [[TMP84:%.*]] = phi <4 x i32> [ [[TMP77]], [[PRED_LOAD_CONTINUE5]] ], [ [[TMP83]], [[PRED_LOAD_IF6]] ]
-; CHECK-NEXT:    [[TMP85:%.*]] = extractelement <4 x i1> [[TMP39]], i32 3
-; CHECK-NEXT:    br i1 [[TMP85]], label [[PRED_LOAD_IF8:%.*]], label [[PRED_LOAD_CONTINUE9:%.*]]
-; CHECK:       pred.load.if8:
-; CHECK-NEXT:    [[TMP86:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP87:%.*]] = getelementptr inbounds i16, i16* [[TMP86]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP88:%.*]] = bitcast i16* [[TMP87]] to i32*
-; CHECK-NEXT:    [[TMP89:%.*]] = load i32, i32* [[TMP88]], align 4
-; CHECK-NEXT:    [[TMP90:%.*]] = insertelement <4 x i32> [[TMP84]], i32 [[TMP89]], i32 3
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE9]]
-; CHECK:       pred.load.continue9:
-; CHECK-NEXT:    [[TMP91:%.*]] = phi <4 x i32> [ [[TMP84]], [[PRED_LOAD_CONTINUE7]] ], [ [[TMP90]], [[PRED_LOAD_IF8]] ]
-; CHECK-NEXT:    [[TMP92:%.*]] = extractelement <4 x i1> [[TMP47]], i32 0
-; CHECK-NEXT:    br i1 [[TMP92]], label [[PRED_LOAD_IF10:%.*]], label [[PRED_LOAD_CONTINUE11:%.*]]
-; CHECK:       pred.load.if10:
-; CHECK-NEXT:    [[TMP93:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP94:%.*]] = getelementptr inbounds i16, i16* [[TMP93]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP95:%.*]] = bitcast i16* [[TMP94]] to i32*
-; CHECK-NEXT:    [[TMP96:%.*]] = load i32, i32* [[TMP95]], align 4
-; CHECK-NEXT:    [[TMP97:%.*]] = insertelement <4 x i32> poison, i32 [[TMP96]], i32 0
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE11]]
-; CHECK:       pred.load.continue11:
-; CHECK-NEXT:    [[TMP98:%.*]] = phi <4 x i32> [ poison, [[PRED_LOAD_CONTINUE9]] ], [ [[TMP97]], [[PRED_LOAD_IF10]] ]
-; CHECK-NEXT:    [[TMP99:%.*]] = extractelement <4 x i1> [[TMP47]], i32 1
-; CHECK-NEXT:    br i1 [[TMP99]], label [[PRED_LOAD_IF12:%.*]], label [[PRED_LOAD_CONTINUE13:%.*]]
-; CHECK:       pred.load.if12:
-; CHECK-NEXT:    [[TMP100:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP101:%.*]] = getelementptr inbounds i16, i16* [[TMP100]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP102:%.*]] = bitcast i16* [[TMP101]] to i32*
-; CHECK-NEXT:    [[TMP103:%.*]] = load i32, i32* [[TMP102]], align 4
-; CHECK-NEXT:    [[TMP104:%.*]] = insertelement <4 x i32> [[TMP98]], i32 [[TMP103]], i32 1
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE13]]
-; CHECK:       pred.load.continue13:
-; CHECK-NEXT:    [[TMP105:%.*]] = phi <4 x i32> [ [[TMP98]], [[PRED_LOAD_CONTINUE11]] ], [ [[TMP104]], [[PRED_LOAD_IF12]] ]
-; CHECK-NEXT:    [[TMP106:%.*]] = extractelement <4 x i1> [[TMP47]], i32 2
-; CHECK-NEXT:    br i1 [[TMP106]], label [[PRED_LOAD_IF14:%.*]], label [[PRED_LOAD_CONTINUE15:%.*]]
-; CHECK:       pred.load.if14:
-; CHECK-NEXT:    [[TMP107:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP108:%.*]] = getelementptr inbounds i16, i16* [[TMP107]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP109:%.*]] = bitcast i16* [[TMP108]] to i32*
-; CHECK-NEXT:    [[TMP110:%.*]] = load i32, i32* [[TMP109]], align 4
-; CHECK-NEXT:    [[TMP111:%.*]] = insertelement <4 x i32> [[TMP105]], i32 [[TMP110]], i32 2
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE15]]
-; CHECK:       pred.load.continue15:
-; CHECK-NEXT:    [[TMP112:%.*]] = phi <4 x i32> [ [[TMP105]], [[PRED_LOAD_CONTINUE13]] ], [ [[TMP111]], [[PRED_LOAD_IF14]] ]
-; CHECK-NEXT:    [[TMP113:%.*]] = extractelement <4 x i1> [[TMP47]], i32 3
-; CHECK-NEXT:    br i1 [[TMP113]], label [[PRED_LOAD_IF16:%.*]], label [[PRED_LOAD_CONTINUE17:%.*]]
-; CHECK:       pred.load.if16:
-; CHECK-NEXT:    [[TMP114:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP115:%.*]] = getelementptr inbounds i16, i16* [[TMP114]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP116:%.*]] = bitcast i16* [[TMP115]] to i32*
-; CHECK-NEXT:    [[TMP117:%.*]] = load i32, i32* [[TMP116]], align 4
-; CHECK-NEXT:    [[TMP118:%.*]] = insertelement <4 x i32> [[TMP112]], i32 [[TMP117]], i32 3
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE17]]
-; CHECK:       pred.load.continue17:
-; CHECK-NEXT:    [[TMP119:%.*]] = phi <4 x i32> [ [[TMP112]], [[PRED_LOAD_CONTINUE15]] ], [ [[TMP118]], [[PRED_LOAD_IF16]] ]
-; CHECK-NEXT:    [[TMP120:%.*]] = extractelement <4 x i1> [[TMP55]], i32 0
-; CHECK-NEXT:    br i1 [[TMP120]], label [[PRED_LOAD_IF18:%.*]], label [[PRED_LOAD_CONTINUE19:%.*]]
-; CHECK:       pred.load.if18:
-; CHECK-NEXT:    [[TMP121:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP122:%.*]] = getelementptr inbounds i16, i16* [[TMP121]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP123:%.*]] = bitcast i16* [[TMP122]] to i32*
-; CHECK-NEXT:    [[TMP124:%.*]] = load i32, i32* [[TMP123]], align 4
-; CHECK-NEXT:    [[TMP125:%.*]] = insertelement <4 x i32> poison, i32 [[TMP124]], i32 0
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE19]]
-; CHECK:       pred.load.continue19:
-; CHECK-NEXT:    [[TMP126:%.*]] = phi <4 x i32> [ poison, [[PRED_LOAD_CONTINUE17]] ], [ [[TMP125]], [[PRED_LOAD_IF18]] ]
-; CHECK-NEXT:    [[TMP127:%.*]] = extractelement <4 x i1> [[TMP55]], i32 1
-; CHECK-NEXT:    br i1 [[TMP127]], label [[PRED_LOAD_IF20:%.*]], label [[PRED_LOAD_CONTINUE21:%.*]]
-; CHECK:       pred.load.if20:
-; CHECK-NEXT:    [[TMP128:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP129:%.*]] = getelementptr inbounds i16, i16* [[TMP128]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP130:%.*]] = bitcast i16* [[TMP129]] to i32*
-; CHECK-NEXT:    [[TMP131:%.*]] = load i32, i32* [[TMP130]], align 4
-; CHECK-NEXT:    [[TMP132:%.*]] = insertelement <4 x i32> [[TMP126]], i32 [[TMP131]], i32 1
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE21]]
-; CHECK:       pred.load.continue21:
-; CHECK-NEXT:    [[TMP133:%.*]] = phi <4 x i32> [ [[TMP126]], [[PRED_LOAD_CONTINUE19]] ], [ [[TMP132]], [[PRED_LOAD_IF20]] ]
-; CHECK-NEXT:    [[TMP134:%.*]] = extractelement <4 x i1> [[TMP55]], i32 2
-; CHECK-NEXT:    br i1 [[TMP134]], label [[PRED_LOAD_IF22:%.*]], label [[PRED_LOAD_CONTINUE23:%.*]]
-; CHECK:       pred.load.if22:
-; CHECK-NEXT:    [[TMP135:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP136:%.*]] = getelementptr inbounds i16, i16* [[TMP135]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP137:%.*]] = bitcast i16* [[TMP136]] to i32*
-; CHECK-NEXT:    [[TMP138:%.*]] = load i32, i32* [[TMP137]], align 4
-; CHECK-NEXT:    [[TMP139:%.*]] = insertelement <4 x i32> [[TMP133]], i32 [[TMP138]], i32 2
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE23]]
-; CHECK:       pred.load.continue23:
-; CHECK-NEXT:    [[TMP140:%.*]] = phi <4 x i32> [ [[TMP133]], [[PRED_LOAD_CONTINUE21]] ], [ [[TMP139]], [[PRED_LOAD_IF22]] ]
-; CHECK-NEXT:    [[TMP141:%.*]] = extractelement <4 x i1> [[TMP55]], i32 3
-; CHECK-NEXT:    br i1 [[TMP141]], label [[PRED_LOAD_IF24:%.*]], label [[PRED_LOAD_CONTINUE25:%.*]]
-; CHECK:       pred.load.if24:
-; CHECK-NEXT:    [[TMP142:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP143:%.*]] = getelementptr inbounds i16, i16* [[TMP142]], i64 [[TMP11]]
+; CHECK-NEXT:    [[TMP80:%.*]] = add i64 [[INDEX]], 7
+; CHECK-NEXT:    [[TMP81:%.*]] = getelementptr inbounds i16, i16* [[TMP79]], i64 [[TMP80]]
+; CHECK-NEXT:    [[TMP82:%.*]] = bitcast i16* [[TMP81]] to i32*
+; CHECK-NEXT:    [[TMP83:%.*]] = load i32, i32* [[TMP82]], align 4
+; CHECK-NEXT:    [[TMP84:%.*]] = insertelement <4 x i32> [[TMP77]], i32 [[TMP83]], i32 3
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE20]]
+; CHECK:       pred.load.continue20:
+; CHECK-NEXT:    [[TMP85:%.*]] = phi <4 x i32> [ [[TMP77]], [[PRED_LOAD_CONTINUE18]] ], [ [[TMP84]], [[PRED_LOAD_IF19]] ]
+; CHECK-NEXT:    [[TMP86:%.*]] = extractelement <4 x i1> [[TMP19]], i32 0
+; CHECK-NEXT:    br i1 [[TMP86]], label [[PRED_LOAD_IF21:%.*]], label [[PRED_LOAD_CONTINUE22:%.*]]
+; CHECK:       pred.load.if21:
+; CHECK-NEXT:    [[TMP87:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP88:%.*]] = getelementptr inbounds i16, i16* [[TMP87]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP89:%.*]] = bitcast i16* [[TMP88]] to i32*
+; CHECK-NEXT:    [[TMP90:%.*]] = load i32, i32* [[TMP89]], align 4
+; CHECK-NEXT:    [[TMP91:%.*]] = insertelement <4 x i32> poison, i32 [[TMP90]], i32 0
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE22]]
+; CHECK:       pred.load.continue22:
+; CHECK-NEXT:    [[TMP92:%.*]] = phi <4 x i32> [ poison, [[PRED_LOAD_CONTINUE20]] ], [ [[TMP91]], [[PRED_LOAD_IF21]] ]
+; CHECK-NEXT:    [[TMP93:%.*]] = extractelement <4 x i1> [[TMP19]], i32 1
+; CHECK-NEXT:    br i1 [[TMP93]], label [[PRED_LOAD_IF23:%.*]], label [[PRED_LOAD_CONTINUE24:%.*]]
+; CHECK:       pred.load.if23:
+; CHECK-NEXT:    [[TMP94:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP95:%.*]] = add i64 [[INDEX]], 9
+; CHECK-NEXT:    [[TMP96:%.*]] = getelementptr inbounds i16, i16* [[TMP94]], i64 [[TMP95]]
+; CHECK-NEXT:    [[TMP97:%.*]] = bitcast i16* [[TMP96]] to i32*
+; CHECK-NEXT:    [[TMP98:%.*]] = load i32, i32* [[TMP97]], align 4
+; CHECK-NEXT:    [[TMP99:%.*]] = insertelement <4 x i32> [[TMP92]], i32 [[TMP98]], i32 1
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE24]]
+; CHECK:       pred.load.continue24:
+; CHECK-NEXT:    [[TMP100:%.*]] = phi <4 x i32> [ [[TMP92]], [[PRED_LOAD_CONTINUE22]] ], [ [[TMP99]], [[PRED_LOAD_IF23]] ]
+; CHECK-NEXT:    [[TMP101:%.*]] = extractelement <4 x i1> [[TMP19]], i32 2
+; CHECK-NEXT:    br i1 [[TMP101]], label [[PRED_LOAD_IF25:%.*]], label [[PRED_LOAD_CONTINUE26:%.*]]
+; CHECK:       pred.load.if25:
+; CHECK-NEXT:    [[TMP102:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP103:%.*]] = add i64 [[INDEX]], 10
+; CHECK-NEXT:    [[TMP104:%.*]] = getelementptr inbounds i16, i16* [[TMP102]], i64 [[TMP103]]
+; CHECK-NEXT:    [[TMP105:%.*]] = bitcast i16* [[TMP104]] to i32*
+; CHECK-NEXT:    [[TMP106:%.*]] = load i32, i32* [[TMP105]], align 4
+; CHECK-NEXT:    [[TMP107:%.*]] = insertelement <4 x i32> [[TMP100]], i32 [[TMP106]], i32 2
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE26]]
+; CHECK:       pred.load.continue26:
+; CHECK-NEXT:    [[TMP108:%.*]] = phi <4 x i32> [ [[TMP100]], [[PRED_LOAD_CONTINUE24]] ], [ [[TMP107]], [[PRED_LOAD_IF25]] ]
+; CHECK-NEXT:    [[TMP109:%.*]] = extractelement <4 x i1> [[TMP19]], i32 3
+; CHECK-NEXT:    br i1 [[TMP109]], label [[PRED_LOAD_IF27:%.*]], label [[PRED_LOAD_CONTINUE28:%.*]]
+; CHECK:       pred.load.if27:
+; CHECK-NEXT:    [[TMP110:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP111:%.*]] = add i64 [[INDEX]], 11
+; CHECK-NEXT:    [[TMP112:%.*]] = getelementptr inbounds i16, i16* [[TMP110]], i64 [[TMP111]]
+; CHECK-NEXT:    [[TMP113:%.*]] = bitcast i16* [[TMP112]] to i32*
+; CHECK-NEXT:    [[TMP114:%.*]] = load i32, i32* [[TMP113]], align 4
+; CHECK-NEXT:    [[TMP115:%.*]] = insertelement <4 x i32> [[TMP108]], i32 [[TMP114]], i32 3
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE28]]
+; CHECK:       pred.load.continue28:
+; CHECK-NEXT:    [[TMP116:%.*]] = phi <4 x i32> [ [[TMP108]], [[PRED_LOAD_CONTINUE26]] ], [ [[TMP115]], [[PRED_LOAD_IF27]] ]
+; CHECK-NEXT:    [[TMP117:%.*]] = extractelement <4 x i1> [[TMP23]], i32 0
+; CHECK-NEXT:    br i1 [[TMP117]], label [[PRED_LOAD_IF29:%.*]], label [[PRED_LOAD_CONTINUE30:%.*]]
+; CHECK:       pred.load.if29:
+; CHECK-NEXT:    [[TMP118:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP119:%.*]] = getelementptr inbounds i16, i16* [[TMP118]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP120:%.*]] = bitcast i16* [[TMP119]] to i32*
+; CHECK-NEXT:    [[TMP121:%.*]] = load i32, i32* [[TMP120]], align 4
+; CHECK-NEXT:    [[TMP122:%.*]] = insertelement <4 x i32> poison, i32 [[TMP121]], i32 0
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE30]]
+; CHECK:       pred.load.continue30:
+; CHECK-NEXT:    [[TMP123:%.*]] = phi <4 x i32> [ poison, [[PRED_LOAD_CONTINUE28]] ], [ [[TMP122]], [[PRED_LOAD_IF29]] ]
+; CHECK-NEXT:    [[TMP124:%.*]] = extractelement <4 x i1> [[TMP23]], i32 1
+; CHECK-NEXT:    br i1 [[TMP124]], label [[PRED_LOAD_IF31:%.*]], label [[PRED_LOAD_CONTINUE32:%.*]]
+; CHECK:       pred.load.if31:
+; CHECK-NEXT:    [[TMP125:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP126:%.*]] = add i64 [[INDEX]], 13
+; CHECK-NEXT:    [[TMP127:%.*]] = getelementptr inbounds i16, i16* [[TMP125]], i64 [[TMP126]]
+; CHECK-NEXT:    [[TMP128:%.*]] = bitcast i16* [[TMP127]] to i32*
+; CHECK-NEXT:    [[TMP129:%.*]] = load i32, i32* [[TMP128]], align 4
+; CHECK-NEXT:    [[TMP130:%.*]] = insertelement <4 x i32> [[TMP123]], i32 [[TMP129]], i32 1
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE32]]
+; CHECK:       pred.load.continue32:
+; CHECK-NEXT:    [[TMP131:%.*]] = phi <4 x i32> [ [[TMP123]], [[PRED_LOAD_CONTINUE30]] ], [ [[TMP130]], [[PRED_LOAD_IF31]] ]
+; CHECK-NEXT:    [[TMP132:%.*]] = extractelement <4 x i1> [[TMP23]], i32 2
+; CHECK-NEXT:    br i1 [[TMP132]], label [[PRED_LOAD_IF33:%.*]], label [[PRED_LOAD_CONTINUE34:%.*]]
+; CHECK:       pred.load.if33:
+; CHECK-NEXT:    [[TMP133:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP134:%.*]] = add i64 [[INDEX]], 14
+; CHECK-NEXT:    [[TMP135:%.*]] = getelementptr inbounds i16, i16* [[TMP133]], i64 [[TMP134]]
+; CHECK-NEXT:    [[TMP136:%.*]] = bitcast i16* [[TMP135]] to i32*
+; CHECK-NEXT:    [[TMP137:%.*]] = load i32, i32* [[TMP136]], align 4
+; CHECK-NEXT:    [[TMP138:%.*]] = insertelement <4 x i32> [[TMP131]], i32 [[TMP137]], i32 2
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE34]]
+; CHECK:       pred.load.continue34:
+; CHECK-NEXT:    [[TMP139:%.*]] = phi <4 x i32> [ [[TMP131]], [[PRED_LOAD_CONTINUE32]] ], [ [[TMP138]], [[PRED_LOAD_IF33]] ]
+; CHECK-NEXT:    [[TMP140:%.*]] = extractelement <4 x i1> [[TMP23]], i32 3
+; CHECK-NEXT:    br i1 [[TMP140]], label [[PRED_LOAD_IF35:%.*]], label [[PRED_LOAD_CONTINUE36]]
+; CHECK:       pred.load.if35:
+; CHECK-NEXT:    [[TMP141:%.*]] = bitcast i32* [[BASE]] to i16*
+; CHECK-NEXT:    [[TMP142:%.*]] = add i64 [[INDEX]], 15
+; CHECK-NEXT:    [[TMP143:%.*]] = getelementptr inbounds i16, i16* [[TMP141]], i64 [[TMP142]]
 ; CHECK-NEXT:    [[TMP144:%.*]] = bitcast i16* [[TMP143]] to i32*
 ; CHECK-NEXT:    [[TMP145:%.*]] = load i32, i32* [[TMP144]], align 4
-; CHECK-NEXT:    [[TMP146:%.*]] = insertelement <4 x i32> [[TMP140]], i32 [[TMP145]], i32 3
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE25]]
-; CHECK:       pred.load.continue25:
-; CHECK-NEXT:    [[TMP147:%.*]] = phi <4 x i32> [ [[TMP140]], [[PRED_LOAD_CONTINUE23]] ], [ [[TMP146]], [[PRED_LOAD_IF24]] ]
-; CHECK-NEXT:    [[TMP148:%.*]] = extractelement <4 x i1> [[TMP63]], i32 0
-; CHECK-NEXT:    br i1 [[TMP148]], label [[PRED_LOAD_IF26:%.*]], label [[PRED_LOAD_CONTINUE27:%.*]]
-; CHECK:       pred.load.if26:
-; CHECK-NEXT:    [[TMP149:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP150:%.*]] = getelementptr inbounds i16, i16* [[TMP149]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP151:%.*]] = bitcast i16* [[TMP150]] to i32*
-; CHECK-NEXT:    [[TMP152:%.*]] = load i32, i32* [[TMP151]], align 4
-; CHECK-NEXT:    [[TMP153:%.*]] = insertelement <4 x i32> poison, i32 [[TMP152]], i32 0
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE27]]
-; CHECK:       pred.load.continue27:
-; CHECK-NEXT:    [[TMP154:%.*]] = phi <4 x i32> [ poison, [[PRED_LOAD_CONTINUE25]] ], [ [[TMP153]], [[PRED_LOAD_IF26]] ]
-; CHECK-NEXT:    [[TMP155:%.*]] = extractelement <4 x i1> [[TMP63]], i32 1
-; CHECK-NEXT:    br i1 [[TMP155]], label [[PRED_LOAD_IF28:%.*]], label [[PRED_LOAD_CONTINUE29:%.*]]
-; CHECK:       pred.load.if28:
-; CHECK-NEXT:    [[TMP156:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP157:%.*]] = getelementptr inbounds i16, i16* [[TMP156]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP158:%.*]] = bitcast i16* [[TMP157]] to i32*
-; CHECK-NEXT:    [[TMP159:%.*]] = load i32, i32* [[TMP158]], align 4
-; CHECK-NEXT:    [[TMP160:%.*]] = insertelement <4 x i32> [[TMP154]], i32 [[TMP159]], i32 1
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE29]]
-; CHECK:       pred.load.continue29:
-; CHECK-NEXT:    [[TMP161:%.*]] = phi <4 x i32> [ [[TMP154]], [[PRED_LOAD_CONTINUE27]] ], [ [[TMP160]], [[PRED_LOAD_IF28]] ]
-; CHECK-NEXT:    [[TMP162:%.*]] = extractelement <4 x i1> [[TMP63]], i32 2
-; CHECK-NEXT:    br i1 [[TMP162]], label [[PRED_LOAD_IF30:%.*]], label [[PRED_LOAD_CONTINUE31:%.*]]
-; CHECK:       pred.load.if30:
-; CHECK-NEXT:    [[TMP163:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP164:%.*]] = getelementptr inbounds i16, i16* [[TMP163]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP165:%.*]] = bitcast i16* [[TMP164]] to i32*
-; CHECK-NEXT:    [[TMP166:%.*]] = load i32, i32* [[TMP165]], align 4
-; CHECK-NEXT:    [[TMP167:%.*]] = insertelement <4 x i32> [[TMP161]], i32 [[TMP166]], i32 2
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE31]]
-; CHECK:       pred.load.continue31:
-; CHECK-NEXT:    [[TMP168:%.*]] = phi <4 x i32> [ [[TMP161]], [[PRED_LOAD_CONTINUE29]] ], [ [[TMP167]], [[PRED_LOAD_IF30]] ]
-; CHECK-NEXT:    [[TMP169:%.*]] = extractelement <4 x i1> [[TMP63]], i32 3
-; CHECK-NEXT:    br i1 [[TMP169]], label [[PRED_LOAD_IF32:%.*]], label [[PRED_LOAD_CONTINUE33]]
-; CHECK:       pred.load.if32:
-; CHECK-NEXT:    [[TMP170:%.*]] = bitcast i32* [[BASE]] to i16*
-; CHECK-NEXT:    [[TMP171:%.*]] = getelementptr inbounds i16, i16* [[TMP170]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP172:%.*]] = bitcast i16* [[TMP171]] to i32*
-; CHECK-NEXT:    [[TMP173:%.*]] = load i32, i32* [[TMP172]], align 4
-; CHECK-NEXT:    [[TMP174:%.*]] = insertelement <4 x i32> [[TMP168]], i32 [[TMP173]], i32 3
-; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE33]]
-; CHECK:       pred.load.continue33:
-; CHECK-NEXT:    [[TMP175:%.*]] = phi <4 x i32> [ [[TMP168]], [[PRED_LOAD_CONTINUE31]] ], [ [[TMP174]], [[PRED_LOAD_IF32]] ]
-; CHECK-NEXT:    [[TMP176:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP177:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP178:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP179:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[TMP91]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI34:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[TMP119]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI35:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[TMP147]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI36:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[TMP175]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP180]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP181]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI34]]
-; CHECK-NEXT:    [[TMP182]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI35]]
-; CHECK-NEXT:    [[TMP183]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI36]]
+; CHECK-NEXT:    [[TMP146:%.*]] = insertelement <4 x i32> [[TMP139]], i32 [[TMP145]], i32 3
+; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE36]]
+; CHECK:       pred.load.continue36:
+; CHECK-NEXT:    [[TMP147:%.*]] = phi <4 x i32> [ [[TMP139]], [[PRED_LOAD_CONTINUE34]] ], [ [[TMP146]], [[PRED_LOAD_IF35]] ]
+; CHECK-NEXT:    [[TMP148:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP149:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP150:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP151:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[TMP54]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI37:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[TMP85]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI38:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[TMP116]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI39:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[TMP147]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP152]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP153]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI37]]
+; CHECK-NEXT:    [[TMP154]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI38]]
+; CHECK-NEXT:    [[TMP155]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI39]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP184:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP184]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP8:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP156:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP156]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP8:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP181]], [[TMP180]]
-; CHECK-NEXT:    [[BIN_RDX37:%.*]] = add <4 x i32> [[TMP182]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX38:%.*]] = add <4 x i32> [[TMP183]], [[BIN_RDX37]]
-; CHECK-NEXT:    [[TMP185:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX38]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP153]], [[TMP152]]
+; CHECK-NEXT:    [[BIN_RDX40:%.*]] = add <4 x i32> [[TMP154]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX41:%.*]] = add <4 x i32> [[TMP155]], [[BIN_RDX40]]
+; CHECK-NEXT:    [[TMP157:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX41]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP185]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP157]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[BASE_I16P:%.*]] = bitcast i32* [[BASE]] to i16*
 ; CHECK-NEXT:    [[ADDR_I16P:%.*]] = getelementptr inbounds i16, i16* [[BASE_I16P]], i64 [[IV]]
 ; CHECK-NEXT:    [[ADDR:%.*]] = bitcast i16* [[ADDR_I16P]] to i32*
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP9:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP185]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP157]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [4096 x i32]
   %base = bitcast [4096 x i32]* %alloca to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 16
 ; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 16
 ; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP84:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP44:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP16:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP16]]
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> poison, i1 [[TMP33]], i32 0
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 1
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 2
-; CHECK-NEXT:    [[TMP40:%.*]] = insertelement <4 x i1> [[TMP39]], i1 [[TMP36]], i32 3
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> poison, i1 [[TMP41]], i32 0
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 1
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 2
-; CHECK-NEXT:    [[TMP48:%.*]] = insertelement <4 x i1> [[TMP47]], i1 [[TMP44]], i32 3
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> poison, i1 [[TMP49]], i32 0
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 1
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 2
-; CHECK-NEXT:    [[TMP56:%.*]] = insertelement <4 x i1> [[TMP55]], i1 [[TMP52]], i32 3
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = load i1, i1* [[TMP32]], align 1
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> poison, i1 [[TMP57]], i32 0
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 1
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 2
-; CHECK-NEXT:    [[TMP64:%.*]] = insertelement <4 x i1> [[TMP63]], i1 [[TMP60]], i32 3
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP69:%.*]] = getelementptr inbounds i32, i32* [[TMP65]], i32 0
-; CHECK-NEXT:    [[TMP70:%.*]] = bitcast i32* [[TMP69]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP70]], i32 4, <4 x i1> [[TMP40]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP71:%.*]] = getelementptr inbounds i32, i32* [[TMP65]], i32 4
-; CHECK-NEXT:    [[TMP72:%.*]] = bitcast i32* [[TMP71]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD4:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP72]], i32 4, <4 x i1> [[TMP48]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP73:%.*]] = getelementptr inbounds i32, i32* [[TMP65]], i32 8
-; CHECK-NEXT:    [[TMP74:%.*]] = bitcast i32* [[TMP73]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD5:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP74]], i32 4, <4 x i1> [[TMP56]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP75:%.*]] = getelementptr inbounds i32, i32* [[TMP65]], i32 12
-; CHECK-NEXT:    [[TMP76:%.*]] = bitcast i32* [[TMP75]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD6:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP76]], i32 4, <4 x i1> [[TMP64]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP40]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP48]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP56]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP80:%.*]] = xor <4 x i1> [[TMP64]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP40]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP48]], <4 x i32> [[WIDE_MASKED_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP56]], <4 x i32> [[WIDE_MASKED_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP64]], <4 x i32> [[WIDE_MASKED_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP84]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
+; CHECK-NEXT:    [[TMP9:%.*]] = bitcast i1* [[TMP5]] to i8*
+; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i8, i8* [[TMP9]], i32 0
+; CHECK-NEXT:    [[TMP11:%.*]] = bitcast i8* [[TMP10]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP11]], align 1
+; CHECK-NEXT:    [[TMP12:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP13:%.*]] = bitcast i1* [[TMP5]] to i8*
+; CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, i8* [[TMP13]], i32 4
+; CHECK-NEXT:    [[TMP15:%.*]] = bitcast i8* [[TMP14]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP15]], align 1
+; CHECK-NEXT:    [[TMP16:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP17:%.*]] = bitcast i1* [[TMP5]] to i8*
+; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i8, i8* [[TMP17]], i32 8
+; CHECK-NEXT:    [[TMP19:%.*]] = bitcast i8* [[TMP18]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP19]], align 1
+; CHECK-NEXT:    [[TMP20:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP21:%.*]] = bitcast i1* [[TMP5]] to i8*
+; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i8, i8* [[TMP21]], i32 12
+; CHECK-NEXT:    [[TMP23:%.*]] = bitcast i8* [[TMP22]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP23]], align 1
+; CHECK-NEXT:    [[TMP24:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
+; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i32, i32* [[TMP25]], i32 0
+; CHECK-NEXT:    [[TMP30:%.*]] = bitcast i32* [[TMP29]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP30]], i32 4, <4 x i1> [[TMP12]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i32, i32* [[TMP25]], i32 4
+; CHECK-NEXT:    [[TMP32:%.*]] = bitcast i32* [[TMP31]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD7:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP32]], i32 4, <4 x i1> [[TMP16]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP33:%.*]] = getelementptr inbounds i32, i32* [[TMP25]], i32 8
+; CHECK-NEXT:    [[TMP34:%.*]] = bitcast i32* [[TMP33]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD8:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP34]], i32 4, <4 x i1> [[TMP20]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP35:%.*]] = getelementptr inbounds i32, i32* [[TMP25]], i32 12
+; CHECK-NEXT:    [[TMP36:%.*]] = bitcast i32* [[TMP35]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD9:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP36]], i32 4, <4 x i1> [[TMP24]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP12]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP16]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP20]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP40:%.*]] = xor <4 x i1> [[TMP24]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP12]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI10:%.*]] = select <4 x i1> [[TMP16]], <4 x i32> [[WIDE_MASKED_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP20]], <4 x i32> [[WIDE_MASKED_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP24]], <4 x i32> [[WIDE_MASKED_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI10]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP44]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI12]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP85:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT:    br i1 [[TMP85]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP10:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP45:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT:    br i1 [[TMP45]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP10:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP82]], [[TMP81]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP84]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP86:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP42]], [[TMP41]]
+; CHECK-NEXT:    [[BIN_RDX13:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP44]], [[BIN_RDX13]]
+; CHECK-NEXT:    [[TMP46:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX14]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP86]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP46]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], [[MIN_N]]
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP11:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP86]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP46]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [4096 x i32]
   %base = bitcast [4096 x i32]* %alloca to i32*
   call void @init(i32* %base)
   %min.cmp = icmp ult i64 4096, %n
   %min.n = select i1 %min.cmp, i64 4096, i64 %n

 ; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [4096 x i32], align 4
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast [4096 x i32]* [[ALLOCA]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP80:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP40:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[OFFSET_IDX:%.*]] = add i64 1024, [[INDEX]]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[OFFSET_IDX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[OFFSET_IDX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[OFFSET_IDX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[OFFSET_IDX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[OFFSET_IDX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[OFFSET_IDX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[OFFSET_IDX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[OFFSET_IDX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[OFFSET_IDX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[OFFSET_IDX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[OFFSET_IDX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[OFFSET_IDX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[OFFSET_IDX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[OFFSET_IDX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[OFFSET_IDX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[OFFSET_IDX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = bitcast i32* [[TMP68]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP69]], i32 4, <4 x i1> [[TMP39]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 4
-; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[TMP70]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD4:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP71]], i32 4, <4 x i1> [[TMP47]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 8
-; CHECK-NEXT:    [[TMP73:%.*]] = bitcast i32* [[TMP72]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD5:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP73]], i32 4, <4 x i1> [[TMP55]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 12
-; CHECK-NEXT:    [[TMP75:%.*]] = bitcast i32* [[TMP74]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD6:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP75]], i32 4, <4 x i1> [[TMP63]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP76:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[WIDE_MASKED_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[WIDE_MASKED_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[WIDE_MASKED_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[OFFSET_IDX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[OFFSET_IDX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[OFFSET_IDX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = bitcast i32* [[TMP28]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP29]], i32 4, <4 x i1> [[TMP11]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 4
+; CHECK-NEXT:    [[TMP31:%.*]] = bitcast i32* [[TMP30]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD7:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP31]], i32 4, <4 x i1> [[TMP15]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 8
+; CHECK-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD8:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP33]], i32 4, <4 x i1> [[TMP19]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 12
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i32* [[TMP34]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD9:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP35]], i32 4, <4 x i1> [[TMP23]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP36:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI10:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[WIDE_MASKED_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[WIDE_MASKED_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[WIDE_MASKED_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP40]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI10]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI12]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP84:%.*]] = icmp eq i64 [[INDEX_NEXT]], 3072
-; CHECK-NEXT:    br i1 [[TMP84]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP12:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP44:%.*]] = icmp eq i64 [[INDEX_NEXT]], 3072
+; CHECK-NEXT:    br i1 [[TMP44]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP12:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP81]], [[TMP80]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP82]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP85:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP41]], [[TMP40]]
+; CHECK-NEXT:    [[BIN_RDX13:%.*]] = add <4 x i32> [[TMP42]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX13]]
+; CHECK-NEXT:    [[TMP45:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX14]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 3072, 3072
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 1024, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP13:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [4096 x i32]
   %base = bitcast [4096 x i32]* %alloca to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [1024 x i32], align 4
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast [1024 x i32]* [[ALLOCA]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP80:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP40:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = bitcast i32* [[TMP68]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP69]], i32 4, <4 x i1> [[TMP39]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 4
-; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[TMP70]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD4:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP71]], i32 4, <4 x i1> [[TMP47]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 8
-; CHECK-NEXT:    [[TMP73:%.*]] = bitcast i32* [[TMP72]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD5:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP73]], i32 4, <4 x i1> [[TMP55]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 12
-; CHECK-NEXT:    [[TMP75:%.*]] = bitcast i32* [[TMP74]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD6:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP75]], i32 4, <4 x i1> [[TMP63]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP76:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[WIDE_MASKED_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[WIDE_MASKED_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[WIDE_MASKED_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = bitcast i32* [[TMP28]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP29]], i32 4, <4 x i1> [[TMP11]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 4
+; CHECK-NEXT:    [[TMP31:%.*]] = bitcast i32* [[TMP30]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD7:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP31]], i32 4, <4 x i1> [[TMP15]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 8
+; CHECK-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD8:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP33]], i32 4, <4 x i1> [[TMP19]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 12
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i32* [[TMP34]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD9:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP35]], i32 4, <4 x i1> [[TMP23]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP36:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI10:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[WIDE_MASKED_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[WIDE_MASKED_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[WIDE_MASKED_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP40]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI10]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI12]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP84:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP84]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP16:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP44:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP44]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP16:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP81]], [[TMP80]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP82]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP85:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP41]], [[TMP40]]
+; CHECK-NEXT:    [[BIN_RDX13:%.*]] = add <4 x i32> [[TMP42]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX13]]
+; CHECK-NEXT:    [[TMP45:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX14]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP17:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [1024 x i32]
   %base = bitcast [1024 x i32]* %alloca to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [4095 x i32], align 4
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast [4095 x i32]* [[ALLOCA]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP80:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP40:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = bitcast i32* [[TMP68]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP69]], i32 4, <4 x i1> [[TMP39]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 4
-; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[TMP70]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD4:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP71]], i32 4, <4 x i1> [[TMP47]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 8
-; CHECK-NEXT:    [[TMP73:%.*]] = bitcast i32* [[TMP72]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD5:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP73]], i32 4, <4 x i1> [[TMP55]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 12
-; CHECK-NEXT:    [[TMP75:%.*]] = bitcast i32* [[TMP74]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD6:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP75]], i32 4, <4 x i1> [[TMP63]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP76:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[WIDE_MASKED_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[WIDE_MASKED_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[WIDE_MASKED_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = bitcast i32* [[TMP28]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP29]], i32 4, <4 x i1> [[TMP11]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 4
+; CHECK-NEXT:    [[TMP31:%.*]] = bitcast i32* [[TMP30]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD7:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP31]], i32 4, <4 x i1> [[TMP15]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 8
+; CHECK-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD8:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP33]], i32 4, <4 x i1> [[TMP19]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 12
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i32* [[TMP34]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD9:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP35]], i32 4, <4 x i1> [[TMP23]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP36:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI10:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[WIDE_MASKED_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[WIDE_MASKED_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[WIDE_MASKED_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP40]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI10]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI12]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP84:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP84]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP18:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP44:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP44]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP18:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP81]], [[TMP80]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP82]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP85:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP41]], [[TMP40]]
+; CHECK-NEXT:    [[BIN_RDX13:%.*]] = add <4 x i32> [[TMP42]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX13]]
+; CHECK-NEXT:    [[TMP45:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX14]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP19:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [4095 x i32]
   %base = bitcast [4095 x i32]* %alloca to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [16383 x i8], align 1
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast [16383 x i8]* [[ALLOCA]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP80:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP40:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = bitcast i32* [[TMP68]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP69]], i32 4, <4 x i1> [[TMP39]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 4
-; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[TMP70]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD4:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP71]], i32 4, <4 x i1> [[TMP47]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 8
-; CHECK-NEXT:    [[TMP73:%.*]] = bitcast i32* [[TMP72]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD5:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP73]], i32 4, <4 x i1> [[TMP55]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 12
-; CHECK-NEXT:    [[TMP75:%.*]] = bitcast i32* [[TMP74]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD6:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP75]], i32 4, <4 x i1> [[TMP63]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP76:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[WIDE_MASKED_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[WIDE_MASKED_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[WIDE_MASKED_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = bitcast i32* [[TMP28]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP29]], i32 4, <4 x i1> [[TMP11]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 4
+; CHECK-NEXT:    [[TMP31:%.*]] = bitcast i32* [[TMP30]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD7:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP31]], i32 4, <4 x i1> [[TMP15]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 8
+; CHECK-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD8:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP33]], i32 4, <4 x i1> [[TMP19]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 12
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i32* [[TMP34]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD9:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP35]], i32 4, <4 x i1> [[TMP23]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP36:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI10:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[WIDE_MASKED_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[WIDE_MASKED_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[WIDE_MASKED_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP40]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI10]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI12]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP84:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP84]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP20:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP44:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP44]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP20:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP81]], [[TMP80]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP82]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP85:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP41]], [[TMP40]]
+; CHECK-NEXT:    [[BIN_RDX13:%.*]] = add <4 x i32> [[TMP42]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX13]]
+; CHECK-NEXT:    [[TMP45:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX14]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP21:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [16383 x i8]
   %base = bitcast [16383 x i8]* %alloca to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 16
 ; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 16
 ; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP84:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP44:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP16:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP16]]
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> poison, i1 [[TMP33]], i32 0
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 1
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 2
-; CHECK-NEXT:    [[TMP40:%.*]] = insertelement <4 x i1> [[TMP39]], i1 [[TMP36]], i32 3
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> poison, i1 [[TMP41]], i32 0
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 1
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 2
-; CHECK-NEXT:    [[TMP48:%.*]] = insertelement <4 x i1> [[TMP47]], i1 [[TMP44]], i32 3
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> poison, i1 [[TMP49]], i32 0
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 1
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 2
-; CHECK-NEXT:    [[TMP56:%.*]] = insertelement <4 x i1> [[TMP55]], i1 [[TMP52]], i32 3
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = load i1, i1* [[TMP32]], align 1
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> poison, i1 [[TMP57]], i32 0
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 1
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 2
-; CHECK-NEXT:    [[TMP64:%.*]] = insertelement <4 x i1> [[TMP63]], i1 [[TMP60]], i32 3
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP69:%.*]] = getelementptr inbounds i32, i32* [[TMP65]], i32 0
-; CHECK-NEXT:    [[TMP70:%.*]] = bitcast i32* [[TMP69]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, <4 x i32>* [[TMP70]], align 4
-; CHECK-NEXT:    [[TMP71:%.*]] = getelementptr inbounds i32, i32* [[TMP65]], i32 4
-; CHECK-NEXT:    [[TMP72:%.*]] = bitcast i32* [[TMP71]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i32>, <4 x i32>* [[TMP72]], align 4
-; CHECK-NEXT:    [[TMP73:%.*]] = getelementptr inbounds i32, i32* [[TMP65]], i32 8
-; CHECK-NEXT:    [[TMP74:%.*]] = bitcast i32* [[TMP73]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i32>, <4 x i32>* [[TMP74]], align 4
-; CHECK-NEXT:    [[TMP75:%.*]] = getelementptr inbounds i32, i32* [[TMP65]], i32 12
-; CHECK-NEXT:    [[TMP76:%.*]] = bitcast i32* [[TMP75]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i32>, <4 x i32>* [[TMP76]], align 4
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP40]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP48]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP56]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP80:%.*]] = xor <4 x i1> [[TMP64]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP40]], <4 x i32> [[WIDE_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP48]], <4 x i32> [[WIDE_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP56]], <4 x i32> [[WIDE_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP64]], <4 x i32> [[WIDE_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP84]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
+; CHECK-NEXT:    [[TMP9:%.*]] = bitcast i1* [[TMP5]] to i8*
+; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i8, i8* [[TMP9]], i32 0
+; CHECK-NEXT:    [[TMP11:%.*]] = bitcast i8* [[TMP10]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP11]], align 1
+; CHECK-NEXT:    [[TMP12:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP13:%.*]] = bitcast i1* [[TMP5]] to i8*
+; CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, i8* [[TMP13]], i32 4
+; CHECK-NEXT:    [[TMP15:%.*]] = bitcast i8* [[TMP14]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP15]], align 1
+; CHECK-NEXT:    [[TMP16:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP17:%.*]] = bitcast i1* [[TMP5]] to i8*
+; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i8, i8* [[TMP17]], i32 8
+; CHECK-NEXT:    [[TMP19:%.*]] = bitcast i8* [[TMP18]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP19]], align 1
+; CHECK-NEXT:    [[TMP20:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP21:%.*]] = bitcast i1* [[TMP5]] to i8*
+; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i8, i8* [[TMP21]], i32 12
+; CHECK-NEXT:    [[TMP23:%.*]] = bitcast i8* [[TMP22]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP23]], align 1
+; CHECK-NEXT:    [[TMP24:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
+; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i32, i32* [[TMP25]], i32 0
+; CHECK-NEXT:    [[TMP30:%.*]] = bitcast i32* [[TMP29]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD7:%.*]] = load <4 x i32>, <4 x i32>* [[TMP30]], align 4
+; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i32, i32* [[TMP25]], i32 4
+; CHECK-NEXT:    [[TMP32:%.*]] = bitcast i32* [[TMP31]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD8:%.*]] = load <4 x i32>, <4 x i32>* [[TMP32]], align 4
+; CHECK-NEXT:    [[TMP33:%.*]] = getelementptr inbounds i32, i32* [[TMP25]], i32 8
+; CHECK-NEXT:    [[TMP34:%.*]] = bitcast i32* [[TMP33]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD9:%.*]] = load <4 x i32>, <4 x i32>* [[TMP34]], align 4
+; CHECK-NEXT:    [[TMP35:%.*]] = getelementptr inbounds i32, i32* [[TMP25]], i32 12
+; CHECK-NEXT:    [[TMP36:%.*]] = bitcast i32* [[TMP35]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD10:%.*]] = load <4 x i32>, <4 x i32>* [[TMP36]], align 4
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP12]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP16]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP20]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP40:%.*]] = xor <4 x i1> [[TMP24]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP12]], <4 x i32> [[WIDE_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP16]], <4 x i32> [[WIDE_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP20]], <4 x i32> [[WIDE_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI13:%.*]] = select <4 x i1> [[TMP24]], <4 x i32> [[WIDE_LOAD10]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI12]]
+; CHECK-NEXT:    [[TMP44]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI13]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP85:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT:    br i1 [[TMP85]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP22:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP45:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT:    br i1 [[TMP45]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP22:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP82]], [[TMP81]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP84]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP86:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP42]], [[TMP41]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX15:%.*]] = add <4 x i32> [[TMP44]], [[BIN_RDX14]]
+; CHECK-NEXT:    [[TMP46:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX15]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP86]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP46]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], [[MIN]]
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP23:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP86]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP46]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %alloca = alloca [4096 x i32]
   %base = bitcast [4096 x i32]* %alloca to i32*
   call void @init(i32* %base)
   %cmp = icmp ugt i64 %len, 4094
   %min = select i1 %cmp, i64 4094, i64 %len

 ; CHECK-NEXT:    [[ALLOCATION:%.*]] = call nonnull i8* @my_alloc(i32 16384)
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast i8* [[ALLOCATION]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP80:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP40:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = bitcast i32* [[TMP68]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, <4 x i32>* [[TMP69]], align 4
-; CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 4
-; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[TMP70]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i32>, <4 x i32>* [[TMP71]], align 4
-; CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 8
-; CHECK-NEXT:    [[TMP73:%.*]] = bitcast i32* [[TMP72]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i32>, <4 x i32>* [[TMP73]], align 4
-; CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 12
-; CHECK-NEXT:    [[TMP75:%.*]] = bitcast i32* [[TMP74]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i32>, <4 x i32>* [[TMP75]], align 4
-; CHECK-NEXT:    [[TMP76:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[WIDE_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[WIDE_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[WIDE_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[WIDE_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = bitcast i32* [[TMP28]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD7:%.*]] = load <4 x i32>, <4 x i32>* [[TMP29]], align 4
+; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 4
+; CHECK-NEXT:    [[TMP31:%.*]] = bitcast i32* [[TMP30]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD8:%.*]] = load <4 x i32>, <4 x i32>* [[TMP31]], align 4
+; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 8
+; CHECK-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD9:%.*]] = load <4 x i32>, <4 x i32>* [[TMP33]], align 4
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 12
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i32* [[TMP34]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD10:%.*]] = load <4 x i32>, <4 x i32>* [[TMP35]], align 4
+; CHECK-NEXT:    [[TMP36:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[WIDE_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[WIDE_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[WIDE_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI13:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[WIDE_LOAD10]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP40]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI12]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI13]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP84:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP84]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP24:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP44:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP44]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP24:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP81]], [[TMP80]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP82]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP85:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP41]], [[TMP40]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP42]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX15:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX14]]
+; CHECK-NEXT:    [[TMP45:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX15]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP25:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %allocation = call nonnull i8* @my_alloc(i32 16384)
   %base = bitcast i8* %allocation to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:    [[ALLOCATION:%.*]] = call nonnull i8* @my_array_alloc(i32 4096, i32 4)
 ; CHECK-NEXT:    [[BASE:%.*]] = bitcast i8* [[ALLOCATION]] to i32*
 ; CHECK-NEXT:    call void @init(i32* [[BASE]])
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP80:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP40:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = bitcast i32* [[TMP68]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, <4 x i32>* [[TMP69]], align 4
-; CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 4
-; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[TMP70]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i32>, <4 x i32>* [[TMP71]], align 4
-; CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 8
-; CHECK-NEXT:    [[TMP73:%.*]] = bitcast i32* [[TMP72]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i32>, <4 x i32>* [[TMP73]], align 4
-; CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 12
-; CHECK-NEXT:    [[TMP75:%.*]] = bitcast i32* [[TMP74]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i32>, <4 x i32>* [[TMP75]], align 4
-; CHECK-NEXT:    [[TMP76:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[WIDE_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[WIDE_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[WIDE_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[WIDE_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = bitcast i32* [[TMP28]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD7:%.*]] = load <4 x i32>, <4 x i32>* [[TMP29]], align 4
+; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 4
+; CHECK-NEXT:    [[TMP31:%.*]] = bitcast i32* [[TMP30]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD8:%.*]] = load <4 x i32>, <4 x i32>* [[TMP31]], align 4
+; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 8
+; CHECK-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD9:%.*]] = load <4 x i32>, <4 x i32>* [[TMP33]], align 4
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 12
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i32* [[TMP34]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_LOAD10:%.*]] = load <4 x i32>, <4 x i32>* [[TMP35]], align 4
+; CHECK-NEXT:    [[TMP36:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[WIDE_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[WIDE_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[WIDE_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI13:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[WIDE_LOAD10]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP40]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI12]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI13]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP84:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP84]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP26:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP44:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP44]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP26:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP81]], [[TMP80]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP82]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP85:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP41]], [[TMP40]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP42]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX15:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX14]]
+; CHECK-NEXT:    [[TMP45:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX15]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP27:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %allocation = call nonnull i8* @my_array_alloc(i32 4096, i32 4)
   %base = bitcast i8* %allocation to i32*
   call void @init(i32* %base)
   br label %loop
 loop:

 ; CHECK-NEXT:    call void @my_free(i8* [[ALLOCATION]])
 ; CHECK-NEXT:    br label [[PREHEADER]]
 ; CHECK:       preheader:
 ; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP80:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP81:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP82:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP83:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP40:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP41:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP42:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP43:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
-; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
-; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 5
-; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 6
-; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 7
-; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 9
-; CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 10
-; CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 11
-; CHECK-NEXT:    [[TMP12:%.*]] = add i64 [[INDEX]], 12
-; CHECK-NEXT:    [[TMP13:%.*]] = add i64 [[INDEX]], 13
-; CHECK-NEXT:    [[TMP14:%.*]] = add i64 [[INDEX]], 14
-; CHECK-NEXT:    [[TMP15:%.*]] = add i64 [[INDEX]], 15
-; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
-; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
-; CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP7]]
-; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP10]]
-; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP11]]
-; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP13]]
-; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP14]]
-; CHECK-NEXT:    [[TMP31:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP15]]
-; CHECK-NEXT:    [[TMP32:%.*]] = load i1, i1* [[TMP16]], align 1
-; CHECK-NEXT:    [[TMP33:%.*]] = load i1, i1* [[TMP17]], align 1
-; CHECK-NEXT:    [[TMP34:%.*]] = load i1, i1* [[TMP18]], align 1
-; CHECK-NEXT:    [[TMP35:%.*]] = load i1, i1* [[TMP19]], align 1
-; CHECK-NEXT:    [[TMP36:%.*]] = insertelement <4 x i1> poison, i1 [[TMP32]], i32 0
-; CHECK-NEXT:    [[TMP37:%.*]] = insertelement <4 x i1> [[TMP36]], i1 [[TMP33]], i32 1
-; CHECK-NEXT:    [[TMP38:%.*]] = insertelement <4 x i1> [[TMP37]], i1 [[TMP34]], i32 2
-; CHECK-NEXT:    [[TMP39:%.*]] = insertelement <4 x i1> [[TMP38]], i1 [[TMP35]], i32 3
-; CHECK-NEXT:    [[TMP40:%.*]] = load i1, i1* [[TMP20]], align 1
-; CHECK-NEXT:    [[TMP41:%.*]] = load i1, i1* [[TMP21]], align 1
-; CHECK-NEXT:    [[TMP42:%.*]] = load i1, i1* [[TMP22]], align 1
-; CHECK-NEXT:    [[TMP43:%.*]] = load i1, i1* [[TMP23]], align 1
-; CHECK-NEXT:    [[TMP44:%.*]] = insertelement <4 x i1> poison, i1 [[TMP40]], i32 0
-; CHECK-NEXT:    [[TMP45:%.*]] = insertelement <4 x i1> [[TMP44]], i1 [[TMP41]], i32 1
-; CHECK-NEXT:    [[TMP46:%.*]] = insertelement <4 x i1> [[TMP45]], i1 [[TMP42]], i32 2
-; CHECK-NEXT:    [[TMP47:%.*]] = insertelement <4 x i1> [[TMP46]], i1 [[TMP43]], i32 3
-; CHECK-NEXT:    [[TMP48:%.*]] = load i1, i1* [[TMP24]], align 1
-; CHECK-NEXT:    [[TMP49:%.*]] = load i1, i1* [[TMP25]], align 1
-; CHECK-NEXT:    [[TMP50:%.*]] = load i1, i1* [[TMP26]], align 1
-; CHECK-NEXT:    [[TMP51:%.*]] = load i1, i1* [[TMP27]], align 1
-; CHECK-NEXT:    [[TMP52:%.*]] = insertelement <4 x i1> poison, i1 [[TMP48]], i32 0
-; CHECK-NEXT:    [[TMP53:%.*]] = insertelement <4 x i1> [[TMP52]], i1 [[TMP49]], i32 1
-; CHECK-NEXT:    [[TMP54:%.*]] = insertelement <4 x i1> [[TMP53]], i1 [[TMP50]], i32 2
-; CHECK-NEXT:    [[TMP55:%.*]] = insertelement <4 x i1> [[TMP54]], i1 [[TMP51]], i32 3
-; CHECK-NEXT:    [[TMP56:%.*]] = load i1, i1* [[TMP28]], align 1
-; CHECK-NEXT:    [[TMP57:%.*]] = load i1, i1* [[TMP29]], align 1
-; CHECK-NEXT:    [[TMP58:%.*]] = load i1, i1* [[TMP30]], align 1
-; CHECK-NEXT:    [[TMP59:%.*]] = load i1, i1* [[TMP31]], align 1
-; CHECK-NEXT:    [[TMP60:%.*]] = insertelement <4 x i1> poison, i1 [[TMP56]], i32 0
-; CHECK-NEXT:    [[TMP61:%.*]] = insertelement <4 x i1> [[TMP60]], i1 [[TMP57]], i32 1
-; CHECK-NEXT:    [[TMP62:%.*]] = insertelement <4 x i1> [[TMP61]], i1 [[TMP58]], i32 2
-; CHECK-NEXT:    [[TMP63:%.*]] = insertelement <4 x i1> [[TMP62]], i1 [[TMP59]], i32 3
-; CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
-; CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP4]]
-; CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP8]]
-; CHECK-NEXT:    [[TMP67:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP12]]
-; CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 0
-; CHECK-NEXT:    [[TMP69:%.*]] = bitcast i32* [[TMP68]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP69]], i32 4, <4 x i1> [[TMP39]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 4
-; CHECK-NEXT:    [[TMP71:%.*]] = bitcast i32* [[TMP70]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD4:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP71]], i32 4, <4 x i1> [[TMP47]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 8
-; CHECK-NEXT:    [[TMP73:%.*]] = bitcast i32* [[TMP72]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD5:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP73]], i32 4, <4 x i1> [[TMP55]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds i32, i32* [[TMP64]], i32 12
-; CHECK-NEXT:    [[TMP75:%.*]] = bitcast i32* [[TMP74]] to <4 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD6:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP75]], i32 4, <4 x i1> [[TMP63]], <4 x i32> poison)
-; CHECK-NEXT:    [[TMP76:%.*]] = xor <4 x i1> [[TMP39]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP77:%.*]] = xor <4 x i1> [[TMP47]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP78:%.*]] = xor <4 x i1> [[TMP55]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[TMP79:%.*]] = xor <4 x i1> [[TMP63]], <i1 true, i1 true, i1 true, i1 true>
-; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP39]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI7:%.*]] = select <4 x i1> [[TMP47]], <4 x i32> [[WIDE_MASKED_LOAD4]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI8:%.*]] = select <4 x i1> [[TMP55]], <4 x i32> [[WIDE_MASKED_LOAD5]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[PREDPHI9:%.*]] = select <4 x i1> [[TMP63]], <4 x i32> [[WIDE_MASKED_LOAD6]], <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
-; CHECK-NEXT:    [[TMP81]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI7]]
-; CHECK-NEXT:    [[TMP82]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI8]]
-; CHECK-NEXT:    [[TMP83]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI9]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE:%.*]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i8, i8* [[TMP8]], i32 0
+; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i8* [[TMP9]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP10]], align 1
+; CHECK-NEXT:    [[TMP11:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i1>
+; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, i8* [[TMP12]], i32 4
+; CHECK-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x i8>, <4 x i8>* [[TMP14]], align 1
+; CHECK-NEXT:    [[TMP15:%.*]] = trunc <4 x i8> [[WIDE_LOAD4]] to <4 x i1>
+; CHECK-NEXT:    [[TMP16:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, i8* [[TMP16]], i32 8
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast i8* [[TMP17]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP18]], align 1
+; CHECK-NEXT:    [[TMP19:%.*]] = trunc <4 x i8> [[WIDE_LOAD5]] to <4 x i1>
+; CHECK-NEXT:    [[TMP20:%.*]] = bitcast i1* [[TMP4]] to i8*
+; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, i8* [[TMP20]], i32 12
+; CHECK-NEXT:    [[TMP22:%.*]] = bitcast i8* [[TMP21]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x i8>, <4 x i8>* [[TMP22]], align 1
+; CHECK-NEXT:    [[TMP23:%.*]] = trunc <4 x i8> [[WIDE_LOAD6]] to <4 x i1>
+; CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP0]]
+; CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[TMP3]]
+; CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 0
+; CHECK-NEXT:    [[TMP29:%.*]] = bitcast i32* [[TMP28]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP29]], i32 4, <4 x i1> [[TMP11]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 4
+; CHECK-NEXT:    [[TMP31:%.*]] = bitcast i32* [[TMP30]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD7:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP31]], i32 4, <4 x i1> [[TMP15]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 8
+; CHECK-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD8:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP33]], i32 4, <4 x i1> [[TMP19]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds i32, i32* [[TMP24]], i32 12
+; CHECK-NEXT:    [[TMP35:%.*]] = bitcast i32* [[TMP34]] to <4 x i32>*
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD9:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP35]], i32 4, <4 x i1> [[TMP23]], <4 x i32> poison)
+; CHECK-NEXT:    [[TMP36:%.*]] = xor <4 x i1> [[TMP11]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP37:%.*]] = xor <4 x i1> [[TMP15]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP38:%.*]] = xor <4 x i1> [[TMP19]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[TMP39:%.*]] = xor <4 x i1> [[TMP23]], <i1 true, i1 true, i1 true, i1 true>
+; CHECK-NEXT:    [[PREDPHI:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[WIDE_MASKED_LOAD]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI10:%.*]] = select <4 x i1> [[TMP15]], <4 x i32> [[WIDE_MASKED_LOAD7]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI11:%.*]] = select <4 x i1> [[TMP19]], <4 x i32> [[WIDE_MASKED_LOAD8]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[PREDPHI12:%.*]] = select <4 x i1> [[TMP23]], <4 x i32> [[WIDE_MASKED_LOAD9]], <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP40]] = add <4 x i32> [[VEC_PHI]], [[PREDPHI]]
+; CHECK-NEXT:    [[TMP41]] = add <4 x i32> [[VEC_PHI1]], [[PREDPHI10]]
+; CHECK-NEXT:    [[TMP42]] = add <4 x i32> [[VEC_PHI2]], [[PREDPHI11]]
+; CHECK-NEXT:    [[TMP43]] = add <4 x i32> [[VEC_PHI3]], [[PREDPHI12]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 16
-; CHECK-NEXT:    [[TMP84:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; CHECK-NEXT:    br i1 [[TMP84]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP28:!llvm.loop !.*]]
+; CHECK-NEXT:    [[TMP44:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
+; CHECK-NEXT:    br i1 [[TMP44]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP28:!llvm.loop !.*]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP81]], [[TMP80]]
-; CHECK-NEXT:    [[BIN_RDX10:%.*]] = add <4 x i32> [[TMP82]], [[BIN_RDX]]
-; CHECK-NEXT:    [[BIN_RDX11:%.*]] = add <4 x i32> [[TMP83]], [[BIN_RDX10]]
-; CHECK-NEXT:    [[TMP85:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX11]])
+; CHECK-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP41]], [[TMP40]]
+; CHECK-NEXT:    [[BIN_RDX13:%.*]] = add <4 x i32> [[TMP42]], [[BIN_RDX]]
+; CHECK-NEXT:    [[BIN_RDX14:%.*]] = add <4 x i32> [[TMP43]], [[BIN_RDX13]]
+; CHECK-NEXT:    [[TMP45:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX14]])
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[LOOP_EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 4096, [[MIDDLE_BLOCK]] ], [ 0, [[PREHEADER]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[PREHEADER]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[PREHEADER]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LATCH:%.*]] ]
 ; CHECK-NEXT:    [[ACCUM:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[ACCUM_NEXT:%.*]], [[LATCH]] ]
 ; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
 ; CHECK-NEXT:    [[TEST_ADDR:%.*]] = getelementptr inbounds i1, i1* [[TEST_BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[EARLYCND:%.*]] = load i1, i1* [[TEST_ADDR]], align 1
 ; CHECK-NEXT:    br i1 [[EARLYCND]], label [[PRED:%.*]], label [[LATCH]]
 ; CHECK:       pred:
 ; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr inbounds i32, i32* [[BASE]], i64 [[IV]]
 ; CHECK-NEXT:    [[VAL:%.*]] = load i32, i32* [[ADDR]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[VAL_PHI:%.*]] = phi i32 [ 0, [[LOOP]] ], [ [[VAL]], [[PRED]] ]
 ; CHECK-NEXT:    [[ACCUM_NEXT]] = add i32 [[ACCUM]], [[VAL_PHI]]
 ; CHECK-NEXT:    [[EXIT:%.*]] = icmp ugt i64 [[IV]], 4094
 ; CHECK-NEXT:    br i1 [[EXIT]], label [[LOOP_EXIT]], label [[LOOP]], [[LOOP29:!llvm.loop !.*]]
 ; CHECK:       loop_exit:
-; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP85]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT:    [[ACCUM_NEXT_LCSSA:%.*]] = phi i32 [ [[ACCUM_NEXT]], [[LATCH]] ], [ [[TMP45]], [[MIDDLE_BLOCK]] ]
 ; CHECK-NEXT:    ret i32 [[ACCUM_NEXT_LCSSA]]
 ;
 entry:
   %allocation = call nonnull i8* @my_alloc(i32 16384)
   %base = bitcast i8* %allocation to i32*
   call void @init(i32* %base)
   br i1 %allzero, label %freeit, label %preheader
 freeit:

llvm/test/Transforms/LoopVectorize/irregular_type.ll

-; RUN: opt %s -loop-vectorize -force-vector-width=4 -S | FileCheck %s
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt %s -loop-vectorize -instcombine -force-vector-interleave=1 -enable-interleaved-mem-accesses=true -force-vector-width=4 -S | FileCheck %s
 
-; Ensure the array loads/stores are not optimized into vector operations when
-; the element type has padding bits.
+; Ensure the array loads/stores are vectorized even when the element types have
+; padding between them by widening the memory operation.
+; Also ensure the value is narrowed after being loaded and extended before being
+; stored.
 
-; CHECK: foo
-; CHECK: vector.body
-; CHECK-NOT: load <4 x i7>
-; CHECK-NOT: store <4 x i7>
-; CHECK: for.body
 define void @foo(i7* %a, i64 %n) {
+; CHECK-LABEL: @foo(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N:%.*]], 4
+; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
+; CHECK:       vector.ph:
+; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[N]], -4
+; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
+; CHECK:       vector.body:
+; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds i7, i7* [[A:%.*]], i64 [[INDEX]]
+; CHECK-NEXT:    [[TMP1:%.*]] = bitcast i7* [[TMP0]] to <4 x i8>*
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, <4 x i8>* [[TMP1]], align 1
+; CHECK-NEXT:    [[TMP2:%.*]] = trunc <4 x i8> [[WIDE_LOAD]] to <4 x i7>
+; CHECK-NEXT:    [[TMP3:%.*]] = add nsw <4 x i7> [[TMP2]], <i7 -1, i7 -1, i7 -1, i7 -1>
+; CHECK-NEXT:    [[TMP4:%.*]] = zext <4 x i7> [[TMP3]] to <4 x i8>
+; CHECK-NEXT:    [[TMP5:%.*]] = bitcast i7* [[TMP0]] to <4 x i8>*
+; CHECK-NEXT:    store <4 x i8> [[TMP4]], <4 x i8>* [[TMP5]], align 1
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK:       middle.block:
+; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[N_VEC]], [[N]]
+; CHECK-NEXT:    br i1 [[CMP_N]], label [[FOR_EXIT:%.*]], label [[SCALAR_PH]]
+; CHECK:       scalar.ph:
+; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
+; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
+; CHECK:       for.body:
+; CHECK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
+; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i7, i7* [[A]], i64 [[INDVARS_IV]]
+; CHECK-NEXT:    [[TMP7:%.*]] = load i7, i7* [[ARRAYIDX]], align 1
+; CHECK-NEXT:    [[SUB:%.*]] = add nsw i7 [[TMP7]], -1
+; CHECK-NEXT:    store i7 [[SUB]], i7* [[ARRAYIDX]], align 1
+; CHECK-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
+; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], [[N]]
+; CHECK-NEXT:    br i1 [[CMP]], label [[FOR_EXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP2:![0-9]+]]
+; CHECK:       for.exit:
+; CHECK-NEXT:    ret void
+;
 entry:
   br label %for.body
 
 for.body:
   %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
   %arrayidx = getelementptr inbounds i7, i7* %a, i64 %indvars.iv
   %0 = load i7, i7* %arrayidx, align 1
-  %sub = add nuw nsw i7 %0, 0
+  %sub = sub nuw nsw i7 %0, 1
   store i7 %sub, i7* %arrayidx, align 1
   %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
   %cmp = icmp eq i64 %indvars.iv.next, %n
   br i1 %cmp, label %for.exit, label %for.body
 
 for.exit:
   ret void
 }