[llvm] [LV] Support binary and unary operations with EVL-vectorization (PR #93854)

[Kolya Panchenko via llvm-commits]

https://github.com/nikolaypanchenko updated https://github.com/llvm/llvm-project/pull/93854

>From ed540f2e7460f17fc86a57e8318ade2e69a14e3d Mon Sep 17 00:00:00 2001
From: Kolya Panchenko <kolya.panchenko at sifive.com>
Date: Thu, 30 May 2024 09:56:26 -0700
Subject: [PATCH 1/2] [LV] Support binary and unary operations with
 EVL-vectorization

The patch adds `VPWidenEVLRecipe` which represents `VPWidenRecipe` + EVL
argument. The new recipe replaces `VPWidenRecipe` in
`tryAddExplicitVectorLength` for each binary and unary operations.
Follow up patches will extend support for remaining cases, like `FCmp`
and `ICmp`
---
 llvm/lib/Transforms/Vectorize/VPlan.h         |   54 +-
 .../lib/Transforms/Vectorize/VPlanRecipes.cpp |   71 +
 .../Transforms/Vectorize/VPlanTransforms.cpp  |   82 +-
 llvm/lib/Transforms/Vectorize/VPlanValue.h    |    1 +
 ...-force-tail-with-evl-bin-unary-ops-args.ll | 1799 +++++++++++++++++
 ...ze-force-tail-with-evl-masked-loadstore.ll |   20 +-
 .../RISCV/vectorize-vp-intrinsics.ll          |   22 +-
 .../RISCV/vplan-vp-intrinsics.ll              |    2 +-
 8 files changed, 2001 insertions(+), 50 deletions(-)
 create mode 100644 llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-bin-unary-ops-args.ll

diff --git a/llvm/lib/Transforms/Vectorize/VPlan.h b/llvm/lib/Transforms/Vectorize/VPlan.h
index d6b63a5d43c464..4b22e4983efac8 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.h
+++ b/llvm/lib/Transforms/Vectorize/VPlan.h
@@ -855,6 +855,7 @@ class VPSingleDefRecipe : public VPRecipeBase, public VPValue {
     case VPRecipeBase::VPWidenCastSC:
     case VPRecipeBase::VPWidenGEPSC:
     case VPRecipeBase::VPWidenSC:
+    case VPRecipeBase::VPWidenEVLSC:
     case VPRecipeBase::VPWidenSelectSC:
     case VPRecipeBase::VPBlendSC:
     case VPRecipeBase::VPPredInstPHISC:
@@ -1039,6 +1040,7 @@ class VPRecipeWithIRFlags : public VPSingleDefRecipe {
   static inline bool classof(const VPRecipeBase *R) {
     return R->getVPDefID() == VPRecipeBase::VPInstructionSC ||
            R->getVPDefID() == VPRecipeBase::VPWidenSC ||
+           R->getVPDefID() == VPRecipeBase::VPWidenEVLSC ||
            R->getVPDefID() == VPRecipeBase::VPWidenGEPSC ||
            R->getVPDefID() == VPRecipeBase::VPWidenCastSC ||
            R->getVPDefID() == VPRecipeBase::VPReplicateSC ||
@@ -1333,13 +1335,18 @@ class VPInstruction : public VPRecipeWithIRFlags {
 /// ingredient. This recipe covers most of the traditional vectorization cases
 /// where each ingredient transforms into a vectorized version of itself.
 class VPWidenRecipe : public VPRecipeWithIRFlags {
+protected:
   unsigned Opcode;
 
+  template <typename IterT>
+  VPWidenRecipe(unsigned VPDefOpcode, Instruction &I,
+                iterator_range<IterT> Operands)
+      : VPRecipeWithIRFlags(VPDefOpcode, Operands, I), Opcode(I.getOpcode()) {}
+
 public:
   template <typename IterT>
   VPWidenRecipe(Instruction &I, iterator_range<IterT> Operands)
-      : VPRecipeWithIRFlags(VPDef::VPWidenSC, Operands, I),
-        Opcode(I.getOpcode()) {}
+      : VPWidenRecipe(VPDef::VPWidenSC, I, Operands) {}
 
   ~VPWidenRecipe() override = default;
 
@@ -1363,6 +1370,49 @@ class VPWidenRecipe : public VPRecipeWithIRFlags {
 #endif
 };
 
+class VPWidenEVLRecipe : public VPWidenRecipe {
+private:
+  using VPRecipeWithIRFlags::transferFlags;
+
+public:
+  template <typename IterT>
+  VPWidenEVLRecipe(Instruction &I, iterator_range<IterT> Operands, VPValue &EVL)
+      : VPWidenRecipe(VPDef::VPWidenEVLSC, I, Operands) {
+    addOperand(&EVL);
+  }
+
+  ~VPWidenEVLRecipe() override = default;
+
+  VPWidenRecipe *clone() override final {
+    SmallVector<VPValue *> Ops(operands());
+    VPValue *EVL = Ops.pop_back_val();
+    auto *R = new VPWidenEVLRecipe(*getUnderlyingInstr(),
+                                   make_range(Ops.begin(), Ops.end()), *EVL);
+    R->transferFlags(*this);
+    return R;
+  }
+
+  VP_CLASSOF_IMPL(VPDef::VPWidenEVLSC);
+
+  VPValue *getEVL() { return getOperand(getNumOperands() - 1); }
+  const VPValue *getEVL() const { return getOperand(getNumOperands() - 1); }
+
+  /// A helper function to create widen EVL recipe from regular widen recipe.
+  static VPWidenEVLRecipe *create(VPWidenRecipe *W, VPValue &EVL);
+
+  /// Produce widened copies of all Ingredients.
+  void execute(VPTransformState &State) override final;
+
+  /// Returns true if the recipe only uses the first lane of operand \p Op.
+  bool onlyFirstLaneUsed(const VPValue *Op) const override;
+
+#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
+  /// Print the recipe.
+  void print(raw_ostream &O, const Twine &Indent,
+             VPSlotTracker &SlotTracker) const override final;
+#endif
+};
+
 /// VPWidenCastRecipe is a recipe to create vector cast instructions.
 class VPWidenCastRecipe : public VPRecipeWithIRFlags {
   /// Cast instruction opcode.
diff --git a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
index 5eb99ffd1e10e3..a6fd26b6665013 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
@@ -23,6 +23,7 @@
 #include "llvm/IR/Instructions.h"
 #include "llvm/IR/Type.h"
 #include "llvm/IR/Value.h"
+#include "llvm/IR/VectorBuilder.h"
 #include "llvm/Support/Casting.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Debug.h"
@@ -71,6 +72,7 @@ bool VPRecipeBase::mayWriteToMemory() const {
   case VPWidenLoadSC:
   case VPWidenPHISC:
   case VPWidenSC:
+  case VPWidenEVLSC:
   case VPWidenSelectSC: {
     const Instruction *I =
         dyn_cast_or_null<Instruction>(getVPSingleValue()->getUnderlyingValue());
@@ -110,6 +112,7 @@ bool VPRecipeBase::mayReadFromMemory() const {
   case VPWidenIntOrFpInductionSC:
   case VPWidenPHISC:
   case VPWidenSC:
+  case VPWidenEVLSC:
   case VPWidenSelectSC: {
     const Instruction *I =
         dyn_cast_or_null<Instruction>(getVPSingleValue()->getUnderlyingValue());
@@ -157,6 +160,7 @@ bool VPRecipeBase::mayHaveSideEffects() const {
   case VPWidenPHISC:
   case VPWidenPointerInductionSC:
   case VPWidenSC:
+  case VPWidenEVLSC:
   case VPWidenSelectSC: {
     const Instruction *I =
         dyn_cast_or_null<Instruction>(getVPSingleValue()->getUnderlyingValue());
@@ -993,6 +997,64 @@ void VPWidenRecipe::execute(VPTransformState &State) {
 #endif
 }
 
+VPWidenEVLRecipe *VPWidenEVLRecipe::create(VPWidenRecipe *W, VPValue &EVL) {
+  auto *R = new VPWidenEVLRecipe(*W->getUnderlyingInstr(), W->operands(), EVL);
+  R->transferFlags(*W);
+  return R;
+}
+
+void VPWidenEVLRecipe::execute(VPTransformState &State) {
+  assert(State.UF == 1 && "Expected only UF == 1 when vectorizing with "
+                          "explicit vector length.");
+  VPValue *Op0 = getOperand(0);
+
+  // If it's scalar operation, hand translation over to VPWidenRecipe
+  if (!State.get(Op0, 0)->getType()->isVectorTy())
+    return VPWidenRecipe::execute(State);
+
+  VPValue *EVL = getEVL();
+  Value *EVLArg = State.get(EVL, 0, /*NeedsScalar=*/true);
+  unsigned Opcode = getOpcode();
+  Instruction *I = getUnderlyingInstr();
+  IRBuilderBase &BuilderIR = State.Builder;
+  VectorBuilder Builder(BuilderIR);
+  Value *Mask = BuilderIR.CreateVectorSplat(State.VF, BuilderIR.getTrue());
+  Value *VPInst = nullptr;
+
+  //===------------------- Binary and Unary Ops ---------------------===//
+  if (Instruction::isBinaryOp(Opcode) || Instruction::isUnaryOp(Opcode)) {
+    // Just widen unops and binops.
+
+    SmallVector<Value *, 4> Ops;
+    for (unsigned I = 0, E = getNumOperands() - 1; I < E; ++I) {
+      VPValue *VPOp = getOperand(I);
+      Ops.push_back(State.get(VPOp, 0));
+    }
+
+    Builder.setMask(Mask).setEVL(EVLArg);
+    VPInst = Builder.createVectorInstruction(Opcode, Ops[0]->getType(), Ops,
+                                             "vp.op");
+
+    if (I)
+      if (auto *VecOp = dyn_cast<Instruction>(VPInst))
+        VecOp->copyIRFlags(I);
+  } else {
+    llvm_unreachable("Unsupported opcode in VPWidenEVLRecipe::execute");
+  }
+  State.set(this, VPInst, 0);
+  State.addMetadata(VPInst, I);
+}
+
+bool VPWidenEVLRecipe::onlyFirstLaneUsed(const VPValue *Op) const {
+  assert(is_contained(operands(), Op) && "Op must be an operand of the recipe");
+  // EVL in that recipe is always the last operand, thus any use before means
+  // the VPValue should be vectorized.
+  for (unsigned I = 0, E = getNumOperands() - 1; I != E; ++I)
+    if (getOperand(I) == Op)
+      return false;
+  return true;
+}
+
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
 void VPWidenRecipe::print(raw_ostream &O, const Twine &Indent,
                           VPSlotTracker &SlotTracker) const {
@@ -1002,6 +1064,15 @@ void VPWidenRecipe::print(raw_ostream &O, const Twine &Indent,
   printFlags(O);
   printOperands(O, SlotTracker);
 }
+
+void VPWidenEVLRecipe::print(raw_ostream &O, const Twine &Indent,
+                             VPSlotTracker &SlotTracker) const {
+  O << Indent << "WIDEN vp ";
+  printAsOperand(O, SlotTracker);
+  O << " = " << Instruction::getOpcodeName(Opcode);
+  printFlags(O);
+  printOperands(O, SlotTracker);
+}
 #endif
 
 void VPWidenCastRecipe::execute(VPTransformState &State) {
diff --git a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
index 422579ea8b84f1..aada846473fe9d 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
@@ -20,6 +20,7 @@
 #include "llvm/ADT/PostOrderIterator.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SetVector.h"
+#include "llvm/ADT/TypeSwitch.h"
 #include "llvm/Analysis/IVDescriptors.h"
 #include "llvm/Analysis/VectorUtils.h"
 #include "llvm/IR/Intrinsics.h"
@@ -1219,7 +1220,7 @@ static VPActiveLaneMaskPHIRecipe *addVPLaneMaskPhiAndUpdateExitBranch(
 /// WideCanonicalIV, backedge-taken-count) pattern.
 /// TODO: Introduce explicit recipe for header-mask instead of searching
 /// for the header-mask pattern manually.
-static SmallVector<VPValue *> collectAllHeaderMasks(VPlan &Plan) {
+static DenseSet<VPValue *> collectAllHeaderMasks(VPlan &Plan) {
   SmallVector<VPValue *> WideCanonicalIVs;
   auto *FoundWidenCanonicalIVUser =
       find_if(Plan.getCanonicalIV()->users(),
@@ -1245,7 +1246,7 @@ static SmallVector<VPValue *> collectAllHeaderMasks(VPlan &Plan) {
 
   // Walk users of wide canonical IVs and collect to all compares of the form
   // (ICMP_ULE, WideCanonicalIV, backedge-taken-count).
-  SmallVector<VPValue *> HeaderMasks;
+  DenseSet<VPValue *> HeaderMasks;
   VPValue *BTC = Plan.getOrCreateBackedgeTakenCount();
   for (auto *Wide : WideCanonicalIVs) {
     for (VPUser *U : SmallVector<VPUser *>(Wide->users())) {
@@ -1257,7 +1258,7 @@ static SmallVector<VPValue *> collectAllHeaderMasks(VPlan &Plan) {
 
       assert(HeaderMask->getOperand(0) == Wide &&
              "WidenCanonicalIV must be the first operand of the compare");
-      HeaderMasks.push_back(HeaderMask);
+      HeaderMasks.insert(HeaderMask);
     }
   }
   return HeaderMasks;
@@ -1296,6 +1297,56 @@ void VPlanTransforms::addActiveLaneMask(
     HeaderMask->replaceAllUsesWith(LaneMask);
 }
 
+/// Replace recipes with their EVL variants.
+static void transformRecipestoEVLRecipes(VPlan &Plan, VPValue &EVL) {
+  DenseSet<VPRecipeBase *> ToRemove;
+
+  ReversePostOrderTraversal<VPBlockDeepTraversalWrapper<VPBlockBase *>> RPOT(
+      Plan.getEntry());
+  DenseSet<VPValue *> HeaderMasks = collectAllHeaderMasks(Plan);
+  for (VPBasicBlock *VPBB :
+       reverse(VPBlockUtils::blocksOnly<VPBasicBlock>(RPOT))) {
+    // The recipes in the block are processed in reverse order, to catch chains
+    // of dead recipes.
+    for (VPRecipeBase &R : make_early_inc_range(reverse(*VPBB))) {
+      TypeSwitch<VPRecipeBase *>(&R)
+          .Case<VPWidenLoadRecipe>([&](VPWidenLoadRecipe *L) {
+            VPValue *NewMask =
+                HeaderMasks.contains(L->getMask()) ? nullptr : L->getMask();
+            auto *N = new VPWidenLoadEVLRecipe(L, &EVL, NewMask);
+            N->insertBefore(L);
+            L->replaceAllUsesWith(N);
+            ToRemove.insert(L);
+          })
+          .Case<VPWidenStoreRecipe>([&](VPWidenStoreRecipe *S) {
+            VPValue *NewMask =
+                HeaderMasks.contains(S->getMask()) ? nullptr : S->getMask();
+            auto *N = new VPWidenStoreEVLRecipe(S, &EVL, NewMask);
+            N->insertBefore(S);
+            ToRemove.insert(S);
+          })
+          .Case<VPWidenRecipe>([&](VPWidenRecipe *W) {
+            unsigned Opcode = W->getOpcode();
+            if (!Instruction::isBinaryOp(Opcode) &&
+                !Instruction::isUnaryOp(Opcode))
+              return;
+            auto *N = VPWidenEVLRecipe::create(W, EVL);
+            N->insertBefore(W);
+            W->replaceAllUsesWith(N);
+            ToRemove.insert(W);
+          });
+    }
+  }
+
+  for (VPRecipeBase *R : ToRemove)
+    R->eraseFromParent();
+
+  for (VPValue *HeaderMask : HeaderMasks)
+    recursivelyDeleteDeadRecipes(HeaderMask);
+}
+
+
+
 /// Add a VPEVLBasedIVPHIRecipe and related recipes to \p Plan and
 /// replaces all uses except the canonical IV increment of
 /// VPCanonicalIVPHIRecipe with a VPEVLBasedIVPHIRecipe. VPCanonicalIVPHIRecipe
@@ -1356,29 +1407,8 @@ bool VPlanTransforms::tryAddExplicitVectorLength(VPlan &Plan) {
   NextEVLIV->insertBefore(CanonicalIVIncrement);
   EVLPhi->addOperand(NextEVLIV);
 
-  for (VPValue *HeaderMask : collectAllHeaderMasks(Plan)) {
-    for (VPUser *U : collectUsersRecursively(HeaderMask)) {
-      auto *MemR = dyn_cast<VPWidenMemoryRecipe>(U);
-      if (!MemR)
-        continue;
-      VPValue *OrigMask = MemR->getMask();
-      assert(OrigMask && "Unmasked widen memory recipe when folding tail");
-      VPValue *NewMask = HeaderMask == OrigMask ? nullptr : OrigMask;
-      if (auto *L = dyn_cast<VPWidenLoadRecipe>(MemR)) {
-        auto *N = new VPWidenLoadEVLRecipe(L, VPEVL, NewMask);
-        N->insertBefore(L);
-        L->replaceAllUsesWith(N);
-        L->eraseFromParent();
-      } else if (auto *S = dyn_cast<VPWidenStoreRecipe>(MemR)) {
-        auto *N = new VPWidenStoreEVLRecipe(S, VPEVL, NewMask);
-        N->insertBefore(S);
-        S->eraseFromParent();
-      } else {
-        llvm_unreachable("unsupported recipe");
-      }
-    }
-    recursivelyDeleteDeadRecipes(HeaderMask);
-  }
+  transformRecipestoEVLRecipes(Plan, *VPEVL);
+
   // Replace all uses of VPCanonicalIVPHIRecipe by
   // VPEVLBasedIVPHIRecipe except for the canonical IV increment.
   CanonicalIVPHI->replaceAllUsesWith(EVLPhi);
diff --git a/llvm/lib/Transforms/Vectorize/VPlanValue.h b/llvm/lib/Transforms/Vectorize/VPlanValue.h
index 8d945f6f2b8ea8..a90d42289e5f1d 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanValue.h
+++ b/llvm/lib/Transforms/Vectorize/VPlanValue.h
@@ -356,6 +356,7 @@ class VPDef {
     VPWidenStoreEVLSC,
     VPWidenStoreSC,
     VPWidenSC,
+    VPWidenEVLSC,
     VPWidenSelectSC,
     VPBlendSC,
     // START: Phi-like recipes. Need to be kept together.
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-bin-unary-ops-args.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-bin-unary-ops-args.ll
new file mode 100644
index 00000000000000..77689fc8a76ee6
--- /dev/null
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-bin-unary-ops-args.ll
@@ -0,0 +1,1799 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -passes=loop-vectorize \
+; RUN: -force-tail-folding-style=data-with-evl \
+; RUN: -prefer-predicate-over-epilogue=predicate-dont-vectorize \
+; RUN: -mtriple=riscv64 -mattr=+v -S < %s | FileCheck %s --check-prefix=IF-EVL
+
+; RUN: opt -passes=loop-vectorize \
+; RUN: -force-tail-folding-style=none \
+; RUN: -prefer-predicate-over-epilogue=predicate-dont-vectorize \
+; RUN: -mtriple=riscv64 -mattr=+v -S < %s | FileCheck %s --check-prefix=NO-VP
+
+
+define void @test_and(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_and(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0:[0-9]+]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.and.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 1, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = and i8 [[TMP20]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP3:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_and(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0:[0-9]+]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = and i8 [[TMP0]], 1
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = and i8 %0, 1
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_or(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_or(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.or.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 1, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = or i8 [[TMP20]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP5:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_or(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = or i8 [[TMP0]], 1
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = or i8 %0, 1
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_xor(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_xor(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.xor.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 1, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = xor i8 [[TMP20]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP7:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_xor(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = xor i8 [[TMP0]], 1
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = xor i8 %0, 1
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_shl(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_shl(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.shl.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 1, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = shl i8 [[TMP20]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP9:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_shl(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = shl i8 [[TMP0]], 1
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = shl i8 %0, 1
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_lshr(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_lshr(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.lshr.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 1, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = lshr i8 [[TMP20]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP11:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_lshr(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = lshr i8 [[TMP0]], 1
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = lshr i8 %0, 1
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_ashr(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_ashr(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.ashr.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 1, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = ashr i8 [[TMP20]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP13:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_ashr(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = ashr i8 [[TMP0]], 1
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = ashr i8 %0, 1
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_add(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_add(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.add.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 1, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP14:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = add i8 [[TMP20]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP15:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_add(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = add i8 [[TMP0]], 1
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = add i8 %0, 1
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_sub(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_sub(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.sub.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 1, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP16:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = sub i8 [[TMP20]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP17:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_sub(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = sub i8 [[TMP0]], 1
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = sub i8 %0, 1
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_mul(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_mul(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.mul.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 3, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP18:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = mul i8 [[TMP20]], 3
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP19:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_mul(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = mul i8 [[TMP0]], 3
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = mul i8 %0, 3
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_sdiv(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_sdiv(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.sdiv.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 3, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP20:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = sdiv i8 [[TMP20]], 3
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP21:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_sdiv(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = sdiv i8 [[TMP0]], 3
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = sdiv i8 %0, 3
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_udiv(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_udiv(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.udiv.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 3, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP22:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = udiv i8 [[TMP20]], 3
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP23:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_udiv(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = udiv i8 [[TMP0]], 3
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = udiv i8 %0, 3
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_srem(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_srem(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.srem.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 3, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP24:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = srem i8 [[TMP20]], 3
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP25:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_srem(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = srem i8 [[TMP0]], 3
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = srem i8 %0, 3
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_urem(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_urem(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 16
+; IF-EVL-NEXT:    [[TMP2:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP2]], [[TMP1]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP4:%.*]] = mul i64 [[TMP3]], 16
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 [[TMP6]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP7]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP4]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP9:%.*]] = mul i64 [[TMP8]], 16
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP10:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP10]], i32 16, i1 true)
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[TMP13]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr align 1 [[TMP14]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP15:%.*]] = call <vscale x 16 x i8> @llvm.vp.urem.nxv16i8(<vscale x 16 x i8> [[VP_OP_LOAD]], <vscale x 16 x i8> shufflevector (<vscale x 16 x i8> insertelement (<vscale x 16 x i8> poison, i8 3, i64 0), <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer), <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[TMP12]]
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> [[TMP15]], ptr align 1 [[TMP17]], <vscale x 16 x i1> shufflevector (<vscale x 16 x i1> insertelement (<vscale x 16 x i1> poison, i1 true, i64 0), <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer), i32 [[TMP11]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP11]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP18]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP9]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP19]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP26:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; IF-EVL-NEXT:    [[TMP:%.*]] = urem i8 [[TMP20]], 3
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP27:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_urem(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
+; NO-VP-NEXT:    [[TMP:%.*]] = urem i8 [[TMP0]], 3
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store i8 [[TMP]], ptr [[ARRAYIDX1]], align 1
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds i8, ptr %a, i64 %len
+  %0 = load i8, ptr %arrayidx, align 1
+  %tmp = urem i8 %0, 3
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i64 %len
+  store i8 %tmp, ptr %arrayidx1, align 1
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+; Floating point tests
+
+define void @test_fadd(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_fadd(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 4
+; IF-EVL-NEXT:    [[TMP2:%.*]] = mul i64 [[TMP1]], 4
+; IF-EVL-NEXT:    [[TMP3:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP3]], [[TMP2]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP5:%.*]] = mul i64 [[TMP4]], 4
+; IF-EVL-NEXT:    [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP7:%.*]] = mul i64 [[TMP6]], 4
+; IF-EVL-NEXT:    [[TMP8:%.*]] = sub i64 [[TMP7]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP8]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP5]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP9:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP10:%.*]] = mul i64 [[TMP9]], 4
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP12:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP11]], i32 4, i1 true)
+; IF-EVL-NEXT:    [[TMP13:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP15:%.*]] = getelementptr inbounds float, ptr [[TMP14]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 4 x float> @llvm.vp.load.nxv4f32.p0(ptr align 4 [[TMP15]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = call fast <vscale x 4 x float> @llvm.vp.fadd.nxv4f32(<vscale x 4 x float> [[VP_OP_LOAD]], <vscale x 4 x float> shufflevector (<vscale x 4 x float> insertelement (<vscale x 4 x float> poison, float 3.000000e+00, i64 0), <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer), <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, ptr [[TMP17]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4f32.p0(<vscale x 4 x float> [[TMP16]], ptr align 4 [[TMP18]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP19:%.*]] = zext i32 [[TMP12]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP19]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP20]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP28:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP21:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; IF-EVL-NEXT:    [[TMP:%.*]] = fadd fast float [[TMP21]], 3.000000e+00
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP29:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_fadd(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; NO-VP-NEXT:    [[TMP:%.*]] = fadd fast float [[TMP0]], 3.000000e+00
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds float, ptr %a, i64 %len
+  %0 = load float, ptr %arrayidx, align 4
+  %tmp = fadd fast float %0, 3.000000e+00
+  %arrayidx1 = getelementptr inbounds float, ptr %b, i64 %len
+  store float %tmp, ptr %arrayidx1, align 4
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_fsub(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_fsub(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 4
+; IF-EVL-NEXT:    [[TMP2:%.*]] = mul i64 [[TMP1]], 4
+; IF-EVL-NEXT:    [[TMP3:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP3]], [[TMP2]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP5:%.*]] = mul i64 [[TMP4]], 4
+; IF-EVL-NEXT:    [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP7:%.*]] = mul i64 [[TMP6]], 4
+; IF-EVL-NEXT:    [[TMP8:%.*]] = sub i64 [[TMP7]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP8]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP5]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP9:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP10:%.*]] = mul i64 [[TMP9]], 4
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP12:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP11]], i32 4, i1 true)
+; IF-EVL-NEXT:    [[TMP13:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP15:%.*]] = getelementptr inbounds float, ptr [[TMP14]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 4 x float> @llvm.vp.load.nxv4f32.p0(ptr align 4 [[TMP15]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = call fast <vscale x 4 x float> @llvm.vp.fsub.nxv4f32(<vscale x 4 x float> [[VP_OP_LOAD]], <vscale x 4 x float> shufflevector (<vscale x 4 x float> insertelement (<vscale x 4 x float> poison, float 3.000000e+00, i64 0), <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer), <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, ptr [[TMP17]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4f32.p0(<vscale x 4 x float> [[TMP16]], ptr align 4 [[TMP18]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP19:%.*]] = zext i32 [[TMP12]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP19]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP20]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP30:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP21:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; IF-EVL-NEXT:    [[TMP:%.*]] = fsub fast float [[TMP21]], 3.000000e+00
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP31:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_fsub(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; NO-VP-NEXT:    [[TMP:%.*]] = fsub fast float [[TMP0]], 3.000000e+00
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds float, ptr %a, i64 %len
+  %0 = load float, ptr %arrayidx, align 4
+  %tmp = fsub fast float %0, 3.000000e+00
+  %arrayidx1 = getelementptr inbounds float, ptr %b, i64 %len
+  store float %tmp, ptr %arrayidx1, align 4
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_fmul(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_fmul(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 4
+; IF-EVL-NEXT:    [[TMP2:%.*]] = mul i64 [[TMP1]], 4
+; IF-EVL-NEXT:    [[TMP3:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP3]], [[TMP2]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP5:%.*]] = mul i64 [[TMP4]], 4
+; IF-EVL-NEXT:    [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP7:%.*]] = mul i64 [[TMP6]], 4
+; IF-EVL-NEXT:    [[TMP8:%.*]] = sub i64 [[TMP7]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP8]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP5]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP9:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP10:%.*]] = mul i64 [[TMP9]], 4
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP12:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP11]], i32 4, i1 true)
+; IF-EVL-NEXT:    [[TMP13:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP15:%.*]] = getelementptr inbounds float, ptr [[TMP14]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 4 x float> @llvm.vp.load.nxv4f32.p0(ptr align 4 [[TMP15]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = call fast <vscale x 4 x float> @llvm.vp.fmul.nxv4f32(<vscale x 4 x float> [[VP_OP_LOAD]], <vscale x 4 x float> shufflevector (<vscale x 4 x float> insertelement (<vscale x 4 x float> poison, float 3.000000e+00, i64 0), <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer), <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, ptr [[TMP17]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4f32.p0(<vscale x 4 x float> [[TMP16]], ptr align 4 [[TMP18]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP19:%.*]] = zext i32 [[TMP12]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP19]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP20]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP32:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP21:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; IF-EVL-NEXT:    [[TMP:%.*]] = fmul fast float [[TMP21]], 3.000000e+00
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP33:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_fmul(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; NO-VP-NEXT:    [[TMP:%.*]] = fmul fast float [[TMP0]], 3.000000e+00
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds float, ptr %a, i64 %len
+  %0 = load float, ptr %arrayidx, align 4
+  %tmp = fmul fast float %0, 3.000000e+00
+  %arrayidx1 = getelementptr inbounds float, ptr %b, i64 %len
+  store float %tmp, ptr %arrayidx1, align 4
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_fdiv(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_fdiv(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 4
+; IF-EVL-NEXT:    [[TMP2:%.*]] = mul i64 [[TMP1]], 4
+; IF-EVL-NEXT:    [[TMP3:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP3]], [[TMP2]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP5:%.*]] = mul i64 [[TMP4]], 4
+; IF-EVL-NEXT:    [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP7:%.*]] = mul i64 [[TMP6]], 4
+; IF-EVL-NEXT:    [[TMP8:%.*]] = sub i64 [[TMP7]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP8]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP5]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP9:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP10:%.*]] = mul i64 [[TMP9]], 4
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP12:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP11]], i32 4, i1 true)
+; IF-EVL-NEXT:    [[TMP13:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP15:%.*]] = getelementptr inbounds float, ptr [[TMP14]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 4 x float> @llvm.vp.load.nxv4f32.p0(ptr align 4 [[TMP15]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = call fast <vscale x 4 x float> @llvm.vp.fdiv.nxv4f32(<vscale x 4 x float> [[VP_OP_LOAD]], <vscale x 4 x float> shufflevector (<vscale x 4 x float> insertelement (<vscale x 4 x float> poison, float 3.000000e+00, i64 0), <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer), <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, ptr [[TMP17]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4f32.p0(<vscale x 4 x float> [[TMP16]], ptr align 4 [[TMP18]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP19:%.*]] = zext i32 [[TMP12]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP19]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP20]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP34:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP21:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; IF-EVL-NEXT:    [[TMP:%.*]] = fdiv fast float [[TMP21]], 3.000000e+00
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP35:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_fdiv(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; NO-VP-NEXT:    [[TMP:%.*]] = fdiv fast float [[TMP0]], 3.000000e+00
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds float, ptr %a, i64 %len
+  %0 = load float, ptr %arrayidx, align 4
+  %tmp = fdiv fast float %0, 3.000000e+00
+  %arrayidx1 = getelementptr inbounds float, ptr %b, i64 %len
+  store float %tmp, ptr %arrayidx1, align 4
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_frem(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_frem(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP0:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; IF-EVL-NEXT:    [[TMP:%.*]] = frem fast float [[TMP0]], 3.000000e+00
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_frem(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; NO-VP-NEXT:    [[TMP:%.*]] = frem fast float [[TMP0]], 3.000000e+00
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds float, ptr %a, i64 %len
+  %0 = load float, ptr %arrayidx, align 4
+  %tmp = frem fast float %0, 3.000000e+00
+  %arrayidx1 = getelementptr inbounds float, ptr %b, i64 %len
+  store float %tmp, ptr %arrayidx1, align 4
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
+
+define void @test_fneg(ptr nocapture %a, ptr nocapture readonly %b) {
+; IF-EVL-LABEL: define void @test_fneg(
+; IF-EVL-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; IF-EVL-NEXT:  [[LOOP_PREHEADER:.*]]:
+; IF-EVL-NEXT:    [[A2:%.*]] = ptrtoint ptr [[A]] to i64
+; IF-EVL-NEXT:    [[B1:%.*]] = ptrtoint ptr [[B]] to i64
+; IF-EVL-NEXT:    br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
+; IF-EVL:       [[VECTOR_MEMCHECK]]:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 4
+; IF-EVL-NEXT:    [[TMP2:%.*]] = mul i64 [[TMP1]], 4
+; IF-EVL-NEXT:    [[TMP3:%.*]] = sub i64 [[B1]], [[A2]]
+; IF-EVL-NEXT:    [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP3]], [[TMP2]]
+; IF-EVL-NEXT:    br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]]
+; IF-EVL:       [[VECTOR_PH]]:
+; IF-EVL-NEXT:    [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP5:%.*]] = mul i64 [[TMP4]], 4
+; IF-EVL-NEXT:    [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP7:%.*]] = mul i64 [[TMP6]], 4
+; IF-EVL-NEXT:    [[TMP8:%.*]] = sub i64 [[TMP7]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 100, [[TMP8]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP5]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP9:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP10:%.*]] = mul i64 [[TMP9]], 4
+; IF-EVL-NEXT:    br label %[[VECTOR_BODY:.*]]
+; IF-EVL:       [[VECTOR_BODY]]:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP11:%.*]] = sub i64 100, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP12:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP11]], i32 4, i1 true)
+; IF-EVL-NEXT:    [[TMP13:%.*]] = add i64 [[EVL_BASED_IV]], 0
+; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP15:%.*]] = getelementptr inbounds float, ptr [[TMP14]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 4 x float> @llvm.vp.load.nxv4f32.p0(ptr align 4 [[TMP15]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP16:%.*]] = call fast <vscale x 4 x float> @llvm.vp.fneg.nxv4f32(<vscale x 4 x float> [[VP_OP_LOAD]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, ptr [[TMP17]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4f32.p0(<vscale x 4 x float> [[TMP16]], ptr align 4 [[TMP18]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP19:%.*]] = zext i32 [[TMP12]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP19]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP20]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP36:![0-9]+]]
+; IF-EVL:       [[MIDDLE_BLOCK]]:
+; IF-EVL-NEXT:    br i1 true, label %[[FINISH_LOOPEXIT:.*]], label %[[SCALAR_PH]]
+; IF-EVL:       [[SCALAR_PH]]:
+; IF-EVL-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[LOOP_PREHEADER]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
+; IF-EVL-NEXT:    br label %[[LOOP:.*]]
+; IF-EVL:       [[LOOP]]:
+; IF-EVL-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
+; IF-EVL-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; IF-EVL-NEXT:    [[TMP21:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; IF-EVL-NEXT:    [[TMP:%.*]] = fneg fast float [[TMP21]]
+; IF-EVL-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; IF-EVL-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; IF-EVL-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; IF-EVL-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT]], label %[[LOOP]], !llvm.loop [[LOOP37:![0-9]+]]
+; IF-EVL:       [[FINISH_LOOPEXIT]]:
+; IF-EVL-NEXT:    ret void
+;
+; NO-VP-LABEL: define void @test_fneg(
+; NO-VP-SAME: ptr nocapture [[A:%.*]], ptr nocapture readonly [[B:%.*]]) #[[ATTR0]] {
+; NO-VP-NEXT:  [[LOOP_PREHEADER:.*]]:
+; NO-VP-NEXT:    br label %[[LOOP:.*]]
+; NO-VP:       [[LOOP]]:
+; NO-VP-NEXT:    [[LEN:%.*]] = phi i64 [ [[DEC:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ]
+; NO-VP-NEXT:    [[DEC]] = add nsw i64 [[LEN]], 1
+; NO-VP-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[LEN]]
+; NO-VP-NEXT:    [[TMP0:%.*]] = load float, ptr [[ARRAYIDX]], align 4
+; NO-VP-NEXT:    [[TMP:%.*]] = fneg fast float [[TMP0]]
+; NO-VP-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[LEN]]
+; NO-VP-NEXT:    store float [[TMP]], ptr [[ARRAYIDX1]], align 4
+; NO-VP-NEXT:    [[DOTNOT:%.*]] = icmp eq i64 [[DEC]], 100
+; NO-VP-NEXT:    br i1 [[DOTNOT]], label %[[FINISH_LOOPEXIT:.*]], label %[[LOOP]]
+; NO-VP:       [[FINISH_LOOPEXIT]]:
+; NO-VP-NEXT:    ret void
+;
+loop.preheader:
+  br label %loop
+
+loop:
+  %len = phi i64 [ %dec, %loop ], [ 0, %loop.preheader ]
+  %dec = add nsw i64 %len, 1
+  %arrayidx = getelementptr inbounds float, ptr %a, i64 %len
+  %0 = load float, ptr %arrayidx, align 4
+  %tmp = fneg fast float %0
+  %arrayidx1 = getelementptr inbounds float, ptr %b, i64 %len
+  store float %tmp, ptr %arrayidx1, align 4
+  %.not = icmp eq i64 %dec, 100
+  br i1 %.not, label %finish.loopexit, label %loop
+
+finish.loopexit:
+  ret void
+}
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-masked-loadstore.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-masked-loadstore.ll
index 203d0c977074e9..5ecdb69bcf93fa 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-masked-loadstore.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-masked-loadstore.ll
@@ -52,13 +52,13 @@ define void @masked_loadstore(ptr noalias %a, ptr noalias %b, i64 %n) {
 ; IF-EVL-NEXT:    [[TMP21:%.*]] = getelementptr i32, ptr [[A:%.*]], i64 [[TMP13]]
 ; IF-EVL-NEXT:    [[TMP22:%.*]] = getelementptr i32, ptr [[TMP21]], i32 0
 ; IF-EVL-NEXT:    [[VP_OP_LOAD3:%.*]] = call <vscale x 4 x i32> @llvm.vp.load.nxv4i32.p0(ptr align 4 [[TMP22]], <vscale x 4 x i1> [[TMP20]], i32 [[TMP12]])
-; IF-EVL-NEXT:    [[TMP23:%.*]] = add <vscale x 4 x i32> [[VP_OP_LOAD]], [[VP_OP_LOAD3]]
-; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4i32.p0(<vscale x 4 x i32> [[TMP23]], ptr align 4 [[TMP22]], <vscale x 4 x i1> [[TMP20]], i32 [[TMP12]])
-; IF-EVL-NEXT:    [[TMP24:%.*]] = zext i32 [[TMP12]] to i64
-; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP24]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[VP_OP:%.*]] = call <vscale x 4 x i32> @llvm.vp.add.nxv4i32(<vscale x 4 x i32> [[VP_OP_LOAD]], <vscale x 4 x i32> [[VP_OP_LOAD3]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4i32.p0(<vscale x 4 x i32> [[VP_OP]], ptr align 4 [[TMP22]], <vscale x 4 x i1> [[TMP20]], i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP23:%.*]] = zext i32 [[TMP12]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP23]], [[EVL_BASED_IV]]
 ; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
-; IF-EVL-NEXT:    [[TMP25:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; IF-EVL-NEXT:    br i1 [[TMP25]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; IF-EVL-NEXT:    [[TMP24:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP24]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
 ; IF-EVL:       middle.block:
 ; IF-EVL-NEXT:    br i1 true, label [[EXIT:%.*]], label [[SCALAR_PH]]
 ; IF-EVL:       scalar.ph:
@@ -67,13 +67,13 @@ define void @masked_loadstore(ptr noalias %a, ptr noalias %b, i64 %n) {
 ; IF-EVL:       for.body:
 ; IF-EVL-NEXT:    [[I_011:%.*]] = phi i64 [ [[INC:%.*]], [[FOR_INC:%.*]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
 ; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[B]], i64 [[I_011]]
-; IF-EVL-NEXT:    [[TMP26:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
-; IF-EVL-NEXT:    [[CMP1:%.*]] = icmp ne i32 [[TMP26]], 0
+; IF-EVL-NEXT:    [[TMP25:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
+; IF-EVL-NEXT:    [[CMP1:%.*]] = icmp ne i32 [[TMP25]], 0
 ; IF-EVL-NEXT:    br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
 ; IF-EVL:       if.then:
 ; IF-EVL-NEXT:    [[ARRAYIDX3:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 [[I_011]]
-; IF-EVL-NEXT:    [[TMP27:%.*]] = load i32, ptr [[ARRAYIDX3]], align 4
-; IF-EVL-NEXT:    [[ADD:%.*]] = add i32 [[TMP26]], [[TMP27]]
+; IF-EVL-NEXT:    [[TMP26:%.*]] = load i32, ptr [[ARRAYIDX3]], align 4
+; IF-EVL-NEXT:    [[ADD:%.*]] = add i32 [[TMP25]], [[TMP26]]
 ; IF-EVL-NEXT:    store i32 [[ADD]], ptr [[ARRAYIDX3]], align 4
 ; IF-EVL-NEXT:    br label [[FOR_INC]]
 ; IF-EVL:       for.inc:
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-vp-intrinsics.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-vp-intrinsics.ll
index c69bb17f698aaf..25736fd52466b6 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-vp-intrinsics.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-vp-intrinsics.ll
@@ -41,15 +41,15 @@ define void @foo(ptr noalias %a, ptr noalias %b, ptr noalias %c, i64 %N) {
 ; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i32, ptr [[C:%.*]], i64 [[TMP13]]
 ; IF-EVL-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i32, ptr [[TMP16]], i32 0
 ; IF-EVL-NEXT:    [[VP_OP_LOAD1:%.*]] = call <vscale x 4 x i32> @llvm.vp.load.nxv4i32.p0(ptr align 4 [[TMP17]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
-; IF-EVL-NEXT:    [[TMP18:%.*]] = add nsw <vscale x 4 x i32> [[VP_OP_LOAD1]], [[VP_OP_LOAD]]
-; IF-EVL-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i32, ptr [[A:%.*]], i64 [[TMP13]]
-; IF-EVL-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i32, ptr [[TMP19]], i32 0
-; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4i32.p0(<vscale x 4 x i32> [[TMP18]], ptr align 4 [[TMP20]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
-; IF-EVL-NEXT:    [[TMP21:%.*]] = zext i32 [[TMP12]] to i64
-; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP21]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[VP_OP:%.*]] = call <vscale x 4 x i32> @llvm.vp.add.nxv4i32(<vscale x 4 x i32> [[VP_OP_LOAD1]], <vscale x 4 x i32> [[VP_OP_LOAD]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i32, ptr [[A:%.*]], i64 [[TMP13]]
+; IF-EVL-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i32, ptr [[TMP18]], i32 0
+; IF-EVL-NEXT:    call void @llvm.vp.store.nxv4i32.p0(<vscale x 4 x i32> [[VP_OP]], ptr align 4 [[TMP19]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[TMP20:%.*]] = zext i32 [[TMP12]] to i64
+; IF-EVL-NEXT:    [[INDEX_EVL_NEXT]] = add i64 [[TMP20]], [[EVL_BASED_IV]]
 ; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
-; IF-EVL-NEXT:    [[TMP22:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; IF-EVL-NEXT:    br i1 [[TMP22]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; IF-EVL-NEXT:    [[TMP21:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT:    br i1 [[TMP21]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
 ; IF-EVL:       middle.block:
 ; IF-EVL-NEXT:    br i1 true, label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH]]
 ; IF-EVL:       scalar.ph:
@@ -58,10 +58,10 @@ define void @foo(ptr noalias %a, ptr noalias %b, ptr noalias %c, i64 %N) {
 ; IF-EVL:       for.body:
 ; IF-EVL-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[FOR_BODY]] ]
 ; IF-EVL-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[B]], i64 [[IV]]
-; IF-EVL-NEXT:    [[TMP23:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
+; IF-EVL-NEXT:    [[TMP22:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
 ; IF-EVL-NEXT:    [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, ptr [[C]], i64 [[IV]]
-; IF-EVL-NEXT:    [[TMP24:%.*]] = load i32, ptr [[ARRAYIDX2]], align 4
-; IF-EVL-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP24]], [[TMP23]]
+; IF-EVL-NEXT:    [[TMP23:%.*]] = load i32, ptr [[ARRAYIDX2]], align 4
+; IF-EVL-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP23]], [[TMP22]]
 ; IF-EVL-NEXT:    [[ARRAYIDX4:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 [[IV]]
 ; IF-EVL-NEXT:    store i32 [[ADD]], ptr [[ARRAYIDX4]], align 4
 ; IF-EVL-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vplan-vp-intrinsics.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vplan-vp-intrinsics.ll
index 8caa9368bfde18..0b220a0884b74a 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vplan-vp-intrinsics.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vplan-vp-intrinsics.ll
@@ -31,7 +31,7 @@ define void @foo(ptr noalias %a, ptr noalias %b, ptr noalias %c, i64 %N) {
 ; IF-EVL-NEXT:    CLONE ir<[[GEP2:%.+]]> = getelementptr inbounds ir<%c>, vp<[[ST]]>
 ; IF-EVL-NEXT:    vp<[[PTR2:%[0-9]+]]> = vector-pointer ir<[[GEP2]]>
 ; IF-EVL-NEXT:    WIDEN ir<[[LD2:%.+]]> = vp.load vp<[[PTR2]]>, vp<[[EVL]]>
-; IF-EVL-NEXT:    WIDEN ir<[[ADD:%.+]]> = add nsw ir<[[LD2]]>, ir<[[LD1]]>
+; IF-EVL-NEXT:    WIDEN vp ir<[[ADD:%.+]]> = add nsw ir<[[LD2]]>, ir<[[LD1]]>
 ; IF-EVL-NEXT:    CLONE ir<[[GEP3:%.+]]> = getelementptr inbounds ir<%a>, vp<[[ST]]>
 ; IF-EVL-NEXT:    vp<[[PTR3:%[0-9]+]]> = vector-pointer ir<[[GEP3]]>
 ; IF-EVL-NEXT:    WIDEN vp.store vp<[[PTR3]]>, ir<[[ADD]]>, vp<[[EVL]]>

>From 5311f92977082d200e15dd56cf155732d9afd8ff Mon Sep 17 00:00:00 2001
From: Kolya Panchenko <kolya.panchenko at sifive.com>
Date: Wed, 5 Jun 2024 14:22:01 -0700
Subject: [PATCH 2/2] Addressed comments

---
 .../Transforms/Vectorize/LoopVectorize.cpp    |  5 +-
 llvm/lib/Transforms/Vectorize/VPlan.h         | 10 +--
 .../Transforms/Vectorize/VPlanAnalysis.cpp    |  3 +-
 .../lib/Transforms/Vectorize/VPlanRecipes.cpp | 17 ++---
 .../Transforms/Vectorize/VPlanTransforms.cpp  | 11 +--
 .../Transforms/Vectorize/VPlanVerifier.cpp    | 71 +++++++++++++++++++
 ...-force-tail-with-evl-bin-unary-ops-args.ll |  4 +-
 7 files changed, 93 insertions(+), 28 deletions(-)

diff --git a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
index 5939ce5b917aa2..7fc015510b55fb 100644
--- a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -8549,14 +8549,15 @@ void LoopVectorizationPlanner::buildVPlansWithVPRecipes(ElementCount MinVF,
   for (ElementCount VF = MinVF; ElementCount::isKnownLT(VF, MaxVFTimes2);) {
     VFRange SubRange = {VF, MaxVFTimes2};
     if (auto Plan = tryToBuildVPlanWithVPRecipes(SubRange)) {
+      bool IsScalarVPlan = Plan->hasVF(ElementCount::getFixed(1));
       // Now optimize the initial VPlan.
-      if (!Plan->hasVF(ElementCount::getFixed(1)))
+      if (!IsScalarVPlan)
         VPlanTransforms::truncateToMinimalBitwidths(
             *Plan, CM.getMinimalBitwidths(), PSE.getSE()->getContext());
       VPlanTransforms::optimize(*Plan, *PSE.getSE());
       // TODO: try to put it close to addActiveLaneMask().
       // Discard the plan if it is not EVL-compatible
-      if (CM.foldTailWithEVL() &&
+      if (!IsScalarVPlan && CM.foldTailWithEVL() &&
           !VPlanTransforms::tryAddExplicitVectorLength(*Plan))
         break;
       assert(verifyVPlanIsValid(*Plan) && "VPlan is invalid");
diff --git a/llvm/lib/Transforms/Vectorize/VPlan.h b/llvm/lib/Transforms/Vectorize/VPlan.h
index 4b22e4983efac8..42e10f8a69c42d 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.h
+++ b/llvm/lib/Transforms/Vectorize/VPlan.h
@@ -1335,9 +1335,9 @@ class VPInstruction : public VPRecipeWithIRFlags {
 /// ingredient. This recipe covers most of the traditional vectorization cases
 /// where each ingredient transforms into a vectorized version of itself.
 class VPWidenRecipe : public VPRecipeWithIRFlags {
-protected:
   unsigned Opcode;
 
+protected:
   template <typename IterT>
   VPWidenRecipe(unsigned VPDefOpcode, Instruction &I,
                 iterator_range<IterT> Operands)
@@ -1371,7 +1371,6 @@ class VPWidenRecipe : public VPRecipeWithIRFlags {
 };
 
 class VPWidenEVLRecipe : public VPWidenRecipe {
-private:
   using VPRecipeWithIRFlags::transferFlags;
 
 public:
@@ -1380,6 +1379,10 @@ class VPWidenEVLRecipe : public VPWidenRecipe {
       : VPWidenRecipe(VPDef::VPWidenEVLSC, I, Operands) {
     addOperand(&EVL);
   }
+  VPWidenEVLRecipe(VPWidenRecipe *W, VPValue &EVL)
+      : VPWidenEVLRecipe(*W->getUnderlyingInstr(), W->operands(), EVL) {
+    this->transferFlags(*W);
+  }
 
   ~VPWidenEVLRecipe() override = default;
 
@@ -1397,9 +1400,6 @@ class VPWidenEVLRecipe : public VPWidenRecipe {
   VPValue *getEVL() { return getOperand(getNumOperands() - 1); }
   const VPValue *getEVL() const { return getOperand(getNumOperands() - 1); }
 
-  /// A helper function to create widen EVL recipe from regular widen recipe.
-  static VPWidenEVLRecipe *create(VPWidenRecipe *W, VPValue &EVL);
-
   /// Produce widened copies of all Ingredients.
   void execute(VPTransformState &State) override final;
 
diff --git a/llvm/lib/Transforms/Vectorize/VPlanAnalysis.cpp b/llvm/lib/Transforms/Vectorize/VPlanAnalysis.cpp
index efe8c21874a3ac..cf50fa44b649e4 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanAnalysis.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanAnalysis.cpp
@@ -239,7 +239,8 @@ Type *VPTypeAnalysis::inferScalarType(const VPValue *V) {
             return inferScalarType(R->getOperand(0));
           })
           .Case<VPBlendRecipe, VPInstruction, VPWidenRecipe, VPReplicateRecipe,
-                VPWidenCallRecipe, VPWidenMemoryRecipe, VPWidenSelectRecipe>(
+                VPWidenCallRecipe, VPWidenMemoryRecipe, VPWidenSelectRecipe,
+                VPWidenEVLRecipe>(
               [this](const auto *R) { return inferScalarTypeForRecipe(R); })
           .Case<VPInterleaveRecipe>([V](const VPInterleaveRecipe *R) {
             // TODO: Use info from interleave group.
diff --git a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
index a6fd26b6665013..fb3edaaf1b0e2a 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
@@ -997,20 +997,14 @@ void VPWidenRecipe::execute(VPTransformState &State) {
 #endif
 }
 
-VPWidenEVLRecipe *VPWidenEVLRecipe::create(VPWidenRecipe *W, VPValue &EVL) {
-  auto *R = new VPWidenEVLRecipe(*W->getUnderlyingInstr(), W->operands(), EVL);
-  R->transferFlags(*W);
-  return R;
-}
-
 void VPWidenEVLRecipe::execute(VPTransformState &State) {
   assert(State.UF == 1 && "Expected only UF == 1 when vectorizing with "
                           "explicit vector length.");
   VPValue *Op0 = getOperand(0);
 
   // If it's scalar operation, hand translation over to VPWidenRecipe
-  if (!State.get(Op0, 0)->getType()->isVectorTy())
-    return VPWidenRecipe::execute(State);
+  assert(State.get(Op0, 0)->getType()->isVectorTy() &&
+         "VPWidenEVLRecipe should not be used for scalars");
 
   VPValue *EVL = getEVL();
   Value *EVLArg = State.get(EVL, 0, /*NeedsScalar=*/true);
@@ -1049,10 +1043,7 @@ bool VPWidenEVLRecipe::onlyFirstLaneUsed(const VPValue *Op) const {
   assert(is_contained(operands(), Op) && "Op must be an operand of the recipe");
   // EVL in that recipe is always the last operand, thus any use before means
   // the VPValue should be vectorized.
-  for (unsigned I = 0, E = getNumOperands() - 1; I != E; ++I)
-    if (getOperand(I) == Op)
-      return false;
-  return true;
+  return getEVL() == Op;
 }
 
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
@@ -1069,7 +1060,7 @@ void VPWidenEVLRecipe::print(raw_ostream &O, const Twine &Indent,
                              VPSlotTracker &SlotTracker) const {
   O << Indent << "WIDEN vp ";
   printAsOperand(O, SlotTracker);
-  O << " = " << Instruction::getOpcodeName(Opcode);
+  O << " = " << Instruction::getOpcodeName(getOpcode());
   printFlags(O);
   printOperands(O, SlotTracker);
 }
diff --git a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
index aada846473fe9d..1f5dc376c12b33 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
@@ -1299,6 +1299,8 @@ void VPlanTransforms::addActiveLaneMask(
 
 /// Replace recipes with their EVL variants.
 static void transformRecipestoEVLRecipes(VPlan &Plan, VPValue &EVL) {
+  VPDominatorTree VPDT;
+  VPDT.recalculate(Plan);
   DenseSet<VPRecipeBase *> ToRemove;
 
   ReversePostOrderTraversal<VPBlockDeepTraversalWrapper<VPBlockBase *>> RPOT(
@@ -1306,9 +1308,10 @@ static void transformRecipestoEVLRecipes(VPlan &Plan, VPValue &EVL) {
   DenseSet<VPValue *> HeaderMasks = collectAllHeaderMasks(Plan);
   for (VPBasicBlock *VPBB :
        reverse(VPBlockUtils::blocksOnly<VPBasicBlock>(RPOT))) {
-    // The recipes in the block are processed in reverse order, to catch chains
-    // of dead recipes.
     for (VPRecipeBase &R : make_early_inc_range(reverse(*VPBB))) {
+      if (!properlyDominates(EVL.getDefiningRecipe(), &R, VPDT))
+        continue;
+
       TypeSwitch<VPRecipeBase *>(&R)
           .Case<VPWidenLoadRecipe>([&](VPWidenLoadRecipe *L) {
             VPValue *NewMask =
@@ -1330,7 +1333,7 @@ static void transformRecipestoEVLRecipes(VPlan &Plan, VPValue &EVL) {
             if (!Instruction::isBinaryOp(Opcode) &&
                 !Instruction::isUnaryOp(Opcode))
               return;
-            auto *N = VPWidenEVLRecipe::create(W, EVL);
+            auto *N = new VPWidenEVLRecipe(W, EVL);
             N->insertBefore(W);
             W->replaceAllUsesWith(N);
             ToRemove.insert(W);
@@ -1345,8 +1348,6 @@ static void transformRecipestoEVLRecipes(VPlan &Plan, VPValue &EVL) {
     recursivelyDeleteDeadRecipes(HeaderMask);
 }
 
-
-
 /// Add a VPEVLBasedIVPHIRecipe and related recipes to \p Plan and
 /// replaces all uses except the canonical IV increment of
 /// VPCanonicalIVPHIRecipe with a VPEVLBasedIVPHIRecipe. VPCanonicalIVPHIRecipe
diff --git a/llvm/lib/Transforms/Vectorize/VPlanVerifier.cpp b/llvm/lib/Transforms/Vectorize/VPlanVerifier.cpp
index 56c82b13ccdef3..d83d66198347e3 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanVerifier.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanVerifier.cpp
@@ -18,6 +18,7 @@
 #include "VPlanDominatorTree.h"
 #include "llvm/ADT/DepthFirstIterator.h"
 #include "llvm/ADT/SmallPtrSet.h"
+#include "llvm/ADT/TypeSwitch.h"
 #include "llvm/Support/CommandLine.h"
 
 #define DEBUG_TYPE "loop-vectorize"
@@ -35,6 +36,11 @@ class VPlanVerifier {
   // VPHeaderPHIRecipes.
   bool verifyPhiRecipes(const VPBasicBlock *VPBB);
 
+  // Verify that \p EVL is used correctly. The user must be  either in EVL-based
+  // recipes as a last operand or VPInstruction::Add which is incoming value
+  // into EVL's recipe.
+  bool verifyEVLRecipe(const VPInstruction &EVL) const;
+
   bool verifyVPBasicBlock(const VPBasicBlock *VPBB);
 
   bool verifyBlock(const VPBlockBase *VPB);
@@ -114,6 +120,64 @@ bool VPlanVerifier::verifyPhiRecipes(const VPBasicBlock *VPBB) {
   return true;
 }
 
+bool VPlanVerifier::verifyEVLRecipe(const VPInstruction &EVL) const {
+  if (EVL.getOpcode() != VPInstruction::ExplicitVectorLength) {
+    errs() << "verifyEVLRecipe should only be called on "
+              "VPInstruction::ExplicitVectorLength\n";
+    return false;
+  }
+  auto VerifyEVLUse = [&](const VPRecipeBase &R,
+                          const unsigned ExpectedIdx) -> bool {
+    SmallVector<const VPValue *> Ops(R.operands());
+    unsigned UseCount = count(Ops, &EVL);
+    if (UseCount != 1 || Ops[ExpectedIdx] != &EVL) {
+      errs() << "EVL is used as non-last operand in EVL-based recipe\n";
+      return false;
+    }
+    return true;
+  };
+  for (const VPUser *U : EVL.users()) {
+    if (!TypeSwitch<const VPUser *, bool>(U)
+             .Case<VPWidenStoreEVLRecipe>([&](const VPWidenStoreEVLRecipe *S) {
+               return VerifyEVLUse(*S, 2);
+             })
+             .Case<VPWidenLoadEVLRecipe>([&](const VPWidenLoadEVLRecipe *L) {
+               return VerifyEVLUse(*L, 1);
+             })
+             .Case<VPWidenEVLRecipe>([&](const VPWidenEVLRecipe *W) {
+               return VerifyEVLUse(
+                   *W, Instruction::isUnaryOp(W->getOpcode()) ? 1 : 2);
+             })
+             .Case<VPScalarCastRecipe>(
+                 [&](const VPScalarCastRecipe *S) { return true; })
+             .Case<VPInstruction>([&](const VPInstruction *I) {
+               if (I->getOpcode() != Instruction::Add) {
+                 errs()
+                     << "EVL is used as an operand in non-VPInstruction::Add\n";
+                 return false;
+               }
+               if (I->getNumUsers() != 1) {
+                 errs() << "EVL is used in VPInstruction:Add with multiple "
+                           "users\n";
+                 return false;
+               }
+               if (!isa<VPEVLBasedIVPHIRecipe>(*I->users().begin())) {
+                 errs() << "Result of VPInstruction::Add with EVL operand is "
+                           "not used by VPEVLBasedIVPHIRecipe\n";
+                 return false;
+               }
+               return true;
+             })
+             .Default([&](const VPUser *U) {
+               errs() << "EVL has unexpected user\n";
+               return false;
+             })) {
+      return false;
+    }
+  }
+  return true;
+}
+
 bool VPlanVerifier::verifyVPBasicBlock(const VPBasicBlock *VPBB) {
   if (!verifyPhiRecipes(VPBB))
     return false;
@@ -150,6 +214,13 @@ bool VPlanVerifier::verifyVPBasicBlock(const VPBasicBlock *VPBB) {
         }
       }
     }
+    if (const auto *EVL = dyn_cast<VPInstruction>(&R)) {
+      if (EVL->getOpcode() == VPInstruction::ExplicitVectorLength &&
+          !verifyEVLRecipe(*EVL)) {
+        errs() << "EVL VPValue is not used correctly\n";
+        return false;
+      }
+    }
   }
 
   auto *IRBB = dyn_cast<VPIRBasicBlock>(VPBB);
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-bin-unary-ops-args.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-bin-unary-ops-args.ll
index 77689fc8a76ee6..6927bad301ef1c 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-bin-unary-ops-args.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-bin-unary-ops-args.ll
@@ -2,12 +2,12 @@
 ; RUN: opt -passes=loop-vectorize \
 ; RUN: -force-tail-folding-style=data-with-evl \
 ; RUN: -prefer-predicate-over-epilogue=predicate-dont-vectorize \
-; RUN: -mtriple=riscv64 -mattr=+v -S < %s | FileCheck %s --check-prefix=IF-EVL
+; RUN: -mtriple=riscv64 -mattr=+v -S %s | FileCheck %s --check-prefix=IF-EVL
 
 ; RUN: opt -passes=loop-vectorize \
 ; RUN: -force-tail-folding-style=none \
 ; RUN: -prefer-predicate-over-epilogue=predicate-dont-vectorize \
-; RUN: -mtriple=riscv64 -mattr=+v -S < %s | FileCheck %s --check-prefix=NO-VP
+; RUN: -mtriple=riscv64 -mattr=+v -S %s | FileCheck %s --check-prefix=NO-VP
 
 
 define void @test_and(ptr nocapture %a, ptr nocapture readonly %b) {