[VPlan] Introduce and use BranchOnCount VPInstruction.

[fhahn]

Summary

This patch adds a new BranchOnCount VPInstruction opcode with 2
operands. It first compares its 2 operands (increment of canonical
induction and vector trip count), followed by a branch to either the
exit block or back to the vector header.

It must be the last recipe in the exit block of the topmost vector loop
region.

This extracts parts from D113224 and was discussed in D113223.

[fhahn]

Update addCanonicalIVRecipes comment.

[Ayal]

Combine ExitCheckAndBranch with CanonicalIVIncrement{NUW} into a single BranchOnCount recipe or two BranchOnCount{NUW} VPInstructions?
It is conceptually a single inc-cmp-br idiom, akin to PowerPC's "bdnz" instruction - except instead of decrementing an implicit count register until zero we increment an IV until VectorTripCount.

[fhahn]

Move VectorTripCount printing from D113223 to this patch.

In D116479#3216397, @Ayal wrote:

Combine ExitCheckAndBranch with CanonicalIVIncrement{NUW} into a single BranchOnCount recipe or two BranchOnCount{NUW} VPInstructions?
It is conceptually a single inc-cmp-br idiom, akin to PowerPC's "bdnz" instruction - except instead of decrementing an implicit count register until zero we increment an IV until VectorTripCount.

I'm not sure about the benefits of having the incremenet included in the CheckAndBranch opcode. It would reduce the number of opcodes, but may have the following potential drawbacks:

1. Once we can model live-ins, we may be able to get rid of the specialized increment opcodes and model it as add with the increment as operand,
2. increments for other phi recipes also need to be modeled explicitly, so having an explicit increment for the canonical IV seems more consistent,
3. we may want to remove the 'increment' instruction in a VPlan transform in the future, e.g. because we have a canonical vector induction which could be used instead of the scalar one.

[Ayal]

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

8962

setCondBit under IsVPlanNative?

9424

Setting insert point once before the loop is fine; in addition, selects are now introduced at the beginning of the latch rather than at the end? Make sure live-out instruction Red appears before the latch?
Perhaps insert at/before Last(NonTerminal?) of latch , akin to FirstNonPhi?

llvm/lib/Transforms/Vectorize/VPlan.cpp

734

early-break if Part > 0?

746

Once the exiting block is modelled in VPlan, it should help retrieve its IRBB directly (TODO).

945–946

So if now LastBB is expected to terminate with unreachable rather than branch in native as well, remove the EnableVPlanNativePath from the assert?

947

Should the unreachable terminator be removed from LastBB when the branch is introduced, rather than here? As in VPBranchOnMaskRecipe::execute()'s final ReplaceInstWithInst().

Could the compare stay in LastBB, soon to join its branch in VectorLatchBB when the two BB's merge?

It may be possible to associate LatchVPBB with VectorLatchBB upon construction and avoid createEmptyBasicBlock for it during VPBB::execute(), thereby having all latch recipes code-gen'd directly into VectorLatchBB, rather than having to merge separate LastBB and VectorLatchBB here.

[Ayal]

In D116479#3216412, @fhahn wrote:

Move VectorTripCount printing from D113223 to this patch.

In D116479#3216397, @Ayal wrote:

Combine ExitCheckAndBranch with CanonicalIVIncrement{NUW} into a single BranchOnCount recipe or two BranchOnCount{NUW} VPInstructions?
It is conceptually a single inc-cmp-br idiom, akin to PowerPC's "bdnz" instruction - except instead of decrementing an implicit count register until zero we increment an IV until VectorTripCount.

I'm not sure about the benefits of having the incremenet included in the CheckAndBranch opcode. It would reduce the number of opcodes, but may have the following potential drawbacks:

1. Once we can model live-ins, we may be able to get rid of the specialized increment opcodes and model it as add with the increment as operand,
2. increments for other phi recipes also need to be modeled explicitly, so having an explicit increment for the canonical IV seems more consistent,
3. we may want to remove the 'increment' instruction in a VPlan transform in the future, e.g. because we have a canonical vector induction which could be used instead of the scalar one.

Main motivation behind a BranchOnCount recipe is to capture higher semantics and a meaningful idiom in VPlan - clearly modelling the scalar IV which controls a countable loop, rather than minimizing number of opcodes or lowering to ordinary operations. How other IV's should be modelled and optimized is indeed worth addressing, aiming for clarity as well while avoiding premature optimization best left to other passes, e.g., LSR. It should also be possible to optimize BranchOnCount away into other recipes as a VPlan to VPlan transformation, if and when desired.

[fhahn]

Address latest comments, thanks!

In D116479#3216491, @Ayal wrote:

In D116479#3216412, @fhahn wrote:

Move VectorTripCount printing from D113223 to this patch.

In D116479#3216397, @Ayal wrote:

Main motivation behind a BranchOnCount recipe is to capture higher semantics and a meaningful idiom in VPlan - clearly modelling the scalar IV which controls a countable loop, rather than minimizing number of opcodes or lowering to ordinary operations. How other IV's should be modelled and optimized is indeed worth addressing, aiming for clarity as well while avoiding premature optimization best left to other passes, e.g., LSR. It should also be possible to optimize BranchOnCount away into other recipes as a VPlan to VPlan transformation, if and when desired.

Sounds good. Should I rename CheckExitAndBranch -> BranchOnCount? The scalar IV should already be modeled clearly via the dedicated recipe, irrespective of where the bump happens. I left it as is for now, but folding the count in should be a trivial change.

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

8962

Updated, thanks!

llvm/lib/Transforms/Vectorize/VPlan.cpp

734

Adjusted, thanks!

746

I added the todo, thanks!

945–946

Removed VPlanNativePath from assert, thanks!

947

Should the unreachable terminator be removed from LastBB when the branch is introduced, rather than here? As in VPBranchOnMaskRecipe::execute()'s final ReplaceInstWithInst().

Updated, thanks!

Could the compare stay in LastBB, soon to join its branch in VectorLatchBB when the two BB's merge?

It could, but it would require updating ~50 tests. The reason is that widenIntOrFpInduction will still generate the vector induction increments in VectorLatchBB, which is outside VPlan.

It may be possible to associate LatchVPBB with VectorLatchBB upon construction and avoid createEmptyBasicBlock for it during VPBB::execute(), thereby having all latch recipes code-gen'd directly into VectorLatchBB, rather than having to merge separate LastBB and VectorLatchBB here.

It seems like this would require threading through VectorLatchBB to VPBasicBlock::execute. Not sure if that extra work is justified for the temporary workaround that won't be needed once the latch block is modeled in VPlan directly.

[fhahn]

llvm/lib/Transforms/Vectorize/VPlan.cpp

938

I looked into removing the separate latch block, but unfortunately widenIntOrFpInduction still relies on the latch block being created up-front:

// Move the last step to the end of the latch block. This ensures consistent
// placement of all induction updates.
auto *LoopVectorLatch = LI->getLoopFor(LoopVectorBody)->getLoopLatch();
auto *Br = cast<BranchInst>(LoopVectorLatch->getTerminator());
LastInduction->moveBefore(Br);
LastInduction->setName("vec.ind.next");

Reusing the created vector latch for the exit block of the region is also not straight-forward, because we may need to look through multiple single successors until we reach a the exit VPBB. This is complicated by blocks inside replicate regions, where we would need to check the successor of the region, but *only* for the last iteration in the loop for the region.

So I think it is simpler to keep things as is, until the VPWidenIntOrFpInduction increments are modeled separately.

[Ayal]

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

8962

Fold it under the above "if (IsVPlanNative)"?

9424

Make sure live-out instruction Red appears before the latch?

llvm/lib/Transforms/Vectorize/VPlan.cpp

743

IsVPlanNative? Perhaps it should use a subclass of VPRegionBlock which produces getHeader()/getExit() by bumping from Entry/Exit, in one place.

936

drop "Remove the Unreachable terminator"

938

Fine. If this too-early sinking of LastInduction is the only reason for creating the latch block up-front, then let's remember (TODO?) to remove the latter once the former is modelled with a recipe properly placed in the Exit VPBB, next to ExitCheckAndBranch.

944

moveBefore(LastBranch) ?

llvm/lib/Transforms/Vectorize/VPlanVerifier.cpp

180

Exit->empty()? Check above along with !Exit, to avoid taking prev of Exit->end()?

[fhahn]

Address latest comments, thanks! Also renamed ExitCheckAndBranch to BranchOnCount.

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

9424

Added an assert. Was that what you had in mind?

llvm/lib/Transforms/Vectorize/VPlan.cpp

743

Added an assert. Maybe it would be better to get rid of the difference in the native path & model the pre-header block outside the vector loop region? (as follow-up down the line)

936

Thanks, removed.

938

Added a todo.

944

That's better, thanks!

llvm/lib/Transforms/Vectorize/VPlanVerifier.cpp

180

I added a lambda for the VPInstruction check and updated to use empty(). With the check above the error message may be a bit misleading or would need duplicating.

[fhahn]

Rebase and ping :)

[Ayal]

This looks good to me, adding minor optional suggestions.

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

9424

Yeah, or set the insert point to the end of LatchVPBB (prior to its BranchOnCount terminator) rather than the beginning of LatchVPBB, retaining original behavior; conceptually, the reduction recipe should be allowed to reside (earlier) in the latch.
Above comment "at the end of the latch" deserves updating.

llvm/lib/Transforms/Vectorize/VPlan.cpp

739

Add a comment that after creating the compare we now create the branch?

743

Added an assert. Maybe it would be better to get rid of the difference in the native path & model the pre-header block outside the vector loop region? (as follow-up down the line)

+1

746

Worth setting BasicBlock *ExitBlock = ... next to setting Header above, and place comment there?

754

This is fine, can alternatively use ReplaceInstWithInst() as does VPBranchOnMask::execute.

936

typo: branchand

998–1003

Now we can assert number of users is positive?

llvm/lib/Transforms/Vectorize/VPlanVerifier.cpp

180

Spit out "exit block must not be empty" if Exit.empty() and check the inlined lambda afterwards?

[fhahn]

Address latest comments, thanks! Also rebased. I am planning to commit this once the pre-commit tests come back OK.

[fhahn]

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

9424

Updated the comment, thanks!

llvm/lib/Transforms/Vectorize/VPlan.cpp

754

I left it as is for now.

936

fixed, thanks

998–1003

There's still an edge case where this assert may not hold unfortunately: VPlan-stress testing, which exits before the recipes are introduced. Potentially buuildHCFG could already add the branch & mask recipe, but for now I left it as is.

[bmahjour]

This commit somehow causes the "unroll.disable" metadata to be dropped from the vector epilogue loops. For example:

> cat tmp.ll
target datalayout = "e-m:e-i64:64-n32:64"
target triple = "powerpc64le-unknown-linux-gnu"

; Function Attrs: nounwind
define dso_local void @f1(float* noalias %aa, float* noalias %bb, float* noalias %cc, i32 signext %N) #0 {
entry:
  %cmp1 = icmp sgt i32 %N, 0
  br i1 %cmp1, label %for.body.preheader, label %for.end

for.body.preheader:                               ; preds = %entry
  %wide.trip.count = zext i32 %N to i64
  br label %for.body

for.body:                                         ; preds = %for.body.preheader, %for.body
  %indvars.iv = phi i64 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
  %arrayidx = getelementptr inbounds float, float* %bb, i64 %indvars.iv
  %0 = load float, float* %arrayidx, align 4
  %arrayidx2 = getelementptr inbounds float, float* %cc, i64 %indvars.iv
  %1 = load float, float* %arrayidx2, align 4
  %add = fadd fast float %0, %1
  %arrayidx4 = getelementptr inbounds float, float* %aa, i64 %indvars.iv
  store float %add, float* %arrayidx4, align 4
  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
  %exitcond = icmp ne i64 %indvars.iv.next, %wide.trip.count
  br i1 %exitcond, label %for.body, label %for.end.loopexit

for.end.loopexit:                                 ; preds = %for.body
  br label %for.end

for.end:                                          ; preds = %for.end.loopexit, %entry
  ret void
}

opt -passes=loop-vectorize ./tmp.ll -S -o out.ll -epilogue-vectorization-force-VF=2

> grep "!" out.ll
  br i1 %132, label %middle.block, label %vector.body, !llvm.loop !0
  br i1 %144, label %vec.epilog.middle.block, label %vec.epilog.vector.body, !llvm.loop !2
  br i1 %exitcond, label %for.body, label %for.end.loopexit.loopexit, !llvm.loop !3
!0 = distinct !{!0, !1}
!1 = !{!"llvm.loop.isvectorized", i32 1}
!2 = distinct !{!2, !1}
!3 = distinct !{!3, !4, !1}
!4 = !{!"llvm.loop.unroll.runtime.disable"}

Previously we had:

...
  br i1 %144, label %vec.epilog.middle.block, label %vec.epilog.vector.body, !llvm.loop !2
...
!2 = distinct !{!2, !3, !1}
!3 = !{!"llvm.loop.unroll.runtime.disable"}

[fhahn]

This commit somehow causes the "unroll.disable" metadata to be dropped from the vector epilogue loops. For example:

Thanks for the heads-up. Looks like unfortunately the relevant check lines to guard against the regression have been removed earlier. It should be fixed by 070d1034da87. I also tried to restore the original check lines removed by 278aa65cc495

[bmahjour]

In D116479#3246783, @fhahn wrote:

This commit somehow causes the "unroll.disable" metadata to be dropped from the vector epilogue loops. For example:

Thanks for the heads-up. Looks like unfortunately the relevant check lines to guard against the regression have been removed earlier. It should be fixed by 070d1034da87. I also tried to restore the original check lines removed by 278aa65cc495

Thanks for the fix and the reviving of the MD checks!

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

void LoopVectorizationPlanner::buildVPlansWithVPRecipes(ElementCount MinVF,
   for (ElementCount VF = MinVF; ElementCount::isKnownLT(VF, MaxVFPlusOne);) {
     VFRange SubRange = {VF, MaxVFPlusOne};
     VPlans.push_back(
         buildVPlanWithVPRecipes(SubRange, DeadInstructions, SinkAfter));
     VF = SubRange.End;
   }
 }
 
-// Add a VPCanonicalIVPHIRecipe starting at 0 to the header and a
-// CanonicalIVIncrement{NUW} VPInstruction to increment it by VF * UF to the
-// latch.
+// Add a VPCanonicalIVPHIRecipe starting at 0 to the header, a
+// CanonicalIVIncrement{NUW} VPInstruction to increment it by VF * UF and a
+// BranchOnCount VPInstruction to the latch.
 static void addCanonicalIVRecipes(VPlan &Plan, Type *IdxTy, DebugLoc DL,
                                   bool HasNUW, bool IsVPlanNative) {
   Value *StartIdx = ConstantInt::get(IdxTy, 0);
   auto *StartV = Plan.getOrAddVPValue(StartIdx);
 
   auto *CanonicalIVPHI = new VPCanonicalIVPHIRecipe(StartV, DL);
   VPRegionBlock *TopRegion = Plan.getVectorLoopRegion();
   VPBasicBlock *Header = TopRegion->getEntryBasicBlock();
   if (IsVPlanNative)
     Header = cast<VPBasicBlock>(Header->getSingleSuccessor());
   Header->insert(CanonicalIVPHI, Header->begin());
 
   auto *CanonicalIVIncrement =
       new VPInstruction(HasNUW ? VPInstruction::CanonicalIVIncrementNUW
                                : VPInstruction::CanonicalIVIncrement,
                         {CanonicalIVPHI}, DL);
   CanonicalIVPHI->addOperand(CanonicalIVIncrement);
 
   VPBasicBlock *EB = TopRegion->getExitBasicBlock();
-  if (IsVPlanNative)
+  if (IsVPlanNative) {
     EB = cast<VPBasicBlock>(EB->getSinglePredecessor());
+    EB->setCondBit(nullptr);
+  }
   EB->appendRecipe(CanonicalIVIncrement);
+
+  auto *BranchOnCount =
+      new VPInstruction(VPInstruction::BranchOnCount,
+                        {CanonicalIVIncrement, &Plan.getVectorTripCount()}, DL);
+  EB->appendRecipe(BranchOnCount);
 }
 
 VPlanPtr LoopVectorizationPlanner::buildVPlanWithVPRecipes(
     VFRange &Range, SmallPtrSetImpl<Instruction *> &DeadInstructions,
     const MapVector<Instruction *, Instruction *> &SinkAfter) {
 
   SmallPtrSet<const InterleaveGroup<Instruction> *, 1> InterleaveGroups;
 
   VPRecipeBuilder RecipeBuilder(OrigLoop, TLI, Legal, CM, PSE, Builder);
for (Instruction *R : ReductionOperations) {
                "Expected no remaining users");
         CompareRecipe->eraseFromParent();
       }
       Chain = R;
     }
   }
 
   // If tail is folded by masking, introduce selects between the phi
-  // and the live-out instruction of each reduction, at the end of the latch.
+  // and the live-out instruction of each reduction, at the beginning of the
+  // dedicated latch block.
   if (CM.foldTailByMasking()) {
+    Builder.setInsertPoint(LatchVPBB, LatchVPBB->begin());
     for (VPRecipeBase &R : Plan->getEntry()->getEntryBasicBlock()->phis()) {
       VPReductionPHIRecipe *PhiR = dyn_cast<VPReductionPHIRecipe>(&R);
       if (!PhiR || PhiR->isInLoop())
         continue;
-      Builder.setInsertPoint(LatchVPBB);
       VPValue *Cond =
           RecipeBuilder.createBlockInMask(OrigLoop->getHeader(), Plan);
       VPValue *Red = PhiR->getBackedgeValue();
+      assert(cast<VPRecipeBase>(Red->getDef())->getParent() != LatchVPBB &&
+             "reduction recipe must be defined before latch");
       Builder.createNaryOp(Instruction::Select, {Cond, Red, PhiR});
     }
   }
 }
 
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
 void VPInterleaveRecipe::print(raw_ostream &O, const Twine &Indent,
                                VPSlotTracker &SlotTracker) const {

llvm/lib/Transforms/Vectorize/VPlan.h

FirstOrderRecurrenceSplice =
                                       // values of a first-order recurrence.
     Not,
     ICmpULE,
     SLPLoad,
     SLPStore,
     ActiveLaneMask,
     CanonicalIVIncrement,
     CanonicalIVIncrementNUW,
+    BranchOnCount,
   };
 
 private:
   typedef unsigned char OpcodeTy;
   OpcodeTy Opcode;
   FastMathFlags FMF;
   DebugLoc DL;
 
bool hasResult() const {
     case Instruction::Store:
     case Instruction::Switch:
     case Instruction::IndirectBr:
     case Instruction::Resume:
     case Instruction::CatchRet:
     case Instruction::Unreachable:
     case Instruction::Fence:
     case Instruction::AtomicRMW:
+    case VPInstruction::BranchOnCount:
       return false;
     default:
       return true;
     }
   }
 
   /// Set the fast-math flags.
   void setFastMathFlags(FastMathFlags FMFNew);

llvm/lib/Transforms/Vectorize/VPlan.cpp

if (Part == 0) {
       Next = Builder.CreateAdd(Phi, Step, "index.next", IsNUW, false);
     } else {
       Next = State.get(this, 0);
     }
 
     State.set(this, Next, Part);
     break;
   }
+  case VPInstruction::BranchOnCount: {
+    if (Part != 0)
+      break;
+    // First create the compare.
+    Value *IV = State.get(getOperand(0), Part);
+    Value *TC = State.get(getOperand(1), Part);
+    Value *Cond = Builder.CreateICmpEQ(IV, TC);
+
+    // Now create the branch.
+    auto *Plan = getParent()->getPlan();
+    VPRegionBlock *TopRegion = Plan->getVectorLoopRegion();
+    VPBasicBlock *Header = TopRegion->getEntry()->getEntryBasicBlock();
+    if (Header->empty()) {
+      assert(EnableVPlanNativePath &&
+             "empty entry block only expected in VPlanNativePath");
+      Header = cast<VPBasicBlock>(Header->getSingleSuccessor());
+    }
+    // TODO: Once the exit block is modeled in VPlan, use it instead of going
+    // through State.CFG.LastBB.
+    BasicBlock *Exit =
+        cast<BranchInst>(State.CFG.LastBB->getTerminator())->getSuccessor(0);
+
+    Builder.CreateCondBr(Cond, Exit, State.CFG.VPBB2IRBB[Header]);
+    Builder.GetInsertBlock()->getTerminator()->eraseFromParent();
+    break;
+  }
   default:
     llvm_unreachable("Unsupported opcode for instruction");
   }
 }
 
 void VPInstruction::execute(VPTransformState &State) {
   assert(!State.Instance && "VPInstruction executing an Instance");
   IRBuilderBase::FastMathFlagGuard FMFGuard(State.Builder);
case VPInstruction::FirstOrderRecurrenceSplice:
     O << "first-order splice";
     break;
   case VPInstruction::CanonicalIVIncrement:
     O << "VF * UF + ";
     break;
   case VPInstruction::CanonicalIVIncrementNUW:
     O << "VF * UF +(nuw) ";
     break;
+  case VPInstruction::BranchOnCount:
+    O << "branch-on-count ";
+    break;
   default:
     O << Instruction::getOpcodeName(getOpcode());
   }
 
   O << FMF;
 
   for (const VPValue *Operand : operands()) {
     O << " ";
for (VPBlockBase *SuccVPBlock : VPBB->getHierarchicalSuccessors()) {
       VPBasicBlock *SuccVPBB = SuccVPBlock->getEntryBasicBlock();
       BBTerminator->setSuccessor(Idx, State->CFG.VPBB2IRBB[SuccVPBB]);
       ++Idx;
     }
   }
 
   // 3. Merge the temporary latch created with the last basic-block filled.
   BasicBlock *LastBB = State->CFG.PrevBB;
+  assert(isa<BranchInst>(LastBB->getTerminator()) &&
+         "Expected VPlan CFG to terminate with branch");
+
+  // Move both the branch and check from LastBB to VectorLatchBB.
+  auto *LastBranch = cast<BranchInst>(LastBB->getTerminator());
+  LastBranch->moveBefore(VectorLatchBB->getTerminator());
+  VectorLatchBB->getTerminator()->eraseFromParent();
+  // Move condition so it is guaranteed to be next to branch. This is only done
+  // to avoid excessive test updates.
+  // TODO: Remove special handling once the increments for all inductions are
+  // modeled explicitly in VPlan.
+  cast<Instruction>(LastBranch->getCondition())->moveBefore(LastBranch);
   // Connect LastBB to VectorLatchBB to facilitate their merge.
-  assert((EnableVPlanNativePath ||
-          isa<UnreachableInst>(LastBB->getTerminator())) &&
-         "Expected InnerLoop VPlan CFG to terminate with unreachable");
-  assert((!EnableVPlanNativePath || isa<BranchInst>(LastBB->getTerminator())) &&
-         "Expected VPlan CFG to terminate with branch in NativePath");
-  LastBB->getTerminator()->eraseFromParent();
   BranchInst::Create(VectorLatchBB, LastBB);
 
   // Merge LastBB with Latch.
   bool Merged = MergeBlockIntoPredecessor(VectorLatchBB, nullptr, State->LI);
   (void)Merged;
   assert(Merged && "Could not merge last basic block with latch.");
   VectorLatchBB = LastBB;
 
   // Fix the latch value of canonical, reduction and first-order recurrences
   // phis in the vector loop.
for (VPRecipeBase &R : Header->phis()) {
     for (unsigned Part = 0; Part < LastPartForNewPhi; ++Part) {
       Value *Phi = State->get(PhiR, Part);
       Value *Val = State->get(PhiR->getBackedgeValue(),
                               SinglePartNeeded ? State->UF - 1 : Part);
       cast<PHINode>(Phi)->addIncoming(Val, VectorLatchBB);
     }
   }
 
-  // Add the loop exit condition and branch based on the canonical induction.
-  auto *CanonicalIV = getCanonicalIV();
-  // TODO: Model compare and branch explicitly in VPlan as recipes.
-  auto *Next = State->get(CanonicalIV->getBackedgeValue(), 0);
-  auto *TermBr = cast<BranchInst>(VectorLatchBB->getTerminator());
-  State->Builder.SetInsertPoint(TermBr);
-  auto *ICmp =
-      State->Builder.CreateICmpEQ(Next, State->get(&getVectorTripCount(), 0));
-  TermBr->setCondition(ICmp);
-
   // We do not attempt to preserve DT for outer loop vectorization currently.
   if (!EnableVPlanNativePath)
     updateDominatorTree(State->DT, VectorPreHeaderBB, VectorLatchBB,
                         L->getExitBlock());
 }
 
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
 LLVM_DUMP_METHOD
 void VPlan::print(raw_ostream &O) const {
   VPSlotTracker SlotTracker(this);
 
   O << "VPlan '" << Name << "' {";
 
-  assert(VectorTripCount.getNumUsers() == 0 &&
-         "should not be used yet in VPlan");
+  if (VectorTripCount.getNumUsers() > 0) {
+    O << "\nLive-in ";
+    VectorTripCount.printAsOperand(O, SlotTracker);
+    O << " = vector-trip-count\n";
+  }
+
   if (BackedgeTakenCount && BackedgeTakenCount->getNumUsers()) {
     O << "\nLive-in ";
     BackedgeTakenCount->printAsOperand(O, SlotTracker);
     O << " = backedge-taken count\n";
   }
 
   for (const VPBlockBase *Block : depth_first(getEntry())) {
     O << '\n';

llvm/lib/Transforms/Vectorize/VPlanVerifier.cpp

#endif
   }
 
   const VPBasicBlock *Exit = dyn_cast<VPBasicBlock>(TopRegion->getExit());
   if (!Exit) {
     errs() << "VPlan exit block is not a VPBasicBlock\n";
     return false;
   }
 
+  if (Exit->empty()) {
+    errs() << "VPlan vector loop exit must end with BranchOnCount "
+              "VPInstruction but is empty\n";
+    return false;
+  }
+
+  auto *LastInst = dyn_cast<VPInstruction>(std::prev(Exit->end()));
+  if (!LastInst || LastInst->getOpcode() != VPInstruction::BranchOnCount) {
+    errs() << "VPlan vector loop exit must end with BranchOnCount "
+              "VPInstruction\n";
+    return false;
+  }
+
   for (const VPRegionBlock *Region :
        VPBlockUtils::blocksOnly<const VPRegionBlock>(
            depth_first(VPBlockRecursiveTraversalWrapper<const VPBlockBase *>(
                Plan.getEntry())))) {
     if (Region->getEntry()->getNumPredecessors() != 0) {
       errs() << "region entry block has predecessors\n";
       return false;
     }
     if (Region->getExit()->getNumSuccessors() != 0) {
       errs() << "region exit block has successors\n";
       return false;
     }
   }
   return true;
 }

llvm/test/Transforms/LoopVectorize/AArch64/sve-tail-folding.ll

 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <vscale x 4 x i64> [[BROADCAST_SPLAT4]], [[TMP7]]
 ; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX1]], 0
 ; CHECK-NEXT:    [[TMP9:%.*]] = icmp ule <vscale x 4 x i64> [[INDUCTION]], [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr i32, i32* [[PTR:%.*]], i64 [[TMP8]]
 ; CHECK-NEXT:    [[TMP11:%.*]] = getelementptr i32, i32* [[TMP10]], i32 0
 ; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i32* [[TMP11]] to <vscale x 4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 4 x i32> @llvm.masked.load.nxv4i32.p0nxv4i32(<vscale x 4 x i32>* [[TMP12]], i32 4, <vscale x 4 x i1> [[TMP9]], <vscale x 4 x i32> poison)
 ; CHECK-NEXT:    [[TMP13]] = add <vscale x 4 x i32> [[VEC_PHI]], [[WIDE_MASKED_LOAD]]
+; CHECK-NEXT:    [[TMP16:%.*]] = select <vscale x 4 x i1> [[TMP9]], <vscale x 4 x i32> [[TMP13]], <vscale x 4 x i32> [[VEC_PHI]]
 ; CHECK-NEXT:    [[TMP14:%.*]] = call i64 @llvm.vscale.i64()
 ; CHECK-NEXT:    [[TMP15:%.*]] = mul i64 [[TMP14]], 4
 ; CHECK-NEXT:    [[INDEX_NEXT2]] = add i64 [[INDEX1]], [[TMP15]]
-; CHECK-NEXT:    [[TMP16:%.*]] = select <vscale x 4 x i1> [[TMP9]], <vscale x 4 x i32> [[TMP13]], <vscale x 4 x i32> [[VEC_PHI]]
 ; CHECK-NEXT:    [[TMP17:%.*]] = icmp eq i64 [[INDEX_NEXT2]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP17]], label %middle.block, label %vector.body, !llvm.loop [[LOOP20:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP18:%.*]] = call i32 @llvm.vector.reduce.add.nxv4i32(<vscale x 4 x i32> [[TMP16]])
 ; CHECK-NEXT:    br i1 true, label %while.end.loopexit, label %scalar.ph
 ;
 entry:
   br label %while.body

llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-gep.ll

 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; REQUIRES: asserts
 ; RUN: opt -loop-vectorize -S -mtriple=aarch64 -mattr=+sve -debug-only=loop-vectorize < %s 2>&1 | FileCheck %s
 
 target triple = "aarch64-unknown-linux-gnu"
 
 ; CHECK-LABEL:  LV: Checking a loop in "pointer_induction_used_as_vector"
 ; CHECK-NOT:    LV: Found {{.*}} scalar instruction:   %ptr.iv.2.next = getelementptr inbounds i8, i8* %ptr.iv.2, i64 1
 ;
 ; CHECK:        VPlan 'Initial VPlan for VF={vscale x 2},UF>=1' {
+; CHECK-NEXT:   Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT:   <x1> vector loop: {
 ; CHECK-NEXT:   loop.body:
 ; CHECK-NEXT:     EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:     WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:     WIDEN-PHI %ptr.iv.1 = phi %start.1, %ptr.iv.1.next
 ; CHECK-NEXT:     WIDEN-PHI %ptr.iv.2 = phi %start.2, %ptr.iv.2.next
 ; CHECK-NEXT:     WIDEN-GEP Var[Inv] ir<%ptr.iv.2.next> = getelementptr ir<%ptr.iv.2>, ir<1>
 ; CHECK-NEXT:     WIDEN store ir<%ptr.iv.1>, ir<%ptr.iv.2.next>
 ; CHECK-NEXT:     WIDEN ir<%lv> = load ir<%ptr.iv.2>
 ; CHECK-NEXT:     WIDEN ir<%add> = add ir<%lv>, ir<1>
 ; CHECK-NEXT:     WIDEN store ir<%ptr.iv.2>, ir<%add>
-; CHECK-NEXT:     EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:     EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:     EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT:   }
 
 ; In the test below the pointer phi %ptr.iv.2 is used as
 ;  1. As a uniform address for the load, and
 ;  2. Non-uniform use by the getelementptr which is stored. This requires the
 ;     vector value.
 define void @pointer_induction_used_as_vector(i8** noalias %start.1, i8* noalias %start.2, i64 %N) {

llvm/test/Transforms/LoopVectorize/ARM/mve-reduction-predselect.ll

 ; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds i32, i32* [[A:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[TMP0]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP1]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[B:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD1:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = mul <4 x i32> [[VEC_PHI]], [[WIDE_MASKED_LOAD]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = mul <4 x i32> [[TMP4]], [[WIDE_MASKED_LOAD1]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> [[TMP5]], <4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 260
 ; CHECK-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP8:%.*]] = call i32 @llvm.vector.reduce.mul.v4i32(<4 x i32> [[TMP6]])
 ; CHECK-NEXT:    br i1 true, label [[DOT_CRIT_EDGE:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    br label [[DOTLR_PH:%.*]]
 ; CHECK:       .lr.ph:

 ; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds i32, i32* [[A:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[TMP0]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP1]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[B:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD1:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = and <4 x i32> [[VEC_PHI]], [[WIDE_MASKED_LOAD]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = and <4 x i32> [[TMP4]], [[WIDE_MASKED_LOAD1]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> [[TMP5]], <4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 260
 ; CHECK-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP8:%.*]] = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> [[TMP6]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_END:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
 ; CHECK:       for.body:

 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 [[INDEX]], i32 257)
 ; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds i32, i32* [[A:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[TMP0]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP1]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[B:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD1:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = add nsw <4 x i32> [[WIDE_MASKED_LOAD1]], [[WIDE_MASKED_LOAD]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5:%.*]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> [[TMP4]], <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP6]] = or <4 x i32> [[VEC_PHI]], [[TMP5]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 260
 ; CHECK-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP8:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP6]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_END:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
 ; CHECK:       for.body:

 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 [[INDEX]], i32 257)
 ; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds i32, i32* [[A:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[TMP0]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP1]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[B:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD1:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = add nsw <4 x i32> [[WIDE_MASKED_LOAD1]], [[WIDE_MASKED_LOAD]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5:%.*]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> [[TMP4]], <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP6]] = xor <4 x i32> [[VEC_PHI]], [[TMP5]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 260
 ; CHECK-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP8:%.*]] = call i32 @llvm.vector.reduce.xor.v4i32(<4 x i32> [[TMP6]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_END:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
 ; CHECK:       for.body:

 ; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds float, float* [[A:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[TMP0]] to <4 x float>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* [[TMP1]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> poison)
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds float, float* [[B:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast float* [[TMP2]] to <4 x float>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD1:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = fadd fast <4 x float> [[VEC_PHI]], [[WIDE_MASKED_LOAD]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = fadd fast <4 x float> [[TMP4]], [[WIDE_MASKED_LOAD1]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> [[TMP5]], <4 x float> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 260
 ; CHECK-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP14:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP8:%.*]] = call fast float @llvm.vector.reduce.fadd.v4f32(float -0.000000e+00, <4 x float> [[TMP6]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_END:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
 ; CHECK:       for.body:

 ; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds float, float* [[A:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[TMP0]] to <4 x float>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* [[TMP1]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> poison)
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds float, float* [[B:%.*]], i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast float* [[TMP2]] to <4 x float>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD1:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = fmul fast <4 x float> [[VEC_PHI]], [[WIDE_MASKED_LOAD]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = fmul fast <4 x float> [[TMP4]], [[WIDE_MASKED_LOAD1]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> [[TMP5]], <4 x float> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 260
 ; CHECK-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP16:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP8:%.*]] = call fast float @llvm.vector.reduce.fmul.v4f32(float 1.000000e+00, <4 x float> [[TMP6]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_END:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
 ; CHECK:       for.body:

llvm/test/Transforms/LoopVectorize/ARM/mve-reduction-types.ll

 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[INDEX]], 0
 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 [[TMP0]], i32 [[N]])
 ; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds i32, i32* [[X:%.*]], i32 [[TMP0]]
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i32 0
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = mul <4 x i32> [[WIDE_MASKED_LOAD]], [[VEC_PHI]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> [[TMP4]], <4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.vector.reduce.mul.v4i32(<4 x i32> [[TMP5]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 1, [[FOR_BODY_PREHEADER]] ], [ [[TMP7]], [[MIDDLE_BLOCK]] ]

 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[INDEX]], 0
 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 [[TMP0]], i32 [[N]])
 ; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds i32, i32* [[X:%.*]], i32 [[TMP0]]
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i32 0
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = and <4 x i32> [[WIDE_MASKED_LOAD]], [[VEC_PHI]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> [[TMP4]], <4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> [[TMP5]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ -1, [[FOR_BODY_PREHEADER]] ], [ [[TMP7]], [[MIDDLE_BLOCK]] ]

 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[INDEX]], 0
 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 [[TMP0]], i32 [[N]])
 ; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds i32, i32* [[X:%.*]], i32 [[TMP0]]
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i32 0
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = or <4 x i32> [[WIDE_MASKED_LOAD]], [[VEC_PHI]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> [[TMP4]], <4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> [[TMP5]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[FOR_BODY_PREHEADER]] ], [ [[TMP7]], [[MIDDLE_BLOCK]] ]

 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[INDEX]], 0
 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 [[TMP0]], i32 [[N]])
 ; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds i32, i32* [[X:%.*]], i32 [[TMP0]]
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i32 0
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0v4i32(<4 x i32>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = xor <4 x i32> [[WIDE_MASKED_LOAD]], [[VEC_PHI]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x i32> [[TMP4]], <4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP14:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.vector.reduce.xor.v4i32(<4 x i32> [[TMP5]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[FOR_BODY_PREHEADER]] ], [ [[TMP7]], [[MIDDLE_BLOCK]] ]

 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[INDEX]], 0
 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 [[TMP0]], i32 [[N]])
 ; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds float, float* [[X:%.*]], i32 [[TMP0]]
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds float, float* [[TMP1]], i32 0
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast float* [[TMP2]] to <4 x float>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = fadd fast <4 x float> [[WIDE_MASKED_LOAD]], [[VEC_PHI]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> [[TMP4]], <4 x float> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP16:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP7:%.*]] = call fast float @llvm.vector.reduce.fadd.v4f32(float -0.000000e+00, <4 x float> [[TMP5]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi float [ 0.000000e+00, [[FOR_BODY_PREHEADER]] ], [ [[TMP7]], [[MIDDLE_BLOCK]] ]

 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[INDEX]], 0
 ; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 [[TMP0]], i32 [[N]])
 ; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds float, float* [[X:%.*]], i32 [[TMP0]]
 ; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds float, float* [[TMP1]], i32 0
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast float* [[TMP2]] to <4 x float>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0v4f32(<4 x float>* [[TMP3]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> poison)
 ; CHECK-NEXT:    [[TMP4:%.*]] = fmul fast <4 x float> [[WIDE_MASKED_LOAD]], [[VEC_PHI]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5]] = select <4 x i1> [[ACTIVE_LANE_MASK]], <4 x float> [[TMP4]], <4 x float> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP18:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP7:%.*]] = call fast float @llvm.vector.reduce.fmul.v4f32(float 1.000000e+00, <4 x float> [[TMP5]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi float [ 1.000000e+00, [[FOR_BODY_PREHEADER]] ], [ [[TMP7]], [[MIDDLE_BLOCK]] ]

llvm/test/Transforms/LoopVectorize/X86/invariant-store-vectorization.ll

 ; CHECK:       vec.epilog.iter.check:
 ; CHECK-NEXT:    [[N_VEC_REMAINING:%.*]] = and i64 [[SMAX6]], 8
 ; CHECK-NEXT:    [[MIN_EPILOG_ITERS_CHECK_NOT_NOT:%.*]] = icmp eq i64 [[N_VEC_REMAINING]], 0
 ; CHECK-NEXT:    br i1 [[MIN_EPILOG_ITERS_CHECK_NOT_NOT]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
 ; CHECK:       vec.epilog.ph:
 ; CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
 ; CHECK-NEXT:    [[SMAX12:%.*]] = call i64 @llvm.smax.i64(i64 [[N]], i64 1)
 ; CHECK-NEXT:    [[N_VEC14:%.*]] = and i64 [[SMAX12]], 9223372036854775800
-; CHECK-NEXT:    [[BROADCAST_SPLATINSERT19:%.*]] = insertelement <8 x i32> poison, i32 [[K]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT20:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT19]], <8 x i32> poison, <8 x i32> zeroinitializer
-; CHECK-NEXT:    [[BROADCAST_SPLATINSERT21:%.*]] = insertelement <8 x i32> poison, i32 [[NTRUNC]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT22:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT21]], <8 x i32> poison, <8 x i32> zeroinitializer
-; CHECK-NEXT:    [[BROADCAST_SPLATINSERT23:%.*]] = insertelement <8 x i32*> poison, i32* [[A]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT24:%.*]] = shufflevector <8 x i32*> [[BROADCAST_SPLATINSERT23]], <8 x i32*> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT18:%.*]] = insertelement <8 x i32> poison, i32 [[K]], i64 0
+; CHECK-NEXT:    [[BROADCAST_SPLAT19:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT18]], <8 x i32> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT20:%.*]] = insertelement <8 x i32> poison, i32 [[NTRUNC]], i64 0
+; CHECK-NEXT:    [[BROADCAST_SPLAT21:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT20]], <8 x i32> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT22:%.*]] = insertelement <8 x i32*> poison, i32* [[A]], i64 0
+; CHECK-NEXT:    [[BROADCAST_SPLAT23:%.*]] = shufflevector <8 x i32*> [[BROADCAST_SPLATINSERT22]], <8 x i32*> poison, <8 x i32> zeroinitializer
 ; CHECK-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
 ; CHECK:       vec.epilog.vector.body:
-; CHECK-NEXT:    [[INDEX15:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT16:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
-; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[INDEX15]]
+; CHECK-NEXT:    [[INDEX16:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT24:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[INDEX16]]
 ; CHECK-NEXT:    [[TMP6:%.*]] = bitcast i32* [[TMP5]] to <8 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD18:%.*]] = load <8 x i32>, <8 x i32>* [[TMP6]], align 8
-; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq <8 x i32> [[WIDE_LOAD18]], [[BROADCAST_SPLAT20]]
+; CHECK-NEXT:    [[WIDE_LOAD17:%.*]] = load <8 x i32>, <8 x i32>* [[TMP6]], align 8
+; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq <8 x i32> [[WIDE_LOAD17]], [[BROADCAST_SPLAT19]]
 ; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i32* [[TMP5]] to <8 x i32>*
-; CHECK-NEXT:    store <8 x i32> [[BROADCAST_SPLAT22]], <8 x i32>* [[TMP8]], align 4
-; CHECK-NEXT:    call void @llvm.masked.scatter.v8i32.v8p0i32(<8 x i32> [[BROADCAST_SPLAT22]], <8 x i32*> [[BROADCAST_SPLAT24]], i32 4, <8 x i1> [[TMP7]])
-; CHECK-NEXT:    [[INDEX_NEXT16]] = add nuw i64 [[INDEX15]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT16]], [[N_VEC14]]
+; CHECK-NEXT:    store <8 x i32> [[BROADCAST_SPLAT21]], <8 x i32>* [[TMP8]], align 4
+; CHECK-NEXT:    call void @llvm.masked.scatter.v8i32.v8p0i32(<8 x i32> [[BROADCAST_SPLAT21]], <8 x i32*> [[BROADCAST_SPLAT23]], i32 4, <8 x i1> [[TMP7]])
+; CHECK-NEXT:    [[INDEX_NEXT24]] = add nuw i64 [[INDEX16]], 8
+; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT24]], [[N_VEC14]]
 ; CHECK-NEXT:    br i1 [[TMP9]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP14:![0-9]+]]
 ; CHECK:       vec.epilog.middle.block:
-; CHECK-NEXT:    [[CMP_N17:%.*]] = icmp eq i64 [[SMAX12]], [[N_VEC14]]
-; CHECK-NEXT:    br i1 [[CMP_N17]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
+; CHECK-NEXT:    [[CMP_N15:%.*]] = icmp eq i64 [[SMAX12]], [[N_VEC14]]
+; CHECK-NEXT:    br i1 [[CMP_N15]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
 ; CHECK:       vec.epilog.scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC14]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MEMCHECK]] ], [ 0, [[ITER_CHECK:%.*]] ]
 ; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
 ; CHECK:       for.body:
 ; CHECK-NEXT:    [[I:%.*]] = phi i64 [ [[I_NEXT:%.*]], [[LATCH:%.*]] ], [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ]
 ; CHECK-NEXT:    [[T1:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[I]]
 ; CHECK-NEXT:    [[T2:%.*]] = load i32, i32* [[T1]], align 8
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[T2]], [[K]]
 ; CHECK-NEXT:    store i32 [[NTRUNC]], i32* [[T1]], align 4
 ; CHECK-NEXT:    br i1 [[CMP]], label [[COND_STORE:%.*]], label [[LATCH]]
 ; CHECK:       cond_store:
 ; CHECK-NEXT:    store i32 [[NTRUNC]], i32* [[A]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[I_NEXT]] = add nuw nsw i64 [[I]], 1
 ; CHECK-NEXT:    [[COND:%.*]] = icmp slt i64 [[I_NEXT]], [[N]]
-; CHECK-NEXT:    br i1 [[COND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT]], !llvm.loop [[LOOP16:![0-9]+]]
+; CHECK-NEXT:    br i1 [[COND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT]], !llvm.loop [[LOOP15:![0-9]+]]
 ; CHECK:       for.end.loopexit:
 ; CHECK-NEXT:    br label [[FOR_END]]
 ; CHECK:       for.end:
 ; CHECK-NEXT:    ret void
 ;
 entry:
   %ntrunc = trunc i64 %n to i32
   br label %for.body

 ; CHECK-NEXT:    [[BROADCAST_SPLAT19:%.*]] = shufflevector <16 x i32> [[BROADCAST_SPLATINSERT18]], <16 x i32> poison, <16 x i32> zeroinitializer
 ; CHECK-NEXT:    [[BROADCAST_SPLATINSERT20:%.*]] = insertelement <16 x i32*> poison, i32* [[A]], i64 0
 ; CHECK-NEXT:    [[BROADCAST_SPLAT21:%.*]] = shufflevector <16 x i32*> [[BROADCAST_SPLATINSERT20]], <16 x i32*> poison, <16 x i32> zeroinitializer
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[TMP0]] to <16 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <16 x i32>, <16 x i32>* [[TMP1]], align 8, !alias.scope !17, !noalias !20
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <16 x i32>, <16 x i32>* [[TMP1]], align 8, !alias.scope !16, !noalias !19
 ; CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <16 x i32> [[WIDE_LOAD]], [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast i32* [[TMP0]] to <16 x i32>*
-; CHECK-NEXT:    store <16 x i32> [[BROADCAST_SPLAT19]], <16 x i32>* [[TMP3]], align 4, !alias.scope !17, !noalias !20
+; CHECK-NEXT:    store <16 x i32> [[BROADCAST_SPLAT19]], <16 x i32>* [[TMP3]], align 4, !alias.scope !16, !noalias !19
 ; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr i32, i32* [[C]], i64 [[INDEX]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = bitcast i32* [[TMP4]] to <16 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p0v16i32(<16 x i32>* [[TMP5]], i32 8, <16 x i1> [[TMP2]], <16 x i32> poison), !alias.scope !23
-; CHECK-NEXT:    call void @llvm.masked.scatter.v16i32.v16p0i32(<16 x i32> [[WIDE_MASKED_LOAD]], <16 x i32*> [[BROADCAST_SPLAT21]], i32 4, <16 x i1> [[TMP2]]), !alias.scope !24, !noalias !23
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p0v16i32(<16 x i32>* [[TMP5]], i32 8, <16 x i1> [[TMP2]], <16 x i32> poison), !alias.scope !22
+; CHECK-NEXT:    call void @llvm.masked.scatter.v16i32.v16p0i32(<16 x i32> [[WIDE_MASKED_LOAD]], <16 x i32*> [[BROADCAST_SPLAT21]], i32 4, <16 x i1> [[TMP2]]), !alias.scope !23, !noalias !22
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP25:![0-9]+]]
+; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP24:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[SMAX16]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
 ; CHECK:       vec.epilog.iter.check:
 ; CHECK-NEXT:    [[N_VEC_REMAINING:%.*]] = and i64 [[SMAX16]], 8
 ; CHECK-NEXT:    [[MIN_EPILOG_ITERS_CHECK_NOT_NOT:%.*]] = icmp eq i64 [[N_VEC_REMAINING]], 0
 ; CHECK-NEXT:    br i1 [[MIN_EPILOG_ITERS_CHECK_NOT_NOT]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
 ; CHECK:       vec.epilog.ph:
 ; CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
 ; CHECK-NEXT:    [[SMAX22:%.*]] = call i64 @llvm.smax.i64(i64 [[N]], i64 1)
 ; CHECK-NEXT:    [[N_VEC24:%.*]] = and i64 [[SMAX22]], 9223372036854775800
-; CHECK-NEXT:    [[BROADCAST_SPLATINSERT29:%.*]] = insertelement <8 x i32> poison, i32 [[K]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT30:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT29]], <8 x i32> poison, <8 x i32> zeroinitializer
-; CHECK-NEXT:    [[BROADCAST_SPLATINSERT31:%.*]] = insertelement <8 x i32> poison, i32 [[NTRUNC]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT32:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT31]], <8 x i32> poison, <8 x i32> zeroinitializer
-; CHECK-NEXT:    [[BROADCAST_SPLATINSERT34:%.*]] = insertelement <8 x i32*> poison, i32* [[A]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT35:%.*]] = shufflevector <8 x i32*> [[BROADCAST_SPLATINSERT34]], <8 x i32*> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT28:%.*]] = insertelement <8 x i32> poison, i32 [[K]], i64 0
+; CHECK-NEXT:    [[BROADCAST_SPLAT29:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT28]], <8 x i32> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT30:%.*]] = insertelement <8 x i32> poison, i32 [[NTRUNC]], i64 0
+; CHECK-NEXT:    [[BROADCAST_SPLAT31:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT30]], <8 x i32> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT33:%.*]] = insertelement <8 x i32*> poison, i32* [[A]], i64 0
+; CHECK-NEXT:    [[BROADCAST_SPLAT34:%.*]] = shufflevector <8 x i32*> [[BROADCAST_SPLATINSERT33]], <8 x i32*> poison, <8 x i32> zeroinitializer
 ; CHECK-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
 ; CHECK:       vec.epilog.vector.body:
-; CHECK-NEXT:    [[INDEX25:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT26:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
-; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[INDEX25]]
+; CHECK-NEXT:    [[INDEX26:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT35:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[INDEX26]]
 ; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i32* [[TMP7]] to <8 x i32>*
-; CHECK-NEXT:    [[WIDE_LOAD28:%.*]] = load <8 x i32>, <8 x i32>* [[TMP8]], align 8
-; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq <8 x i32> [[WIDE_LOAD28]], [[BROADCAST_SPLAT30]]
+; CHECK-NEXT:    [[WIDE_LOAD27:%.*]] = load <8 x i32>, <8 x i32>* [[TMP8]], align 8
+; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq <8 x i32> [[WIDE_LOAD27]], [[BROADCAST_SPLAT29]]
 ; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i32* [[TMP7]] to <8 x i32>*
-; CHECK-NEXT:    store <8 x i32> [[BROADCAST_SPLAT32]], <8 x i32>* [[TMP10]], align 4
-; CHECK-NEXT:    [[TMP11:%.*]] = getelementptr i32, i32* [[C]], i64 [[INDEX25]]
+; CHECK-NEXT:    store <8 x i32> [[BROADCAST_SPLAT31]], <8 x i32>* [[TMP10]], align 4
+; CHECK-NEXT:    [[TMP11:%.*]] = getelementptr i32, i32* [[C]], i64 [[INDEX26]]
 ; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i32* [[TMP11]] to <8 x i32>*
-; CHECK-NEXT:    [[WIDE_MASKED_LOAD33:%.*]] = call <8 x i32> @llvm.masked.load.v8i32.p0v8i32(<8 x i32>* [[TMP12]], i32 8, <8 x i1> [[TMP9]], <8 x i32> poison)
-; CHECK-NEXT:    call void @llvm.masked.scatter.v8i32.v8p0i32(<8 x i32> [[WIDE_MASKED_LOAD33]], <8 x i32*> [[BROADCAST_SPLAT35]], i32 4, <8 x i1> [[TMP9]])
-; CHECK-NEXT:    [[INDEX_NEXT26]] = add nuw i64 [[INDEX25]], 8
-; CHECK-NEXT:    [[TMP13:%.*]] = icmp eq i64 [[INDEX_NEXT26]], [[N_VEC24]]
-; CHECK-NEXT:    br i1 [[TMP13]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP26:![0-9]+]]
+; CHECK-NEXT:    [[WIDE_MASKED_LOAD32:%.*]] = call <8 x i32> @llvm.masked.load.v8i32.p0v8i32(<8 x i32>* [[TMP12]], i32 8, <8 x i1> [[TMP9]], <8 x i32> poison)
+; CHECK-NEXT:    call void @llvm.masked.scatter.v8i32.v8p0i32(<8 x i32> [[WIDE_MASKED_LOAD32]], <8 x i32*> [[BROADCAST_SPLAT34]], i32 4, <8 x i1> [[TMP9]])
+; CHECK-NEXT:    [[INDEX_NEXT35]] = add nuw i64 [[INDEX26]], 8
+; CHECK-NEXT:    [[TMP13:%.*]] = icmp eq i64 [[INDEX_NEXT35]], [[N_VEC24]]
+; CHECK-NEXT:    br i1 [[TMP13]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP25:![0-9]+]]
 ; CHECK:       vec.epilog.middle.block:
-; CHECK-NEXT:    [[CMP_N27:%.*]] = icmp eq i64 [[SMAX22]], [[N_VEC24]]
-; CHECK-NEXT:    br i1 [[CMP_N27]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
+; CHECK-NEXT:    [[CMP_N25:%.*]] = icmp eq i64 [[SMAX22]], [[N_VEC24]]
+; CHECK-NEXT:    br i1 [[CMP_N25]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
 ; CHECK:       vec.epilog.scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC24]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MEMCHECK]] ], [ 0, [[ITER_CHECK:%.*]] ]
 ; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
 ; CHECK:       for.body:
 ; CHECK-NEXT:    [[I:%.*]] = phi i64 [ [[I_NEXT:%.*]], [[LATCH:%.*]] ], [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ]
 ; CHECK-NEXT:    [[T1:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[I]]
 ; CHECK-NEXT:    [[T2:%.*]] = load i32, i32* [[T1]], align 8
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[T2]], [[K]]
 ; CHECK-NEXT:    store i32 [[NTRUNC]], i32* [[T1]], align 4
 ; CHECK-NEXT:    br i1 [[CMP]], label [[COND_STORE:%.*]], label [[LATCH]]
 ; CHECK:       cond_store:
 ; CHECK-NEXT:    [[T3:%.*]] = getelementptr inbounds i32, i32* [[C]], i64 [[I]]
 ; CHECK-NEXT:    [[T4:%.*]] = load i32, i32* [[T3]], align 8
 ; CHECK-NEXT:    store i32 [[T4]], i32* [[A]], align 4
 ; CHECK-NEXT:    br label [[LATCH]]
 ; CHECK:       latch:
 ; CHECK-NEXT:    [[I_NEXT]] = add nuw nsw i64 [[I]], 1
 ; CHECK-NEXT:    [[COND:%.*]] = icmp slt i64 [[I_NEXT]], [[N]]
-; CHECK-NEXT:    br i1 [[COND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT]], !llvm.loop [[LOOP27:![0-9]+]]
+; CHECK-NEXT:    br i1 [[COND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT]], !llvm.loop [[LOOP26:![0-9]+]]
 ; CHECK:       for.end.loopexit:
 ; CHECK-NEXT:    br label [[FOR_END]]
 ; CHECK:       for.end:
 ; CHECK-NEXT:    ret void
 ;
 entry:
   %ntrunc = trunc i64 %n to i32
   br label %for.body

llvm/test/Transforms/LoopVectorize/X86/masked_load_store.ll

 ; AVX512-NEXT:    [[CMP_N:%.*]] = icmp eq i64 10000, 9984
 ; AVX512-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
 ; AVX512:       vec.epilog.iter.check:
 ; AVX512-NEXT:    br i1 false, label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
 ; AVX512:       vec.epilog.ph:
 ; AVX512-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ 9984, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
 ; AVX512-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
 ; AVX512:       vec.epilog.vector.body:
-; AVX512-NEXT:    [[INDEX18:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT19:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
-; AVX512-NEXT:    [[TMP49:%.*]] = add i64 [[INDEX18]], 0
+; AVX512-NEXT:    [[INDEX19:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT22:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
+; AVX512-NEXT:    [[TMP49:%.*]] = add i64 [[INDEX19]], 0
 ; AVX512-NEXT:    [[TMP50:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP49]]
 ; AVX512-NEXT:    [[TMP51:%.*]] = getelementptr inbounds i32, i32* [[TMP50]], i32 0
 ; AVX512-NEXT:    [[TMP52:%.*]] = bitcast i32* [[TMP51]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD21:%.*]] = load <8 x i32>, <8 x i32>* [[TMP52]], align 4
-; AVX512-NEXT:    [[TMP53:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD21]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
+; AVX512-NEXT:    [[WIDE_LOAD20:%.*]] = load <8 x i32>, <8 x i32>* [[TMP52]], align 4
+; AVX512-NEXT:    [[TMP53:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD20]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP54:%.*]] = getelementptr i32, i32* [[B]], i64 [[TMP49]]
 ; AVX512-NEXT:    [[TMP55:%.*]] = getelementptr i32, i32* [[TMP54]], i32 0
 ; AVX512-NEXT:    [[TMP56:%.*]] = bitcast i32* [[TMP55]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD22:%.*]] = call <8 x i32> @llvm.masked.load.v8i32.p0v8i32(<8 x i32>* [[TMP56]], i32 4, <8 x i1> [[TMP53]], <8 x i32> poison)
-; AVX512-NEXT:    [[TMP57:%.*]] = add nsw <8 x i32> [[WIDE_MASKED_LOAD22]], [[WIDE_LOAD21]]
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD21:%.*]] = call <8 x i32> @llvm.masked.load.v8i32.p0v8i32(<8 x i32>* [[TMP56]], i32 4, <8 x i1> [[TMP53]], <8 x i32> poison)
+; AVX512-NEXT:    [[TMP57:%.*]] = add nsw <8 x i32> [[WIDE_MASKED_LOAD21]], [[WIDE_LOAD20]]
 ; AVX512-NEXT:    [[TMP58:%.*]] = getelementptr i32, i32* [[A]], i64 [[TMP49]]
 ; AVX512-NEXT:    [[TMP59:%.*]] = getelementptr i32, i32* [[TMP58]], i32 0
 ; AVX512-NEXT:    [[TMP60:%.*]] = bitcast i32* [[TMP59]] to <8 x i32>*
 ; AVX512-NEXT:    call void @llvm.masked.store.v8i32.p0v8i32(<8 x i32> [[TMP57]], <8 x i32>* [[TMP60]], i32 4, <8 x i1> [[TMP53]])
-; AVX512-NEXT:    [[INDEX_NEXT19]] = add nuw i64 [[INDEX18]], 8
-; AVX512-NEXT:    [[TMP61:%.*]] = icmp eq i64 [[INDEX_NEXT19]], 10000
+; AVX512-NEXT:    [[INDEX_NEXT22]] = add nuw i64 [[INDEX19]], 8
+; AVX512-NEXT:    [[TMP61:%.*]] = icmp eq i64 [[INDEX_NEXT22]], 10000
 ; AVX512-NEXT:    br i1 [[TMP61]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
 ; AVX512:       vec.epilog.middle.block:
-; AVX512-NEXT:    [[CMP_N20:%.*]] = icmp eq i64 10000, 10000
-; AVX512-NEXT:    br i1 [[CMP_N20]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
+; AVX512-NEXT:    [[CMP_N18:%.*]] = icmp eq i64 10000, 10000
+; AVX512-NEXT:    br i1 [[CMP_N18]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
 ; AVX512:       vec.epilog.scalar.ph:
 ; AVX512-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 10000, [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ 9984, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MEMCHECK]] ], [ 0, [[ITER_CHECK:%.*]] ]
 ; AVX512-NEXT:    br label [[FOR_BODY:%.*]]
 ; AVX512:       for.body:
 ; AVX512-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_INC:%.*]] ]
 ; AVX512-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP62:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
 ; AVX512-NEXT:    [[CMP1:%.*]] = icmp slt i32 [[TMP62]], 100
 ; AVX512-NEXT:    br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
 ; AVX512:       if.then:
 ; AVX512-NEXT:    [[ARRAYIDX3:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP63:%.*]] = load i32, i32* [[ARRAYIDX3]], align 4
 ; AVX512-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP63]], [[TMP62]]
 ; AVX512-NEXT:    [[ARRAYIDX7:%.*]] = getelementptr inbounds i32, i32* [[A]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    store i32 [[ADD]], i32* [[ARRAYIDX7]], align 4
 ; AVX512-NEXT:    br label [[FOR_INC]]
 ; AVX512:       for.inc:
 ; AVX512-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
 ; AVX512-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 10000
-; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
+; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP11:![0-9]+]]
 ; AVX512:       for.end.loopexit:
 ; AVX512-NEXT:    br label [[FOR_END]]
 ; AVX512:       for.end:
 ; AVX512-NEXT:    ret void
 ;
 entry:
   br label %for.body
 

 ; AVX512-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 32
 ; AVX512-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 48
 ; AVX512-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TRIGGER]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TRIGGER]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TRIGGER]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TRIGGER]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TMP4]], i32 0
 ; AVX512-NEXT:    [[TMP9:%.*]] = bitcast i32 addrspace(1)* [[TMP8]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_LOAD:%.*]] = load <16 x i32>, <16 x i32> addrspace(1)* [[TMP9]], align 4, !alias.scope !13
+; AVX512-NEXT:    [[WIDE_LOAD:%.*]] = load <16 x i32>, <16 x i32> addrspace(1)* [[TMP9]], align 4, !alias.scope !12
 ; AVX512-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TMP4]], i32 16
 ; AVX512-NEXT:    [[TMP11:%.*]] = bitcast i32 addrspace(1)* [[TMP10]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_LOAD12:%.*]] = load <16 x i32>, <16 x i32> addrspace(1)* [[TMP11]], align 4, !alias.scope !13
+; AVX512-NEXT:    [[WIDE_LOAD12:%.*]] = load <16 x i32>, <16 x i32> addrspace(1)* [[TMP11]], align 4, !alias.scope !12
 ; AVX512-NEXT:    [[TMP12:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TMP4]], i32 32
 ; AVX512-NEXT:    [[TMP13:%.*]] = bitcast i32 addrspace(1)* [[TMP12]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_LOAD13:%.*]] = load <16 x i32>, <16 x i32> addrspace(1)* [[TMP13]], align 4, !alias.scope !13
+; AVX512-NEXT:    [[WIDE_LOAD13:%.*]] = load <16 x i32>, <16 x i32> addrspace(1)* [[TMP13]], align 4, !alias.scope !12
 ; AVX512-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TMP4]], i32 48
 ; AVX512-NEXT:    [[TMP15:%.*]] = bitcast i32 addrspace(1)* [[TMP14]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_LOAD14:%.*]] = load <16 x i32>, <16 x i32> addrspace(1)* [[TMP15]], align 4, !alias.scope !13
+; AVX512-NEXT:    [[WIDE_LOAD14:%.*]] = load <16 x i32>, <16 x i32> addrspace(1)* [[TMP15]], align 4, !alias.scope !12
 ; AVX512-NEXT:    [[TMP16:%.*]] = icmp slt <16 x i32> [[WIDE_LOAD]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP17:%.*]] = icmp slt <16 x i32> [[WIDE_LOAD12]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP18:%.*]] = icmp slt <16 x i32> [[WIDE_LOAD13]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP19:%.*]] = icmp slt <16 x i32> [[WIDE_LOAD14]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP20:%.*]] = getelementptr i32, i32 addrspace(1)* [[B]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP21:%.*]] = getelementptr i32, i32 addrspace(1)* [[B]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP22:%.*]] = getelementptr i32, i32 addrspace(1)* [[B]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP23:%.*]] = getelementptr i32, i32 addrspace(1)* [[B]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP24:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP20]], i32 0
 ; AVX512-NEXT:    [[TMP25:%.*]] = bitcast i32 addrspace(1)* [[TMP24]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p1v16i32(<16 x i32> addrspace(1)* [[TMP25]], i32 4, <16 x i1> [[TMP16]], <16 x i32> poison), !alias.scope !16
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p1v16i32(<16 x i32> addrspace(1)* [[TMP25]], i32 4, <16 x i1> [[TMP16]], <16 x i32> poison), !alias.scope !15
 ; AVX512-NEXT:    [[TMP26:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP20]], i32 16
 ; AVX512-NEXT:    [[TMP27:%.*]] = bitcast i32 addrspace(1)* [[TMP26]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD15:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p1v16i32(<16 x i32> addrspace(1)* [[TMP27]], i32 4, <16 x i1> [[TMP17]], <16 x i32> poison), !alias.scope !16
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD15:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p1v16i32(<16 x i32> addrspace(1)* [[TMP27]], i32 4, <16 x i1> [[TMP17]], <16 x i32> poison), !alias.scope !15
 ; AVX512-NEXT:    [[TMP28:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP20]], i32 32
 ; AVX512-NEXT:    [[TMP29:%.*]] = bitcast i32 addrspace(1)* [[TMP28]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD16:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p1v16i32(<16 x i32> addrspace(1)* [[TMP29]], i32 4, <16 x i1> [[TMP18]], <16 x i32> poison), !alias.scope !16
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD16:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p1v16i32(<16 x i32> addrspace(1)* [[TMP29]], i32 4, <16 x i1> [[TMP18]], <16 x i32> poison), !alias.scope !15
 ; AVX512-NEXT:    [[TMP30:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP20]], i32 48
 ; AVX512-NEXT:    [[TMP31:%.*]] = bitcast i32 addrspace(1)* [[TMP30]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD17:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p1v16i32(<16 x i32> addrspace(1)* [[TMP31]], i32 4, <16 x i1> [[TMP19]], <16 x i32> poison), !alias.scope !16
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD17:%.*]] = call <16 x i32> @llvm.masked.load.v16i32.p1v16i32(<16 x i32> addrspace(1)* [[TMP31]], i32 4, <16 x i1> [[TMP19]], <16 x i32> poison), !alias.scope !15
 ; AVX512-NEXT:    [[TMP32:%.*]] = add nsw <16 x i32> [[WIDE_MASKED_LOAD]], [[WIDE_LOAD]]
 ; AVX512-NEXT:    [[TMP33:%.*]] = add nsw <16 x i32> [[WIDE_MASKED_LOAD15]], [[WIDE_LOAD12]]
 ; AVX512-NEXT:    [[TMP34:%.*]] = add nsw <16 x i32> [[WIDE_MASKED_LOAD16]], [[WIDE_LOAD13]]
 ; AVX512-NEXT:    [[TMP35:%.*]] = add nsw <16 x i32> [[WIDE_MASKED_LOAD17]], [[WIDE_LOAD14]]
 ; AVX512-NEXT:    [[TMP36:%.*]] = getelementptr i32, i32 addrspace(1)* [[A]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP37:%.*]] = getelementptr i32, i32 addrspace(1)* [[A]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP38:%.*]] = getelementptr i32, i32 addrspace(1)* [[A]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP39:%.*]] = getelementptr i32, i32 addrspace(1)* [[A]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP40:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP36]], i32 0
 ; AVX512-NEXT:    [[TMP41:%.*]] = bitcast i32 addrspace(1)* [[TMP40]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    call void @llvm.masked.store.v16i32.p1v16i32(<16 x i32> [[TMP32]], <16 x i32> addrspace(1)* [[TMP41]], i32 4, <16 x i1> [[TMP16]]), !alias.scope !18, !noalias !20
+; AVX512-NEXT:    call void @llvm.masked.store.v16i32.p1v16i32(<16 x i32> [[TMP32]], <16 x i32> addrspace(1)* [[TMP41]], i32 4, <16 x i1> [[TMP16]]), !alias.scope !17, !noalias !19
 ; AVX512-NEXT:    [[TMP42:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP36]], i32 16
 ; AVX512-NEXT:    [[TMP43:%.*]] = bitcast i32 addrspace(1)* [[TMP42]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    call void @llvm.masked.store.v16i32.p1v16i32(<16 x i32> [[TMP33]], <16 x i32> addrspace(1)* [[TMP43]], i32 4, <16 x i1> [[TMP17]]), !alias.scope !18, !noalias !20
+; AVX512-NEXT:    call void @llvm.masked.store.v16i32.p1v16i32(<16 x i32> [[TMP33]], <16 x i32> addrspace(1)* [[TMP43]], i32 4, <16 x i1> [[TMP17]]), !alias.scope !17, !noalias !19
 ; AVX512-NEXT:    [[TMP44:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP36]], i32 32
 ; AVX512-NEXT:    [[TMP45:%.*]] = bitcast i32 addrspace(1)* [[TMP44]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    call void @llvm.masked.store.v16i32.p1v16i32(<16 x i32> [[TMP34]], <16 x i32> addrspace(1)* [[TMP45]], i32 4, <16 x i1> [[TMP18]]), !alias.scope !18, !noalias !20
+; AVX512-NEXT:    call void @llvm.masked.store.v16i32.p1v16i32(<16 x i32> [[TMP34]], <16 x i32> addrspace(1)* [[TMP45]], i32 4, <16 x i1> [[TMP18]]), !alias.scope !17, !noalias !19
 ; AVX512-NEXT:    [[TMP46:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP36]], i32 48
 ; AVX512-NEXT:    [[TMP47:%.*]] = bitcast i32 addrspace(1)* [[TMP46]] to <16 x i32> addrspace(1)*
-; AVX512-NEXT:    call void @llvm.masked.store.v16i32.p1v16i32(<16 x i32> [[TMP35]], <16 x i32> addrspace(1)* [[TMP47]], i32 4, <16 x i1> [[TMP19]]), !alias.scope !18, !noalias !20
+; AVX512-NEXT:    call void @llvm.masked.store.v16i32.p1v16i32(<16 x i32> [[TMP35]], <16 x i32> addrspace(1)* [[TMP47]], i32 4, <16 x i1> [[TMP19]]), !alias.scope !17, !noalias !19
 ; AVX512-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 64
 ; AVX512-NEXT:    [[TMP48:%.*]] = icmp eq i64 [[INDEX_NEXT]], 9984
-; AVX512-NEXT:    br i1 [[TMP48]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP21:![0-9]+]]
+; AVX512-NEXT:    br i1 [[TMP48]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP20:![0-9]+]]
 ; AVX512:       middle.block:
 ; AVX512-NEXT:    [[CMP_N:%.*]] = icmp eq i64 10000, 9984
 ; AVX512-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
 ; AVX512:       vec.epilog.iter.check:
 ; AVX512-NEXT:    br i1 false, label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
 ; AVX512:       vec.epilog.ph:
 ; AVX512-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ 9984, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
 ; AVX512-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
 ; AVX512:       vec.epilog.vector.body:
-; AVX512-NEXT:    [[INDEX18:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT19:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
-; AVX512-NEXT:    [[TMP49:%.*]] = add i64 [[INDEX18]], 0
+; AVX512-NEXT:    [[INDEX19:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT22:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
+; AVX512-NEXT:    [[TMP49:%.*]] = add i64 [[INDEX19]], 0
 ; AVX512-NEXT:    [[TMP50:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TRIGGER]], i64 [[TMP49]]
 ; AVX512-NEXT:    [[TMP51:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TMP50]], i32 0
 ; AVX512-NEXT:    [[TMP52:%.*]] = bitcast i32 addrspace(1)* [[TMP51]] to <8 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_LOAD21:%.*]] = load <8 x i32>, <8 x i32> addrspace(1)* [[TMP52]], align 4
-; AVX512-NEXT:    [[TMP53:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD21]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
+; AVX512-NEXT:    [[WIDE_LOAD20:%.*]] = load <8 x i32>, <8 x i32> addrspace(1)* [[TMP52]], align 4
+; AVX512-NEXT:    [[TMP53:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD20]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP54:%.*]] = getelementptr i32, i32 addrspace(1)* [[B]], i64 [[TMP49]]
 ; AVX512-NEXT:    [[TMP55:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP54]], i32 0
 ; AVX512-NEXT:    [[TMP56:%.*]] = bitcast i32 addrspace(1)* [[TMP55]] to <8 x i32> addrspace(1)*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD22:%.*]] = call <8 x i32> @llvm.masked.load.v8i32.p1v8i32(<8 x i32> addrspace(1)* [[TMP56]], i32 4, <8 x i1> [[TMP53]], <8 x i32> poison)
-; AVX512-NEXT:    [[TMP57:%.*]] = add nsw <8 x i32> [[WIDE_MASKED_LOAD22]], [[WIDE_LOAD21]]
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD21:%.*]] = call <8 x i32> @llvm.masked.load.v8i32.p1v8i32(<8 x i32> addrspace(1)* [[TMP56]], i32 4, <8 x i1> [[TMP53]], <8 x i32> poison)
+; AVX512-NEXT:    [[TMP57:%.*]] = add nsw <8 x i32> [[WIDE_MASKED_LOAD21]], [[WIDE_LOAD20]]
 ; AVX512-NEXT:    [[TMP58:%.*]] = getelementptr i32, i32 addrspace(1)* [[A]], i64 [[TMP49]]
 ; AVX512-NEXT:    [[TMP59:%.*]] = getelementptr i32, i32 addrspace(1)* [[TMP58]], i32 0
 ; AVX512-NEXT:    [[TMP60:%.*]] = bitcast i32 addrspace(1)* [[TMP59]] to <8 x i32> addrspace(1)*
 ; AVX512-NEXT:    call void @llvm.masked.store.v8i32.p1v8i32(<8 x i32> [[TMP57]], <8 x i32> addrspace(1)* [[TMP60]], i32 4, <8 x i1> [[TMP53]])
-; AVX512-NEXT:    [[INDEX_NEXT19]] = add nuw i64 [[INDEX18]], 8
-; AVX512-NEXT:    [[TMP61:%.*]] = icmp eq i64 [[INDEX_NEXT19]], 10000
-; AVX512-NEXT:    br i1 [[TMP61]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP22:![0-9]+]]
+; AVX512-NEXT:    [[INDEX_NEXT22]] = add nuw i64 [[INDEX19]], 8
+; AVX512-NEXT:    [[TMP61:%.*]] = icmp eq i64 [[INDEX_NEXT22]], 10000
+; AVX512-NEXT:    br i1 [[TMP61]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP21:![0-9]+]]
 ; AVX512:       vec.epilog.middle.block:
-; AVX512-NEXT:    [[CMP_N20:%.*]] = icmp eq i64 10000, 10000
-; AVX512-NEXT:    br i1 [[CMP_N20]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
+; AVX512-NEXT:    [[CMP_N18:%.*]] = icmp eq i64 10000, 10000
+; AVX512-NEXT:    br i1 [[CMP_N18]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
 ; AVX512:       vec.epilog.scalar.ph:
 ; AVX512-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 10000, [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ 9984, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MEMCHECK]] ], [ 0, [[ITER_CHECK:%.*]] ]
 ; AVX512-NEXT:    br label [[FOR_BODY:%.*]]
 ; AVX512:       for.body:
 ; AVX512-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_INC:%.*]] ]
 ; AVX512-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[TRIGGER]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP62:%.*]] = load i32, i32 addrspace(1)* [[ARRAYIDX]], align 4
 ; AVX512-NEXT:    [[CMP1:%.*]] = icmp slt i32 [[TMP62]], 100
 ; AVX512-NEXT:    br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
 ; AVX512:       if.then:
 ; AVX512-NEXT:    [[ARRAYIDX3:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[B]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP63:%.*]] = load i32, i32 addrspace(1)* [[ARRAYIDX3]], align 4
 ; AVX512-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP63]], [[TMP62]]
 ; AVX512-NEXT:    [[ARRAYIDX7:%.*]] = getelementptr inbounds i32, i32 addrspace(1)* [[A]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    store i32 [[ADD]], i32 addrspace(1)* [[ARRAYIDX7]], align 4
 ; AVX512-NEXT:    br label [[FOR_INC]]
 ; AVX512:       for.inc:
 ; AVX512-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
 ; AVX512-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 10000
-; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP23:![0-9]+]]
+; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP22:![0-9]+]]
 ; AVX512:       for.end.loopexit:
 ; AVX512-NEXT:    br label [[FOR_END]]
 ; AVX512:       for.end:
 ; AVX512-NEXT:    ret void
 ;
 entry:
   br label %for.body
 

 ; AVX512-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 32
 ; AVX512-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 48
 ; AVX512-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 0
 ; AVX512-NEXT:    [[TMP9:%.*]] = bitcast i32* [[TMP8]] to <16 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD:%.*]] = load <16 x i32>, <16 x i32>* [[TMP9]], align 4, !alias.scope !24
+; AVX512-NEXT:    [[WIDE_LOAD:%.*]] = load <16 x i32>, <16 x i32>* [[TMP9]], align 4, !alias.scope !23
 ; AVX512-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 16
 ; AVX512-NEXT:    [[TMP11:%.*]] = bitcast i32* [[TMP10]] to <16 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD12:%.*]] = load <16 x i32>, <16 x i32>* [[TMP11]], align 4, !alias.scope !24
+; AVX512-NEXT:    [[WIDE_LOAD12:%.*]] = load <16 x i32>, <16 x i32>* [[TMP11]], align 4, !alias.scope !23
 ; AVX512-NEXT:    [[TMP12:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 32
 ; AVX512-NEXT:    [[TMP13:%.*]] = bitcast i32* [[TMP12]] to <16 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD13:%.*]] = load <16 x i32>, <16 x i32>* [[TMP13]], align 4, !alias.scope !24
+; AVX512-NEXT:    [[WIDE_LOAD13:%.*]] = load <16 x i32>, <16 x i32>* [[TMP13]], align 4, !alias.scope !23
 ; AVX512-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 48
 ; AVX512-NEXT:    [[TMP15:%.*]] = bitcast i32* [[TMP14]] to <16 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD14:%.*]] = load <16 x i32>, <16 x i32>* [[TMP15]], align 4, !alias.scope !24
+; AVX512-NEXT:    [[WIDE_LOAD14:%.*]] = load <16 x i32>, <16 x i32>* [[TMP15]], align 4, !alias.scope !23
 ; AVX512-NEXT:    [[TMP16:%.*]] = icmp slt <16 x i32> [[WIDE_LOAD]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP17:%.*]] = icmp slt <16 x i32> [[WIDE_LOAD12]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP18:%.*]] = icmp slt <16 x i32> [[WIDE_LOAD13]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP19:%.*]] = icmp slt <16 x i32> [[WIDE_LOAD14]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP20:%.*]] = getelementptr float, float* [[B]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP21:%.*]] = getelementptr float, float* [[B]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP22:%.*]] = getelementptr float, float* [[B]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP23:%.*]] = getelementptr float, float* [[B]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP24:%.*]] = getelementptr float, float* [[TMP20]], i32 0
 ; AVX512-NEXT:    [[TMP25:%.*]] = bitcast float* [[TMP24]] to <16 x float>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* [[TMP25]], i32 4, <16 x i1> [[TMP16]], <16 x float> poison), !alias.scope !27
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* [[TMP25]], i32 4, <16 x i1> [[TMP16]], <16 x float> poison), !alias.scope !26
 ; AVX512-NEXT:    [[TMP26:%.*]] = getelementptr float, float* [[TMP20]], i32 16
 ; AVX512-NEXT:    [[TMP27:%.*]] = bitcast float* [[TMP26]] to <16 x float>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD15:%.*]] = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* [[TMP27]], i32 4, <16 x i1> [[TMP17]], <16 x float> poison), !alias.scope !27
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD15:%.*]] = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* [[TMP27]], i32 4, <16 x i1> [[TMP17]], <16 x float> poison), !alias.scope !26
 ; AVX512-NEXT:    [[TMP28:%.*]] = getelementptr float, float* [[TMP20]], i32 32
 ; AVX512-NEXT:    [[TMP29:%.*]] = bitcast float* [[TMP28]] to <16 x float>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD16:%.*]] = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* [[TMP29]], i32 4, <16 x i1> [[TMP18]], <16 x float> poison), !alias.scope !27
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD16:%.*]] = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* [[TMP29]], i32 4, <16 x i1> [[TMP18]], <16 x float> poison), !alias.scope !26
 ; AVX512-NEXT:    [[TMP30:%.*]] = getelementptr float, float* [[TMP20]], i32 48
 ; AVX512-NEXT:    [[TMP31:%.*]] = bitcast float* [[TMP30]] to <16 x float>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD17:%.*]] = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* [[TMP31]], i32 4, <16 x i1> [[TMP19]], <16 x float> poison), !alias.scope !27
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD17:%.*]] = call <16 x float> @llvm.masked.load.v16f32.p0v16f32(<16 x float>* [[TMP31]], i32 4, <16 x i1> [[TMP19]], <16 x float> poison), !alias.scope !26
 ; AVX512-NEXT:    [[TMP32:%.*]] = sitofp <16 x i32> [[WIDE_LOAD]] to <16 x float>
 ; AVX512-NEXT:    [[TMP33:%.*]] = sitofp <16 x i32> [[WIDE_LOAD12]] to <16 x float>
 ; AVX512-NEXT:    [[TMP34:%.*]] = sitofp <16 x i32> [[WIDE_LOAD13]] to <16 x float>
 ; AVX512-NEXT:    [[TMP35:%.*]] = sitofp <16 x i32> [[WIDE_LOAD14]] to <16 x float>
 ; AVX512-NEXT:    [[TMP36:%.*]] = fadd <16 x float> [[WIDE_MASKED_LOAD]], [[TMP32]]
 ; AVX512-NEXT:    [[TMP37:%.*]] = fadd <16 x float> [[WIDE_MASKED_LOAD15]], [[TMP33]]
 ; AVX512-NEXT:    [[TMP38:%.*]] = fadd <16 x float> [[WIDE_MASKED_LOAD16]], [[TMP34]]
 ; AVX512-NEXT:    [[TMP39:%.*]] = fadd <16 x float> [[WIDE_MASKED_LOAD17]], [[TMP35]]
 ; AVX512-NEXT:    [[TMP40:%.*]] = getelementptr float, float* [[A]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP41:%.*]] = getelementptr float, float* [[A]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP42:%.*]] = getelementptr float, float* [[A]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP43:%.*]] = getelementptr float, float* [[A]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP44:%.*]] = getelementptr float, float* [[TMP40]], i32 0
 ; AVX512-NEXT:    [[TMP45:%.*]] = bitcast float* [[TMP44]] to <16 x float>*
-; AVX512-NEXT:    call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP36]], <16 x float>* [[TMP45]], i32 4, <16 x i1> [[TMP16]]), !alias.scope !29, !noalias !31
+; AVX512-NEXT:    call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP36]], <16 x float>* [[TMP45]], i32 4, <16 x i1> [[TMP16]]), !alias.scope !28, !noalias !30
 ; AVX512-NEXT:    [[TMP46:%.*]] = getelementptr float, float* [[TMP40]], i32 16
 ; AVX512-NEXT:    [[TMP47:%.*]] = bitcast float* [[TMP46]] to <16 x float>*
-; AVX512-NEXT:    call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP37]], <16 x float>* [[TMP47]], i32 4, <16 x i1> [[TMP17]]), !alias.scope !29, !noalias !31
+; AVX512-NEXT:    call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP37]], <16 x float>* [[TMP47]], i32 4, <16 x i1> [[TMP17]]), !alias.scope !28, !noalias !30
 ; AVX512-NEXT:    [[TMP48:%.*]] = getelementptr float, float* [[TMP40]], i32 32
 ; AVX512-NEXT:    [[TMP49:%.*]] = bitcast float* [[TMP48]] to <16 x float>*
-; AVX512-NEXT:    call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP38]], <16 x float>* [[TMP49]], i32 4, <16 x i1> [[TMP18]]), !alias.scope !29, !noalias !31
+; AVX512-NEXT:    call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP38]], <16 x float>* [[TMP49]], i32 4, <16 x i1> [[TMP18]]), !alias.scope !28, !noalias !30
 ; AVX512-NEXT:    [[TMP50:%.*]] = getelementptr float, float* [[TMP40]], i32 48
 ; AVX512-NEXT:    [[TMP51:%.*]] = bitcast float* [[TMP50]] to <16 x float>*
-; AVX512-NEXT:    call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP39]], <16 x float>* [[TMP51]], i32 4, <16 x i1> [[TMP19]]), !alias.scope !29, !noalias !31
+; AVX512-NEXT:    call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> [[TMP39]], <16 x float>* [[TMP51]], i32 4, <16 x i1> [[TMP19]]), !alias.scope !28, !noalias !30
 ; AVX512-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 64
 ; AVX512-NEXT:    [[TMP52:%.*]] = icmp eq i64 [[INDEX_NEXT]], 9984
-; AVX512-NEXT:    br i1 [[TMP52]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP32:![0-9]+]]
+; AVX512-NEXT:    br i1 [[TMP52]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP31:![0-9]+]]
 ; AVX512:       middle.block:
 ; AVX512-NEXT:    [[CMP_N:%.*]] = icmp eq i64 10000, 9984
 ; AVX512-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
 ; AVX512:       vec.epilog.iter.check:
 ; AVX512-NEXT:    br i1 false, label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
 ; AVX512:       vec.epilog.ph:
 ; AVX512-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ 9984, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
 ; AVX512-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
 ; AVX512:       vec.epilog.vector.body:
-; AVX512-NEXT:    [[INDEX18:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT19:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
-; AVX512-NEXT:    [[TMP53:%.*]] = add i64 [[INDEX18]], 0
+; AVX512-NEXT:    [[INDEX19:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT22:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
+; AVX512-NEXT:    [[TMP53:%.*]] = add i64 [[INDEX19]], 0
 ; AVX512-NEXT:    [[TMP54:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP53]]
 ; AVX512-NEXT:    [[TMP55:%.*]] = getelementptr inbounds i32, i32* [[TMP54]], i32 0
 ; AVX512-NEXT:    [[TMP56:%.*]] = bitcast i32* [[TMP55]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD21:%.*]] = load <8 x i32>, <8 x i32>* [[TMP56]], align 4
-; AVX512-NEXT:    [[TMP57:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD21]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
+; AVX512-NEXT:    [[WIDE_LOAD20:%.*]] = load <8 x i32>, <8 x i32>* [[TMP56]], align 4
+; AVX512-NEXT:    [[TMP57:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD20]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP58:%.*]] = getelementptr float, float* [[B]], i64 [[TMP53]]
 ; AVX512-NEXT:    [[TMP59:%.*]] = getelementptr float, float* [[TMP58]], i32 0
 ; AVX512-NEXT:    [[TMP60:%.*]] = bitcast float* [[TMP59]] to <8 x float>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD22:%.*]] = call <8 x float> @llvm.masked.load.v8f32.p0v8f32(<8 x float>* [[TMP60]], i32 4, <8 x i1> [[TMP57]], <8 x float> poison)
-; AVX512-NEXT:    [[TMP61:%.*]] = sitofp <8 x i32> [[WIDE_LOAD21]] to <8 x float>
-; AVX512-NEXT:    [[TMP62:%.*]] = fadd <8 x float> [[WIDE_MASKED_LOAD22]], [[TMP61]]
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD21:%.*]] = call <8 x float> @llvm.masked.load.v8f32.p0v8f32(<8 x float>* [[TMP60]], i32 4, <8 x i1> [[TMP57]], <8 x float> poison)
+; AVX512-NEXT:    [[TMP61:%.*]] = sitofp <8 x i32> [[WIDE_LOAD20]] to <8 x float>
+; AVX512-NEXT:    [[TMP62:%.*]] = fadd <8 x float> [[WIDE_MASKED_LOAD21]], [[TMP61]]
 ; AVX512-NEXT:    [[TMP63:%.*]] = getelementptr float, float* [[A]], i64 [[TMP53]]
 ; AVX512-NEXT:    [[TMP64:%.*]] = getelementptr float, float* [[TMP63]], i32 0
 ; AVX512-NEXT:    [[TMP65:%.*]] = bitcast float* [[TMP64]] to <8 x float>*
 ; AVX512-NEXT:    call void @llvm.masked.store.v8f32.p0v8f32(<8 x float> [[TMP62]], <8 x float>* [[TMP65]], i32 4, <8 x i1> [[TMP57]])
-; AVX512-NEXT:    [[INDEX_NEXT19]] = add nuw i64 [[INDEX18]], 8
-; AVX512-NEXT:    [[TMP66:%.*]] = icmp eq i64 [[INDEX_NEXT19]], 10000
-; AVX512-NEXT:    br i1 [[TMP66]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP33:![0-9]+]]
+; AVX512-NEXT:    [[INDEX_NEXT22]] = add nuw i64 [[INDEX19]], 8
+; AVX512-NEXT:    [[TMP66:%.*]] = icmp eq i64 [[INDEX_NEXT22]], 10000
+; AVX512-NEXT:    br i1 [[TMP66]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP32:![0-9]+]]
 ; AVX512:       vec.epilog.middle.block:
-; AVX512-NEXT:    [[CMP_N20:%.*]] = icmp eq i64 10000, 10000
-; AVX512-NEXT:    br i1 [[CMP_N20]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
+; AVX512-NEXT:    [[CMP_N18:%.*]] = icmp eq i64 10000, 10000
+; AVX512-NEXT:    br i1 [[CMP_N18]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_SCALAR_PH]]
 ; AVX512:       vec.epilog.scalar.ph:
 ; AVX512-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 10000, [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ 9984, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MEMCHECK]] ], [ 0, [[ITER_CHECK:%.*]] ]
 ; AVX512-NEXT:    br label [[FOR_BODY:%.*]]
 ; AVX512:       for.body:
 ; AVX512-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_INC:%.*]] ]
 ; AVX512-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP67:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
 ; AVX512-NEXT:    [[CMP1:%.*]] = icmp slt i32 [[TMP67]], 100
 ; AVX512-NEXT:    br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
 ; AVX512:       if.then:
 ; AVX512-NEXT:    [[ARRAYIDX3:%.*]] = getelementptr inbounds float, float* [[B]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP68:%.*]] = load float, float* [[ARRAYIDX3]], align 4
 ; AVX512-NEXT:    [[CONV:%.*]] = sitofp i32 [[TMP67]] to float
 ; AVX512-NEXT:    [[ADD:%.*]] = fadd float [[TMP68]], [[CONV]]
 ; AVX512-NEXT:    [[ARRAYIDX7:%.*]] = getelementptr inbounds float, float* [[A]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    store float [[ADD]], float* [[ARRAYIDX7]], align 4
 ; AVX512-NEXT:    br label [[FOR_INC]]
 ; AVX512:       for.inc:
 ; AVX512-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
 ; AVX512-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 10000
-; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP34:![0-9]+]]
+; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP33:![0-9]+]]
 ; AVX512:       for.end.loopexit:
 ; AVX512-NEXT:    br label [[FOR_END]]
 ; AVX512:       for.end:
 ; AVX512-NEXT:    ret void
 ;
 entry:
   br label %for.body
 

 ; AVX512-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 16
 ; AVX512-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 24
 ; AVX512-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 0
 ; AVX512-NEXT:    [[TMP9:%.*]] = bitcast i32* [[TMP8]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD:%.*]] = load <8 x i32>, <8 x i32>* [[TMP9]], align 4, !alias.scope !35
+; AVX512-NEXT:    [[WIDE_LOAD:%.*]] = load <8 x i32>, <8 x i32>* [[TMP9]], align 4, !alias.scope !34
 ; AVX512-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 8
 ; AVX512-NEXT:    [[TMP11:%.*]] = bitcast i32* [[TMP10]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD12:%.*]] = load <8 x i32>, <8 x i32>* [[TMP11]], align 4, !alias.scope !35
+; AVX512-NEXT:    [[WIDE_LOAD12:%.*]] = load <8 x i32>, <8 x i32>* [[TMP11]], align 4, !alias.scope !34
 ; AVX512-NEXT:    [[TMP12:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 16
 ; AVX512-NEXT:    [[TMP13:%.*]] = bitcast i32* [[TMP12]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD13:%.*]] = load <8 x i32>, <8 x i32>* [[TMP13]], align 4, !alias.scope !35
+; AVX512-NEXT:    [[WIDE_LOAD13:%.*]] = load <8 x i32>, <8 x i32>* [[TMP13]], align 4, !alias.scope !34
 ; AVX512-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 24
 ; AVX512-NEXT:    [[TMP15:%.*]] = bitcast i32* [[TMP14]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD14:%.*]] = load <8 x i32>, <8 x i32>* [[TMP15]], align 4, !alias.scope !35
+; AVX512-NEXT:    [[WIDE_LOAD14:%.*]] = load <8 x i32>, <8 x i32>* [[TMP15]], align 4, !alias.scope !34
 ; AVX512-NEXT:    [[TMP16:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP17:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD12]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP18:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD13]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP19:%.*]] = icmp slt <8 x i32> [[WIDE_LOAD14]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP20:%.*]] = getelementptr double, double* [[B]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP21:%.*]] = getelementptr double, double* [[B]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP22:%.*]] = getelementptr double, double* [[B]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP23:%.*]] = getelementptr double, double* [[B]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP24:%.*]] = getelementptr double, double* [[TMP20]], i32 0
 ; AVX512-NEXT:    [[TMP25:%.*]] = bitcast double* [[TMP24]] to <8 x double>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP25]], i32 8, <8 x i1> [[TMP16]], <8 x double> poison), !alias.scope !38
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP25]], i32 8, <8 x i1> [[TMP16]], <8 x double> poison), !alias.scope !37
 ; AVX512-NEXT:    [[TMP26:%.*]] = getelementptr double, double* [[TMP20]], i32 8
 ; AVX512-NEXT:    [[TMP27:%.*]] = bitcast double* [[TMP26]] to <8 x double>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD15:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP27]], i32 8, <8 x i1> [[TMP17]], <8 x double> poison), !alias.scope !38
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD15:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP27]], i32 8, <8 x i1> [[TMP17]], <8 x double> poison), !alias.scope !37
 ; AVX512-NEXT:    [[TMP28:%.*]] = getelementptr double, double* [[TMP20]], i32 16
 ; AVX512-NEXT:    [[TMP29:%.*]] = bitcast double* [[TMP28]] to <8 x double>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD16:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP29]], i32 8, <8 x i1> [[TMP18]], <8 x double> poison), !alias.scope !38
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD16:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP29]], i32 8, <8 x i1> [[TMP18]], <8 x double> poison), !alias.scope !37
 ; AVX512-NEXT:    [[TMP30:%.*]] = getelementptr double, double* [[TMP20]], i32 24
 ; AVX512-NEXT:    [[TMP31:%.*]] = bitcast double* [[TMP30]] to <8 x double>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD17:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP31]], i32 8, <8 x i1> [[TMP19]], <8 x double> poison), !alias.scope !38
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD17:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP31]], i32 8, <8 x i1> [[TMP19]], <8 x double> poison), !alias.scope !37
 ; AVX512-NEXT:    [[TMP32:%.*]] = sitofp <8 x i32> [[WIDE_LOAD]] to <8 x double>
 ; AVX512-NEXT:    [[TMP33:%.*]] = sitofp <8 x i32> [[WIDE_LOAD12]] to <8 x double>
 ; AVX512-NEXT:    [[TMP34:%.*]] = sitofp <8 x i32> [[WIDE_LOAD13]] to <8 x double>
 ; AVX512-NEXT:    [[TMP35:%.*]] = sitofp <8 x i32> [[WIDE_LOAD14]] to <8 x double>
 ; AVX512-NEXT:    [[TMP36:%.*]] = fadd <8 x double> [[WIDE_MASKED_LOAD]], [[TMP32]]
 ; AVX512-NEXT:    [[TMP37:%.*]] = fadd <8 x double> [[WIDE_MASKED_LOAD15]], [[TMP33]]
 ; AVX512-NEXT:    [[TMP38:%.*]] = fadd <8 x double> [[WIDE_MASKED_LOAD16]], [[TMP34]]
 ; AVX512-NEXT:    [[TMP39:%.*]] = fadd <8 x double> [[WIDE_MASKED_LOAD17]], [[TMP35]]
 ; AVX512-NEXT:    [[TMP40:%.*]] = getelementptr double, double* [[A]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP41:%.*]] = getelementptr double, double* [[A]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP42:%.*]] = getelementptr double, double* [[A]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP43:%.*]] = getelementptr double, double* [[A]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP44:%.*]] = getelementptr double, double* [[TMP40]], i32 0
 ; AVX512-NEXT:    [[TMP45:%.*]] = bitcast double* [[TMP44]] to <8 x double>*
-; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[TMP36]], <8 x double>* [[TMP45]], i32 8, <8 x i1> [[TMP16]]), !alias.scope !40, !noalias !42
+; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[TMP36]], <8 x double>* [[TMP45]], i32 8, <8 x i1> [[TMP16]]), !alias.scope !39, !noalias !41
 ; AVX512-NEXT:    [[TMP46:%.*]] = getelementptr double, double* [[TMP40]], i32 8
 ; AVX512-NEXT:    [[TMP47:%.*]] = bitcast double* [[TMP46]] to <8 x double>*
-; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[TMP37]], <8 x double>* [[TMP47]], i32 8, <8 x i1> [[TMP17]]), !alias.scope !40, !noalias !42
+; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[TMP37]], <8 x double>* [[TMP47]], i32 8, <8 x i1> [[TMP17]]), !alias.scope !39, !noalias !41
 ; AVX512-NEXT:    [[TMP48:%.*]] = getelementptr double, double* [[TMP40]], i32 16
 ; AVX512-NEXT:    [[TMP49:%.*]] = bitcast double* [[TMP48]] to <8 x double>*
-; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[TMP38]], <8 x double>* [[TMP49]], i32 8, <8 x i1> [[TMP18]]), !alias.scope !40, !noalias !42
+; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[TMP38]], <8 x double>* [[TMP49]], i32 8, <8 x i1> [[TMP18]]), !alias.scope !39, !noalias !41
 ; AVX512-NEXT:    [[TMP50:%.*]] = getelementptr double, double* [[TMP40]], i32 24
 ; AVX512-NEXT:    [[TMP51:%.*]] = bitcast double* [[TMP50]] to <8 x double>*
-; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[TMP39]], <8 x double>* [[TMP51]], i32 8, <8 x i1> [[TMP19]]), !alias.scope !40, !noalias !42
+; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[TMP39]], <8 x double>* [[TMP51]], i32 8, <8 x i1> [[TMP19]]), !alias.scope !39, !noalias !41
 ; AVX512-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 32
 ; AVX512-NEXT:    [[TMP52:%.*]] = icmp eq i64 [[INDEX_NEXT]], 9984
-; AVX512-NEXT:    br i1 [[TMP52]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP43:![0-9]+]]
+; AVX512-NEXT:    br i1 [[TMP52]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP42:![0-9]+]]
 ; AVX512:       middle.block:
 ; AVX512-NEXT:    [[CMP_N:%.*]] = icmp eq i64 10000, 9984
 ; AVX512-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
 ; AVX512:       scalar.ph:
 ; AVX512-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 9984, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ], [ 0, [[VECTOR_MEMCHECK]] ]
 ; AVX512-NEXT:    br label [[FOR_BODY:%.*]]
 ; AVX512:       for.body:
 ; AVX512-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_INC:%.*]] ]
 ; AVX512-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP53:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
 ; AVX512-NEXT:    [[CMP1:%.*]] = icmp slt i32 [[TMP53]], 100
 ; AVX512-NEXT:    br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
 ; AVX512:       if.then:
 ; AVX512-NEXT:    [[ARRAYIDX3:%.*]] = getelementptr inbounds double, double* [[B]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP54:%.*]] = load double, double* [[ARRAYIDX3]], align 8
 ; AVX512-NEXT:    [[CONV:%.*]] = sitofp i32 [[TMP53]] to double
 ; AVX512-NEXT:    [[ADD:%.*]] = fadd double [[TMP54]], [[CONV]]
 ; AVX512-NEXT:    [[ARRAYIDX7:%.*]] = getelementptr inbounds double, double* [[A]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    store double [[ADD]], double* [[ARRAYIDX7]], align 8
 ; AVX512-NEXT:    br label [[FOR_INC]]
 ; AVX512:       for.inc:
 ; AVX512-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
 ; AVX512-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 10000
-; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP44:![0-9]+]]
+; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP43:![0-9]+]]
 ; AVX512:       for.end:
 ; AVX512-NEXT:    ret void
 ;
 entry:
   br label %for.body
 
 for.body:                                         ; preds = %for.inc, %entry
   %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.inc ]

 ; AVX512-NEXT:    [[CONFLICT_RDX:%.*]] = or i1 [[FOUND_CONFLICT]], [[FOUND_CONFLICT11]]
 ; AVX512-NEXT:    br i1 [[CONFLICT_RDX]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
 ; AVX512:       vector.ph:
 ; AVX512-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; AVX512:       vector.body:
 ; AVX512-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; AVX512-NEXT:    [[VEC_IND:%.*]] = phi <8 x i64> [ <i64 0, i64 16, i64 32, i64 48, i64 64, i64 80, i64 96, i64 112>, [[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; AVX512-NEXT:    [[TMP0:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], <8 x i64> [[VEC_IND]]
-; AVX512-NEXT:    [[WIDE_MASKED_GATHER:%.*]] = call <8 x i32> @llvm.masked.gather.v8i32.v8p0i32(<8 x i32*> [[TMP0]], i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x i32> undef), !alias.scope !45
+; AVX512-NEXT:    [[WIDE_MASKED_GATHER:%.*]] = call <8 x i32> @llvm.masked.gather.v8i32.v8p0i32(<8 x i32*> [[TMP0]], i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x i32> undef), !alias.scope !44
 ; AVX512-NEXT:    [[TMP1:%.*]] = icmp slt <8 x i32> [[WIDE_MASKED_GATHER]], <i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100, i32 100>
 ; AVX512-NEXT:    [[TMP2:%.*]] = shl nuw nsw <8 x i64> [[VEC_IND]], <i64 1, i64 1, i64 1, i64 1, i64 1, i64 1, i64 1, i64 1>
 ; AVX512-NEXT:    [[TMP3:%.*]] = getelementptr inbounds double, double* [[B]], <8 x i64> [[TMP2]]
-; AVX512-NEXT:    [[WIDE_MASKED_GATHER12:%.*]] = call <8 x double> @llvm.masked.gather.v8f64.v8p0f64(<8 x double*> [[TMP3]], i32 8, <8 x i1> [[TMP1]], <8 x double> undef), !alias.scope !48
+; AVX512-NEXT:    [[WIDE_MASKED_GATHER12:%.*]] = call <8 x double> @llvm.masked.gather.v8f64.v8p0f64(<8 x double*> [[TMP3]], i32 8, <8 x i1> [[TMP1]], <8 x double> undef), !alias.scope !47
 ; AVX512-NEXT:    [[TMP4:%.*]] = sitofp <8 x i32> [[WIDE_MASKED_GATHER]] to <8 x double>
 ; AVX512-NEXT:    [[TMP5:%.*]] = fadd <8 x double> [[WIDE_MASKED_GATHER12]], [[TMP4]]
 ; AVX512-NEXT:    [[TMP6:%.*]] = getelementptr inbounds double, double* [[A]], <8 x i64> [[VEC_IND]]
-; AVX512-NEXT:    call void @llvm.masked.scatter.v8f64.v8p0f64(<8 x double> [[TMP5]], <8 x double*> [[TMP6]], i32 8, <8 x i1> [[TMP1]]), !alias.scope !50, !noalias !52
+; AVX512-NEXT:    call void @llvm.masked.scatter.v8f64.v8p0f64(<8 x double> [[TMP5]], <8 x double*> [[TMP6]], i32 8, <8 x i1> [[TMP1]]), !alias.scope !49, !noalias !51
 ; AVX512-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
 ; AVX512-NEXT:    [[VEC_IND_NEXT]] = add <8 x i64> [[VEC_IND]], <i64 128, i64 128, i64 128, i64 128, i64 128, i64 128, i64 128, i64 128>
 ; AVX512-NEXT:    [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], 624
-; AVX512-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP53:![0-9]+]]
+; AVX512-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP52:![0-9]+]]
 ; AVX512:       middle.block:
 ; AVX512-NEXT:    [[CMP_N:%.*]] = icmp eq i64 625, 624
 ; AVX512-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
 ; AVX512:       scalar.ph:
 ; AVX512-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 9984, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ], [ 0, [[VECTOR_MEMCHECK]] ]
 ; AVX512-NEXT:    br label [[FOR_BODY:%.*]]
 ; AVX512:       for.body:
 ; AVX512-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_INC:%.*]] ]
 ; AVX512-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP8:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
 ; AVX512-NEXT:    [[CMP1:%.*]] = icmp slt i32 [[TMP8]], 100
 ; AVX512-NEXT:    br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
 ; AVX512:       if.then:
 ; AVX512-NEXT:    [[TMP9:%.*]] = shl nuw nsw i64 [[INDVARS_IV]], 1
 ; AVX512-NEXT:    [[ARRAYIDX3:%.*]] = getelementptr inbounds double, double* [[B]], i64 [[TMP9]]
 ; AVX512-NEXT:    [[TMP10:%.*]] = load double, double* [[ARRAYIDX3]], align 8
 ; AVX512-NEXT:    [[CONV:%.*]] = sitofp i32 [[TMP8]] to double
 ; AVX512-NEXT:    [[ADD:%.*]] = fadd double [[TMP10]], [[CONV]]
 ; AVX512-NEXT:    [[ARRAYIDX7:%.*]] = getelementptr inbounds double, double* [[A]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    store double [[ADD]], double* [[ARRAYIDX7]], align 8
 ; AVX512-NEXT:    br label [[FOR_INC]]
 ; AVX512:       for.inc:
 ; AVX512-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 16
 ; AVX512-NEXT:    [[CMP:%.*]] = icmp ult i64 [[INDVARS_IV_NEXT]], 10000
-; AVX512-NEXT:    br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_END]], !llvm.loop [[LOOP54:![0-9]+]]
+; AVX512-NEXT:    br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_END]], !llvm.loop [[LOOP53:![0-9]+]]
 ; AVX512:       for.end:
 ; AVX512-NEXT:    ret void
 ;
 entry:
   br label %for.body
 
 for.body:                                         ; preds = %entry, %for.inc
   %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.inc ]

 ; AVX512-NEXT:    [[TMP3:%.*]] = add i64 [[OFFSET_IDX]], -24
 ; AVX512-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 0
 ; AVX512-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i32, i32* [[TMP8]], i32 -7
 ; AVX512-NEXT:    [[TMP10:%.*]] = bitcast i32* [[TMP9]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD:%.*]] = load <8 x i32>, <8 x i32>* [[TMP10]], align 4, !alias.scope !55
+; AVX512-NEXT:    [[WIDE_LOAD:%.*]] = load <8 x i32>, <8 x i32>* [[TMP10]], align 4, !alias.scope !54
 ; AVX512-NEXT:    [[REVERSE:%.*]] = shufflevector <8 x i32> [[WIDE_LOAD]], <8 x i32> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP11:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 -8
 ; AVX512-NEXT:    [[TMP12:%.*]] = getelementptr inbounds i32, i32* [[TMP11]], i32 -7
 ; AVX512-NEXT:    [[TMP13:%.*]] = bitcast i32* [[TMP12]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD12:%.*]] = load <8 x i32>, <8 x i32>* [[TMP13]], align 4, !alias.scope !55
+; AVX512-NEXT:    [[WIDE_LOAD12:%.*]] = load <8 x i32>, <8 x i32>* [[TMP13]], align 4, !alias.scope !54
 ; AVX512-NEXT:    [[REVERSE13:%.*]] = shufflevector <8 x i32> [[WIDE_LOAD12]], <8 x i32> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 -16
 ; AVX512-NEXT:    [[TMP15:%.*]] = getelementptr inbounds i32, i32* [[TMP14]], i32 -7
 ; AVX512-NEXT:    [[TMP16:%.*]] = bitcast i32* [[TMP15]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD14:%.*]] = load <8 x i32>, <8 x i32>* [[TMP16]], align 4, !alias.scope !55
+; AVX512-NEXT:    [[WIDE_LOAD14:%.*]] = load <8 x i32>, <8 x i32>* [[TMP16]], align 4, !alias.scope !54
 ; AVX512-NEXT:    [[REVERSE15:%.*]] = shufflevector <8 x i32> [[WIDE_LOAD14]], <8 x i32> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i32, i32* [[TMP4]], i32 -24
 ; AVX512-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i32, i32* [[TMP17]], i32 -7
 ; AVX512-NEXT:    [[TMP19:%.*]] = bitcast i32* [[TMP18]] to <8 x i32>*
-; AVX512-NEXT:    [[WIDE_LOAD16:%.*]] = load <8 x i32>, <8 x i32>* [[TMP19]], align 4, !alias.scope !55
+; AVX512-NEXT:    [[WIDE_LOAD16:%.*]] = load <8 x i32>, <8 x i32>* [[TMP19]], align 4, !alias.scope !54
 ; AVX512-NEXT:    [[REVERSE17:%.*]] = shufflevector <8 x i32> [[WIDE_LOAD16]], <8 x i32> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP20:%.*]] = icmp sgt <8 x i32> [[REVERSE]], zeroinitializer
 ; AVX512-NEXT:    [[TMP21:%.*]] = icmp sgt <8 x i32> [[REVERSE13]], zeroinitializer
 ; AVX512-NEXT:    [[TMP22:%.*]] = icmp sgt <8 x i32> [[REVERSE15]], zeroinitializer
 ; AVX512-NEXT:    [[TMP23:%.*]] = icmp sgt <8 x i32> [[REVERSE17]], zeroinitializer
 ; AVX512-NEXT:    [[TMP24:%.*]] = getelementptr double, double* [[IN]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP25:%.*]] = getelementptr double, double* [[IN]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP26:%.*]] = getelementptr double, double* [[IN]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP27:%.*]] = getelementptr double, double* [[IN]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[TMP28:%.*]] = getelementptr double, double* [[TMP24]], i32 0
 ; AVX512-NEXT:    [[TMP29:%.*]] = getelementptr double, double* [[TMP28]], i32 -7
 ; AVX512-NEXT:    [[REVERSE18:%.*]] = shufflevector <8 x i1> [[TMP20]], <8 x i1> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP30:%.*]] = bitcast double* [[TMP29]] to <8 x double>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP30]], i32 8, <8 x i1> [[REVERSE18]], <8 x double> poison), !alias.scope !58
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP30]], i32 8, <8 x i1> [[REVERSE18]], <8 x double> poison), !alias.scope !57
 ; AVX512-NEXT:    [[REVERSE19:%.*]] = shufflevector <8 x double> [[WIDE_MASKED_LOAD]], <8 x double> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP31:%.*]] = getelementptr double, double* [[TMP24]], i32 -8
 ; AVX512-NEXT:    [[TMP32:%.*]] = getelementptr double, double* [[TMP31]], i32 -7
 ; AVX512-NEXT:    [[REVERSE20:%.*]] = shufflevector <8 x i1> [[TMP21]], <8 x i1> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP33:%.*]] = bitcast double* [[TMP32]] to <8 x double>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD21:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP33]], i32 8, <8 x i1> [[REVERSE20]], <8 x double> poison), !alias.scope !58
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD21:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP33]], i32 8, <8 x i1> [[REVERSE20]], <8 x double> poison), !alias.scope !57
 ; AVX512-NEXT:    [[REVERSE22:%.*]] = shufflevector <8 x double> [[WIDE_MASKED_LOAD21]], <8 x double> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP34:%.*]] = getelementptr double, double* [[TMP24]], i32 -16
 ; AVX512-NEXT:    [[TMP35:%.*]] = getelementptr double, double* [[TMP34]], i32 -7
 ; AVX512-NEXT:    [[REVERSE23:%.*]] = shufflevector <8 x i1> [[TMP22]], <8 x i1> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP36:%.*]] = bitcast double* [[TMP35]] to <8 x double>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD24:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP36]], i32 8, <8 x i1> [[REVERSE23]], <8 x double> poison), !alias.scope !58
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD24:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP36]], i32 8, <8 x i1> [[REVERSE23]], <8 x double> poison), !alias.scope !57
 ; AVX512-NEXT:    [[REVERSE25:%.*]] = shufflevector <8 x double> [[WIDE_MASKED_LOAD24]], <8 x double> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP37:%.*]] = getelementptr double, double* [[TMP24]], i32 -24
 ; AVX512-NEXT:    [[TMP38:%.*]] = getelementptr double, double* [[TMP37]], i32 -7
 ; AVX512-NEXT:    [[REVERSE26:%.*]] = shufflevector <8 x i1> [[TMP23]], <8 x i1> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP39:%.*]] = bitcast double* [[TMP38]] to <8 x double>*
-; AVX512-NEXT:    [[WIDE_MASKED_LOAD27:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP39]], i32 8, <8 x i1> [[REVERSE26]], <8 x double> poison), !alias.scope !58
+; AVX512-NEXT:    [[WIDE_MASKED_LOAD27:%.*]] = call <8 x double> @llvm.masked.load.v8f64.p0v8f64(<8 x double>* [[TMP39]], i32 8, <8 x i1> [[REVERSE26]], <8 x double> poison), !alias.scope !57
 ; AVX512-NEXT:    [[REVERSE28:%.*]] = shufflevector <8 x double> [[WIDE_MASKED_LOAD27]], <8 x double> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP40:%.*]] = fadd <8 x double> [[REVERSE19]], <double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01>
 ; AVX512-NEXT:    [[TMP41:%.*]] = fadd <8 x double> [[REVERSE22]], <double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01>
 ; AVX512-NEXT:    [[TMP42:%.*]] = fadd <8 x double> [[REVERSE25]], <double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01>
 ; AVX512-NEXT:    [[TMP43:%.*]] = fadd <8 x double> [[REVERSE28]], <double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01>
 ; AVX512-NEXT:    [[TMP44:%.*]] = getelementptr double, double* [[OUT]], i64 [[TMP0]]
 ; AVX512-NEXT:    [[TMP45:%.*]] = getelementptr double, double* [[OUT]], i64 [[TMP1]]
 ; AVX512-NEXT:    [[TMP46:%.*]] = getelementptr double, double* [[OUT]], i64 [[TMP2]]
 ; AVX512-NEXT:    [[TMP47:%.*]] = getelementptr double, double* [[OUT]], i64 [[TMP3]]
 ; AVX512-NEXT:    [[REVERSE29:%.*]] = shufflevector <8 x double> [[TMP40]], <8 x double> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP48:%.*]] = getelementptr double, double* [[TMP44]], i32 0
 ; AVX512-NEXT:    [[TMP49:%.*]] = getelementptr double, double* [[TMP48]], i32 -7
 ; AVX512-NEXT:    [[TMP50:%.*]] = bitcast double* [[TMP49]] to <8 x double>*
-; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[REVERSE29]], <8 x double>* [[TMP50]], i32 8, <8 x i1> [[REVERSE18]]), !alias.scope !60, !noalias !62
+; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[REVERSE29]], <8 x double>* [[TMP50]], i32 8, <8 x i1> [[REVERSE18]]), !alias.scope !59, !noalias !61
 ; AVX512-NEXT:    [[REVERSE31:%.*]] = shufflevector <8 x double> [[TMP41]], <8 x double> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP51:%.*]] = getelementptr double, double* [[TMP44]], i32 -8
 ; AVX512-NEXT:    [[TMP52:%.*]] = getelementptr double, double* [[TMP51]], i32 -7
 ; AVX512-NEXT:    [[TMP53:%.*]] = bitcast double* [[TMP52]] to <8 x double>*
-; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[REVERSE31]], <8 x double>* [[TMP53]], i32 8, <8 x i1> [[REVERSE20]]), !alias.scope !60, !noalias !62
+; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[REVERSE31]], <8 x double>* [[TMP53]], i32 8, <8 x i1> [[REVERSE20]]), !alias.scope !59, !noalias !61
 ; AVX512-NEXT:    [[REVERSE33:%.*]] = shufflevector <8 x double> [[TMP42]], <8 x double> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP54:%.*]] = getelementptr double, double* [[TMP44]], i32 -16
 ; AVX512-NEXT:    [[TMP55:%.*]] = getelementptr double, double* [[TMP54]], i32 -7
 ; AVX512-NEXT:    [[TMP56:%.*]] = bitcast double* [[TMP55]] to <8 x double>*
-; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[REVERSE33]], <8 x double>* [[TMP56]], i32 8, <8 x i1> [[REVERSE23]]), !alias.scope !60, !noalias !62
+; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[REVERSE33]], <8 x double>* [[TMP56]], i32 8, <8 x i1> [[REVERSE23]]), !alias.scope !59, !noalias !61
 ; AVX512-NEXT:    [[REVERSE35:%.*]] = shufflevector <8 x double> [[TMP43]], <8 x double> poison, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:    [[TMP57:%.*]] = getelementptr double, double* [[TMP44]], i32 -24
 ; AVX512-NEXT:    [[TMP58:%.*]] = getelementptr double, double* [[TMP57]], i32 -7
 ; AVX512-NEXT:    [[TMP59:%.*]] = bitcast double* [[TMP58]] to <8 x double>*
-; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[REVERSE35]], <8 x double>* [[TMP59]], i32 8, <8 x i1> [[REVERSE26]]), !alias.scope !60, !noalias !62
+; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> [[REVERSE35]], <8 x double>* [[TMP59]], i32 8, <8 x i1> [[REVERSE26]]), !alias.scope !59, !noalias !61
 ; AVX512-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 32
 ; AVX512-NEXT:    [[TMP60:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4096
-; AVX512-NEXT:    br i1 [[TMP60]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP63:![0-9]+]]
+; AVX512-NEXT:    br i1 [[TMP60]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP62:![0-9]+]]
 ; AVX512:       middle.block:
 ; AVX512-NEXT:    [[CMP_N:%.*]] = icmp eq i64 4096, 4096
 ; AVX512-NEXT:    br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
 ; AVX512:       scalar.ph:
 ; AVX512-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ -1, [[MIDDLE_BLOCK]] ], [ 4095, [[ENTRY:%.*]] ], [ 4095, [[VECTOR_MEMCHECK]] ]
 ; AVX512-NEXT:    br label [[FOR_BODY:%.*]]
 ; AVX512:       for.body:
 ; AVX512-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_INC:%.*]] ]
 ; AVX512-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, i32* [[TRIGGER]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP61:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
 ; AVX512-NEXT:    [[CMP1:%.*]] = icmp sgt i32 [[TMP61]], 0
 ; AVX512-NEXT:    br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
 ; AVX512:       if.then:
 ; AVX512-NEXT:    [[ARRAYIDX3:%.*]] = getelementptr inbounds double, double* [[IN]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    [[TMP62:%.*]] = load double, double* [[ARRAYIDX3]], align 8
 ; AVX512-NEXT:    [[ADD:%.*]] = fadd double [[TMP62]], 5.000000e-01
 ; AVX512-NEXT:    [[ARRAYIDX5:%.*]] = getelementptr inbounds double, double* [[OUT]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    store double [[ADD]], double* [[ARRAYIDX5]], align 8
 ; AVX512-NEXT:    br label [[FOR_INC]]
 ; AVX512:       for.inc:
 ; AVX512-NEXT:    [[INDVARS_IV_NEXT]] = add nsw i64 [[INDVARS_IV]], -1
 ; AVX512-NEXT:    [[CMP:%.*]] = icmp eq i64 [[INDVARS_IV]], 0
-; AVX512-NEXT:    br i1 [[CMP]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP64:![0-9]+]]
+; AVX512-NEXT:    br i1 [[CMP]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP63:![0-9]+]]
 ; AVX512:       for.end:
 ; AVX512-NEXT:    ret void
 ;
 entry:
   br label %for.body
 
 for.body:                                         ; preds = %for.inc, %entry
   %indvars.iv = phi i64 [ 4095, %entry ], [ %indvars.iv.next, %for.inc ]

 ; AVX512-NEXT:    [[TMP60:%.*]] = getelementptr double, double* [[TMP44]], i32 16
 ; AVX512-NEXT:    [[TMP61:%.*]] = bitcast double* [[TMP60]] to <8 x double>*
 ; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> <double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01>, <8 x double>* [[TMP61]], i32 8, <8 x i1> [[TMP54]])
 ; AVX512-NEXT:    [[TMP62:%.*]] = getelementptr double, double* [[TMP44]], i32 24
 ; AVX512-NEXT:    [[TMP63:%.*]] = bitcast double* [[TMP62]] to <8 x double>*
 ; AVX512-NEXT:    call void @llvm.masked.store.v8f64.p0v8f64(<8 x double> <double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01, double 5.000000e-01>, <8 x double>* [[TMP63]], i32 8, <8 x i1> [[TMP55]])
 ; AVX512-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 32
 ; AVX512-NEXT:    [[TMP64:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; AVX512-NEXT:    br i1 [[TMP64]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP65:![0-9]+]]
+; AVX512-NEXT:    br i1 [[TMP64]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP64:![0-9]+]]
 ; AVX512:       middle.block:
 ; AVX512-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
 ; AVX512-NEXT:    br i1 [[CMP_N]], label [[FOR_END_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; AVX512:       scalar.ph:
 ; AVX512-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
 ; AVX512-NEXT:    br label [[FOR_BODY:%.*]]
 ; AVX512:       for.body:
 ; AVX512-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_INC:%.*]] ]

 ; AVX512-NEXT:    br i1 [[CMP3]], label [[FOR_INC]], label [[IF_THEN:%.*]]
 ; AVX512:       if.then:
 ; AVX512-NEXT:    [[ARRAYIDX5:%.*]] = getelementptr inbounds double, double* [[OUT]], i64 [[INDVARS_IV]]
 ; AVX512-NEXT:    store double 5.000000e-01, double* [[ARRAYIDX5]], align 8
 ; AVX512-NEXT:    br label [[FOR_INC]]
 ; AVX512:       for.inc:
 ; AVX512-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
 ; AVX512-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], [[WIDE_TRIP_COUNT]]
-; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP66:![0-9]+]]
+; AVX512-NEXT:    br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP65:![0-9]+]]
 ; AVX512:       for.end.loopexit:
 ; AVX512-NEXT:    br label [[FOR_END]]
 ; AVX512:       for.end:
 ; AVX512-NEXT:    ret void
 ;
 entry:
   %cmp5 = icmp eq i32 %size, 0
   br i1 %cmp5, label %for.end, label %for.body.preheader

llvm/test/Transforms/LoopVectorize/X86/tail_loop_folding.ll

 ; CHECK-NEXT:    [[TMP7:%.*]] = bitcast i32* [[TMP6]] to <8 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <8 x i32> @llvm.masked.load.v8i32.p0v8i32(<8 x i32>* [[TMP7]], i32 4, <8 x i1> [[TMP4]], <8 x i32> poison)
 ; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, i32* [[B:%.*]], i64 [[TMP3]]
 ; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i32, i32* [[TMP8]], i32 0
 ; CHECK-NEXT:    [[TMP10:%.*]] = bitcast i32* [[TMP9]] to <8 x i32>*
 ; CHECK-NEXT:    [[WIDE_MASKED_LOAD3:%.*]] = call <8 x i32> @llvm.masked.load.v8i32.p0v8i32(<8 x i32>* [[TMP10]], i32 4, <8 x i1> [[TMP4]], <8 x i32> poison)
 ; CHECK-NEXT:    [[TMP11:%.*]] = add nsw <8 x i32> [[WIDE_MASKED_LOAD3]], [[WIDE_MASKED_LOAD]]
 ; CHECK-NEXT:    [[TMP12]] = add <8 x i32> [[TMP11]], [[VEC_PHI]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 8
 ; CHECK-NEXT:    [[TMP13:%.*]] = select <8 x i1> [[TMP4]], <8 x i32> [[TMP12]], <8 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 8
 ; CHECK-NEXT:    [[TMP14:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP14]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP6:!llvm.loop !.*]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP15:%.*]] = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> [[TMP13]])
 ; CHECK-NEXT:    br i1 true, label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[TMP15]], [[MIDDLE_BLOCK]] ]

llvm/test/Transforms/LoopVectorize/first-order-recurrence-sink-replicate-region.ll

 ; REQUIRES: asserts
 ; RUN: opt < %s -loop-vectorize -force-vector-width=2 -force-vector-interleave=1 -disable-output -debug-only=loop-vectorize 2>&1 | FileCheck %s
 
 target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
 
 ; Test cases for PR50009, which require sinking a replicate-region due to a
 ; first-order recurrence.
 
 define void @sink_replicate_region_1(i32 %x, i8* %ptr) optsize {
 ; CHECK-LABEL: sink_replicate_region_1
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   FIRST-ORDER-RECURRENCE-PHI ir<%0> = phi ir<0>, ir<%conv>
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>

 ; CHECK-NEXT:   pred.srem.continue:
 ; CHECK-NEXT:     PHI-PREDICATED-INSTRUCTION vp<[[PRED2:%.+]]> = ir<%rem>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): loop.1.split
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.1.split:
 ; CHECK-NEXT:   WIDEN ir<%add> = add ir<%conv>, vp<[[PRED2]]>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 

 
 exit:
   ret void
 }
 
 define void @sink_replicate_region_2(i32 %x, i8 %y, i32* %ptr) optsize {
 ; CHECK-LABEL: sink_replicate_region_2
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   FIRST-ORDER-RECURRENCE-PHI ir<%recur> = phi ir<0>, ir<%recur.next>
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>

 ; CHECK-EMPTY:
 ; CHECK-NEXT:   pred.store.continue:
 ; CHECK-NEXT:     PHI-PREDICATED-INSTRUCTION vp<[[PRED:%.+]]> = ir<%rem>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): loop.1
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.1:
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 

 
 exit:
   ret void
 }
 
 define i32 @sink_replicate_region_3_reduction(i32 %x, i8 %y, i32* %ptr) optsize {
 ; CHECK-LABEL: sink_replicate_region_3_reduction
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   FIRST-ORDER-RECURRENCE-PHI ir<%recur> = phi ir<0>, ir<%recur.next>
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   WIDEN-REDUCTION-PHI ir<%and.red> = phi ir<1234>, ir<%and.red.next>

 ; CHECK-NEXT:     PHI-PREDICATED-INSTRUCTION vp<[[PRED:%.+]]> = ir<%rem>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): loop.0.split
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.0.split:
 ; CHECK-NEXT:   WIDEN ir<%add> = add vp<[[PRED]]>, ir<%recur.next>
 ; CHECK-NEXT:   WIDEN ir<%and.red.next> = and ir<%and.red>, ir<%add>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
 ; CHECK-NEXT:   EMIT vp<[[SEL:%.+]]> = select vp<[[MASK]]> ir<%and.red.next> ir<%and.red>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 
exit:
   ret i32 %res
 }
 
 ; To sink the replicate region containing %rem, we need to split the block
 ; containing %conv at the end, because %conv is the last recipe in the block.
 define void @sink_replicate_region_4_requires_split_at_end_of_block(i32 %x, i8* %ptr) optsize {
 ; CHECK-LABEL: sink_replicate_region_4_requires_split_at_end_of_block
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   FIRST-ORDER-RECURRENCE-PHI ir<%0> = phi ir<0>, ir<%conv>
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>

 ; CHECK-NEXT:     PHI-PREDICATED-INSTRUCTION vp<[[PRED2:%.+]]> = ir<%lv.2>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 
 ; CHECK:      loop.2:
 ; CHECK-NEXT:   WIDEN ir<%add.1> = add ir<%conv>, vp<[[PRED1]]>
 ; CHECK-NEXT:   WIDEN ir<%conv.lv.2> = sext vp<[[PRED2]]>
 ; CHECK-NEXT:   WIDEN ir<%add> = add ir<%add.1>, ir<%conv.lv.2>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 

 exit:
   ret void
 }
 
 ; Test case that requires sinking a recipe in a replicate region after another replicate region.
 define void @sink_replicate_region_after_replicate_region(i32* %ptr, i32 %x, i8 %y) optsize {
 ; CHECK-LABEL: sink_replicate_region_after_replicate_region
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   FIRST-ORDER-RECURRENCE-PHI ir<%recur> = phi ir<0>, ir<%recur.next>
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>

 ; CHECK-EMPTY:
 ; CHECK-NEXT:   pred.store.continue:
 ; CHECK-NEXT:     PHI-PREDICATED-INSTRUCTION vp<[[PRED2:%.+]]> = ir<%rem.div>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): loop.2
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.2:
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 

llvm/test/Transforms/LoopVectorize/first-order-recurrence.ll

 ; UNROLL-NO-IC-NEXT:    [[TMP42:%.*]] = insertelement <4 x i32> [[TMP38]], i32 [[TMP41]], i32 3
 ; UNROLL-NO-IC-NEXT:    br label [[PRED_UDIV_CONTINUE19]]
 ; UNROLL-NO-IC:       pred.udiv.continue19:
 ; UNROLL-NO-IC-NEXT:    [[TMP43]] = phi <4 x i32> [ [[TMP38]], [[PRED_UDIV_CONTINUE17]] ], [ [[TMP42]], [[PRED_UDIV_IF18]] ]
 ; UNROLL-NO-IC-NEXT:    [[TMP44:%.*]] = shufflevector <4 x i32> [[VECTOR_RECUR]], <4 x i32> [[TMP23]], <4 x i32> <i32 3, i32 4, i32 5, i32 6>
 ; UNROLL-NO-IC-NEXT:    [[TMP45:%.*]] = shufflevector <4 x i32> [[TMP23]], <4 x i32> [[TMP43]], <4 x i32> <i32 3, i32 4, i32 5, i32 6>
 ; UNROLL-NO-IC-NEXT:    [[TMP46]] = add <4 x i32> [[VEC_PHI]], [[TMP44]]
 ; UNROLL-NO-IC-NEXT:    [[TMP47]] = add <4 x i32> [[VEC_PHI2]], [[TMP45]]
-; UNROLL-NO-IC-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 8
 ; UNROLL-NO-IC-NEXT:    [[TMP48:%.*]] = select <4 x i1> [[TMP2]], <4 x i32> [[TMP46]], <4 x i32> [[VEC_PHI]]
 ; UNROLL-NO-IC-NEXT:    [[TMP49:%.*]] = select <4 x i1> [[TMP3]], <4 x i32> [[TMP47]], <4 x i32> [[VEC_PHI2]]
+; UNROLL-NO-IC-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 8
 ; UNROLL-NO-IC-NEXT:    [[VEC_IND_NEXT]] = add <4 x i32> [[STEP_ADD]], <i32 -4, i32 -4, i32 -4, i32 -4>
 ; UNROLL-NO-IC-NEXT:    [[TMP50:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; UNROLL-NO-IC-NEXT:    br i1 [[TMP50]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !prof [[PROF52:![0-9]+]], !llvm.loop [[LOOP53:![0-9]+]]
 ; UNROLL-NO-IC:       middle.block:
 ; UNROLL-NO-IC-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP49]], [[TMP48]]
 ; UNROLL-NO-IC-NEXT:    [[TMP51:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX]])
 ; UNROLL-NO-IC-NEXT:    [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <4 x i32> [[TMP43]], i32 3
 ; UNROLL-NO-IC-NEXT:    [[VECTOR_RECUR_EXTRACT_FOR_PHI:%.*]] = extractelement <4 x i32> [[TMP43]], i32 2

 ; UNROLL-NO-VF:       pred.udiv.if4:
 ; UNROLL-NO-VF-NEXT:    [[INDUCTION1:%.*]] = add i32 [[OFFSET_IDX]], -1
 ; UNROLL-NO-VF-NEXT:    [[TMP6:%.*]] = udiv i32 219220132, [[INDUCTION1]]
 ; UNROLL-NO-VF-NEXT:    br label [[PRED_UDIV_CONTINUE5]]
 ; UNROLL-NO-VF:       pred.udiv.continue5:
 ; UNROLL-NO-VF-NEXT:    [[TMP7]] = phi i32 [ poison, [[PRED_UDIV_CONTINUE]] ], [ [[TMP6]], [[PRED_UDIV_IF4]] ]
 ; UNROLL-NO-VF-NEXT:    [[TMP8]] = add i32 [[VEC_PHI]], [[VECTOR_RECUR]]
 ; UNROLL-NO-VF-NEXT:    [[TMP9]] = add i32 [[VEC_PHI2]], [[TMP5]]
-; UNROLL-NO-VF-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 2
 ; UNROLL-NO-VF-NEXT:    [[TMP10:%.*]] = select i1 [[TMP2]], i32 [[TMP8]], i32 [[VEC_PHI]]
 ; UNROLL-NO-VF-NEXT:    [[TMP11:%.*]] = select i1 [[TMP3]], i32 [[TMP9]], i32 [[VEC_PHI2]]
+; UNROLL-NO-VF-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 2
 ; UNROLL-NO-VF-NEXT:    [[TMP12:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; UNROLL-NO-VF-NEXT:    br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !prof [[PROF51:![0-9]+]], !llvm.loop [[LOOP52:![0-9]+]]
 ; UNROLL-NO-VF:       middle.block:
 ; UNROLL-NO-VF-NEXT:    [[BIN_RDX:%.*]] = add i32 [[TMP11]], [[TMP10]]
 ; UNROLL-NO-VF-NEXT:    br i1 true, label [[BB1:%.*]], label [[SCALAR_PH]]
 ; UNROLL-NO-VF:       scalar.ph:
 ; UNROLL-NO-VF-NEXT:    [[SCALAR_RECUR_INIT:%.*]] = phi i32 [ 0, [[BB:%.*]] ], [ [[TMP7]], [[MIDDLE_BLOCK]] ]
 ; UNROLL-NO-VF-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[Y]], [[BB]] ]

 ; SINK-AFTER-NEXT:    [[TMP19:%.*]] = add i32 [[OFFSET_IDX]], -3
 ; SINK-AFTER-NEXT:    [[TMP20:%.*]] = udiv i32 219220132, [[TMP19]]
 ; SINK-AFTER-NEXT:    [[TMP21:%.*]] = insertelement <4 x i32> [[TMP17]], i32 [[TMP20]], i32 3
 ; SINK-AFTER-NEXT:    br label [[PRED_UDIV_CONTINUE8]]
 ; SINK-AFTER:       pred.udiv.continue8:
 ; SINK-AFTER-NEXT:    [[TMP22]] = phi <4 x i32> [ [[TMP17]], [[PRED_UDIV_CONTINUE6]] ], [ [[TMP21]], [[PRED_UDIV_IF7]] ]
 ; SINK-AFTER-NEXT:    [[TMP23:%.*]] = shufflevector <4 x i32> [[VECTOR_RECUR]], <4 x i32> [[TMP22]], <4 x i32> <i32 3, i32 4, i32 5, i32 6>
 ; SINK-AFTER-NEXT:    [[TMP24]] = add <4 x i32> [[VEC_PHI]], [[TMP23]]
-; SINK-AFTER-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; SINK-AFTER-NEXT:    [[TMP25:%.*]] = select <4 x i1> [[TMP2]], <4 x i32> [[TMP24]], <4 x i32> [[VEC_PHI]]
+; SINK-AFTER-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; SINK-AFTER-NEXT:    [[VEC_IND_NEXT]] = add <4 x i32> [[VEC_IND]], <i32 -4, i32 -4, i32 -4, i32 -4>
 ; SINK-AFTER-NEXT:    [[TMP26:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; SINK-AFTER-NEXT:    br i1 [[TMP26]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !prof [[PROF52:![0-9]+]], !llvm.loop [[LOOP53:![0-9]+]]
 ; SINK-AFTER:       middle.block:
 ; SINK-AFTER-NEXT:    [[TMP27:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[TMP25]])
 ; SINK-AFTER-NEXT:    [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <4 x i32> [[TMP22]], i32 3
 ; SINK-AFTER-NEXT:    [[VECTOR_RECUR_EXTRACT_FOR_PHI:%.*]] = extractelement <4 x i32> [[TMP22]], i32 2
 ; SINK-AFTER-NEXT:    br i1 true, label [[BB1:%.*]], label [[SCALAR_PH]]

 ; NO-SINK-AFTER-NEXT:    [[TMP19:%.*]] = add i32 [[OFFSET_IDX]], -3
 ; NO-SINK-AFTER-NEXT:    [[TMP20:%.*]] = udiv i32 219220132, [[TMP19]]
 ; NO-SINK-AFTER-NEXT:    [[TMP21:%.*]] = insertelement <4 x i32> [[TMP17]], i32 [[TMP20]], i32 3
 ; NO-SINK-AFTER-NEXT:    br label [[PRED_UDIV_CONTINUE8]]
 ; NO-SINK-AFTER:       pred.udiv.continue8:
 ; NO-SINK-AFTER-NEXT:    [[TMP22]] = phi <4 x i32> [ [[TMP17]], [[PRED_UDIV_CONTINUE6]] ], [ [[TMP21]], [[PRED_UDIV_IF7]] ]
 ; NO-SINK-AFTER-NEXT:    [[TMP23:%.*]] = shufflevector <4 x i32> [[VECTOR_RECUR]], <4 x i32> [[TMP22]], <4 x i32> <i32 3, i32 4, i32 5, i32 6>
 ; NO-SINK-AFTER-NEXT:    [[TMP24]] = add <4 x i32> [[VEC_PHI]], [[TMP23]]
-; NO-SINK-AFTER-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; NO-SINK-AFTER-NEXT:    [[TMP25:%.*]] = select <4 x i1> [[TMP2]], <4 x i32> [[TMP24]], <4 x i32> [[VEC_PHI]]
+; NO-SINK-AFTER-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; NO-SINK-AFTER-NEXT:    [[VEC_IND_NEXT]] = add <4 x i32> [[VEC_IND]], <i32 -4, i32 -4, i32 -4, i32 -4>
 ; NO-SINK-AFTER-NEXT:    [[TMP26:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; NO-SINK-AFTER-NEXT:    br i1 [[TMP26]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !prof [[PROF52:![0-9]+]], !llvm.loop [[LOOP53:![0-9]+]]
 ; NO-SINK-AFTER:       middle.block:
 ; NO-SINK-AFTER-NEXT:    [[TMP27:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[TMP25]])
 ; NO-SINK-AFTER-NEXT:    [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <4 x i32> [[TMP22]], i32 3
 ; NO-SINK-AFTER-NEXT:    [[VECTOR_RECUR_EXTRACT_FOR_PHI:%.*]] = extractelement <4 x i32> [[TMP22]], i32 2
 ; NO-SINK-AFTER-NEXT:    br i1 true, label [[BB1:%.*]], label [[SCALAR_PH]]

 ; UNROLL-NO-IC-NEXT:    [[TMP69:%.*]] = extractelement <4 x i1> [[TMP11]], i32 3
 ; UNROLL-NO-IC-NEXT:    br i1 [[TMP69]], label [[PRED_STORE_IF34:%.*]], label [[PRED_STORE_CONTINUE35]]
 ; UNROLL-NO-IC:       pred.store.if34:
 ; UNROLL-NO-IC-NEXT:    [[TMP70:%.*]] = add i32 [[INDEX]], 7
 ; UNROLL-NO-IC-NEXT:    [[TMP71:%.*]] = getelementptr inbounds i32, i32* [[X]], i32 [[TMP70]]
 ; UNROLL-NO-IC-NEXT:    store i32 [[TMP9]], i32* [[TMP71]], align 4
 ; UNROLL-NO-IC-NEXT:    br label [[PRED_STORE_CONTINUE35]]
 ; UNROLL-NO-IC:       pred.store.continue35:
-; UNROLL-NO-IC-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 8
 ; UNROLL-NO-IC-NEXT:    [[TMP72:%.*]] = select <4 x i1> [[TMP10]], <4 x i32> [[TMP46]], <4 x i32> [[VEC_PHI]]
 ; UNROLL-NO-IC-NEXT:    [[TMP73:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[TMP47]], <4 x i32> [[VEC_PHI7]]
+; UNROLL-NO-IC-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 8
 ; UNROLL-NO-IC-NEXT:    [[VEC_IND_NEXT]] = add <4 x i32> [[STEP_ADD]], <i32 -4, i32 -4, i32 -4, i32 -4>
 ; UNROLL-NO-IC-NEXT:    [[VEC_IND_NEXT6]] = add <4 x i32> [[STEP_ADD4]], <i32 4, i32 4, i32 4, i32 4>
 ; UNROLL-NO-IC-NEXT:    [[TMP74:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; UNROLL-NO-IC-NEXT:    br i1 [[TMP74]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !prof [[PROF52]], !llvm.loop [[LOOP56:![0-9]+]]
 ; UNROLL-NO-IC:       middle.block:
 ; UNROLL-NO-IC-NEXT:    [[BIN_RDX:%.*]] = add <4 x i32> [[TMP73]], [[TMP72]]
 ; UNROLL-NO-IC-NEXT:    [[TMP75:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX]])
 ; UNROLL-NO-IC-NEXT:    [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <4 x i32> [[TMP43]], i32 3

 ; UNROLL-NO-VF-NEXT:    br label [[PRED_STORE_CONTINUE]]
 ; UNROLL-NO-VF:       pred.store.continue:
 ; UNROLL-NO-VF-NEXT:    br i1 [[TMP3]], label [[PRED_STORE_IF8:%.*]], label [[PRED_STORE_CONTINUE9]]
 ; UNROLL-NO-VF:       pred.store.if8:
 ; UNROLL-NO-VF-NEXT:    [[TMP11:%.*]] = getelementptr inbounds i32, i32* [[X]], i32 [[INDUCTION4]]
 ; UNROLL-NO-VF-NEXT:    store i32 [[INDUCTION2]], i32* [[TMP11]], align 4
 ; UNROLL-NO-VF-NEXT:    br label [[PRED_STORE_CONTINUE9]]
 ; UNROLL-NO-VF:       pred.store.continue9:
-; UNROLL-NO-VF-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 2
 ; UNROLL-NO-VF-NEXT:    [[TMP12:%.*]] = select i1 [[TMP2]], i32 [[TMP8]], i32 [[VEC_PHI]]
 ; UNROLL-NO-VF-NEXT:    [[TMP13:%.*]] = select i1 [[TMP3]], i32 [[TMP9]], i32 [[VEC_PHI5]]
+; UNROLL-NO-VF-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 2
 ; UNROLL-NO-VF-NEXT:    [[TMP14:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; UNROLL-NO-VF-NEXT:    br i1 [[TMP14]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !prof [[PROF51]], !llvm.loop [[LOOP55:![0-9]+]]
 ; UNROLL-NO-VF:       middle.block:
 ; UNROLL-NO-VF-NEXT:    [[BIN_RDX:%.*]] = add i32 [[TMP13]], [[TMP12]]
 ; UNROLL-NO-VF-NEXT:    br i1 true, label [[BB1:%.*]], label [[SCALAR_PH]]
 ; UNROLL-NO-VF:       scalar.ph:
 ; UNROLL-NO-VF-NEXT:    [[SCALAR_RECUR_INIT:%.*]] = phi i32 [ 0, [[BB:%.*]] ], [ [[TMP7]], [[MIDDLE_BLOCK]] ]
 ; UNROLL-NO-VF-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[Y]], [[BB]] ]

 ; SINK-AFTER-NEXT:    [[TMP34:%.*]] = extractelement <4 x i1> [[TMP6]], i32 3
 ; SINK-AFTER-NEXT:    br i1 [[TMP34]], label [[PRED_STORE_IF14:%.*]], label [[PRED_STORE_CONTINUE15]]
 ; SINK-AFTER:       pred.store.if14:
 ; SINK-AFTER-NEXT:    [[TMP35:%.*]] = add i32 [[INDEX]], 3
 ; SINK-AFTER-NEXT:    [[TMP36:%.*]] = getelementptr inbounds i32, i32* [[X]], i32 [[TMP35]]
 ; SINK-AFTER-NEXT:    store i32 [[TMP5]], i32* [[TMP36]], align 4
 ; SINK-AFTER-NEXT:    br label [[PRED_STORE_CONTINUE15]]
 ; SINK-AFTER:       pred.store.continue15:
-; SINK-AFTER-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; SINK-AFTER-NEXT:    [[TMP37:%.*]] = select <4 x i1> [[TMP6]], <4 x i32> [[TMP24]], <4 x i32> [[VEC_PHI]]
+; SINK-AFTER-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; SINK-AFTER-NEXT:    [[VEC_IND_NEXT]] = add <4 x i32> [[VEC_IND]], <i32 -4, i32 -4, i32 -4, i32 -4>
 ; SINK-AFTER-NEXT:    [[VEC_IND_NEXT3]] = add <4 x i32> [[VEC_IND2]], <i32 4, i32 4, i32 4, i32 4>
 ; SINK-AFTER-NEXT:    [[TMP38:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; SINK-AFTER-NEXT:    br i1 [[TMP38]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !prof [[PROF52]], !llvm.loop [[LOOP56:![0-9]+]]
 ; SINK-AFTER:       middle.block:
 ; SINK-AFTER-NEXT:    [[TMP39:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[TMP37]])
 ; SINK-AFTER-NEXT:    [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <4 x i32> [[TMP22]], i32 3
 ; SINK-AFTER-NEXT:    [[VECTOR_RECUR_EXTRACT_FOR_PHI:%.*]] = extractelement <4 x i32> [[TMP22]], i32 2

 ; NO-SINK-AFTER-NEXT:    [[TMP34:%.*]] = extractelement <4 x i1> [[TMP6]], i32 3
 ; NO-SINK-AFTER-NEXT:    br i1 [[TMP34]], label [[PRED_STORE_IF14:%.*]], label [[PRED_STORE_CONTINUE15]]
 ; NO-SINK-AFTER:       pred.store.if14:
 ; NO-SINK-AFTER-NEXT:    [[TMP35:%.*]] = add i32 [[INDEX]], 3
 ; NO-SINK-AFTER-NEXT:    [[TMP36:%.*]] = getelementptr inbounds i32, i32* [[X]], i32 [[TMP35]]
 ; NO-SINK-AFTER-NEXT:    store i32 [[TMP5]], i32* [[TMP36]], align 4
 ; NO-SINK-AFTER-NEXT:    br label [[PRED_STORE_CONTINUE15]]
 ; NO-SINK-AFTER:       pred.store.continue15:
-; NO-SINK-AFTER-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; NO-SINK-AFTER-NEXT:    [[TMP37:%.*]] = select <4 x i1> [[TMP6]], <4 x i32> [[TMP24]], <4 x i32> [[VEC_PHI]]
+; NO-SINK-AFTER-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 4
 ; NO-SINK-AFTER-NEXT:    [[VEC_IND_NEXT]] = add <4 x i32> [[VEC_IND]], <i32 -4, i32 -4, i32 -4, i32 -4>
 ; NO-SINK-AFTER-NEXT:    [[VEC_IND_NEXT3]] = add <4 x i32> [[VEC_IND2]], <i32 4, i32 4, i32 4, i32 4>
 ; NO-SINK-AFTER-NEXT:    [[TMP38:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
 ; NO-SINK-AFTER-NEXT:    br i1 [[TMP38]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !prof [[PROF52]], !llvm.loop [[LOOP56:![0-9]+]]
 ; NO-SINK-AFTER:       middle.block:
 ; NO-SINK-AFTER-NEXT:    [[TMP39:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[TMP37]])
 ; NO-SINK-AFTER-NEXT:    [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <4 x i32> [[TMP22]], i32 3
 ; NO-SINK-AFTER-NEXT:    [[VECTOR_RECUR_EXTRACT_FOR_PHI:%.*]] = extractelement <4 x i32> [[TMP22]], i32 2

llvm/test/Transforms/LoopVectorize/icmp-uniforms.ll

 
 for.end:
   %tmp4 = phi i32 [ %tmp3, %for.body ]
   ret i32 %tmp4
 }
 
 ; Check for crash exposed by D76992.
 ; CHECK:      VPlan 'Initial VPlan for VF={4},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   WIDEN ir<%cond0> = icmp ir<%iv>, ir<13>
 ; CHECK-NEXT:   WIDEN-SELECT ir<%s> = select ir<%cond0>, ir<10>, ir<20>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successor
 ; CHECK-NEXT: }
 define void @test() {
 entry:
   br label %loop
 
 loop:                       ; preds = %loop, %entry
   %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]

llvm/test/Transforms/LoopVectorize/outer-loop-vec-phi-predecessor-order.ll

 ; CHECK-NEXT:    [[TMP5]] = add nuw nsw <4 x i64> [[VEC_PHI]], <i64 1, i64 1, i64 1, i64 1>
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq <4 x i64> [[TMP5]], [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <4 x i1> [[TMP6]], i32 0
 ; CHECK-NEXT:    br i1 [[TMP7]], label [[LOOP_1_LATCH5]], label [[LOOP_2_HEADER1]]
 ; CHECK:       loop.1.latch5:
 ; CHECK-NEXT:    [[TMP8:%.*]] = add nuw nsw <4 x i64> [[VEC_IND]], <i64 1, i64 1, i64 1, i64 1>
 ; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq <4 x i64> [[TMP8]], [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP10:%.*]] = extractelement <4 x i1> [[TMP9]], i32 0
 ; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <4 x i64> [[VEC_IND]], <i64 4, i64 4, i64 4, i64 4>
 ; CHECK-NEXT:    [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP11]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP0:!llvm.loop !.*]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]

llvm/test/Transforms/LoopVectorize/pr51614-fold-tail-by-masking.ll

 ; CHECK-NEXT:    [[TMP20:%.*]] = load i16, i16* [[TMP19]], align 1
 ; CHECK-NEXT:    [[TMP21:%.*]] = insertelement <2 x i16> [[TMP12]], i16 [[TMP20]], i32 1
 ; CHECK-NEXT:    br label [[PRED_LOAD_CONTINUE4]]
 ; CHECK:       pred.load.continue4:
 ; CHECK-NEXT:    [[TMP22:%.*]] = phi <2 x i16> [ [[TMP11]], [[PRED_LOAD_CONTINUE]] ], [ [[TMP18]], [[PRED_LOAD_IF3]] ]
 ; CHECK-NEXT:    [[TMP23:%.*]] = phi <2 x i16> [ [[TMP12]], [[PRED_LOAD_CONTINUE]] ], [ [[TMP21]], [[PRED_LOAD_IF3]] ]
 ; CHECK-NEXT:    [[TMP24:%.*]] = add nsw <2 x i16> [[TMP22]], [[TMP23]]
 ; CHECK-NEXT:    [[TMP25]] = add <2 x i16> [[VEC_PHI]], [[TMP24]]
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 2
 ; CHECK-NEXT:    [[TMP26:%.*]] = select <2 x i1> [[TMP1]], <2 x i16> [[TMP25]], <2 x i16> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i32 [[INDEX]], 2
 ; CHECK-NEXT:    [[TMP27:%.*]] = icmp eq i32 [[INDEX_NEXT]], 42
 ; CHECK-NEXT:    br i1 [[TMP27]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP28:%.*]] = call i16 @llvm.vector.reduce.add.v2i16(<2 x i16> [[TMP26]])
 ; CHECK-NEXT:    br i1 true, label [[EXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i16 [ -1, [[MIDDLE_BLOCK]] ], [ 41, [[ENTRY:%.*]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i16 [ 0, [[ENTRY]] ], [ [[TMP28]], [[MIDDLE_BLOCK]] ]

llvm/test/Transforms/LoopVectorize/select-reduction.ll

 ; CHECK-NEXT:    [[INDUCTION:%.*]] = add <4 x i64> [[BROADCAST_SPLAT2]], <i64 0, i64 -1, i64 -2, i64 -3>
 ; CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[OFFSET_IDX]], 0
 ; CHECK-NEXT:    [[BROADCAST_SPLATINSERT3:%.*]] = insertelement <4 x i64> poison, i64 [[INDEX]], i32 0
 ; CHECK-NEXT:    [[BROADCAST_SPLAT4:%.*]] = shufflevector <4 x i64> [[BROADCAST_SPLATINSERT3]], <4 x i64> poison, <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[VEC_IV:%.*]] = add <4 x i64> [[BROADCAST_SPLAT4]], <i64 0, i64 1, i64 2, i64 3>
 ; CHECK-NEXT:    [[TMP1:%.*]] = icmp ule <4 x i64> [[VEC_IV]], [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[TMP2:%.*]] = icmp sgt <4 x i32> [[VEC_PHI]], <i32 10, i32 10, i32 10, i32 10>
 ; CHECK-NEXT:    [[TMP3]] = select <4 x i1> [[TMP2]], <4 x i32> [[VEC_PHI]], <4 x i32> <i32 10, i32 10, i32 10, i32 10>
-; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP4:%.*]] = select <4 x i1> [[TMP1]], <4 x i32> [[TMP3]], <4 x i32> [[VEC_PHI]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP5]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP0:!llvm.loop !.*]]
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[TMP6:%.*]] = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> [[TMP4]])
 ; CHECK-NEXT:    br i1 true, label [[EXIT_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[EXTRA_ITER]], [[LOOP_PREHEADER]] ]
 ; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[LOOP_PREHEADER]] ], [ [[TMP6]], [[MIDDLE_BLOCK]] ]

llvm/test/Transforms/LoopVectorize/vplan-dot-printing.ll

 ; CHECK-NEXT:    "for.body:\l" +
 ; CHECK-NEXT:    "  EMIT vp\<[[CAN_IV:%.+]]\> = CANONICAL-INDUCTION\l" +
 ; CHECK-NEXT:    "  WIDEN-INDUCTION %iv = phi %iv.next, 0\l" +
 ; CHECK-NEXT:    "  CLONE ir\<%arrayidx\> = getelementptr ir\<%y\>, ir\<%iv\>\l" +
 ; CHECK-NEXT:    "  WIDEN ir\<%lv\> = load ir\<%arrayidx\>\l" +
 ; CHECK-NEXT:    "  WIDEN-CALL ir\<%call\> = call @llvm.sqrt.f32(ir\<%lv\>)\l" +
 ; CHECK-NEXT:    "  CLONE ir\<%arrayidx2\> = getelementptr ir\<%x\>, ir\<%iv\>\l" +
 ; CHECK-NEXT:    "  WIDEN store ir\<%arrayidx2\>, ir\<%call\>\l" +
-; CHECK-NEXT:    "  EMIT vp\<{{.+}}\> = VF * UF +(nuw) vp\<[[CAN_IV]]\>\l" +
+; CHECK-NEXT:    "  EMIT vp\<[[CAN_IV_NEXT:%.+]]\> = VF * UF +(nuw) vp\<[[CAN_IV]]\>\l" +
+; CHECK-NEXT:    "  EMIT branch-on-count vp\<[[CAN_IV_NEXT]]\> vp\<{{.+}}\>\l" +
 ; CHECK-NEXT:    "No successors\l"
 ; CHECK-NEXT:  ]
 ;
 entry:
   %cmp6 = icmp sgt i64 %n, 0
   br i1 %cmp6, label %for.body, label %for.end
 
 for.body:                                         ; preds = %entry, %for.body

llvm/test/Transforms/LoopVectorize/vplan-printing.ll

 ; REQUIRES: asserts
 
 ; RUN: opt -loop-vectorize -debug-only=loop-vectorize -force-vector-interleave=1 -force-vector-width=4 -prefer-inloop-reductions -enable-interleaved-mem-accesses=true -enable-masked-interleaved-mem-accesses -disable-output %s 2>&1 | FileCheck %s
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
 
 ; Tests for printing VPlans.
 
 define void @print_call_and_memory(i64 %n, float* noalias %y, float* noalias %x) nounwind uwtable {
 ; CHECK-LABEL: Checking a loop in "print_call_and_memory"
 ; CHECK:      VPlan 'Initial VPlan for VF={4},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: for.body:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi %iv.next, 0
 ; CHECK-NEXT:   CLONE ir<%arrayidx> = getelementptr ir<%y>, ir<%iv>
 ; CHECK-NEXT:   WIDEN ir<%lv> = load ir<%arrayidx>
 ; CHECK-NEXT:   WIDEN-CALL ir<%call> = call @llvm.sqrt.f32(ir<%lv>)
 ; CHECK-NEXT:   CLONE ir<%arrayidx2> = getelementptr ir<%x>, ir<%iv>
 ; CHECK-NEXT:   WIDEN store ir<%arrayidx2>, ir<%call>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   %cmp6 = icmp sgt i64 %n, 0
   br i1 %cmp6, label %for.body, label %for.end

 
 for.end:                                          ; preds = %for.body, %entry
   ret void
 }
 
 define void @print_widen_gep_and_select(i64 %n, float* noalias %y, float* noalias %x, float* %z) nounwind uwtable {
 ; CHECK-LABEL: Checking a loop in "print_widen_gep_and_select"
 ; CHECK:      VPlan 'Initial VPlan for VF={4},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: for.body:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi %iv.next, 0
 ; CHECK-NEXT:   WIDEN-GEP Inv[Var] ir<%arrayidx> = getelementptr ir<%y>, ir<%iv>
 ; CHECK-NEXT:   WIDEN ir<%lv> = load ir<%arrayidx>
 ; CHECK-NEXT:   WIDEN ir<%cmp> = icmp ir<%arrayidx>, ir<%z>
 ; CHECK-NEXT:   WIDEN-SELECT ir<%sel> = select ir<%cmp>, ir<1.000000e+01>, ir<2.000000e+01>
 ; CHECK-NEXT:   WIDEN ir<%add> = fadd ir<%lv>, ir<%sel>
 ; CHECK-NEXT:   CLONE ir<%arrayidx2> = getelementptr ir<%x>, ir<%iv>
 ; CHECK-NEXT:   WIDEN store ir<%arrayidx2>, ir<%add>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   %cmp6 = icmp sgt i64 %n, 0
   br i1 %cmp6, label %for.body, label %for.end

 
 for.end:                                          ; preds = %for.body, %entry
   ret void
 }
 
 define float @print_reduction(i64 %n, float* noalias %y) {
 ; CHECK-LABEL: Checking a loop in "print_reduction"
 ; CHECK:      VPlan 'Initial VPlan for VF={4},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: for.body:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi %iv.next, 0
 ; CHECK-NEXT:   WIDEN-REDUCTION-PHI ir<%red> = phi ir<0.000000e+00>, ir<%red.next>
 ; CHECK-NEXT:   CLONE ir<%arrayidx> = getelementptr ir<%y>, ir<%iv>
 ; CHECK-NEXT:   WIDEN ir<%lv> = load ir<%arrayidx>
 ; CHECK-NEXT:   REDUCE ir<%red.next> = ir<%red> + fast reduce.fadd (ir<%lv>)
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %for.body
 

 
 for.end:                                          ; preds = %for.body, %entry
   ret float %red.next
 }
 
 define void @print_replicate_predicated_phi(i64 %n, i64* %x) {
 ; CHECK-LABEL: Checking a loop in "print_replicate_predicated_phi"
 ; CHECK:      VPlan 'Initial VPlan for VF={4},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: for.body:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %i = phi 0, %i.next
 ; CHECK-NEXT:   WIDEN ir<%cmp> = icmp ir<%i>, ir<5>
 ; CHECK-NEXT: Successor(s): if.then
 ; CHECK-EMPTY:
 ; CHECK-NEXT: if.then:

 ; CHECK-NEXT: if.then.0:
 ; CHECK-NEXT: Successor(s): for.inc
 ; CHECK-EMPTY:
 ; CHECK-NEXT: for.inc:
 ; CHECK-NEXT:   EMIT vp<[[NOT:%.+]]> = not ir<%cmp>
 ; CHECK-NEXT:   BLEND %d = ir<0>/vp<[[NOT]]> vp<[[PRED]]>/ir<%cmp>
 ; CHECK-NEXT:   CLONE ir<%idx> = getelementptr ir<%x>, ir<%i>
 ; CHECK-NEXT:   WIDEN store ir<%idx>, ir<%d>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %for.body
 

 }
 
 @AB = common global [1024 x i32] zeroinitializer, align 4
 @CD = common global [1024 x i32] zeroinitializer, align 4
 
 define void @print_interleave_groups(i32 %C, i32 %D) {
 ; CHECK-LABEL: Checking a loop in "print_interleave_groups"
 ; CHECK:       VPlan 'Initial VPlan for VF={4},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT:  for.body:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   CLONE ir<%gep.AB.0> = getelementptr ir<@AB>, ir<0>, ir<%iv>
 ; CHECK-NEXT:   INTERLEAVE-GROUP with factor 4 at %AB.0, ir<%gep.AB.0>
 ; CHECK-NEXT:     ir<%AB.0> = load from index 0
 ; CHECK-NEXT:     ir<%AB.1> = load from index 1
 ; CHECK-NEXT:     ir<%AB.3> = load from index 3
 ; CHECK-NEXT:   CLONE ir<%iv.plus.1> = add ir<%iv>, ir<1>
 ; CHECK-NEXT:   CLONE ir<%gep.AB.1> = getelementptr ir<@AB>, ir<0>, ir<%iv.plus.1>
 ; CHECK-NEXT:   CLONE ir<%iv.plus.2> = add ir<%iv>, ir<2>
 ; CHECK-NEXT:   CLONE ir<%iv.plus.3> = add ir<%iv>, ir<3>
 ; CHECK-NEXT:   CLONE ir<%gep.AB.3> = getelementptr ir<@AB>, ir<0>, ir<%iv.plus.3>
 ; CHECK-NEXT:   WIDEN ir<%add> = add ir<%AB.0>, ir<%AB.1>
 ; CHECK-NEXT:   CLONE ir<%gep.CD.0> = getelementptr ir<@CD>, ir<0>, ir<%iv>
 ; CHECK-NEXT:   CLONE ir<%gep.CD.1> = getelementptr ir<@CD>, ir<0>, ir<%iv.plus.1>
 ; CHECK-NEXT:   CLONE ir<%gep.CD.2> = getelementptr ir<@CD>, ir<0>, ir<%iv.plus.2>
 ; CHECK-NEXT:   CLONE ir<%gep.CD.3> = getelementptr ir<@CD>, ir<0>, ir<%iv.plus.3>
 ; CHECK-NEXT:   INTERLEAVE-GROUP with factor 4 at <badref>, ir<%gep.CD.3>
 ; CHECK-NEXT:     store ir<%add> to index 0
 ; CHECK-NEXT:     store ir<1> to index 1
 ; CHECK-NEXT:     store ir<2> to index 2
 ; CHECK-NEXT:     store ir<%AB.3> to index 3
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %for.body
 

 
 for.end:
   ret void
 }
 
 define float @print_fmuladd_strict(float* %a, float* %b, i64 %n) {
 ; CHECK-LABEL: Checking a loop in "print_fmuladd_strict"
 ; CHECK:      VPlan 'Initial VPlan for VF={4},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: for.body:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   WIDEN-REDUCTION-PHI ir<%sum.07> = phi ir<0.000000e+00>, ir<%muladd>
 ; CHECK-NEXT:   CLONE ir<%arrayidx> = getelementptr ir<%a>, ir<%iv>
 ; CHECK-NEXT:   WIDEN ir<%l.a> = load ir<%arrayidx>
 ; CHECK-NEXT:   CLONE ir<%arrayidx2> = getelementptr ir<%b>, ir<%iv>
 ; CHECK-NEXT:   WIDEN ir<%l.b> = load ir<%arrayidx2>
 ; CHECK-NEXT:   EMIT vp<[[FMUL:%.]]> = fmul nnan ninf nsz ir<%l.a> ir<%l.b>
 ; CHECK-NEXT:   REDUCE ir<[[MULADD:%.+]]> = ir<%sum.07> + nnan ninf nsz reduce.fadd (vp<[[FMUL]]>)
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 
 entry:
   br label %for.body
 
 for.body:
   %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]

 
 for.end:
   ret float %muladd
 }
 
 define void @debug_loc_vpinstruction(i32* nocapture %asd, i32* nocapture %bsd) !dbg !5 {
 ; CHECK-LABEL: Checking a loop in "debug_loc_vpinstruction"
 ; CHECK:    VPlan 'Initial VPlan for VF={4},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT:  loop:
 ; CHECK-NEXT:    EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:    WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:    CLONE ir<%isd> = getelementptr ir<%asd>, ir<%iv>
 ; CHECK-NEXT:    WIDEN ir<%lsd> = load ir<%isd>
 ; CHECK-NEXT:    WIDEN ir<%psd> = add ir<%lsd>, ir<23>
 ; CHECK-NEXT:    WIDEN ir<%cmp1> = icmp ir<%lsd>, ir<100>

 ; CHECK-NEXT:  if.then.0:
 ; CHECK-NEXT:  Successor(s): if.end
 ; CHECK-EMPTY:
 ; CHECK-NEXT:  if.end:
 ; CHECK-NEXT:    EMIT vp<[[NOT2:%.+]]> = not ir<%cmp2>
 ; CHECK-NEXT:    EMIT vp<[[SEL2:%.+]]> = select vp<[[NOT1]]> vp<[[NOT2]]> ir<false>
 ; CHECK-NEXT:    BLEND %ysd.0 = vp<[[PHI]]>/vp<[[OR1]]> ir<%psd>/vp<[[SEL2]]>
 ; CHECK-NEXT:    WIDEN store ir<%isd>, ir<%ysd.0>
-; CHECK-NEXT:    EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:    EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:    EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT:  No successors
 ; CHECK-NEXT:}
 ; CHECK-NEXT:No successors
 ; CHECK-NEXT:}
 ;
 entry:
   br label %loop
 

llvm/test/Transforms/LoopVectorize/vplan-sink-scalars-and-merge-vf1.ll

 ; REQUIRES: asserts
 
 ; RUN: opt -loop-vectorize -force-vector-interleave=2 -force-vector-width=1 -debug -disable-output %s 2>&1 | FileCheck %s
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
 
 ; Make sure recipes with side-effects are not sunk.
 define void @sink_with_sideeffects(i1 %c, i8* %ptr) {
 ; CHECK-LABEL: sink_with_sideeffects
 ; CHECK:      VPlan 'Initial VPlan for VF={1},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: for.body:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %tmp0 = phi %tmp6, 0
 ; CHECK-NEXT:   WIDEN-INDUCTION %tmp1 = phi %tmp7, 0
 ; CHECK-NEXT:   CLONE ir<%tmp2> = getelementptr ir<%ptr>, ir<%tmp0>
 ; CHECK-NEXT:   CLONE ir<%tmp3> = load ir<%tmp2>
 ; CHECK-NEXT:   CLONE store ir<0>, ir<%tmp2>

 ; CHECK:      pred.store.continue:
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 
 ; CHECK:      if.then.0:
 ; CHECK-NEXT: Successor(s): for.inc
 
 ; CHECK:      for.inc:
-; CHECK-NEXT:  EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:  EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:  EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %for.body
 

llvm/test/Transforms/LoopVectorize/vplan-sink-scalars-and-merge.ll

 ; REQUIRES: asserts
 
 ; RUN: opt -loop-vectorize -force-vector-interleave=1 -force-vector-width=2 -debug -disable-output %s 2>&1 | FileCheck %s
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
 
 @a = common global [2048 x i32] zeroinitializer, align 16
 @b = common global [2048 x i32] zeroinitializer, align 16
 @c = common global [2048 x i32] zeroinitializer, align 16
 
 
 ; CHECK-LABEL: LV: Checking a loop in "sink1"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>
 ; CHECK-NEXT: Successor(s): loop.0

 ; CHECK:      pred.store.continue:
 ; CHECK-NEXT:   PHI-PREDICATED-INSTRUCTION vp<[[PRED:%.+]]> = ir<%lv.b>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 
 ; CHECK:      loop.1:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 define void @sink1(i32 %k) {
 entry:
   br label %loop
 
 loop:
loop:
   br i1 %realexit, label %exit, label %loop
 
 exit:
   ret void
 }
 
 ; CHECK-LABEL: LV: Checking a loop in "sink2"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>
 ; CHECK-NEXT: Successor(s): pred.load

 
 ; CHECK:      pred.store.continue:
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 
 ; CHECK:       loop.1:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 define void @sink2(i32 %k) {
 entry:
   br label %loop
 
 loop:
loop:
   br i1 %realexit, label %exit, label %loop
 
 exit:
   ret void
 }
 
 ; CHECK-LABEL: LV: Checking a loop in "sink3"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>
 ; CHECK-NEXT: Successor(s): pred.load

 
 ; CHECK:      pred.store.continue:
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 
 ; CHECK:      loop.1:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 define void @sink3(i32 %k) {
 entry:
   br label %loop
 
 loop:

 exit:
   ret void
 }
 
 ; Make sure we do not sink uniform instructions.
 define void @uniform_gep(i64 %k, i16* noalias %A, i16* noalias %B) {
 ; CHECK-LABEL: LV: Checking a loop in "uniform_gep"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 21, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[WIDE_CAN_IV:%.+]]> = WIDEN-CANONICAL-INDUCTION
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule vp<[[WIDE_CAN_IV]]> vp<[[BTC]]>

 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): loop.then.0
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.then.0:
 ; CHECK-NEXT: Successor(s): loop.latch
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.latch:
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 
 loop:
   %iv = phi i64 [ 21, %entry ], [ %iv.next, %loop.latch ]

 exit:
   ret void
 }
 
 ; Loop with predicated load.
 define void @pred_cfg1(i32 %k, i32 %j) {
 ; CHECK-LABEL: LV: Checking a loop in "pred_cfg1"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK1:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>
 ; CHECK-NEXT:   WIDEN ir<%c.1> = icmp ir<%iv>, ir<%j>

 ; CHECK-NEXT:   pred.store.continue:
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): next.0.0
 ; CHECK-EMPTY:
 ; CHECK-NEXT: next.0.0:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 
 loop:
   %iv = phi i32 [ 0, %entry ], [ %iv.next, %next.0 ]
exit:
   ret void
 }
 
 ; Loop with predicated load and store in separate blocks, store depends on
 ; loaded value.
 define void @pred_cfg2(i32 %k, i32 %j) {
 ; CHECK-LABEL: LV: Checking a loop in "pred_cfg2"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK1:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>
 ; CHECK-NEXT:   WIDEN ir<%mul> = mul ir<%iv>, ir<10>

 ; CHECK-NEXT: Successor(s): then.1.0
 ; CHECK-EMPTY:
 ; CHECK-NEXT: then.1.0:
 ; CHECK-NEXT: Successor(s): next.1
 ; CHECK-EMPTY:
 ; CHECK-NEXT: next.1:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 
 loop:
   %iv = phi i32 [ 0, %entry ], [ %iv.next, %next.1 ]
exit:
   ret void
 }
 
 ; Loop with predicated load and store in separate blocks, store does not depend
 ; on loaded value.
 define void @pred_cfg3(i32 %k, i32 %j) {
 ; CHECK-LABEL: LV: Checking a loop in "pred_cfg3"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK1:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>
 ; CHECK-NEXT:   WIDEN ir<%mul> = mul ir<%iv>, ir<10>

 ; CHECK-NEXT: Successor(s): then.1.0
 ; CHECK-EMPTY:
 ; CHECK-NEXT: then.1.0:
 ; CHECK-NEXT: Successor(s): next.1
 ; CHECK-EMPTY:
 ; CHECK-NEXT: next.1:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 
 loop:
   %iv = phi i32 [ 0, %entry ], [ %iv.next, %next.1 ]

 
 exit:
   ret void
 }
 
 define void @merge_3_replicate_region(i32 %k, i32 %j) {
 ; CHECK-LABEL: LV: Checking a loop in "merge_3_replicate_region"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>
 ; CHECK-NEXT:   REPLICATE ir<%gep.a> = getelementptr ir<@a>, ir<0>, ir<%iv>

 ; CHECK-NEXT: Successor(s): then.0.0
 ; CHECK-EMPTY:
 ; CHECK-NEXT: then.0.0:
 ; CHECK-NEXT: Successor(s): latch
 ; CHECK-EMPTY:
 ; CHECK-NEXT: latch:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 
 loop:
   %iv = phi i32 [ 0, %entry ], [ %iv.next, %latch ]

 exit:
   ret void
 }
 
 
 define void @update_2_uses_in_same_recipe_in_merged_block(i32 %k) {
 ; CHECK-LABEL: LV: Checking a loop in "update_2_uses_in_same_recipe_in_merged_block"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>
 ; CHECK-NEXT:   REPLICATE ir<%gep.a> = getelementptr ir<@a>, ir<0>, ir<%iv>

 ; CHECK-NEXT:     PHI-PREDICATED-INSTRUCTION vp<[[PRED2:%.+]]> = ir<%div>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): loop.2
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.2:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 
 loop:
   %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ]

 
 exit:
   ret void
 }
 
 define void @recipe_in_merge_candidate_used_by_first_order_recurrence(i32 %k) {
 ; CHECK-LABEL: LV: Checking a loop in "recipe_in_merge_candidate_used_by_first_order_recurrence"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: Live-in vp<[[BTC:%.+]]> = backedge-taken count
 ; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   FIRST-ORDER-RECURRENCE-PHI ir<%for> = phi ir<0>, ir<%lv.a>
 ; CHECK-NEXT:   EMIT vp<[[MASK:%.+]]> = icmp ule ir<%iv> vp<[[BTC]]>

 ; CHECK-NEXT:     PHI-PREDICATED-INSTRUCTION vp<[[PRED2:%.+]]> = ir<%div>
 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): loop.2
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.2:
 ; CHECK-NEXT:   CLONE ir<%large> = icmp ir<%iv>, ir<8>
 ; CHECK-NEXT:   CLONE ir<%exitcond> = icmp ir<%iv>, ir<%k>
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF + vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop
 
 loop:
   %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ]

 
 exit:
   ret void
 }
 
 define void @update_multiple_users(i16* noalias %src, i8* noalias %dst, i1 %c) {
 ; CHECK-LABEL: LV: Checking a loop in "update_multiple_users"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop.header:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT: Successor(s): loop.then
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.then:
 ; CHECK-NEXT: Successor(s): loop.then.0

 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): loop.then.1
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.then.1:
 ; CHECK-NEXT:   WIDEN ir<%sext.l1> = sext vp<[[PRED]]>
 ; CHECK-NEXT: Successor(s): loop.latch
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.latch:
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop.header
 
 loop.header:
   %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]

 
 exit:
   ret void
 }
 
 define void @sinking_requires_duplication(float* %addr) {
 ; CHECK-LABEL: LV: Checking a loop in "sinking_requires_duplication"
 ; CHECK:      VPlan 'Initial VPlan for VF={2},UF>=1' {
+; CHECK-NEXT: Live-in vp<[[VEC_TC:%.+]]> = vector-trip-count
+; CHECK-EMPTY:
 ; CHECK-NEXT: <x1> vector loop: {
 ; CHECK-NEXT: loop.header:
 ; CHECK-NEXT:   EMIT vp<[[CAN_IV:%.+]]> = CANONICAL-INDUCTION
 ; CHECK-NEXT:   WIDEN-INDUCTION %iv = phi 0, %iv.next
 ; CHECK-NEXT:   CLONE ir<%gep> = getelementptr ir<%addr>, ir<%iv>
 ; CHECK-NEXT: Successor(s): loop.body
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.body:

 ; CHECK-NEXT:   No successors
 ; CHECK-NEXT: }
 ; CHECK-NEXT: Successor(s): then.0
 ; CHECK-EMPTY:
 ; CHECK-NEXT: then.0:
 ; CHECK-NEXT: Successor(s): loop.latch
 ; CHECK-EMPTY:
 ; CHECK-NEXT: loop.latch:
-; CHECK-NEXT:   EMIT vp<{{.+}}> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT vp<[[CAN_IV_NEXT:%.+]]> = VF * UF +(nuw) vp<[[CAN_IV]]>
+; CHECK-NEXT:   EMIT branch-on-count vp<[[CAN_IV_NEXT]]> vp<[[VEC_TC]]>
 ; CHECK-NEXT: No successors
 ; CHECK-NEXT: }
 ;
 entry:
   br label %loop.header
 
 loop.header:
   %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]

llvm/test/Transforms/LoopVectorize/vplan-vectorize-inner-loop-reduction.ll

 
 ; CHECK: [[FOR1_LATCH]]:
 ; CHECK-NEXT: %[[REDUCTION:.*]] = phi <4 x double> [ %[[REDUCTION_NEXT]], %[[FOR2_HEADER]] ]
 ; CHECK-NEXT: %[[C_PTR:.*]] = getelementptr inbounds double, double* %c.out, <4 x i64> %[[VEC_INDEX]]
 ; CHECK-NEXT: call void @llvm.masked.scatter.v4f64.v4p0f64(<4 x double> %[[REDUCTION]], <4 x double*> %[[C_PTR]], i32 8, <4 x i1> <i1 true, i1 true, i1 true, i1 true>)
 ; CHECK-NEXT: %[[VEC_INDEX_NEXT:.*]] = add nuw nsw <4 x i64> %[[VEC_INDEX]], <i64 1, i64 1, i64 1, i64 1>
 ; CHECK-NEXT: %[[VEC_PTR:.*]] = icmp eq <4 x i64> %[[VEC_INDEX_NEXT]], <i64 1000, i64 1000, i64 1000, i64 1000>
 ; CHECK-NEXT: %[[FOR1_INDEX_NEXT:.*]] = add nuw i64 %[[FOR1_INDEX]], 4
-; CHECK-NEXT: %{{.*}} = extractelement <4 x i1> %[[VEC_PTR]], i32 0
 ; CHECK-NEXT: %{{.*}} = add <4 x i64> %[[VEC_INDEX]], <i64 4, i64 4, i64 4, i64 4>
 ; CHECK-NEXT: %[[EXIT_COND:.*]] = icmp eq i64 %[[FOR1_INDEX_NEXT]], 1000
 ; CHECK-NEXT: br i1 %[[EXIT_COND]], label %{{.*}}, label %vector.body
 
 entry:
   br label %for1.header
 
 for1.header:                                              ; preds = %entry

llvm/test/Transforms/LoopVectorize/vplan-widen-call-instruction.ll

 
 ; CHECK: [[FOR1_LATCH]]:
 ; CHECK-NEXT: %[[REDUCTION:.*]] = phi <4 x double> [ %[[REDUCTION_NEXT]], %[[FOR2_HEADER]] ]
 ; CHECK-NEXT: %[[C_PTR:.*]] = getelementptr inbounds double, double* %c.out, <4 x i64> %[[VEC_INDEX]]
 ; CHECK-NEXT: call void @llvm.masked.scatter.v4f64.v4p0f64(<4 x double> %[[REDUCTION]], <4 x double*> %[[C_PTR]], i32 8, <4 x i1> <i1 true, i1 true, i1 true, i1 true>)
 ; CHECK-NEXT: %[[VEC_INDEX_NEXT:.*]] = add nuw nsw <4 x i64> %[[VEC_INDEX]], <i64 1, i64 1, i64 1, i64 1>
 ; CHECK-NEXT: %[[VEC_PTR:.*]] = icmp eq <4 x i64> %[[VEC_INDEX_NEXT]], <i64 1000, i64 1000, i64 1000, i64 1000>
 ; CHECK-NEXT: %[[FOR1_INDEX_NEXT:.*]] = add nuw i64 %[[FOR1_INDEX]], 4
-; CHECK-NEXT: %{{.*}} = extractelement <4 x i1> %[[VEC_PTR]], i32 0
 ; CHECK-NEXT: %{{.*}} = add <4 x i64> %[[VEC_INDEX]], <i64 4, i64 4, i64 4, i64 4>
 ; CHECK-NEXT: %[[EXIT_COND:.*]] = icmp eq i64 %[[FOR1_INDEX_NEXT]], 1000
 ; CHECK-NEXT: br i1 %[[EXIT_COND]], label %{{.*}}, label %vector.body
 
 entry:
   br label %for1.header
 
 for1.header: