[LV] Keep track of cost-based ScalarAfterVec in VPWidenPointerInd.
Epilogue vectorization uses isScalarAfterVectorization to check if
widened versions for inductions need to be generated and bails out in
those cases.
At the moment, there are scenarios where isScalarAfterVectorization
returns true but VPWidenPointerInduction::onlyScalarsGenerated would
return false, causing widening.
This can lead to widened phis with incorrect start values being created
in the epilogue vector body.
This patch addresses the issue by storing the cost-model decision in
VPWidenPointerInductionRecipe and restoring the behavior before 151c144.
This effectively reverts 151c144, but the long-term fix is to properly
support widened inductions during epilogue vectorization
Fixes #57712.
diff --git a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
index 91bc7db..2cdae02 100644
--- a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -8149,9 +8149,15 @@
*PSE.getSE(), *OrigLoop, Range);
// Check if this is pointer induction. If so, build the recipe for it.
- if (auto *II = Legal->getPointerInductionDescriptor(Phi))
- return new VPWidenPointerInductionRecipe(Phi, Operands[0], *II,
- *PSE.getSE());
+ if (auto *II = Legal->getPointerInductionDescriptor(Phi)) {
+ return new VPWidenPointerInductionRecipe(
+ Phi, Operands[0], *II, *PSE.getSE(),
+ LoopVectorizationPlanner::getDecisionAndClampRange(
+ [&](ElementCount VF) {
+ return !VF.isScalable() && CM.isScalarAfterVectorization(Phi, VF);
+ },
+ Range));
+ }
return nullptr;
}
@@ -9332,7 +9338,7 @@
auto *IVR = getParent()->getPlan()->getCanonicalIV();
PHINode *CanonicalIV = cast<PHINode>(State.get(IVR, 0));
- if (onlyScalarsGenerated(State.VF)) {
+ if (onlyScalarsGenerated()) {
// This is the normalized GEP that starts counting at zero.
Value *PtrInd = State.Builder.CreateSExtOrTrunc(
CanonicalIV, IndDesc.getStep()->getType());
diff --git a/llvm/lib/Transforms/Vectorize/VPlan.cpp b/llvm/lib/Transforms/Vectorize/VPlan.cpp
index 30032dd..0662ca8 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlan.cpp
@@ -698,7 +698,7 @@
auto *WidenPhi = cast<VPWidenPointerInductionRecipe>(&R);
// TODO: Split off the case that all users of a pointer phi are scalar
// from the VPWidenPointerInductionRecipe.
- if (WidenPhi->onlyScalarsGenerated(State->VF))
+ if (WidenPhi->onlyScalarsGenerated())
continue;
auto *GEP = cast<GetElementPtrInst>(State->get(WidenPhi, 0));
diff --git a/llvm/lib/Transforms/Vectorize/VPlan.h b/llvm/lib/Transforms/Vectorize/VPlan.h
index f009a7e..329843b 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.h
+++ b/llvm/lib/Transforms/Vectorize/VPlan.h
@@ -1187,15 +1187,19 @@
/// explicitly.
ScalarEvolution &SE;
+ bool IsScalarAfterVectorization;
+
public:
/// Create a new VPWidenPointerInductionRecipe for \p Phi with start value \p
/// Start.
VPWidenPointerInductionRecipe(PHINode *Phi, VPValue *Start,
const InductionDescriptor &IndDesc,
- ScalarEvolution &SE)
+ ScalarEvolution &SE,
+ bool IsScalarAfterVectorization)
: VPHeaderPHIRecipe(VPVWidenPointerInductionSC, VPWidenPointerInductionSC,
Phi),
- IndDesc(IndDesc), SE(SE) {
+ IndDesc(IndDesc), SE(SE),
+ IsScalarAfterVectorization(IsScalarAfterVectorization) {
addOperand(Start);
}
@@ -1216,7 +1220,7 @@
void execute(VPTransformState &State) override;
/// Returns true if only scalar values will be generated.
- bool onlyScalarsGenerated(ElementCount VF);
+ bool onlyScalarsGenerated();
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
/// Print the recipe.
diff --git a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
index cb75072..f214563 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp
@@ -982,11 +982,8 @@
}
#endif
-bool VPWidenPointerInductionRecipe::onlyScalarsGenerated(ElementCount VF) {
- bool IsUniform = vputils::onlyFirstLaneUsed(this);
- return all_of(users(),
- [&](const VPUser *U) { return U->usesScalars(this); }) &&
- (IsUniform || !VF.isScalable());
+bool VPWidenPointerInductionRecipe::onlyScalarsGenerated() {
+ return IsScalarAfterVectorization;
}
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/epilog-vectorization-widen-inductions.ll b/llvm/test/Transforms/LoopVectorize/AArch64/epilog-vectorization-widen-inductions.ll
index 2200929..b9294a1 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/epilog-vectorization-widen-inductions.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/epilog-vectorization-widen-inductions.ll
@@ -13,65 +13,74 @@
; CHECK: vector.ph:
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
-; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi ptr [ [[PTR_START_1:%.*]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT: [[TMP0:%.*]] = getelementptr i8, ptr [[POINTER_PHI]], <2 x i64> <i64 0, i64 1>
-; CHECK-NEXT: [[TMP1:%.*]] = getelementptr i8, ptr [[POINTER_PHI]], <2 x i64> <i64 2, i64 3>
-; CHECK-NEXT: [[TMP2:%.*]] = icmp ne <2 x ptr> [[TMP0]], zeroinitializer
-; CHECK-NEXT: [[TMP3:%.*]] = icmp ne <2 x ptr> [[TMP1]], zeroinitializer
-; CHECK-NEXT: [[TMP4:%.*]] = extractelement <2 x i1> [[TMP2]], i32 0
-; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP4]])
-; CHECK-NEXT: [[TMP5:%.*]] = extractelement <2 x i1> [[TMP2]], i32 1
-; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP5]])
-; CHECK-NEXT: [[TMP6:%.*]] = extractelement <2 x i1> [[TMP3]], i32 0
-; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP6]])
-; CHECK-NEXT: [[TMP7:%.*]] = extractelement <2 x i1> [[TMP3]], i32 1
-; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP7]])
-; CHECK-NEXT: [[TMP8:%.*]] = extractelement <2 x ptr> [[TMP0]], i32 0
-; CHECK-NEXT: [[TMP9:%.*]] = getelementptr i8, ptr [[TMP8]], i32 0
-; CHECK-NEXT: store <2 x i8> zeroinitializer, ptr [[TMP9]], align 1
-; CHECK-NEXT: [[TMP10:%.*]] = getelementptr i8, ptr [[TMP8]], i32 2
-; CHECK-NEXT: store <2 x i8> zeroinitializer, ptr [[TMP10]], align 1
+; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 0
+; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i8, ptr [[PTR_START_1:%.*]], i64 [[TMP0]]
+; CHECK-NEXT: [[TMP1:%.*]] = add i64 [[INDEX]], 1
+; CHECK-NEXT: [[NEXT_GEP1:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 [[TMP1]]
+; CHECK-NEXT: [[TMP2:%.*]] = insertelement <2 x ptr> poison, ptr [[NEXT_GEP]], i32 0
+; CHECK-NEXT: [[TMP3:%.*]] = insertelement <2 x ptr> [[TMP2]], ptr [[NEXT_GEP1]], i32 1
+; CHECK-NEXT: [[TMP4:%.*]] = add i64 [[INDEX]], 2
+; CHECK-NEXT: [[NEXT_GEP2:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 [[TMP4]]
+; CHECK-NEXT: [[TMP5:%.*]] = add i64 [[INDEX]], 3
+; CHECK-NEXT: [[NEXT_GEP3:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 [[TMP5]]
+; CHECK-NEXT: [[TMP6:%.*]] = insertelement <2 x ptr> poison, ptr [[NEXT_GEP2]], i32 0
+; CHECK-NEXT: [[TMP7:%.*]] = insertelement <2 x ptr> [[TMP6]], ptr [[NEXT_GEP3]], i32 1
+; CHECK-NEXT: [[TMP8:%.*]] = icmp ne <2 x ptr> [[TMP3]], zeroinitializer
+; CHECK-NEXT: [[TMP9:%.*]] = icmp ne <2 x ptr> [[TMP7]], zeroinitializer
+; CHECK-NEXT: [[TMP10:%.*]] = extractelement <2 x i1> [[TMP8]], i32 0
+; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP10]])
+; CHECK-NEXT: [[TMP11:%.*]] = extractelement <2 x i1> [[TMP8]], i32 1
+; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP11]])
+; CHECK-NEXT: [[TMP12:%.*]] = extractelement <2 x i1> [[TMP9]], i32 0
+; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP12]])
+; CHECK-NEXT: [[TMP13:%.*]] = extractelement <2 x i1> [[TMP9]], i32 1
+; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP13]])
+; CHECK-NEXT: [[TMP14:%.*]] = getelementptr i8, ptr [[NEXT_GEP]], i32 0
+; CHECK-NEXT: store <2 x i8> zeroinitializer, ptr [[TMP14]], align 1
+; CHECK-NEXT: [[TMP15:%.*]] = getelementptr i8, ptr [[NEXT_GEP]], i32 2
+; CHECK-NEXT: store <2 x i8> zeroinitializer, ptr [[TMP15]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
-; CHECK-NEXT: [[PTR_IND]] = getelementptr i8, ptr [[POINTER_PHI]], i64 4
-; CHECK-NEXT: [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], 10000
-; CHECK-NEXT: br i1 [[TMP11]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT: [[TMP16:%.*]] = icmp eq i64 [[INDEX_NEXT]], 10000
+; CHECK-NEXT: br i1 [[TMP16]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 10001, 10000
; CHECK-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
; CHECK: vec.epilog.iter.check:
-; CHECK-NEXT: [[IND_END1:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 10000
+; CHECK-NEXT: [[IND_END4:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 10000
; CHECK-NEXT: br i1 true, label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ 10000, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT: [[IND_END:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 10000
; CHECK-NEXT: br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
; CHECK: vec.epilog.vector.body:
-; CHECK-NEXT: [[POINTER_PHI5:%.*]] = phi ptr [ [[PTR_START_1]], [[VEC_EPILOG_PH]] ], [ [[PTR_IND6:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
-; CHECK-NEXT: [[INDEX4:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT7:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
-; CHECK-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[POINTER_PHI5]], <2 x i64> <i64 0, i64 1>
-; CHECK-NEXT: [[TMP13:%.*]] = icmp ne <2 x ptr> [[TMP12]], zeroinitializer
-; CHECK-NEXT: [[TMP14:%.*]] = extractelement <2 x i1> [[TMP13]], i32 0
-; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP14]])
-; CHECK-NEXT: [[TMP15:%.*]] = extractelement <2 x i1> [[TMP13]], i32 1
-; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP15]])
-; CHECK-NEXT: [[TMP16:%.*]] = extractelement <2 x ptr> [[TMP12]], i32 0
-; CHECK-NEXT: [[TMP17:%.*]] = getelementptr i8, ptr [[TMP16]], i32 0
-; CHECK-NEXT: store <2 x i8> zeroinitializer, ptr [[TMP17]], align 1
-; CHECK-NEXT: [[INDEX_NEXT7]] = add nuw i64 [[INDEX4]], 2
-; CHECK-NEXT: [[PTR_IND6]] = getelementptr i8, ptr [[POINTER_PHI5]], i64 2
-; CHECK-NEXT: [[TMP18:%.*]] = icmp eq i64 [[INDEX_NEXT7]], 10000
-; CHECK-NEXT: br i1 [[TMP18]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP2:![0-9]+]]
+; CHECK-NEXT: [[INDEX7:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT10:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
+; CHECK-NEXT: [[TMP17:%.*]] = add i64 [[INDEX7]], 0
+; CHECK-NEXT: [[NEXT_GEP8:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 [[TMP17]]
+; CHECK-NEXT: [[TMP18:%.*]] = add i64 [[INDEX7]], 1
+; CHECK-NEXT: [[NEXT_GEP9:%.*]] = getelementptr i8, ptr [[PTR_START_1]], i64 [[TMP18]]
+; CHECK-NEXT: [[TMP19:%.*]] = insertelement <2 x ptr> poison, ptr [[NEXT_GEP8]], i32 0
+; CHECK-NEXT: [[TMP20:%.*]] = insertelement <2 x ptr> [[TMP19]], ptr [[NEXT_GEP9]], i32 1
+; CHECK-NEXT: [[TMP21:%.*]] = icmp ne <2 x ptr> [[TMP20]], zeroinitializer
+; CHECK-NEXT: [[TMP22:%.*]] = extractelement <2 x i1> [[TMP21]], i32 0
+; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP22]])
+; CHECK-NEXT: [[TMP23:%.*]] = extractelement <2 x i1> [[TMP21]], i32 1
+; CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP23]])
+; CHECK-NEXT: [[TMP24:%.*]] = getelementptr i8, ptr [[NEXT_GEP8]], i32 0
+; CHECK-NEXT: store <2 x i8> zeroinitializer, ptr [[TMP24]], align 1
+; CHECK-NEXT: [[INDEX_NEXT10]] = add nuw i64 [[INDEX7]], 2
+; CHECK-NEXT: [[TMP25:%.*]] = icmp eq i64 [[INDEX_NEXT10]], 10000
+; CHECK-NEXT: br i1 [[TMP25]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP2:![0-9]+]]
; CHECK: vec.epilog.middle.block:
-; CHECK-NEXT: [[CMP_N3:%.*]] = icmp eq i64 10001, 10000
-; CHECK-NEXT: br i1 [[CMP_N3]], label [[EXIT]], label [[VEC_EPILOG_SCALAR_PH]]
+; CHECK-NEXT: [[CMP_N6:%.*]] = icmp eq i64 10001, 10000
+; CHECK-NEXT: br i1 [[CMP_N6]], label [[EXIT]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ 10000, [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ 10000, [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK:%.*]] ]
-; CHECK-NEXT: [[BC_RESUME_VAL2:%.*]] = phi ptr [ [[IND_END]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END1]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[PTR_START_1]], [[ITER_CHECK]] ]
+; CHECK-NEXT: [[BC_RESUME_VAL5:%.*]] = phi ptr [ [[IND_END]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END4]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[PTR_START_1]], [[ITER_CHECK]] ]
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
-; CHECK-NEXT: [[PTR_IV:%.*]] = phi ptr [ [[BC_RESUME_VAL2]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[PTR_IV_NEXT:%.*]], [[LOOP]] ]
+; CHECK-NEXT: [[PTR_IV:%.*]] = phi ptr [ [[BC_RESUME_VAL5]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[PTR_IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[CMP_I_I_I_I:%.*]] = icmp ne ptr [[PTR_IV]], null
; CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I_I_I]])
; CHECK-NEXT: store i8 0, ptr [[PTR_IV]], align 1
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/sve-live-out-pointer-induction.ll b/llvm/test/Transforms/LoopVectorize/AArch64/sve-live-out-pointer-induction.ll
index 718b2e8..fbf347f 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/sve-live-out-pointer-induction.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/sve-live-out-pointer-induction.ll
@@ -26,6 +26,7 @@
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi ptr [ [[START_1]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT: [[POINTER_PHI8:%.*]] = phi ptr [ [[START_2]], [[VECTOR_PH]] ], [ [[PTR_IND9:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP10:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT: [[TMP11:%.*]] = mul i64 [[TMP10]], 2
@@ -45,32 +46,43 @@
; CHECK-NEXT: [[TMP20:%.*]] = add <vscale x 2 x i64> [[DOTSPLAT6]], [[TMP19]]
; CHECK-NEXT: [[VECTOR_GEP7:%.*]] = mul <vscale x 2 x i64> [[TMP20]], shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 8, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
; CHECK-NEXT: [[TMP21:%.*]] = getelementptr i8, ptr [[POINTER_PHI]], <vscale x 2 x i64> [[VECTOR_GEP7]]
-; CHECK-NEXT: [[TMP22:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT: [[TMP23:%.*]] = mul i64 [[TMP22]], 8
-; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i8, ptr [[START_2]], i64 [[TMP23]]
-; CHECK-NEXT: [[TMP24:%.*]] = call i64 @llvm.vscale.i64()
-; CHECK-NEXT: [[TMP25:%.*]] = mul i64 [[TMP24]], 2
-; CHECK-NEXT: [[TMP26:%.*]] = add i64 [[TMP25]], 0
-; CHECK-NEXT: [[TMP27:%.*]] = add i64 [[INDEX]], [[TMP26]]
-; CHECK-NEXT: [[TMP28:%.*]] = mul i64 [[TMP27]], 8
-; CHECK-NEXT: [[NEXT_GEP8:%.*]] = getelementptr i8, ptr [[START_2]], i64 [[TMP28]]
-; CHECK-NEXT: [[TMP29:%.*]] = getelementptr i64, ptr [[NEXT_GEP]], i32 0
-; CHECK-NEXT: store <vscale x 2 x i64> zeroinitializer, ptr [[TMP29]], align 4
-; CHECK-NEXT: [[TMP30:%.*]] = call i32 @llvm.vscale.i32()
-; CHECK-NEXT: [[TMP31:%.*]] = mul i32 [[TMP30]], 2
-; CHECK-NEXT: [[TMP32:%.*]] = getelementptr i64, ptr [[NEXT_GEP]], i32 [[TMP31]]
-; CHECK-NEXT: store <vscale x 2 x i64> zeroinitializer, ptr [[TMP32]], align 4
-; CHECK-NEXT: [[TMP33:%.*]] = call i64 @llvm.vscale.i64()
-; CHECK-NEXT: [[TMP34:%.*]] = mul i64 [[TMP33]], 4
-; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP34]]
+; CHECK-NEXT: [[TMP22:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP23:%.*]] = mul i64 [[TMP22]], 2
+; CHECK-NEXT: [[TMP24:%.*]] = mul i64 [[TMP23]], 2
+; CHECK-NEXT: [[TMP25:%.*]] = mul i64 8, [[TMP24]]
+; CHECK-NEXT: [[TMP26:%.*]] = mul i64 [[TMP23]], 0
+; CHECK-NEXT: [[DOTSPLATINSERT10:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[TMP26]], i32 0
+; CHECK-NEXT: [[DOTSPLAT11:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT10]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
+; CHECK-NEXT: [[TMP27:%.*]] = call <vscale x 2 x i64> @llvm.experimental.stepvector.nxv2i64()
+; CHECK-NEXT: [[TMP28:%.*]] = add <vscale x 2 x i64> [[DOTSPLAT11]], [[TMP27]]
+; CHECK-NEXT: [[VECTOR_GEP12:%.*]] = mul <vscale x 2 x i64> [[TMP28]], shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 8, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP29:%.*]] = getelementptr i8, ptr [[POINTER_PHI8]], <vscale x 2 x i64> [[VECTOR_GEP12]]
+; CHECK-NEXT: [[TMP30:%.*]] = mul i64 [[TMP23]], 1
+; CHECK-NEXT: [[DOTSPLATINSERT13:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[TMP30]], i32 0
+; CHECK-NEXT: [[DOTSPLAT14:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT13]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
+; CHECK-NEXT: [[TMP31:%.*]] = call <vscale x 2 x i64> @llvm.experimental.stepvector.nxv2i64()
+; CHECK-NEXT: [[TMP32:%.*]] = add <vscale x 2 x i64> [[DOTSPLAT14]], [[TMP31]]
+; CHECK-NEXT: [[VECTOR_GEP15:%.*]] = mul <vscale x 2 x i64> [[TMP32]], shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 8, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP33:%.*]] = getelementptr i8, ptr [[POINTER_PHI8]], <vscale x 2 x i64> [[VECTOR_GEP15]]
+; CHECK-NEXT: [[TMP34:%.*]] = extractelement <vscale x 2 x ptr> [[TMP29]], i32 0
+; CHECK-NEXT: [[TMP35:%.*]] = getelementptr i64, ptr [[TMP34]], i32 0
+; CHECK-NEXT: store <vscale x 2 x i64> zeroinitializer, ptr [[TMP35]], align 4
+; CHECK-NEXT: [[TMP36:%.*]] = call i32 @llvm.vscale.i32()
+; CHECK-NEXT: [[TMP37:%.*]] = mul i32 [[TMP36]], 2
+; CHECK-NEXT: [[TMP38:%.*]] = getelementptr i64, ptr [[TMP34]], i32 [[TMP37]]
+; CHECK-NEXT: store <vscale x 2 x i64> zeroinitializer, ptr [[TMP38]], align 4
+; CHECK-NEXT: [[TMP39:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP40:%.*]] = mul i64 [[TMP39]], 4
+; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP40]]
; CHECK-NEXT: [[PTR_IND]] = getelementptr i8, ptr [[POINTER_PHI]], i64 [[TMP13]]
-; CHECK-NEXT: [[TMP35:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT: br i1 [[TMP35]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT: [[PTR_IND9]] = getelementptr i8, ptr [[POINTER_PHI8]], i64 [[TMP25]]
+; CHECK-NEXT: [[TMP41:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[TMP41]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP3]], [[N_VEC]]
; CHECK-NEXT: [[CAST_CMO:%.*]] = sub i64 [[N_VEC]], 1
-; CHECK-NEXT: [[TMP36:%.*]] = mul i64 [[CAST_CMO]], 8
-; CHECK-NEXT: [[IND_ESCAPE:%.*]] = getelementptr i8, ptr [[START_1]], i64 [[TMP36]]
+; CHECK-NEXT: [[TMP42:%.*]] = mul i64 [[CAST_CMO]], 8
+; CHECK-NEXT: [[IND_ESCAPE:%.*]] = getelementptr i8, ptr [[START_1]], i64 [[TMP42]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi ptr [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[START_1]], [[ENTRY:%.*]] ]
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-gep.ll b/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-gep.ll
index bb36224..ca27d83 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-gep.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-gep.ll
@@ -46,53 +46,65 @@
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP3]]
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-NEXT: [[IND_END:%.*]] = getelementptr i8*, i8** [[START_1:%.*]], i64 [[N_VEC]]
-; CHECK-NEXT: [[IND_END3:%.*]] = getelementptr i8, i8* [[START_2:%.*]], i64 [[N_VEC]]
+; CHECK-NEXT: [[IND_END2:%.*]] = getelementptr i8, i8* [[START_2:%.*]], i64 [[N_VEC]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
-; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi i8* [ [[START_2]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi i8** [ [[START_1]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT: [[POINTER_PHI4:%.*]] = phi i8* [ [[START_2]], [[VECTOR_PH]] ], [ [[PTR_IND5:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT: [[TMP4:%.*]] = add i64 [[INDEX]], 0
-; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i8*, i8** [[START_1]], i64 [[TMP4]]
-; CHECK-NEXT: [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
-; CHECK-NEXT: [[TMP7:%.*]] = mul i64 [[TMP6]], 2
-; CHECK-NEXT: [[TMP8:%.*]] = mul i64 [[TMP7]], 1
-; CHECK-NEXT: [[TMP9:%.*]] = mul i64 1, [[TMP8]]
-; CHECK-NEXT: [[TMP10:%.*]] = mul i64 [[TMP7]], 0
-; CHECK-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[TMP10]], i32 0
+; CHECK-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP5:%.*]] = mul i64 [[TMP4]], 2
+; CHECK-NEXT: [[TMP6:%.*]] = mul i64 [[TMP5]], 1
+; CHECK-NEXT: [[TMP7:%.*]] = mul i64 1, [[TMP6]]
+; CHECK-NEXT: [[TMP8:%.*]] = mul i64 [[TMP5]], 0
+; CHECK-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[TMP8]], i32 0
; CHECK-NEXT: [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-; CHECK-NEXT: [[TMP11:%.*]] = call <vscale x 2 x i64> @llvm.experimental.stepvector.nxv2i64()
-; CHECK-NEXT: [[TMP12:%.*]] = add <vscale x 2 x i64> [[DOTSPLAT]], [[TMP11]]
-; CHECK-NEXT: [[VECTOR_GEP:%.*]] = mul <vscale x 2 x i64> [[TMP12]], shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 1, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
-; CHECK-NEXT: [[TMP13:%.*]] = getelementptr i8, i8* [[POINTER_PHI]], <vscale x 2 x i64> [[VECTOR_GEP]]
-; CHECK-NEXT: [[TMP14:%.*]] = getelementptr inbounds i8, <vscale x 2 x i8*> [[TMP13]], i64 1
-; CHECK-NEXT: [[TMP15:%.*]] = getelementptr i8*, i8** [[NEXT_GEP]], i32 0
-; CHECK-NEXT: [[TMP16:%.*]] = bitcast i8** [[TMP15]] to <vscale x 2 x i8*>*
-; CHECK-NEXT: store <vscale x 2 x i8*> [[TMP14]], <vscale x 2 x i8*>* [[TMP16]], align 8
-; CHECK-NEXT: [[TMP17:%.*]] = extractelement <vscale x 2 x i8*> [[TMP13]], i32 0
-; CHECK-NEXT: [[TMP18:%.*]] = getelementptr i8, i8* [[TMP17]], i32 0
-; CHECK-NEXT: [[TMP19:%.*]] = bitcast i8* [[TMP18]] to <vscale x 2 x i8>*
-; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP19]], align 1
-; CHECK-NEXT: [[TMP20:%.*]] = add <vscale x 2 x i8> [[WIDE_LOAD]], shufflevector (<vscale x 2 x i8> insertelement (<vscale x 2 x i8> poison, i8 1, i32 0), <vscale x 2 x i8> poison, <vscale x 2 x i32> zeroinitializer)
-; CHECK-NEXT: [[TMP21:%.*]] = bitcast i8* [[TMP18]] to <vscale x 2 x i8>*
-; CHECK-NEXT: store <vscale x 2 x i8> [[TMP20]], <vscale x 2 x i8>* [[TMP21]], align 1
-; CHECK-NEXT: [[TMP22:%.*]] = call i64 @llvm.vscale.i64()
-; CHECK-NEXT: [[TMP23:%.*]] = mul i64 [[TMP22]], 2
-; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP23]]
-; CHECK-NEXT: [[PTR_IND]] = getelementptr i8, i8* [[POINTER_PHI]], i64 [[TMP9]]
-; CHECK-NEXT: [[TMP24:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT: br i1 [[TMP24]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT: [[TMP9:%.*]] = call <vscale x 2 x i64> @llvm.experimental.stepvector.nxv2i64()
+; CHECK-NEXT: [[TMP10:%.*]] = add <vscale x 2 x i64> [[DOTSPLAT]], [[TMP9]]
+; CHECK-NEXT: [[VECTOR_GEP:%.*]] = mul <vscale x 2 x i64> [[TMP10]], shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 1, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP11:%.*]] = getelementptr i8*, i8** [[POINTER_PHI]], <vscale x 2 x i64> [[VECTOR_GEP]]
+; CHECK-NEXT: [[TMP12:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP13:%.*]] = mul i64 [[TMP12]], 2
+; CHECK-NEXT: [[TMP14:%.*]] = mul i64 [[TMP13]], 1
+; CHECK-NEXT: [[TMP15:%.*]] = mul i64 1, [[TMP14]]
+; CHECK-NEXT: [[TMP16:%.*]] = mul i64 [[TMP13]], 0
+; CHECK-NEXT: [[DOTSPLATINSERT6:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[TMP16]], i32 0
+; CHECK-NEXT: [[DOTSPLAT7:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT6]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
+; CHECK-NEXT: [[TMP17:%.*]] = call <vscale x 2 x i64> @llvm.experimental.stepvector.nxv2i64()
+; CHECK-NEXT: [[TMP18:%.*]] = add <vscale x 2 x i64> [[DOTSPLAT7]], [[TMP17]]
+; CHECK-NEXT: [[VECTOR_GEP8:%.*]] = mul <vscale x 2 x i64> [[TMP18]], shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 1, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP19:%.*]] = getelementptr i8, i8* [[POINTER_PHI4]], <vscale x 2 x i64> [[VECTOR_GEP8]]
+; CHECK-NEXT: [[TMP20:%.*]] = getelementptr inbounds i8, <vscale x 2 x i8*> [[TMP19]], i64 1
+; CHECK-NEXT: [[TMP21:%.*]] = extractelement <vscale x 2 x i8**> [[TMP11]], i32 0
+; CHECK-NEXT: [[TMP22:%.*]] = getelementptr i8*, i8** [[TMP21]], i32 0
+; CHECK-NEXT: [[TMP23:%.*]] = bitcast i8** [[TMP22]] to <vscale x 2 x i8*>*
+; CHECK-NEXT: store <vscale x 2 x i8*> [[TMP20]], <vscale x 2 x i8*>* [[TMP23]], align 8
+; CHECK-NEXT: [[TMP24:%.*]] = extractelement <vscale x 2 x i8*> [[TMP19]], i32 0
+; CHECK-NEXT: [[TMP25:%.*]] = getelementptr i8, i8* [[TMP24]], i32 0
+; CHECK-NEXT: [[TMP26:%.*]] = bitcast i8* [[TMP25]] to <vscale x 2 x i8>*
+; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP26]], align 1
+; CHECK-NEXT: [[TMP27:%.*]] = add <vscale x 2 x i8> [[WIDE_LOAD]], shufflevector (<vscale x 2 x i8> insertelement (<vscale x 2 x i8> poison, i8 1, i32 0), <vscale x 2 x i8> poison, <vscale x 2 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP28:%.*]] = bitcast i8* [[TMP25]] to <vscale x 2 x i8>*
+; CHECK-NEXT: store <vscale x 2 x i8> [[TMP27]], <vscale x 2 x i8>* [[TMP28]], align 1
+; CHECK-NEXT: [[TMP29:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP30:%.*]] = mul i64 [[TMP29]], 2
+; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP30]]
+; CHECK-NEXT: [[PTR_IND]] = getelementptr i8*, i8** [[POINTER_PHI]], i64 [[TMP7]]
+; CHECK-NEXT: [[PTR_IND5]] = getelementptr i8, i8* [[POINTER_PHI4]], i64 [[TMP15]]
+; CHECK-NEXT: [[TMP31:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[TMP31]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; 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:%.*]] ]
; CHECK-NEXT: [[BC_RESUME_VAL1:%.*]] = phi i8** [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[START_1]], [[ENTRY]] ]
-; CHECK-NEXT: [[BC_RESUME_VAL2:%.*]] = phi i8* [ [[IND_END3]], [[MIDDLE_BLOCK]] ], [ [[START_2]], [[ENTRY]] ]
+; CHECK-NEXT: [[BC_RESUME_VAL3:%.*]] = phi i8* [ [[IND_END2]], [[MIDDLE_BLOCK]] ], [ [[START_2]], [[ENTRY]] ]
; CHECK-NEXT: br label [[LOOP_BODY:%.*]]
; CHECK: loop.body:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LOOP_BODY]] ]
; CHECK-NEXT: [[PTR_IV_1:%.*]] = phi i8** [ [[BC_RESUME_VAL1]], [[SCALAR_PH]] ], [ [[PTR_IV_1_NEXT:%.*]], [[LOOP_BODY]] ]
-; CHECK-NEXT: [[PTR_IV_2:%.*]] = phi i8* [ [[BC_RESUME_VAL2]], [[SCALAR_PH]] ], [ [[PTR_IV_2_NEXT:%.*]], [[LOOP_BODY]] ]
+; CHECK-NEXT: [[PTR_IV_2:%.*]] = phi i8* [ [[BC_RESUME_VAL3]], [[SCALAR_PH]] ], [ [[PTR_IV_2_NEXT:%.*]], [[LOOP_BODY]] ]
; CHECK-NEXT: [[PTR_IV_1_NEXT]] = getelementptr inbounds i8*, i8** [[PTR_IV_1]], i64 1
; CHECK-NEXT: [[PTR_IV_2_NEXT]] = getelementptr inbounds i8, i8* [[PTR_IV_2]], i64 1
; CHECK-NEXT: store i8* [[PTR_IV_2_NEXT]], i8** [[PTR_IV_1]], align 8
@@ -144,20 +156,32 @@
; CHECK-NEXT: [[IND_END:%.*]] = getelementptr i8, i8* [[START:%.*]], i64 [[N_VEC]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
+; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi i8* [ [[START]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[INDEX2:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT: [[INDEX2_0:%.*]] = add i64 [[INDEX2]], 0
-; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i8, i8* [[START]], i64 [[INDEX2_0]]
-; CHECK-NEXT: [[NEXT_GEP_0:%.*]] = getelementptr i8, i8* [[NEXT_GEP]], i32 0
-; CHECK-NEXT: [[TMP16:%.*]] = bitcast i8* [[NEXT_GEP_0]] to <vscale x 2 x i8>*
-; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP16]], align 1
-; CHECK-NEXT: [[TMP17:%.*]] = add <vscale x 2 x i8> [[WIDE_LOAD]], shufflevector (<vscale x 2 x i8> insertelement (<vscale x 2 x i8> poison, i8 1, i32 0), <vscale x 2 x i8> poison, <vscale x 2 x i32> zeroinitializer)
-; CHECK-NEXT: [[TMP18:%.*]] = bitcast i8* [[NEXT_GEP_0]] to <vscale x 2 x i8>*
-; CHECK-NEXT: store <vscale x 2 x i8> [[TMP17]], <vscale x 2 x i8>* [[TMP18]], align 1
-; CHECK-NEXT: [[TMP21:%.*]] = call i64 @llvm.vscale.i64()
-; CHECK-NEXT: [[TMP22:%.*]] = mul i64 [[TMP21]], 2
-; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX2]], [[TMP22]]
-; CHECK-NEXT: [[TMP23:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT: br i1 [[TMP23]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
+; CHECK-NEXT: [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP6:%.*]] = mul i64 [[TMP5]], 2
+; CHECK-NEXT: [[TMP7:%.*]] = mul i64 [[TMP6]], 1
+; CHECK-NEXT: [[TMP8:%.*]] = mul i64 1, [[TMP7]]
+; CHECK-NEXT: [[TMP9:%.*]] = mul i64 [[TMP6]], 0
+; CHECK-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[TMP9]], i32 0
+; CHECK-NEXT: [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
+; CHECK-NEXT: [[TMP10:%.*]] = call <vscale x 2 x i64> @llvm.experimental.stepvector.nxv2i64()
+; CHECK-NEXT: [[TMP11:%.*]] = add <vscale x 2 x i64> [[DOTSPLAT]], [[TMP10]]
+; CHECK-NEXT: [[VECTOR_GEP:%.*]] = mul <vscale x 2 x i64> [[TMP11]], shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 1, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP12:%.*]] = getelementptr i8, i8* [[POINTER_PHI]], <vscale x 2 x i64> [[VECTOR_GEP]]
+; CHECK-NEXT: [[TMP13:%.*]] = extractelement <vscale x 2 x i8*> [[TMP12]], i32 0
+; CHECK-NEXT: [[TMP14:%.*]] = getelementptr i8, i8* [[TMP13]], i32 0
+; CHECK-NEXT: [[TMP15:%.*]] = bitcast i8* [[TMP14]] to <vscale x 2 x i8>*
+; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP15]], align 1
+; CHECK-NEXT: [[TMP16:%.*]] = add <vscale x 2 x i8> [[WIDE_LOAD]], shufflevector (<vscale x 2 x i8> insertelement (<vscale x 2 x i8> poison, i8 1, i32 0), <vscale x 2 x i8> poison, <vscale x 2 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP17:%.*]] = bitcast i8* [[TMP14]] to <vscale x 2 x i8>*
+; CHECK-NEXT: store <vscale x 2 x i8> [[TMP16]], <vscale x 2 x i8>* [[TMP17]], align 1
+; CHECK-NEXT: [[TMP18:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP19:%.*]] = mul i64 [[TMP18]], 2
+; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX2]], [[TMP19]]
+; CHECK-NEXT: [[PTR_IND]] = getelementptr i8, i8* [[POINTER_PHI]], i64 [[TMP8]]
+; CHECK-NEXT: [[TMP20:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[TMP20]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[END:%.*]], label [[SCALAR_PH]]
@@ -169,8 +193,8 @@
; CHECK-NEXT: [[PTR_PHI:%.*]] = phi i8* [ [[PTR_PHI_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ [[INDEX_NXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL1]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[INDEX_NXT]] = add i64 [[INDEX]], 1
-; CHECK-NEXT: [[TMP24:%.*]] = load i8, i8* [[PTR_PHI]], align 1
-; CHECK-NEXT: [[ADD:%.*]] = add i8 [[TMP24]], 1
+; CHECK-NEXT: [[TMP21:%.*]] = load i8, i8* [[PTR_PHI]], align 1
+; CHECK-NEXT: [[ADD:%.*]] = add i8 [[TMP21]], 1
; CHECK-NEXT: store i8 [[ADD]], i8* [[PTR_PHI]], align 1
; CHECK-NEXT: [[PTR_PHI_NEXT]] = getelementptr inbounds i8, i8* [[PTR_PHI]], i64 1
; CHECK-NEXT: [[CMP_I_NOT:%.*]] = icmp eq i8* [[PTR_PHI_NEXT]], [[START]]
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-phi.ll b/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-phi.ll
index 658bcce..2e6f1a0 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-phi.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-phi.ll
@@ -153,49 +153,55 @@
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP3]]
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-NEXT: [[IND_END:%.*]] = getelementptr i32, i32* [[SRC:%.*]], i64 [[N_VEC]]
-; CHECK-NEXT: [[IND_END3:%.*]] = getelementptr i32, i32* [[DST:%.*]], i64 [[N_VEC]]
+; CHECK-NEXT: [[IND_END2:%.*]] = getelementptr i32, i32* [[DST:%.*]], i64 [[N_VEC]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
+; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi i32* [ [[SRC]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT: [[POINTER_PHI7:%.*]] = phi i32* [ [[DST]], [[VECTOR_PH]] ], [ [[PTR_IND8:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i32, i32* [[SRC]], i64 [[INDEX]]
-; CHECK-NEXT: [[NEXT_GEP5:%.*]] = getelementptr i32, i32* [[DST]], i64 [[INDEX]]
-; CHECK-NEXT: [[TMP4:%.*]] = bitcast i32* [[NEXT_GEP]] to <vscale x 4 x i32>*
-; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP4]], align 4
-; CHECK-NEXT: [[TMP5:%.*]] = call i32 @llvm.vscale.i32()
-; CHECK-NEXT: [[TMP6:%.*]] = shl nuw nsw i32 [[TMP5]], 2
-; CHECK-NEXT: [[TMP7:%.*]] = zext i32 [[TMP6]] to i64
-; CHECK-NEXT: [[TMP8:%.*]] = getelementptr i32, i32* [[NEXT_GEP]], i64 [[TMP7]]
-; CHECK-NEXT: [[TMP9:%.*]] = bitcast i32* [[TMP8]] to <vscale x 4 x i32>*
-; CHECK-NEXT: [[WIDE_LOAD7:%.*]] = load <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP9]], align 4
-; CHECK-NEXT: [[TMP10:%.*]] = shl nsw <vscale x 4 x i32> [[WIDE_LOAD]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 1, i32 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
-; CHECK-NEXT: [[TMP11:%.*]] = shl nsw <vscale x 4 x i32> [[WIDE_LOAD7]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 1, i32 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
-; CHECK-NEXT: [[TMP12:%.*]] = bitcast i32* [[NEXT_GEP5]] to <vscale x 4 x i32>*
-; CHECK-NEXT: store <vscale x 4 x i32> [[TMP10]], <vscale x 4 x i32>* [[TMP12]], align 4
-; CHECK-NEXT: [[TMP13:%.*]] = call i32 @llvm.vscale.i32()
-; CHECK-NEXT: [[TMP14:%.*]] = shl nuw nsw i32 [[TMP13]], 2
-; CHECK-NEXT: [[TMP15:%.*]] = zext i32 [[TMP14]] to i64
-; CHECK-NEXT: [[TMP16:%.*]] = getelementptr i32, i32* [[NEXT_GEP5]], i64 [[TMP15]]
-; CHECK-NEXT: [[TMP17:%.*]] = bitcast i32* [[TMP16]] to <vscale x 4 x i32>*
-; CHECK-NEXT: store <vscale x 4 x i32> [[TMP11]], <vscale x 4 x i32>* [[TMP17]], align 4
-; CHECK-NEXT: [[TMP18:%.*]] = call i64 @llvm.vscale.i64()
-; CHECK-NEXT: [[TMP19:%.*]] = shl nuw nsw i64 [[TMP18]], 3
-; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP19]]
-; CHECK-NEXT: [[TMP20:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT: br i1 [[TMP20]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
+; CHECK-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP5:%.*]] = shl nuw nsw i64 [[TMP4]], 3
+; CHECK-NEXT: [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP7:%.*]] = shl nuw nsw i64 [[TMP6]], 3
+; CHECK-NEXT: [[TMP8:%.*]] = bitcast i32* [[POINTER_PHI]] to <vscale x 4 x i32>*
+; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP8]], align 4
+; CHECK-NEXT: [[TMP9:%.*]] = call i32 @llvm.vscale.i32()
+; CHECK-NEXT: [[TMP10:%.*]] = shl nuw nsw i32 [[TMP9]], 2
+; CHECK-NEXT: [[TMP11:%.*]] = zext i32 [[TMP10]] to i64
+; CHECK-NEXT: [[TMP12:%.*]] = getelementptr i32, i32* [[POINTER_PHI]], i64 [[TMP11]]
+; CHECK-NEXT: [[TMP13:%.*]] = bitcast i32* [[TMP12]] to <vscale x 4 x i32>*
+; CHECK-NEXT: [[WIDE_LOAD15:%.*]] = load <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP13]], align 4
+; CHECK-NEXT: [[TMP14:%.*]] = shl nsw <vscale x 4 x i32> [[WIDE_LOAD]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 1, i32 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP15:%.*]] = shl nsw <vscale x 4 x i32> [[WIDE_LOAD15]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 1, i32 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
+; CHECK-NEXT: [[TMP16:%.*]] = bitcast i32* [[POINTER_PHI7]] to <vscale x 4 x i32>*
+; CHECK-NEXT: store <vscale x 4 x i32> [[TMP14]], <vscale x 4 x i32>* [[TMP16]], align 4
+; CHECK-NEXT: [[TMP17:%.*]] = call i32 @llvm.vscale.i32()
+; CHECK-NEXT: [[TMP18:%.*]] = shl nuw nsw i32 [[TMP17]], 2
+; CHECK-NEXT: [[TMP19:%.*]] = zext i32 [[TMP18]] to i64
+; CHECK-NEXT: [[TMP20:%.*]] = getelementptr i32, i32* [[POINTER_PHI7]], i64 [[TMP19]]
+; CHECK-NEXT: [[TMP21:%.*]] = bitcast i32* [[TMP20]] to <vscale x 4 x i32>*
+; CHECK-NEXT: store <vscale x 4 x i32> [[TMP15]], <vscale x 4 x i32>* [[TMP21]], align 4
+; CHECK-NEXT: [[TMP22:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP23:%.*]] = shl nuw nsw i64 [[TMP22]], 3
+; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP23]]
+; CHECK-NEXT: [[PTR_IND]] = getelementptr i32, i32* [[POINTER_PHI]], i64 [[TMP5]]
+; CHECK-NEXT: [[PTR_IND8]] = getelementptr i32, i32* [[POINTER_PHI7]], i64 [[TMP7]]
+; CHECK-NEXT: [[TMP24:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[TMP24]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N_MOD_VF]], 0
; CHECK-NEXT: br i1 [[CMP_N]], 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_RESUME_VAL1:%.*]] = phi i32* [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[SRC]], [[ENTRY]] ]
-; CHECK-NEXT: [[BC_RESUME_VAL2:%.*]] = phi i32* [ [[IND_END3]], [[MIDDLE_BLOCK]] ], [ [[DST]], [[ENTRY]] ]
+; CHECK-NEXT: [[BC_RESUME_VAL3:%.*]] = phi i32* [ [[IND_END2]], [[MIDDLE_BLOCK]] ], [ [[DST]], [[ENTRY]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[I_011:%.*]] = phi i64 [ [[INC:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[S_010:%.*]] = phi i32* [ [[INCDEC_PTR1:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL1]], [[SCALAR_PH]] ]
-; CHECK-NEXT: [[D_09:%.*]] = phi i32* [ [[INCDEC_PTR:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL2]], [[SCALAR_PH]] ]
-; CHECK-NEXT: [[TMP21:%.*]] = load i32, i32* [[S_010]], align 4
-; CHECK-NEXT: [[MUL:%.*]] = shl nsw i32 [[TMP21]], 1
+; CHECK-NEXT: [[D_09:%.*]] = phi i32* [ [[INCDEC_PTR:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL3]], [[SCALAR_PH]] ]
+; CHECK-NEXT: [[TMP25:%.*]] = load i32, i32* [[S_010]], align 4
+; CHECK-NEXT: [[MUL:%.*]] = shl nsw i32 [[TMP25]], 1
; CHECK-NEXT: store i32 [[MUL]], i32* [[D_09]], align 4
; CHECK-NEXT: [[INCDEC_PTR]] = getelementptr inbounds i32, i32* [[D_09]], i64 1
; CHECK-NEXT: [[INCDEC_PTR1]] = getelementptr inbounds i32, i32* [[S_010]], i64 1
@@ -247,43 +253,46 @@
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[SMAX]], [[TMP3]]
; CHECK-NEXT: [[N_VEC:%.*]] = sub nsw i64 [[SMAX]], [[N_MOD_VF]]
; CHECK-NEXT: [[IND_END:%.*]] = getelementptr i32, i32* [[A:%.*]], i64 [[N_VEC]]
-; CHECK-NEXT: [[IND_END3:%.*]] = getelementptr i32*, i32** [[B:%.*]], i64 [[N_VEC]]
+; CHECK-NEXT: [[IND_END2:%.*]] = getelementptr i32*, i32** [[B:%.*]], i64 [[N_VEC]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi i32* [ [[A]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT: [[POINTER_PHI4:%.*]] = phi i32** [ [[B]], [[VECTOR_PH]] ], [ [[PTR_IND5:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT: [[VEC_PHI:%.*]] = phi <vscale x 2 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP9:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT: [[VEC_PHI:%.*]] = phi <vscale x 2 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP11:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT: [[TMP5:%.*]] = shl nuw nsw i64 [[TMP4]], 1
; CHECK-NEXT: [[TMP6:%.*]] = call <vscale x 2 x i64> @llvm.experimental.stepvector.nxv2i64()
; CHECK-NEXT: [[TMP7:%.*]] = getelementptr i32, i32* [[POINTER_PHI]], <vscale x 2 x i64> [[TMP6]]
-; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i32*, i32** [[B]], i64 [[INDEX]]
+; CHECK-NEXT: [[TMP8:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP9:%.*]] = shl nuw nsw i64 [[TMP8]], 1
; CHECK-NEXT: [[BC:%.*]] = bitcast <vscale x 2 x i32*> [[TMP7]] to <vscale x 2 x <vscale x 2 x i32>*>
-; CHECK-NEXT: [[TMP8:%.*]] = extractelement <vscale x 2 x <vscale x 2 x i32>*> [[BC]], i64 0
-; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP8]], align 8
-; CHECK-NEXT: [[TMP9]] = add <vscale x 2 x i32> [[WIDE_LOAD]], [[VEC_PHI]]
-; CHECK-NEXT: [[TMP10:%.*]] = bitcast i32** [[NEXT_GEP]] to <vscale x 2 x i32*>*
-; CHECK-NEXT: store <vscale x 2 x i32*> [[TMP7]], <vscale x 2 x i32*>* [[TMP10]], align 8
-; CHECK-NEXT: [[TMP11:%.*]] = call i64 @llvm.vscale.i64()
-; CHECK-NEXT: [[TMP12:%.*]] = shl nuw nsw i64 [[TMP11]], 1
-; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP12]]
+; CHECK-NEXT: [[TMP10:%.*]] = extractelement <vscale x 2 x <vscale x 2 x i32>*> [[BC]], i64 0
+; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP10]], align 8
+; CHECK-NEXT: [[TMP11]] = add <vscale x 2 x i32> [[WIDE_LOAD]], [[VEC_PHI]]
+; CHECK-NEXT: [[TMP12:%.*]] = bitcast i32** [[POINTER_PHI4]] to <vscale x 2 x i32*>*
+; CHECK-NEXT: store <vscale x 2 x i32*> [[TMP7]], <vscale x 2 x i32*>* [[TMP12]], align 8
+; CHECK-NEXT: [[TMP13:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP14:%.*]] = shl nuw nsw i64 [[TMP13]], 1
+; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP14]]
; CHECK-NEXT: [[PTR_IND]] = getelementptr i32, i32* [[POINTER_PHI]], i64 [[TMP5]]
-; CHECK-NEXT: [[TMP13:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT: br i1 [[TMP13]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
+; CHECK-NEXT: [[PTR_IND5]] = getelementptr i32*, i32** [[POINTER_PHI4]], i64 [[TMP9]]
+; CHECK-NEXT: [[TMP15:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[TMP15]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
; CHECK: middle.block:
-; CHECK-NEXT: [[TMP14:%.*]] = call i32 @llvm.vector.reduce.add.nxv2i32(<vscale x 2 x i32> [[TMP9]])
+; CHECK-NEXT: [[TMP16:%.*]] = call i32 @llvm.vector.reduce.add.nxv2i32(<vscale x 2 x i32> [[TMP11]])
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N_MOD_VF]], 0
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[BC_RESUME_VAL1:%.*]] = phi i32* [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[A]], [[ENTRY]] ]
-; CHECK-NEXT: [[BC_RESUME_VAL2:%.*]] = phi i32** [ [[IND_END3]], [[MIDDLE_BLOCK]] ], [ [[B]], [[ENTRY]] ]
-; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ [[TMP14]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY]] ]
+; CHECK-NEXT: [[BC_RESUME_VAL3:%.*]] = phi i32** [ [[IND_END2]], [[MIDDLE_BLOCK]] ], [ [[B]], [[ENTRY]] ]
+; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ [[TMP16]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[I:%.*]] = phi i64 [ [[I_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[P:%.*]] = phi i32* [ [[VAR3:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL1]], [[SCALAR_PH]] ]
-; CHECK-NEXT: [[Q:%.*]] = phi i32** [ [[VAR4:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL2]], [[SCALAR_PH]] ]
+; CHECK-NEXT: [[Q:%.*]] = phi i32** [ [[VAR4:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL3]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[VAR0:%.*]] = phi i32 [ [[VAR2:%.*]], [[FOR_BODY]] ], [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[VAR1:%.*]] = load i32, i32* [[P]], align 8
; CHECK-NEXT: [[VAR2]] = add i32 [[VAR1]], [[VAR0]]
@@ -294,7 +303,7 @@
; CHECK-NEXT: [[COND:%.*]] = icmp slt i64 [[I_NEXT]], [[N]]
; CHECK-NEXT: br i1 [[COND]], label [[FOR_BODY]], label [[FOR_END]], !llvm.loop [[LOOP8:![0-9]+]]
; CHECK: for.end:
-; CHECK-NEXT: [[VAR5:%.*]] = phi i32 [ [[VAR2]], [[FOR_BODY]] ], [ [[TMP14]], [[MIDDLE_BLOCK]] ]
+; CHECK-NEXT: [[VAR5:%.*]] = phi i32 [ [[VAR2]], [[FOR_BODY]] ], [ [[TMP16]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: ret i32 [[VAR5]]
;
entry:
diff --git a/llvm/test/Transforms/LoopVectorize/X86/interleave-opaque-pointers.ll b/llvm/test/Transforms/LoopVectorize/X86/interleave-opaque-pointers.ll
index 5018be1..022912f 100644
--- a/llvm/test/Transforms/LoopVectorize/X86/interleave-opaque-pointers.ll
+++ b/llvm/test/Transforms/LoopVectorize/X86/interleave-opaque-pointers.ll
@@ -1,3 +1,4 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -passes=loop-vectorize -force-vector-width=2 -force-vector-interleave=1 -S %s | FileCheck %s
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
@@ -7,19 +8,54 @@
define void @test_pr55375_interleave_opaque_ptr(ptr %start, ptr %end) {
; CHECK-LABEL: @test_pr55375_interleave_opaque_ptr(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[START2:%.*]] = ptrtoint ptr [[START:%.*]] to i64
+; CHECK-NEXT: [[END1:%.*]] = ptrtoint ptr [[END:%.*]] to i64
+; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[END1]], -16
+; CHECK-NEXT: [[TMP1:%.*]] = sub i64 [[TMP0]], [[START2]]
+; CHECK-NEXT: [[TMP2:%.*]] = lshr i64 [[TMP1]], 4
+; CHECK-NEXT: [[TMP3:%.*]] = add nuw nsw i64 [[TMP2]], 1
+; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP3]], 2
+; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
+; CHECK: vector.ph:
+; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP3]], 2
+; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP3]], [[N_MOD_VF]]
+; CHECK-NEXT: [[TMP4:%.*]] = mul i64 [[N_VEC]], 16
+; CHECK-NEXT: [[IND_END:%.*]] = getelementptr i8, ptr [[START]], i64 [[TMP4]]
+; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
-; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi ptr [ %start, %vector.ph ], [ [[PTR_IND:%.*]], %vector.body ]
-; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %vector.ph ], [ [[INDEX_NEXT:%.*]], %vector.body ]
-; CHECK-NEXT: [[TMP5:%.*]] = getelementptr i8, ptr [[POINTER_PHI]], <2 x i64> <i64 0, i64 16>
-; CHECK-NEXT: [[TMP7:%.*]] = extractelement <2 x ptr> [[TMP5]], i32 0
-; CHECK-NEXT: [[TMP8:%.*]] = getelementptr ptr, ptr [[TMP7]], i32 0
-; CHECK-NEXT: [[TMP11:%.*]] = shufflevector <2 x ptr> zeroinitializer, <2 x ptr> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT: [[INTERLEAVED_VEC:%.*]] = shufflevector <4 x ptr> [[TMP11]], <4 x ptr> poison, <4 x i32> <i32 0, i32 2, i32 1, i32 3>
-; CHECK-NEXT: store <4 x ptr> [[INTERLEAVED_VEC]], ptr [[TMP8]], align 8
+; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT: [[TMP5:%.*]] = add i64 [[INDEX]], 0
+; CHECK-NEXT: [[TMP6:%.*]] = mul i64 [[TMP5]], 16
+; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i8, ptr [[START]], i64 [[TMP6]]
+; CHECK-NEXT: [[TMP7:%.*]] = add i64 [[INDEX]], 1
+; CHECK-NEXT: [[TMP8:%.*]] = mul i64 [[TMP7]], 16
+; CHECK-NEXT: [[NEXT_GEP3:%.*]] = getelementptr i8, ptr [[START]], i64 [[TMP8]]
+; CHECK-NEXT: [[TMP9:%.*]] = insertelement <2 x ptr> poison, ptr [[NEXT_GEP]], i32 0
+; CHECK-NEXT: [[TMP10:%.*]] = insertelement <2 x ptr> [[TMP9]], ptr [[NEXT_GEP3]], i32 1
+; CHECK-NEXT: [[TMP11:%.*]] = getelementptr ptr, ptr [[NEXT_GEP]], i32 0
+; CHECK-NEXT: [[TMP12:%.*]] = shufflevector <2 x ptr> zeroinitializer, <2 x ptr> [[TMP10]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT: [[INTERLEAVED_VEC:%.*]] = shufflevector <4 x ptr> [[TMP12]], <4 x ptr> poison, <4 x i32> <i32 0, i32 2, i32 1, i32 3>
+; CHECK-NEXT: store <4 x ptr> [[INTERLEAVED_VEC]], ptr [[TMP11]], align 8
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 2
-; CHECK-NEXT: [[PTR_IND]] = getelementptr i8, ptr [[POINTER_PHI]], i64 32
-; CHECK-NEXT: [[TMP13:%.*]] = icmp eq i64 [[INDEX_NEXT]], %n.vec
-; CHECK-NEXT: br i1 [[TMP13]], label %middle.block, label %vector.body
+; CHECK-NEXT: [[TMP13:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[TMP13]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK: middle.block:
+; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP3]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
+; CHECK: scalar.ph:
+; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi ptr [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[START]], [[ENTRY:%.*]] ]
+; CHECK-NEXT: br label [[LOOP:%.*]]
+; CHECK: loop:
+; CHECK-NEXT: [[IV:%.*]] = phi ptr [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
+; CHECK-NEXT: [[IV_1:%.*]] = getelementptr inbounds [[PAIR:%.*]], ptr [[IV]], i64 0, i32 1
+; CHECK-NEXT: store ptr [[IV]], ptr [[IV_1]], align 8
+; CHECK-NEXT: store ptr null, ptr [[IV]], align 8
+; CHECK-NEXT: [[IV_NEXT]] = getelementptr inbounds [[PAIR]], ptr [[IV]], i64 1
+; CHECK-NEXT: [[EC:%.*]] = icmp eq ptr [[IV_NEXT]], [[END]]
+; CHECK-NEXT: br i1 [[EC]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP2:![0-9]+]]
+; CHECK: exit:
+; CHECK-NEXT: ret void
;
entry:
br label %loop
diff --git a/llvm/test/Transforms/LoopVectorize/opaque-ptr.ll b/llvm/test/Transforms/LoopVectorize/opaque-ptr.ll
index 7beb609..5ec6535 100644
--- a/llvm/test/Transforms/LoopVectorize/opaque-ptr.ll
+++ b/llvm/test/Transforms/LoopVectorize/opaque-ptr.ll
@@ -57,16 +57,20 @@
; CHECK-NEXT: [[IND_END:%.*]] = getelementptr i8, ptr [[START]], i64 [[TMP4]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
-; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi ptr [ [[START]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT: [[TMP5:%.*]] = getelementptr i8, ptr [[POINTER_PHI]], <2 x i64> <i64 0, i64 8>
-; CHECK-NEXT: [[TMP6:%.*]] = extractelement <2 x ptr> [[TMP5]], i32 0
-; CHECK-NEXT: [[TMP7:%.*]] = getelementptr ptr, ptr [[TMP6]], i32 0
-; CHECK-NEXT: store <2 x ptr> [[TMP5]], ptr [[TMP7]], align 4
+; CHECK-NEXT: [[TMP5:%.*]] = add i64 [[INDEX]], 0
+; CHECK-NEXT: [[TMP6:%.*]] = mul i64 [[TMP5]], 8
+; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i8, ptr [[START]], i64 [[TMP6]]
+; CHECK-NEXT: [[TMP7:%.*]] = add i64 [[INDEX]], 1
+; CHECK-NEXT: [[TMP8:%.*]] = mul i64 [[TMP7]], 8
+; CHECK-NEXT: [[NEXT_GEP3:%.*]] = getelementptr i8, ptr [[START]], i64 [[TMP8]]
+; CHECK-NEXT: [[TMP9:%.*]] = insertelement <2 x ptr> poison, ptr [[NEXT_GEP]], i32 0
+; CHECK-NEXT: [[TMP10:%.*]] = insertelement <2 x ptr> [[TMP9]], ptr [[NEXT_GEP3]], i32 1
+; CHECK-NEXT: [[TMP11:%.*]] = getelementptr ptr, ptr [[NEXT_GEP]], i32 0
+; CHECK-NEXT: store <2 x ptr> [[TMP10]], ptr [[TMP11]], align 4
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 2
-; CHECK-NEXT: [[PTR_IND]] = getelementptr i8, ptr [[POINTER_PHI]], i64 16
-; CHECK-NEXT: [[TMP8:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT: br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT: [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP3]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
