[VectorCombine] Add foldShuffleToIdentity (#88693)

This patch adds a basic version of a combine that attempts to remove
shuffles that when combined simplify away to an identity shuffle. For
example:
%ab = shufflevector <8 x half> %a, <8 x half> poison, <4 x i32> <i32 3,
i32 2, i32 1, i32 0>
%at = shufflevector <8 x half> %a, <8 x half> poison, <4 x i32> <i32 7,
i32 6, i32 5, i32 4>
  %abt = fneg <4 x half> %at
  %abb = fneg <4 x half> %ab
%r = shufflevector <4 x half> %abt, <4 x half> %abb, <8 x i32> <i32 7,
i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
By looking through the shuffles and fneg, it can be simplified to:
  %r = fneg <8 x half> %a

The code tracks each lane starting from the original shuffle, keeping a
track of a vector of {src, idx}. As we propagate up through the
instructions we will either look through intermediate instructions
(binops and unops) or see a collections of lanes that all have the same
src and incrementing idx (an identity). We can also see a single value
with identical lanes, which we can treat like a splat.

Only the basic version is added here, handling identities, splats,
binops and unops. In follow-up patches other instructions can be added
such as constants, intrinsics, cmp/sel and zext/sext/trunc.

4 files changed, 212 insertions, 146 deletions

diff --git a/llvm/lib/Transforms/Vectorize/VectorCombine.cpp b/llvm/lib/Transforms/Vectorize/VectorCombine.cpp
index da3c780550a0..04d2ff594bf5 100644
--- a/llvm/lib/Transforms/Vectorize/VectorCombine.cpp
+++ b/llvm/lib/Transforms/Vectorize/VectorCombine.cpp
@@ -114,6 +114,7 @@ private:
   bool foldShuffleOfBinops(Instruction &I);
   bool foldShuffleOfCastops(Instruction &I);
   bool foldShuffleOfShuffles(Instruction &I);
+  bool foldShuffleToIdentity(Instruction &I);
   bool foldShuffleFromReductions(Instruction &I);
   bool foldTruncFromReductions(Instruction &I);
   bool foldSelectShuffle(Instruction &I, bool FromReduction = false);
@@ -1667,6 +1668,151 @@ bool VectorCombine::foldShuffleOfShuffles(Instruction &I) {
   return true;
 }
 
+// Starting from a shuffle, look up through operands tracking the shuffled index
+// of each lane. If we can simplify away the shuffles to identities then
+// do so.
+bool VectorCombine::foldShuffleToIdentity(Instruction &I) {
+  auto *Ty = dyn_cast<FixedVectorType>(I.getType());
+  if (!Ty || !isa<Instruction>(I.getOperand(0)) ||
+      !isa<Instruction>(I.getOperand(1)))
+    return false;
+
+  using InstLane = std::pair<Value *, int>;
+
+  auto LookThroughShuffles = [](Value *V, int Lane) -> InstLane {
+    while (auto *SV = dyn_cast<ShuffleVectorInst>(V)) {
+      unsigned NumElts =
+          cast<FixedVectorType>(SV->getOperand(0)->getType())->getNumElements();
+      int M = SV->getMaskValue(Lane);
+      if (M < 0)
+        return {nullptr, PoisonMaskElem};
+      else if (M < (int)NumElts) {
+        V = SV->getOperand(0);
+        Lane = M;
+      } else {
+        V = SV->getOperand(1);
+        Lane = M - NumElts;
+      }
+    }
+    return InstLane{V, Lane};
+  };
+
+  auto GenerateInstLaneVectorFromOperand =
+      [&LookThroughShuffles](ArrayRef<InstLane> Item, int Op) {
+        SmallVector<InstLane> NItem;
+        for (InstLane V : Item) {
+          NItem.emplace_back(
+              !V.first
+                  ? InstLane{nullptr, PoisonMaskElem}
+                  : LookThroughShuffles(
+                        cast<Instruction>(V.first)->getOperand(Op), V.second));
+        }
+        return NItem;
+      };
+
+  SmallVector<InstLane> Start(Ty->getNumElements());
+  for (unsigned M = 0, E = Ty->getNumElements(); M < E; ++M)
+    Start[M] = LookThroughShuffles(&I, M);
+
+  SmallVector<SmallVector<InstLane>> Worklist;
+  Worklist.push_back(Start);
+  SmallPtrSet<Value *, 4> IdentityLeafs, SplatLeafs;
+  unsigned NumVisited = 0;
+
+  while (!Worklist.empty()) {
+    SmallVector<InstLane> Item = Worklist.pop_back_val();
+    if (++NumVisited > MaxInstrsToScan)
+      return false;
+
+    // If we found an undef first lane then bail out to keep things simple.
+    if (!Item[0].first)
+      return false;
+
+    // Look for an identity value.
+    if (Item[0].second == 0 && Item[0].first->getType() == Ty &&
+        all_of(drop_begin(enumerate(Item)), [&](const auto &E) {
+          return !E.value().first || (E.value().first == Item[0].first &&
+                                      E.value().second == (int)E.index());
+        })) {
+      IdentityLeafs.insert(Item[0].first);
+      continue;
+    }
+    // Look for a splat value.
+    if (all_of(drop_begin(Item), [&](InstLane &IL) {
+          return !IL.first ||
+                 (IL.first == Item[0].first && IL.second == Item[0].second);
+        })) {
+      SplatLeafs.insert(Item[0].first);
+      continue;
+    }
+
+    // We need each element to be the same type of value, and check that each
+    // element has a single use.
+    if (!all_of(drop_begin(Item), [&](InstLane IL) {
+          if (!IL.first)
+            return true;
+          if (auto *I = dyn_cast<Instruction>(IL.first); I && !I->hasOneUse())
+            return false;
+          if (IL.first->getValueID() != Item[0].first->getValueID())
+            return false;
+          auto *II = dyn_cast<IntrinsicInst>(IL.first);
+          return !II ||
+                 II->getIntrinsicID() ==
+                     cast<IntrinsicInst>(Item[0].first)->getIntrinsicID();
+        }))
+      return false;
+
+    // Check the operator is one that we support. We exclude div/rem in case
+    // they hit UB from poison lanes.
+    if (isa<BinaryOperator>(Item[0].first) &&
+        !cast<BinaryOperator>(Item[0].first)->isIntDivRem()) {
+      Worklist.push_back(GenerateInstLaneVectorFromOperand(Item, 0));
+      Worklist.push_back(GenerateInstLaneVectorFromOperand(Item, 1));
+    } else if (isa<UnaryOperator>(Item[0].first)) {
+      Worklist.push_back(GenerateInstLaneVectorFromOperand(Item, 0));
+    } else {
+      return false;
+    }
+  }
+
+  // If we got this far, we know the shuffles are superfluous and can be
+  // removed. Scan through again and generate the new tree of instructions.
+  std::function<Value *(ArrayRef<InstLane>)> Generate =
+      [&](ArrayRef<InstLane> Item) -> Value * {
+    if (IdentityLeafs.contains(Item[0].first) &&
+        all_of(drop_begin(enumerate(Item)), [&](const auto &E) {
+          return !E.value().first || (E.value().first == Item[0].first &&
+                                      E.value().second == (int)E.index());
+        })) {
+      return Item[0].first;
+    }
+    if (SplatLeafs.contains(Item[0].first)) {
+      if (auto ILI = dyn_cast<Instruction>(Item[0].first))
+        Builder.SetInsertPoint(*ILI->getInsertionPointAfterDef());
+      else if (isa<Argument>(Item[0].first))
+        Builder.SetInsertPointPastAllocas(I.getParent()->getParent());
+      SmallVector<int, 16> Mask(Ty->getNumElements(), Item[0].second);
+      return Builder.CreateShuffleVector(Item[0].first, Mask);
+    }
+
+    auto *I = cast<Instruction>(Item[0].first);
+    SmallVector<Value *> Ops(I->getNumOperands());
+    for (unsigned Idx = 0, E = I->getNumOperands(); Idx < E; Idx++)
+      Ops[Idx] = Generate(GenerateInstLaneVectorFromOperand(Item, Idx));
+    Builder.SetInsertPoint(I);
+    if (auto BI = dyn_cast<BinaryOperator>(I))
+      return Builder.CreateBinOp((Instruction::BinaryOps)BI->getOpcode(),
+                                 Ops[0], Ops[1]);
+    assert(isa<UnaryInstruction>(I) &&
+           "Unexpected instruction type in Generate");
+    return Builder.CreateUnOp((Instruction::UnaryOps)I->getOpcode(), Ops[0]);
+  };
+
+  Value *V = Generate(Start);
+  replaceValue(I, *V);
+  return true;
+}
+
 /// Given a commutative reduction, the order of the input lanes does not alter
 /// the results. We can use this to remove certain shuffles feeding the
 /// reduction, removing the need to shuffle at all.
@@ -2224,6 +2370,7 @@ bool VectorCombine::run() {
         MadeChange |= foldShuffleOfCastops(I);
         MadeChange |= foldShuffleOfShuffles(I);
         MadeChange |= foldSelectShuffle(I);
+        MadeChange |= foldShuffleToIdentity(I);
         break;
       case Instruction::BitCast:
         MadeChange |= foldBitcastShuffle(I);
diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/interleavevectorization.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/interleavevectorization.ll
index f1d7c0e0c412..c085e10c049a 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/interleavevectorization.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/interleavevectorization.ll
@@ -22,19 +22,13 @@ define void @add4(ptr noalias noundef %x, ptr noalias noundef %y, i32 noundef %n
 ; CHECK-NEXT:    [[WIDE_VEC:%.*]] = load <32 x i16>, ptr [[TMP0]], align 2
 ; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds i16, ptr [[X]], i64 [[OFFSET_IDX]]
 ; CHECK-NEXT:    [[WIDE_VEC24:%.*]] = load <32 x i16>, ptr [[TMP1]], align 2
-; CHECK-NEXT:    [[TMP2:%.*]] = add <32 x i16> [[WIDE_VEC24]], [[WIDE_VEC]]
-; CHECK-NEXT:    [[TMP3:%.*]] = add <32 x i16> [[WIDE_VEC24]], [[WIDE_VEC]]
-; CHECK-NEXT:    [[TMP4:%.*]] = add <32 x i16> [[WIDE_VEC24]], [[WIDE_VEC]]
-; CHECK-NEXT:    [[TMP5:%.*]] = or disjoint i64 [[OFFSET_IDX]], 3
-; CHECK-NEXT:    [[TMP6:%.*]] = add <32 x i16> [[WIDE_VEC24]], [[WIDE_VEC]]
-; CHECK-NEXT:    [[GEP:%.*]] = getelementptr i16, ptr [[INVARIANT_GEP]], i64 [[TMP5]]
-; CHECK-NEXT:    [[TMP7:%.*]] = shufflevector <32 x i16> [[TMP2]], <32 x i16> [[TMP3]], <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 16, i32 20, i32 24, i32 28, i32 33, i32 37, i32 41, i32 45, i32 49, i32 53, i32 57, i32 61>
-; CHECK-NEXT:    [[TMP8:%.*]] = shufflevector <32 x i16> [[TMP4]], <32 x i16> [[TMP6]], <16 x i32> <i32 2, i32 6, i32 10, i32 14, i32 18, i32 22, i32 26, i32 30, i32 35, i32 39, i32 43, i32 47, i32 51, i32 55, i32 59, i32 63>
-; CHECK-NEXT:    [[INTERLEAVED_VEC:%.*]] = shufflevector <16 x i16> [[TMP7]], <16 x i16> [[TMP8]], <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
+; CHECK-NEXT:    [[INTERLEAVED_VEC:%.*]] = add <32 x i16> [[WIDE_VEC24]], [[WIDE_VEC]]
+; CHECK-NEXT:    [[TMP2:%.*]] = or disjoint i64 [[OFFSET_IDX]], 3
+; CHECK-NEXT:    [[GEP:%.*]] = getelementptr i16, ptr [[INVARIANT_GEP]], i64 [[TMP2]]
 ; CHECK-NEXT:    store <32 x i16> [[INTERLEAVED_VEC]], ptr [[GEP]], align 2
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT]], 256
-; CHECK-NEXT:    br i1 [[TMP9]], label [[FOR_END:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT:    [[TMP3:%.*]] = icmp eq i64 [[INDEX_NEXT]], 256
+; CHECK-NEXT:    br i1 [[TMP3]], label [[FOR_END:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
 ; CHECK:       for.end:
 ; CHECK-NEXT:    ret void
 ;
@@ -412,22 +406,13 @@ define void @addmul(ptr noalias noundef %x, ptr noundef %y, ptr noundef %z, i32
 ; CHECK-NEXT:    [[TMP2:%.*]] = mul <32 x i16> [[WIDE_VEC31]], [[WIDE_VEC]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i16, ptr [[X]], i64 [[OFFSET_IDX]]
 ; CHECK-NEXT:    [[WIDE_VEC36:%.*]] = load <32 x i16>, ptr [[TMP3]], align 2
-; CHECK-NEXT:    [[TMP4:%.*]] = add <32 x i16> [[TMP2]], [[WIDE_VEC36]]
-; CHECK-NEXT:    [[TMP5:%.*]] = mul <32 x i16> [[WIDE_VEC31]], [[WIDE_VEC]]
-; CHECK-NEXT:    [[TMP6:%.*]] = add <32 x i16> [[TMP5]], [[WIDE_VEC36]]
-; CHECK-NEXT:    [[TMP7:%.*]] = mul <32 x i16> [[WIDE_VEC31]], [[WIDE_VEC]]
-; CHECK-NEXT:    [[TMP8:%.*]] = add <32 x i16> [[TMP7]], [[WIDE_VEC36]]
-; CHECK-NEXT:    [[TMP9:%.*]] = or disjoint i64 [[OFFSET_IDX]], 3
-; CHECK-NEXT:    [[TMP10:%.*]] = mul <32 x i16> [[WIDE_VEC31]], [[WIDE_VEC]]
-; CHECK-NEXT:    [[TMP11:%.*]] = add <32 x i16> [[TMP10]], [[WIDE_VEC36]]
-; CHECK-NEXT:    [[GEP:%.*]] = getelementptr i16, ptr [[INVARIANT_GEP]], i64 [[TMP9]]
-; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <32 x i16> [[TMP4]], <32 x i16> [[TMP6]], <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 16, i32 20, i32 24, i32 28, i32 33, i32 37, i32 41, i32 45, i32 49, i32 53, i32 57, i32 61>
-; CHECK-NEXT:    [[TMP13:%.*]] = shufflevector <32 x i16> [[TMP8]], <32 x i16> [[TMP11]], <16 x i32> <i32 2, i32 6, i32 10, i32 14, i32 18, i32 22, i32 26, i32 30, i32 35, i32 39, i32 43, i32 47, i32 51, i32 55, i32 59, i32 63>
-; CHECK-NEXT:    [[INTERLEAVED_VEC:%.*]] = shufflevector <16 x i16> [[TMP12]], <16 x i16> [[TMP13]], <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
+; CHECK-NEXT:    [[INTERLEAVED_VEC:%.*]] = add <32 x i16> [[TMP2]], [[WIDE_VEC36]]
+; CHECK-NEXT:    [[TMP4:%.*]] = or disjoint i64 [[OFFSET_IDX]], 3
+; CHECK-NEXT:    [[GEP:%.*]] = getelementptr i16, ptr [[INVARIANT_GEP]], i64 [[TMP4]]
 ; CHECK-NEXT:    store <32 x i16> [[INTERLEAVED_VEC]], ptr [[GEP]], align 2
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
-; CHECK-NEXT:    [[TMP14:%.*]] = icmp eq i64 [[INDEX_NEXT]], 256
-; CHECK-NEXT:    br i1 [[TMP14]], label [[FOR_END:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
+; CHECK-NEXT:    [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], 256
+; CHECK-NEXT:    br i1 [[TMP5]], label [[FOR_END:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
 ; CHECK:       for.end:
 ; CHECK-NEXT:    ret void
 ;
diff --git a/llvm/test/Transforms/VectorCombine/AArch64/shuffletoidentity.ll b/llvm/test/Transforms/VectorCombine/AArch64/shuffletoidentity.ll
index 6a81964b917e..b96732e24ce4 100644
--- a/llvm/test/Transforms/VectorCombine/AArch64/shuffletoidentity.ll
+++ b/llvm/test/Transforms/VectorCombine/AArch64/shuffletoidentity.ll
@@ -15,13 +15,7 @@ define <8 x i8> @trivial(<8 x i8> %a) {
 
 define <8 x i8> @add(<8 x i8> %a, <8 x i8> %b) {
 ; CHECK-LABEL: @add(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x i8> [[A:%.*]], <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x i8> [[A]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[BB:%.*]] = shufflevector <8 x i8> [[B:%.*]], <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[BT:%.*]] = shufflevector <8 x i8> [[B]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[ABT:%.*]] = add <4 x i8> [[AT]], [[BT]]
-; CHECK-NEXT:    [[ABB:%.*]] = add <4 x i8> [[AB]], [[BB]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x i8> [[ABT]], <4 x i8> [[ABB]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[R:%.*]] = add <8 x i8> [[A:%.*]], [[B:%.*]]
 ; CHECK-NEXT:    ret <8 x i8> [[R]]
 ;
   %ab = shufflevector <8 x i8> %a, <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
@@ -80,13 +74,7 @@ define <8 x i8> @wrong_lanes(<8 x i8> %a, <8 x i8> %b) {
 
 define <8 x half> @fadd(<8 x half> %a, <8 x half> %b) {
 ; CHECK-LABEL: @fadd(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x half> [[A:%.*]], <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x half> [[A]], <8 x half> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[BB:%.*]] = shufflevector <8 x half> [[B:%.*]], <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[BT:%.*]] = shufflevector <8 x half> [[B]], <8 x half> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[ABT:%.*]] = fadd <4 x half> [[AT]], [[BT]]
-; CHECK-NEXT:    [[ABB:%.*]] = fadd <4 x half> [[AB]], [[BB]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x half> [[ABT]], <4 x half> [[ABB]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[R:%.*]] = fadd <8 x half> [[A:%.*]], [[B:%.*]]
 ; CHECK-NEXT:    ret <8 x half> [[R]]
 ;
   %ab = shufflevector <8 x half> %a, <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
@@ -101,11 +89,7 @@ define <8 x half> @fadd(<8 x half> %a, <8 x half> %b) {
 
 define <8 x half> @fneg(<8 x half> %a, <8 x half> %b) {
 ; CHECK-LABEL: @fneg(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x half> [[A:%.*]], <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x half> [[A]], <8 x half> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[ABT:%.*]] = fneg <4 x half> [[AT]]
-; CHECK-NEXT:    [[ABB:%.*]] = fneg <4 x half> [[AB]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x half> [[ABT]], <4 x half> [[ABB]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[R:%.*]] = fneg <8 x half> [[A:%.*]]
 ; CHECK-NEXT:    ret <8 x half> [[R]]
 ;
   %ab = shufflevector <8 x half> %a, <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
@@ -135,12 +119,8 @@ define <8 x i8> @abs(<8 x i8> %a) {
 
 define <8 x half> @splat0(<8 x half> %a, <8 x half> %b) {
 ; CHECK-LABEL: @splat0(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x half> [[A:%.*]], <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x half> [[A]], <8 x half> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[BS:%.*]] = shufflevector <8 x half> [[B:%.*]], <8 x half> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[ABT:%.*]] = fadd <4 x half> [[AT]], [[BS]]
-; CHECK-NEXT:    [[ABB:%.*]] = fadd <4 x half> [[AB]], [[BS]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x half> [[ABT]], <4 x half> [[ABB]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <8 x half> [[B:%.*]], <8 x half> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[R:%.*]] = fadd <8 x half> [[A:%.*]], [[TMP1]]
 ; CHECK-NEXT:    ret <8 x half> [[R]]
 ;
   %ab = shufflevector <8 x half> %a, <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
@@ -154,12 +134,8 @@ define <8 x half> @splat0(<8 x half> %a, <8 x half> %b) {
 
 define <8 x half> @splat2(<8 x half> %a, <8 x half> %b) {
 ; CHECK-LABEL: @splat2(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x half> [[A:%.*]], <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x half> [[A]], <8 x half> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[BS:%.*]] = shufflevector <8 x half> [[B:%.*]], <8 x half> poison, <4 x i32> <i32 2, i32 2, i32 2, i32 2>
-; CHECK-NEXT:    [[ABT:%.*]] = fadd <4 x half> [[AT]], [[BS]]
-; CHECK-NEXT:    [[ABB:%.*]] = fadd <4 x half> [[AB]], [[BS]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x half> [[ABT]], <4 x half> [[ABB]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <8 x half> [[B:%.*]], <8 x half> poison, <8 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
+; CHECK-NEXT:    [[R:%.*]] = fadd <8 x half> [[A:%.*]], [[TMP1]]
 ; CHECK-NEXT:    ret <8 x half> [[R]]
 ;
   %ab = shufflevector <8 x half> %a, <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
@@ -173,12 +149,8 @@ define <8 x half> @splat2(<8 x half> %a, <8 x half> %b) {
 
 define <8 x half> @splatandidentity(<8 x half> %a, <8 x half> %b) {
 ; CHECK-LABEL: @splatandidentity(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x half> [[A:%.*]], <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x half> [[A]], <8 x half> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[BS:%.*]] = shufflevector <8 x half> [[A]], <8 x half> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[ABT:%.*]] = fadd <4 x half> [[AT]], [[BS]]
-; CHECK-NEXT:    [[ABB:%.*]] = fadd <4 x half> [[AB]], [[BS]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x half> [[ABT]], <4 x half> [[ABB]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <8 x half> [[A:%.*]], <8 x half> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[R:%.*]] = fadd <8 x half> [[A]], [[TMP1]]
 ; CHECK-NEXT:    ret <8 x half> [[R]]
 ;
   %ab = shufflevector <8 x half> %a, <8 x half> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
@@ -192,11 +164,9 @@ define <8 x half> @splatandidentity(<8 x half> %a, <8 x half> %b) {
 
 define <8 x half> @splattwice(<8 x half> %a, <8 x half> %b) {
 ; CHECK-LABEL: @splattwice(
-; CHECK-NEXT:    [[AS:%.*]] = shufflevector <8 x half> [[A:%.*]], <8 x half> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[BS:%.*]] = shufflevector <8 x half> [[B:%.*]], <8 x half> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[AB1:%.*]] = fadd <4 x half> [[AS]], [[BS]]
-; CHECK-NEXT:    [[AB2:%.*]] = fadd <4 x half> [[AS]], [[BS]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x half> [[AB1]], <4 x half> [[AB2]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <8 x half> [[B:%.*]], <8 x half> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <8 x half> [[A:%.*]], <8 x half> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[R:%.*]] = fadd <8 x half> [[TMP2]], [[TMP1]]
 ; CHECK-NEXT:    ret <8 x half> [[R]]
 ;
   %as = shufflevector <8 x half> %a, <8 x half> poison, <4 x i32> zeroinitializer
@@ -209,13 +179,7 @@ define <8 x half> @splattwice(<8 x half> %a, <8 x half> %b) {
 
 define <8 x i8> @undeflane(<8 x i8> %a, <8 x i8> %b) {
 ; CHECK-LABEL: @undeflane(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x i8> [[A:%.*]], <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x i8> [[A]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[BB:%.*]] = shufflevector <8 x i8> [[B:%.*]], <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[BT:%.*]] = shufflevector <8 x i8> [[B]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[ABT:%.*]] = add <4 x i8> [[AT]], [[BT]]
-; CHECK-NEXT:    [[ABB:%.*]] = add <4 x i8> [[AB]], [[BB]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x i8> [[ABT]], <4 x i8> [[ABB]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 poison, i32 1, i32 0>
+; CHECK-NEXT:    [[R:%.*]] = add <8 x i8> [[A:%.*]], [[B:%.*]]
 ; CHECK-NEXT:    ret <8 x i8> [[R]]
 ;
   %ab = shufflevector <8 x i8> %a, <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
@@ -281,18 +245,9 @@ define <8 x i8> @constantdiff2(<8 x i8> %a) {
 
 define <8 x i8> @inner_shuffle(<8 x i8> %a, <8 x i8> %b, <8 x i8> %c) {
 ; CHECK-LABEL: @inner_shuffle(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x i8> [[A:%.*]], <8 x i8> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x i8> [[A]], <8 x i8> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[BB:%.*]] = shufflevector <8 x i8> [[B:%.*]], <8 x i8> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[BT:%.*]] = shufflevector <8 x i8> [[B]], <8 x i8> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[CS:%.*]] = shufflevector <8 x i8> [[C:%.*]], <8 x i8> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[ABT:%.*]] = mul <4 x i8> [[AT]], [[BT]]
-; CHECK-NEXT:    [[ABB:%.*]] = mul <4 x i8> [[AB]], [[BB]]
-; CHECK-NEXT:    [[ABT2:%.*]] = shufflevector <4 x i8> [[ABT]], <4 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[ABB2:%.*]] = shufflevector <4 x i8> [[ABB]], <4 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[ABT3:%.*]] = add <4 x i8> [[ABT2]], [[CS]]
-; CHECK-NEXT:    [[ABB3:%.*]] = add <4 x i8> [[ABB2]], [[CS]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x i8> [[ABT3]], <4 x i8> [[ABB3]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <8 x i8> [[C:%.*]], <8 x i8> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP2:%.*]] = mul <8 x i8> [[A:%.*]], [[B:%.*]]
+; CHECK-NEXT:    [[R:%.*]] = add <8 x i8> [[TMP2]], [[TMP1]]
 ; CHECK-NEXT:    ret <8 x i8> [[R]]
 ;
   %ab = shufflevector <8 x i8> %a, <8 x i8> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
@@ -335,14 +290,9 @@ define <8 x i8> @extrause_add(<8 x i8> %a, <8 x i8> %b) {
 
 define <8 x i8> @extrause_shuffle(<8 x i8> %a, <8 x i8> %b) {
 ; CHECK-LABEL: @extrause_shuffle(
-; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x i8> [[A:%.*]], <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x i8> [[A]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
-; CHECK-NEXT:    [[BB:%.*]] = shufflevector <8 x i8> [[B:%.*]], <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:    [[BT:%.*]] = shufflevector <8 x i8> [[B]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
+; CHECK-NEXT:    [[BT:%.*]] = shufflevector <8 x i8> [[B:%.*]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
 ; CHECK-NEXT:    call void @use(<4 x i8> [[BT]])
-; CHECK-NEXT:    [[ABT:%.*]] = add <4 x i8> [[AT]], [[BT]]
-; CHECK-NEXT:    [[ABB:%.*]] = add <4 x i8> [[AB]], [[BB]]
-; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x i8> [[ABT]], <4 x i8> [[ABB]], <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[R:%.*]] = add <8 x i8> [[A:%.*]], [[B]]
 ; CHECK-NEXT:    ret <8 x i8> [[R]]
 ;
   %ab = shufflevector <8 x i8> %a, <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
@@ -457,7 +407,7 @@ define <8 x i8> @intrinsics_minmax(<8 x i8> %a, <8 x i8> %b) {
 ; CHECK-NEXT:    [[BT:%.*]] = shufflevector <8 x i8> [[B]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
 ; CHECK-NEXT:    [[ABT:%.*]] = call <4 x i8> @llvm.smin.v4i8(<4 x i8> [[AT]], <4 x i8> [[BT]])
 ; CHECK-NEXT:    [[ABB:%.*]] = call <4 x i8> @llvm.smin.v4i8(<4 x i8> [[AB]], <4 x i8> [[BB]])
-; CHECK-NEXT:    [[ABT1:%.*]] = call <4 x i8> @llvm.smin.v4i8(<4 x i8> [[ABT]], <4 x i8> [[BT]])
+; CHECK-NEXT:    [[ABT1:%.*]] = call <4 x i8> @llvm.smax.v4i8(<4 x i8> [[ABT]], <4 x i8> [[BT]])
 ; CHECK-NEXT:    [[ABB1:%.*]] = call <4 x i8> @llvm.smax.v4i8(<4 x i8> [[ABB]], <4 x i8> [[BB]])
 ; CHECK-NEXT:    [[ABT2:%.*]] = call <4 x i8> @llvm.umin.v4i8(<4 x i8> [[ABT1]], <4 x i8> [[BT]])
 ; CHECK-NEXT:    [[ABB2:%.*]] = call <4 x i8> @llvm.umin.v4i8(<4 x i8> [[ABB1]], <4 x i8> [[BB]])
@@ -472,7 +422,7 @@ define <8 x i8> @intrinsics_minmax(<8 x i8> %a, <8 x i8> %b) {
   %bt = shufflevector <8 x i8> %b, <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
   %abt = call <4 x i8> @llvm.smin.v4i8(<4 x i8> %at, <4 x i8> %bt)
   %abb = call <4 x i8> @llvm.smin.v4i8(<4 x i8> %ab, <4 x i8> %bb)
-  %abt1 = call <4 x i8> @llvm.smin.v4i8(<4 x i8> %abt, <4 x i8> %bt)
+  %abt1 = call <4 x i8> @llvm.smax.v4i8(<4 x i8> %abt, <4 x i8> %bt)
   %abb1 = call <4 x i8> @llvm.smax.v4i8(<4 x i8> %abb, <4 x i8> %bb)
   %abt2 = call <4 x i8> @llvm.umin.v4i8(<4 x i8> %abt1, <4 x i8> %bt)
   %abb2 = call <4 x i8> @llvm.umin.v4i8(<4 x i8> %abb1, <4 x i8> %bb)
@@ -536,57 +486,42 @@ define <8 x i8> @intrinsics_different(<8 x i8> %a, <8 x i8> %b) {
   ret <8 x i8> %r
 }
 
+; div and rem are currently excluded.
+define <8 x i8> @div(<8 x i8> %a, <8 x i8> %b) {
+; CHECK-LABEL: @div(
+; CHECK-NEXT:    [[AB:%.*]] = shufflevector <8 x i8> [[A:%.*]], <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[AT:%.*]] = shufflevector <8 x i8> [[A]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
+; CHECK-NEXT:    [[BB:%.*]] = shufflevector <8 x i8> [[B:%.*]], <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    [[BT:%.*]] = shufflevector <8 x i8> [[B]], <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
+; CHECK-NEXT:    [[ABT:%.*]] = udiv <4 x i8> [[AT]], [[BT]]
+; CHECK-NEXT:    [[ABB:%.*]] = udiv <4 x i8> [[AB]], [[BB]]
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x i8> [[ABT]], <4 x i8> [[ABB]], <8 x i32> <i32 7, i32 poison, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; CHECK-NEXT:    ret <8 x i8> [[R]]
+;
+  %ab = shufflevector <8 x i8> %a, <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  %at = shufflevector <8 x i8> %a, <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
+  %bb = shufflevector <8 x i8> %b, <8 x i8> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  %bt = shufflevector <8 x i8> %b, <8 x i8> poison, <4 x i32> <i32 7, i32 6, i32 5, i32 4>
+  %abt = udiv <4 x i8> %at, %bt
+  %abb = udiv <4 x i8> %ab, %bb
+  %r = shufflevector <4 x i8> %abt, <4 x i8> %abb, <8 x i32> <i32 7, i32 poison, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+  ret <8 x i8> %r
+}
+
 define void @v8f64interleave(i64 %0, ptr %1, ptr %x, double %z) {
 ; CHECK-LABEL: @v8f64interleave(
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[Z:%.*]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <2 x double> [[BROADCAST_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <2 x double> [[BROADCAST_SPLATINSERT]], <2 x double> poison, <16 x i32> zeroinitializer
 ; CHECK-NEXT:    [[WIDE_VEC:%.*]] = load <16 x double>, ptr [[TMP1:%.*]], align 8
-; CHECK-NEXT:    [[STRIDED_VEC:%.*]] = shufflevector <16 x double> [[WIDE_VEC]], <16 x double> poison, <2 x i32> <i32 0, i32 8>
-; CHECK-NEXT:    [[STRIDED_VEC27:%.*]] = shufflevector <16 x double> [[WIDE_VEC]], <16 x double> poison, <2 x i32> <i32 1, i32 9>
-; CHECK-NEXT:    [[STRIDED_VEC28:%.*]] = shufflevector <16 x double> [[WIDE_VEC]], <16 x double> poison, <2 x i32> <i32 2, i32 10>
-; CHECK-NEXT:    [[STRIDED_VEC29:%.*]] = shufflevector <16 x double> [[WIDE_VEC]], <16 x double> poison, <2 x i32> <i32 3, i32 11>
-; CHECK-NEXT:    [[STRIDED_VEC30:%.*]] = shufflevector <16 x double> [[WIDE_VEC]], <16 x double> poison, <2 x i32> <i32 4, i32 12>
-; CHECK-NEXT:    [[STRIDED_VEC31:%.*]] = shufflevector <16 x double> [[WIDE_VEC]], <16 x double> poison, <2 x i32> <i32 5, i32 13>
-; CHECK-NEXT:    [[STRIDED_VEC32:%.*]] = shufflevector <16 x double> [[WIDE_VEC]], <16 x double> poison, <2 x i32> <i32 6, i32 14>
-; CHECK-NEXT:    [[STRIDED_VEC33:%.*]] = shufflevector <16 x double> [[WIDE_VEC]], <16 x double> poison, <2 x i32> <i32 7, i32 15>
-; CHECK-NEXT:    [[TMP2:%.*]] = fmul fast <2 x double> [[STRIDED_VEC]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds double, ptr [[X:%.*]], i64 [[TMP0:%.*]]
-; CHECK-NEXT:    [[WIDE_VEC34:%.*]] = load <16 x double>, ptr [[TMP3]], align 8
-; CHECK-NEXT:    [[STRIDED_VEC35:%.*]] = shufflevector <16 x double> [[WIDE_VEC34]], <16 x double> poison, <2 x i32> <i32 0, i32 8>
-; CHECK-NEXT:    [[STRIDED_VEC36:%.*]] = shufflevector <16 x double> [[WIDE_VEC34]], <16 x double> poison, <2 x i32> <i32 1, i32 9>
-; CHECK-NEXT:    [[STRIDED_VEC37:%.*]] = shufflevector <16 x double> [[WIDE_VEC34]], <16 x double> poison, <2 x i32> <i32 2, i32 10>
-; CHECK-NEXT:    [[STRIDED_VEC38:%.*]] = shufflevector <16 x double> [[WIDE_VEC34]], <16 x double> poison, <2 x i32> <i32 3, i32 11>
-; CHECK-NEXT:    [[STRIDED_VEC39:%.*]] = shufflevector <16 x double> [[WIDE_VEC34]], <16 x double> poison, <2 x i32> <i32 4, i32 12>
-; CHECK-NEXT:    [[STRIDED_VEC40:%.*]] = shufflevector <16 x double> [[WIDE_VEC34]], <16 x double> poison, <2 x i32> <i32 5, i32 13>
-; CHECK-NEXT:    [[STRIDED_VEC41:%.*]] = shufflevector <16 x double> [[WIDE_VEC34]], <16 x double> poison, <2 x i32> <i32 6, i32 14>
-; CHECK-NEXT:    [[STRIDED_VEC42:%.*]] = shufflevector <16 x double> [[WIDE_VEC34]], <16 x double> poison, <2 x i32> <i32 7, i32 15>
-; CHECK-NEXT:    [[TMP4:%.*]] = fadd fast <2 x double> [[STRIDED_VEC35]], [[TMP2]]
-; CHECK-NEXT:    [[TMP5:%.*]] = fmul fast <2 x double> [[STRIDED_VEC27]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP6:%.*]] = fadd fast <2 x double> [[STRIDED_VEC36]], [[TMP5]]
-; CHECK-NEXT:    [[TMP7:%.*]] = fmul fast <2 x double> [[STRIDED_VEC28]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP8:%.*]] = fadd fast <2 x double> [[STRIDED_VEC37]], [[TMP7]]
-; CHECK-NEXT:    [[TMP9:%.*]] = fmul fast <2 x double> [[STRIDED_VEC29]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP10:%.*]] = fadd fast <2 x double> [[STRIDED_VEC38]], [[TMP9]]
-; CHECK-NEXT:    [[TMP11:%.*]] = fmul fast <2 x double> [[STRIDED_VEC30]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP12:%.*]] = fadd fast <2 x double> [[STRIDED_VEC39]], [[TMP11]]
-; CHECK-NEXT:    [[TMP13:%.*]] = fmul fast <2 x double> [[STRIDED_VEC31]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP14:%.*]] = fadd fast <2 x double> [[STRIDED_VEC40]], [[TMP13]]
-; CHECK-NEXT:    [[TMP15:%.*]] = fmul fast <2 x double> [[STRIDED_VEC32]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP16:%.*]] = fadd fast <2 x double> [[STRIDED_VEC41]], [[TMP15]]
-; CHECK-NEXT:    [[TMP17:%.*]] = or disjoint i64 [[TMP0]], 7
-; CHECK-NEXT:    [[TMP18:%.*]] = fmul fast <2 x double> [[STRIDED_VEC33]], [[BROADCAST_SPLAT]]
-; CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds double, ptr [[X]], i64 [[TMP17]]
-; CHECK-NEXT:    [[TMP20:%.*]] = fadd fast <2 x double> [[STRIDED_VEC42]], [[TMP18]]
-; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i8, ptr [[TMP19]], i64 -56
-; CHECK-NEXT:    [[TMP22:%.*]] = shufflevector <2 x double> [[TMP4]], <2 x double> [[TMP6]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[TMP23:%.*]] = shufflevector <2 x double> [[TMP8]], <2 x double> [[TMP10]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[TMP24:%.*]] = shufflevector <2 x double> [[TMP12]], <2 x double> [[TMP14]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[TMP25:%.*]] = shufflevector <2 x double> [[TMP16]], <2 x double> [[TMP20]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[TMP26:%.*]] = shufflevector <4 x double> [[TMP22]], <4 x double> [[TMP23]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[TMP27:%.*]] = shufflevector <4 x double> [[TMP24]], <4 x double> [[TMP25]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[INTERLEAVED_VEC:%.*]] = shufflevector <8 x double> [[TMP26]], <8 x double> [[TMP27]], <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
-; CHECK-NEXT:    store <16 x double> [[INTERLEAVED_VEC]], ptr [[TMP21]], align 8
+; CHECK-NEXT:    [[TMP3:%.*]] = fmul <16 x double> [[WIDE_VEC]], [[TMP2]]
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds double, ptr [[X:%.*]], i64 [[TMP0:%.*]]
+; CHECK-NEXT:    [[WIDE_VEC34:%.*]] = load <16 x double>, ptr [[TMP4]], align 8
+; CHECK-NEXT:    [[INTERLEAVED_VEC:%.*]] = fadd <16 x double> [[WIDE_VEC34]], [[TMP3]]
+; CHECK-NEXT:    [[TMP5:%.*]] = or disjoint i64 [[TMP0]], 7
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds double, ptr [[X]], i64 [[TMP5]]
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i8, ptr [[TMP6]], i64 -56
+; CHECK-NEXT:    store <16 x double> [[INTERLEAVED_VEC]], ptr [[TMP7]], align 8
 ; CHECK-NEXT:    ret void
 ;
 entry:
diff --git a/llvm/test/Transforms/VectorCombine/X86/select-shuffle.ll b/llvm/test/Transforms/VectorCombine/X86/select-shuffle.ll
index 685d661ea6bc..60a6c4b1d9b9 100644
--- a/llvm/test/Transforms/VectorCombine/X86/select-shuffle.ll
+++ b/llvm/test/Transforms/VectorCombine/X86/select-shuffle.ll
@@ -12,12 +12,11 @@ define <4 x double> @PR60649() {
 ; CHECK:       unreachable:
 ; CHECK-NEXT:    br label [[END]]
 ; CHECK:       end:
-; CHECK-NEXT:    [[T0:%.*]] = phi <4 x double> [ zeroinitializer, [[ENTRY:%.*]] ], [ zeroinitializer, [[UNREACHABLE:%.*]] ]
+; CHECK-NEXT:    [[TMP0:%.*]] = phi <4 x double> [ zeroinitializer, [[ENTRY:%.*]] ], [ zeroinitializer, [[UNREACHABLE:%.*]] ]
 ; CHECK-NEXT:    [[T1:%.*]] = phi <4 x double> [ zeroinitializer, [[ENTRY]] ], [ zeroinitializer, [[UNREACHABLE]] ]
-; CHECK-NEXT:    [[TMP0:%.*]] = shufflevector <4 x double> [[T0]], <4 x double> [[T0]], <4 x i32> <i32 2, i32 3, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[T0]], <4 x double> [[T0]], <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP2:%.*]] = fdiv <4 x double> [[TMP1]], <double 0.000000e+00, double 0.000000e+00, double undef, double undef>
-; CHECK-NEXT:    [[TMP3:%.*]] = fmul <4 x double> [[TMP0]], <double 0.000000e+00, double 0.000000e+00, double undef, double undef>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[TMP0]], <4 x double> [[TMP0]], <4 x i32> <i32 2, i32 3, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP2:%.*]] = fdiv <4 x double> [[TMP0]], <double 0.000000e+00, double 0.000000e+00, double undef, double undef>
+; CHECK-NEXT:    [[TMP3:%.*]] = fmul <4 x double> [[TMP1]], <double 0.000000e+00, double 0.000000e+00, double undef, double undef>
 ; CHECK-NEXT:    [[T5:%.*]] = shufflevector <4 x double> [[TMP2]], <4 x double> [[TMP3]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
 ; CHECK-NEXT:    ret <4 x double> [[T5]]
 ;