Index: lib/Transforms/Vectorize/SLPVectorizer.cpp
===================================================================
--- lib/Transforms/Vectorize/SLPVectorizer.cpp
+++ lib/Transforms/Vectorize/SLPVectorizer.cpp
@@ -1352,8 +1352,7 @@
 
   /// Set the Builder insert point to one after the last instruction in
   /// the bundle
-  void setInsertPointAfterBundle(ArrayRef<Value *> VL,
-                                 const InstructionsState &S);
+  void setInsertPointAfterBundle(TreeEntry *E);
 
   /// \returns a vector from a collection of scalars in \p VL.
   Value *Gather(ArrayRef<Value *> VL, VectorType *Ty);
@@ -1390,6 +1389,10 @@
     /// The Scalars are vectorized into this value. It is initialized to Null.
     Value *VectorizedValue = nullptr;
 
+    /// The main/alternate instruction.
+    Instruction *MainOp = nullptr;
+    Instruction *AltOp = nullptr;
+
     /// Do we need to gather this sequence ?
     bool NeedToGather = false;
 
@@ -1456,6 +1459,24 @@
       return Operands[OpIdx][0];
     }
 
+    /// Some of the instructions in the list have alternate opcodes.
+    bool isAltShuffle() const { return MainOp->getOpcode() != AltOp->getOpcode(); }
+
+    bool isOpcodeOrAlt(Instruction *I) const {
+      unsigned CheckedOpcode = I->getOpcode();
+      return (MainOp->getOpcode() == CheckedOpcode || AltOp->getOpcode() == CheckedOpcode);
+    }
+
+    /// Chooses the correct key for scheduling data. If \p Op has the same (or
+    /// alternate) opcode as \p OpValue, the key is \p Op. Otherwise the key is \p
+    /// OpValue.
+    Value *isOneOf(Value *Op) {
+      auto *I = dyn_cast<Instruction>(Op);
+      if (I && isOpcodeOrAlt(I))
+        return Op;
+      return MainOp;
+    }
+
 #ifndef NDEBUG
     /// Debug printer.
     LLVM_DUMP_METHOD void dump() const {
@@ -1496,7 +1517,7 @@
 
   /// Create a new VectorizableTree entry.
   TreeEntry *newTreeEntry(ArrayRef<Value *> VL, bool Vectorized,
-                          const EdgeInfo &UserTreeIdx,
+                          InstructionsState *S, const EdgeInfo &UserTreeIdx,
                           ArrayRef<unsigned> ReuseShuffleIndices = None,
                           ArrayRef<unsigned> ReorderIndices = None) {
     VectorizableTree.push_back(llvm::make_unique<TreeEntry>(VectorizableTree));
@@ -1507,10 +1528,12 @@
     Last->ReuseShuffleIndices.append(ReuseShuffleIndices.begin(),
                                      ReuseShuffleIndices.end());
     Last->ReorderIndices = ReorderIndices;
+    Last->MainOp = S->MainOp;
+    Last->AltOp = S->AltOp;
     if (Vectorized) {
       for (int i = 0, e = VL.size(); i != e; ++i) {
         assert(!getTreeEntry(VL[i]) && "Scalar already in tree!");
-        ScalarToTreeEntry[VL[i]] = Last->Idx;
+        ScalarToTreeEntry[cast<Instruction>(VL[i])] = Last;
       }
     } else {
       MustGather.insert(VL.begin(), VL.end());
@@ -1538,21 +1561,23 @@
 #endif
 
   TreeEntry *getTreeEntry(Value *V) {
-    auto I = ScalarToTreeEntry.find(V);
-    if (I != ScalarToTreeEntry.end())
-      return VectorizableTree[I->second].get();
+    auto *I = dyn_cast<Instruction>(V);
+    auto It = ScalarToTreeEntry.find(I);
+    if (It != ScalarToTreeEntry.end())
+      return It->second;
     return nullptr;
   }
 
   const TreeEntry *getTreeEntry(Value *V) const {
-    auto I = ScalarToTreeEntry.find(V);
-    if (I != ScalarToTreeEntry.end())
-      return VectorizableTree[I->second].get();
+    auto *I = dyn_cast<Instruction>(V);
+    auto It = ScalarToTreeEntry.find(I);
+    if (It != ScalarToTreeEntry.end())
+      return It->second;
     return nullptr;
   }
 
   /// Maps a specific scalar to its tree entry.
-  SmallDenseMap<Value*, int> ScalarToTreeEntry;
+  SmallDenseMap<Instruction *, TreeEntry *> ScalarToTreeEntry;
 
   /// A list of scalars that we found that we need to keep as scalars.
   ValueSet MustGather;
@@ -2226,28 +2251,28 @@
   InstructionsState S = getSameOpcode(VL);
   if (Depth == RecursionMaxDepth) {
     LLVM_DEBUG(dbgs() << "SLP: Gathering due to max recursion depth.\n");
-    newTreeEntry(VL, false, UserTreeIdx);
+    newTreeEntry(VL, false, &S, UserTreeIdx);
     return;
   }
 
   // Don't handle vectors.
   if (S.OpValue->getType()->isVectorTy()) {
     LLVM_DEBUG(dbgs() << "SLP: Gathering due to vector type.\n");
-    newTreeEntry(VL, false, UserTreeIdx);
+    newTreeEntry(VL, false, &S, UserTreeIdx);
     return;
   }
 
   if (StoreInst *SI = dyn_cast<StoreInst>(S.OpValue))
     if (SI->getValueOperand()->getType()->isVectorTy()) {
       LLVM_DEBUG(dbgs() << "SLP: Gathering due to store vector type.\n");
-      newTreeEntry(VL, false, UserTreeIdx);
+      newTreeEntry(VL, false, &S, UserTreeIdx);
       return;
     }
 
   // If all of the operands are identical or constant we have a simple solution.
   if (allConstant(VL) || isSplat(VL) || !allSameBlock(VL) || !S.getOpcode()) {
     LLVM_DEBUG(dbgs() << "SLP: Gathering due to C,S,B,O. \n");
-    newTreeEntry(VL, false, UserTreeIdx);
+    newTreeEntry(VL, false, &S, UserTreeIdx);
     return;
   }
 
@@ -2259,7 +2284,7 @@
     if (EphValues.count(VL[i])) {
       LLVM_DEBUG(dbgs() << "SLP: The instruction (" << *VL[i]
                         << ") is ephemeral.\n");
-      newTreeEntry(VL, false, UserTreeIdx);
+      newTreeEntry(VL, false, &S, UserTreeIdx);
       return;
     }
   }
@@ -2269,7 +2294,7 @@
     LLVM_DEBUG(dbgs() << "SLP: \tChecking bundle: " << *S.OpValue << ".\n");
     if (!E->isSame(VL)) {
       LLVM_DEBUG(dbgs() << "SLP: Gathering due to partial overlap.\n");
-      newTreeEntry(VL, false, UserTreeIdx);
+      newTreeEntry(VL, false, &S, UserTreeIdx);
       return;
     }
     // Record the reuse of the tree node.  FIXME, currently this is only used to
@@ -2289,7 +2314,7 @@
     if (getTreeEntry(I)) {
       LLVM_DEBUG(dbgs() << "SLP: The instruction (" << *VL[i]
                         << ") is already in tree.\n");
-      newTreeEntry(VL, false, UserTreeIdx);
+      newTreeEntry(VL, false, &S, UserTreeIdx);
       return;
     }
   }
@@ -2300,7 +2325,7 @@
   for (unsigned i = 0, e = VL.size(); i != e; ++i) {
     if (MustGather.count(VL[i]) || is_contained(UserIgnoreList, VL[i])) {
       LLVM_DEBUG(dbgs() << "SLP: Gathering due to gathered scalar.\n");
-      newTreeEntry(VL, false, UserTreeIdx);
+      newTreeEntry(VL, false, &S, UserTreeIdx);
       return;
     }
   }
@@ -2314,7 +2339,7 @@
     // Don't go into unreachable blocks. They may contain instructions with
     // dependency cycles which confuse the final scheduling.
     LLVM_DEBUG(dbgs() << "SLP: bundle in unreachable block.\n");
-    newTreeEntry(VL, false, UserTreeIdx);
+    newTreeEntry(VL, false, &S, UserTreeIdx);
     return;
   }
 
@@ -2334,7 +2359,7 @@
     LLVM_DEBUG(dbgs() << "SLP: Shuffle for reused scalars.\n");
     if (UniqueValues.size() <= 1 || !llvm::isPowerOf2_32(UniqueValues.size())) {
       LLVM_DEBUG(dbgs() << "SLP: Scalar used twice in bundle.\n");
-      newTreeEntry(VL, false, UserTreeIdx);
+      newTreeEntry(VL, false, &S, UserTreeIdx);
       return;
     }
     VL = UniqueValues;
@@ -2351,7 +2376,7 @@
     assert((!BS.getScheduleData(VL0) ||
             !BS.getScheduleData(VL0)->isPartOfBundle()) &&
            "tryScheduleBundle should cancelScheduling on failure");
-    newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+    newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
     return;
   }
   LLVM_DEBUG(dbgs() << "SLP: We are able to schedule this bundle.\n");
@@ -2372,12 +2397,12 @@
             LLVM_DEBUG(dbgs()
                        << "SLP: Need to swizzle PHINodes (terminator use).\n");
             BS.cancelScheduling(VL, VL0);
-            newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+            newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
             return;
           }
         }
 
-      auto *TE = newTreeEntry(VL, true, UserTreeIdx, ReuseShuffleIndicies);
+      auto *TE = newTreeEntry(VL, true, &S, UserTreeIdx, ReuseShuffleIndicies);
       LLVM_DEBUG(dbgs() << "SLP: added a vector of PHINodes.\n");
 
       for (unsigned i = 0, e = PH->getNumIncomingValues(); i < e; ++i) {
@@ -2398,7 +2423,7 @@
       if (Reuse) {
         LLVM_DEBUG(dbgs() << "SLP: Reusing or shuffling extract sequence.\n");
         ++NumOpsWantToKeepOriginalOrder;
-        newTreeEntry(VL, /*Vectorized=*/true, UserTreeIdx,
+        newTreeEntry(VL, /*Vectorized=*/true, &S, UserTreeIdx,
                      ReuseShuffleIndicies);
         // This is a special case, as it does not gather, but at the same time
         // we are not extending buildTree_rec() towards the operands.
@@ -2420,7 +2445,7 @@
         auto StoredCurrentOrderAndNum =
             NumOpsWantToKeepOrder.try_emplace(CurrentOrder).first;
         ++StoredCurrentOrderAndNum->getSecond();
-        newTreeEntry(VL, /*Vectorized=*/true, UserTreeIdx, ReuseShuffleIndicies,
+        newTreeEntry(VL, /*Vectorized=*/true, &S, UserTreeIdx, ReuseShuffleIndicies,
                      StoredCurrentOrderAndNum->getFirst());
         // This is a special case, as it does not gather, but at the same time
         // we are not extending buildTree_rec() towards the operands.
@@ -2430,7 +2455,7 @@
         return;
       }
       LLVM_DEBUG(dbgs() << "SLP: Gather extract sequence.\n");
-      newTreeEntry(VL, /*Vectorized=*/false, UserTreeIdx, ReuseShuffleIndicies);
+      newTreeEntry(VL, /*Vectorized=*/false, &S, UserTreeIdx, ReuseShuffleIndicies);
       BS.cancelScheduling(VL, VL0);
       return;
     }
@@ -2446,7 +2471,7 @@
       if (DL->getTypeSizeInBits(ScalarTy) !=
           DL->getTypeAllocSizeInBits(ScalarTy)) {
         BS.cancelScheduling(VL, VL0);
-        newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+        newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
         LLVM_DEBUG(dbgs() << "SLP: Gathering loads of non-packed type.\n");
         return;
       }
@@ -2459,7 +2484,7 @@
         auto *L = cast<LoadInst>(V);
         if (!L->isSimple()) {
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           LLVM_DEBUG(dbgs() << "SLP: Gathering non-simple loads.\n");
           return;
         }
@@ -2489,14 +2514,14 @@
           if (CurrentOrder.empty()) {
             // Original loads are consecutive and does not require reordering.
             ++NumOpsWantToKeepOriginalOrder;
-            newTreeEntry(VL, /*Vectorized=*/true, UserTreeIdx,
+            newTreeEntry(VL, /*Vectorized=*/true, &S, UserTreeIdx,
                          ReuseShuffleIndicies);
             LLVM_DEBUG(dbgs() << "SLP: added a vector of loads.\n");
           } else {
             // Need to reorder.
             auto I = NumOpsWantToKeepOrder.try_emplace(CurrentOrder).first;
             ++I->getSecond();
-            newTreeEntry(VL, /*Vectorized=*/true, UserTreeIdx,
+            newTreeEntry(VL, /*Vectorized=*/true, &S, UserTreeIdx,
                          ReuseShuffleIndicies, I->getFirst());
             LLVM_DEBUG(dbgs() << "SLP: added a vector of jumbled loads.\n");
           }
@@ -2506,7 +2531,7 @@
 
       LLVM_DEBUG(dbgs() << "SLP: Gathering non-consecutive loads.\n");
       BS.cancelScheduling(VL, VL0);
-      newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+      newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
       return;
     }
     case Instruction::ZExt:
@@ -2526,13 +2551,13 @@
         Type *Ty = cast<Instruction>(VL[i])->getOperand(0)->getType();
         if (Ty != SrcTy || !isValidElementType(Ty)) {
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           LLVM_DEBUG(dbgs()
                      << "SLP: Gathering casts with different src types.\n");
           return;
         }
       }
-      auto *TE = newTreeEntry(VL, true, UserTreeIdx, ReuseShuffleIndicies);
+      auto *TE = newTreeEntry(VL, true, &S, UserTreeIdx, ReuseShuffleIndicies);
       LLVM_DEBUG(dbgs() << "SLP: added a vector of casts.\n");
 
       for (unsigned i = 0, e = VL0->getNumOperands(); i < e; ++i) {
@@ -2556,14 +2581,14 @@
         if ((Cmp->getPredicate() != P0 && Cmp->getPredicate() != SwapP0) ||
             Cmp->getOperand(0)->getType() != ComparedTy) {
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           LLVM_DEBUG(dbgs()
                      << "SLP: Gathering cmp with different predicate.\n");
           return;
         }
       }
 
-      auto *TE = newTreeEntry(VL, true, UserTreeIdx, ReuseShuffleIndicies);
+      auto *TE = newTreeEntry(VL, true, &S, UserTreeIdx, ReuseShuffleIndicies);
       LLVM_DEBUG(dbgs() << "SLP: added a vector of compares.\n");
 
       ValueList Left, Right;
@@ -2609,7 +2634,7 @@
     case Instruction::And:
     case Instruction::Or:
     case Instruction::Xor: {
-      auto *TE = newTreeEntry(VL, true, UserTreeIdx, ReuseShuffleIndicies);
+      auto *TE = newTreeEntry(VL, true, &S, UserTreeIdx, ReuseShuffleIndicies);
       LLVM_DEBUG(dbgs() << "SLP: added a vector of un/bin op.\n");
 
       // Sort operands of the instructions so that each side is more likely to
@@ -2638,7 +2663,7 @@
         if (cast<Instruction>(VL[j])->getNumOperands() != 2) {
           LLVM_DEBUG(dbgs() << "SLP: not-vectorizable GEP (nested indexes).\n");
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           return;
         }
       }
@@ -2652,7 +2677,7 @@
           LLVM_DEBUG(dbgs()
                      << "SLP: not-vectorizable GEP (different types).\n");
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           return;
         }
       }
@@ -2664,12 +2689,12 @@
           LLVM_DEBUG(dbgs()
                      << "SLP: not-vectorizable GEP (non-constant indexes).\n");
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           return;
         }
       }
 
-      auto *TE = newTreeEntry(VL, true, UserTreeIdx, ReuseShuffleIndicies);
+      auto *TE = newTreeEntry(VL, true, &S, UserTreeIdx, ReuseShuffleIndicies);
       LLVM_DEBUG(dbgs() << "SLP: added a vector of GEPs.\n");
       for (unsigned i = 0, e = 2; i < e; ++i) {
         ValueList Operands;
@@ -2686,12 +2711,12 @@
       for (unsigned i = 0, e = VL.size() - 1; i < e; ++i)
         if (!isConsecutiveAccess(VL[i], VL[i + 1], *DL, *SE)) {
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           LLVM_DEBUG(dbgs() << "SLP: Non-consecutive store.\n");
           return;
         }
 
-      auto *TE = newTreeEntry(VL, true, UserTreeIdx, ReuseShuffleIndicies);
+      auto *TE = newTreeEntry(VL, true, &S, UserTreeIdx, ReuseShuffleIndicies);
       LLVM_DEBUG(dbgs() << "SLP: added a vector of stores.\n");
 
       ValueList Operands;
@@ -2709,7 +2734,7 @@
       Intrinsic::ID ID = getVectorIntrinsicIDForCall(CI, TLI);
       if (!isTriviallyVectorizable(ID)) {
         BS.cancelScheduling(VL, VL0);
-        newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+        newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
         LLVM_DEBUG(dbgs() << "SLP: Non-vectorizable call.\n");
         return;
       }
@@ -2725,7 +2750,7 @@
             getVectorIntrinsicIDForCall(CI2, TLI) != ID ||
             !CI->hasIdenticalOperandBundleSchema(*CI2)) {
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           LLVM_DEBUG(dbgs() << "SLP: mismatched calls:" << *CI << "!=" << *VL[i]
                             << "\n");
           return;
@@ -2737,7 +2762,7 @@
             Value *A1J = CI2->getArgOperand(j);
             if (ScalarArgs[j] != A1J) {
               BS.cancelScheduling(VL, VL0);
-              newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+              newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
               LLVM_DEBUG(dbgs() << "SLP: mismatched arguments in call:" << *CI
                                 << " argument " << ScalarArgs[j] << "!=" << A1J
                                 << "\n");
@@ -2751,14 +2776,14 @@
                         CI->op_begin() + CI->getBundleOperandsEndIndex(),
                         CI2->op_begin() + CI2->getBundleOperandsStartIndex())) {
           BS.cancelScheduling(VL, VL0);
-          newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+          newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
           LLVM_DEBUG(dbgs() << "SLP: mismatched bundle operands in calls:"
                             << *CI << "!=" << *VL[i] << '\n');
           return;
         }
       }
 
-      auto *TE = newTreeEntry(VL, true, UserTreeIdx, ReuseShuffleIndicies);
+      auto *TE = newTreeEntry(VL, true, &S, UserTreeIdx, ReuseShuffleIndicies);
       for (unsigned i = 0, e = CI->getNumArgOperands(); i != e; ++i) {
         ValueList Operands;
         // Prepare the operand vector.
@@ -2775,11 +2800,11 @@
       // then do not vectorize this instruction.
       if (!S.isAltShuffle()) {
         BS.cancelScheduling(VL, VL0);
-        newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+        newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
         LLVM_DEBUG(dbgs() << "SLP: ShuffleVector are not vectorized.\n");
         return;
       }
-      auto *TE = newTreeEntry(VL, true, UserTreeIdx, ReuseShuffleIndicies);
+      auto *TE = newTreeEntry(VL, true, &S, UserTreeIdx, ReuseShuffleIndicies);
       LLVM_DEBUG(dbgs() << "SLP: added a ShuffleVector op.\n");
 
       // Reorder operands if reordering would enable vectorization.
@@ -2803,7 +2828,7 @@
     }
     default:
       BS.cancelScheduling(VL, VL0);
-      newTreeEntry(VL, false, UserTreeIdx, ReuseShuffleIndicies);
+      newTreeEntry(VL, false, &S, UserTreeIdx, ReuseShuffleIndicies);
       LLVM_DEBUG(dbgs() << "SLP: Gathering unknown instruction.\n");
       return;
   }
@@ -2904,7 +2929,7 @@
 bool BoUpSLP::areAllUsersVectorized(Instruction *I) const {
   return I->hasOneUse() ||
          std::all_of(I->user_begin(), I->user_end(), [this](User *U) {
-           return ScalarToTreeEntry.count(U) > 0;
+           return ScalarToTreeEntry.count(cast<Instruction>(U)) > 0;
          });
 }
 
@@ -2948,10 +2973,10 @@
           // instruction itself is not going to be vectorized, consider this
           // instruction as dead and remove its cost from the final cost of the
           // vectorized tree.
-          if (areAllUsersVectorized(cast<Instruction>(V)) &&
-              !ScalarToTreeEntry.count(V)) {
+          auto *I = cast<Instruction>(V);
+          if (areAllUsersVectorized(I) && !ScalarToTreeEntry.count(I)) {
             auto *IO = cast<ConstantInt>(
-                cast<ExtractElementInst>(V)->getIndexOperand());
+                cast<ExtractElementInst>(I)->getIndexOperand());
             Cost -= TTI->getVectorInstrCost(Instruction::ExtractElement, VecTy,
                                             IO->getZExtValue());
           }
@@ -3517,16 +3542,16 @@
   Right = Ops.getVL(1);
 }
 
-void BoUpSLP::setInsertPointAfterBundle(ArrayRef<Value *> VL,
-                                        const InstructionsState &S) {
+void BoUpSLP::setInsertPointAfterBundle(TreeEntry *E) {
   // Get the basic block this bundle is in. All instructions in the bundle
   // should be in this block.
-  auto *Front = cast<Instruction>(S.OpValue);
+  auto *Front = E->MainOp;
   auto *BB = Front->getParent();
-  assert(llvm::all_of(make_range(VL.begin(), VL.end()), [=](Value *V) -> bool {
-    auto *I = cast<Instruction>(V);
-    return !S.isOpcodeOrAlt(I) || I->getParent() == BB;
-  }));
+  assert(llvm::all_of(make_range(E->Scalars.begin(), E->Scalars.end()),
+                      [=](Value *V) -> bool {
+                        auto *I = cast<Instruction>(V);
+                        return !E->isOpcodeOrAlt(I) || I->getParent() == BB;
+                      }));
 
   // The last instruction in the bundle in program order.
   Instruction *LastInst = nullptr;
@@ -3537,7 +3562,7 @@
   // bundle. The end of the bundle is marked by null ScheduleData.
   if (BlocksSchedules.count(BB)) {
     auto *Bundle =
-        BlocksSchedules[BB]->getScheduleData(isOneOf(S, VL.back()));
+        BlocksSchedules[BB]->getScheduleData(E->isOneOf(E->Scalars.back()));
     if (Bundle && Bundle->isPartOfBundle())
       for (; Bundle; Bundle = Bundle->NextInBundle)
         if (Bundle->OpValue == Bundle->Inst)
@@ -3563,9 +3588,9 @@
   // we both exit early from buildTree_rec and that the bundle be out-of-order
   // (causing us to iterate all the way to the end of the block).
   if (!LastInst) {
-    SmallPtrSet<Value *, 16> Bundle(VL.begin(), VL.end());
+    SmallPtrSet<Value *, 16> Bundle(E->Scalars.begin(), E->Scalars.end());
     for (auto &I : make_range(BasicBlock::iterator(Front), BB->end())) {
-      if (Bundle.erase(&I) && S.isOpcodeOrAlt(&I))
+      if (Bundle.erase(&I) && E->isOpcodeOrAlt(&I))
         LastInst = &I;
       if (Bundle.empty())
         break;
@@ -3692,8 +3717,7 @@
     return E->VectorizedValue;
   }
 
-  InstructionsState S = getSameOpcode(E->Scalars);
-  Instruction *VL0 = cast<Instruction>(S.OpValue);
+  Instruction *VL0 = E->MainOp;
   Type *ScalarTy = VL0->getType();
   if (StoreInst *SI = dyn_cast<StoreInst>(VL0))
     ScalarTy = SI->getValueOperand()->getType();
@@ -3702,7 +3726,7 @@
   bool NeedToShuffleReuses = !E->ReuseShuffleIndices.empty();
 
   if (E->NeedToGather) {
-    setInsertPointAfterBundle(E->Scalars, S);
+    setInsertPointAfterBundle(E);
     auto *V = Gather(E->Scalars, VecTy);
     if (NeedToShuffleReuses) {
       V = Builder.CreateShuffleVector(V, UndefValue::get(VecTy),
@@ -3716,8 +3740,8 @@
     return V;
   }
 
-  unsigned ShuffleOrOp = S.isAltShuffle() ?
-           (unsigned) Instruction::ShuffleVector : S.getOpcode();
+  unsigned ShuffleOrOp = E->isAltShuffle() ?
+           (unsigned) Instruction::ShuffleVector : E->MainOp->getOpcode();
   switch (ShuffleOrOp) {
     case Instruction::PHI: {
       PHINode *PH = dyn_cast<PHINode>(VL0);
@@ -3775,7 +3799,7 @@
         E->VectorizedValue = V;
         return V;
       }
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
       auto *V = Gather(E->Scalars, VecTy);
       if (NeedToShuffleReuses) {
         V = Builder.CreateShuffleVector(V, UndefValue::get(VecTy),
@@ -3810,7 +3834,7 @@
         E->VectorizedValue = NewV;
         return NewV;
       }
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
       auto *V = Gather(E->Scalars, VecTy);
       if (NeedToShuffleReuses) {
         V = Builder.CreateShuffleVector(V, UndefValue::get(VecTy),
@@ -3835,7 +3859,7 @@
     case Instruction::Trunc:
     case Instruction::FPTrunc:
     case Instruction::BitCast: {
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
 
       Value *InVec = vectorizeTree(E->getOperand(0));
 
@@ -3856,7 +3880,7 @@
     }
     case Instruction::FCmp:
     case Instruction::ICmp: {
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
 
       Value *L = vectorizeTree(E->getOperand(0));
       Value *R = vectorizeTree(E->getOperand(1));
@@ -3868,7 +3892,7 @@
 
       CmpInst::Predicate P0 = cast<CmpInst>(VL0)->getPredicate();
       Value *V;
-      if (S.getOpcode() == Instruction::FCmp)
+      if (E->MainOp->getOpcode() == Instruction::FCmp)
         V = Builder.CreateFCmp(P0, L, R);
       else
         V = Builder.CreateICmp(P0, L, R);
@@ -3883,7 +3907,7 @@
       return V;
     }
     case Instruction::Select: {
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
 
       Value *Cond = vectorizeTree(E->getOperand(0));
       Value *True = vectorizeTree(E->getOperand(1));
@@ -3904,7 +3928,7 @@
       return V;
     }
     case Instruction::FNeg: {
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
 
       Value *Op = vectorizeTree(E->getOperand(0));
 
@@ -3914,7 +3938,7 @@
       }
 
       Value *V = Builder.CreateUnOp(
-          static_cast<Instruction::UnaryOps>(S.getOpcode()), Op);
+          static_cast<Instruction::UnaryOps>(E->MainOp->getOpcode()), Op);
       propagateIRFlags(V, E->Scalars, VL0);
       if (auto *I = dyn_cast<Instruction>(V))
         V = propagateMetadata(I, E->Scalars);
@@ -3946,7 +3970,7 @@
     case Instruction::And:
     case Instruction::Or:
     case Instruction::Xor: {
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
 
       Value *LHS = vectorizeTree(E->getOperand(0));
       Value *RHS = vectorizeTree(E->getOperand(1));
@@ -3957,7 +3981,8 @@
       }
 
       Value *V = Builder.CreateBinOp(
-          static_cast<Instruction::BinaryOps>(S.getOpcode()), LHS, RHS);
+          static_cast<Instruction::BinaryOps>(E->MainOp->getOpcode()), LHS,
+          RHS);
       propagateIRFlags(V, E->Scalars, VL0);
       if (auto *I = dyn_cast<Instruction>(V))
         V = propagateMetadata(I, E->Scalars);
@@ -3976,10 +4001,13 @@
       // sink them all the way down past store instructions.
       bool IsReorder = !E->ReorderIndices.empty();
       if (IsReorder) {
-        S = getSameOpcode(E->Scalars, E->ReorderIndices.front());
-        VL0 = cast<Instruction>(S.OpValue);
+        InstructionsState S =
+            getSameOpcode(E->Scalars, E->ReorderIndices.front());
+        E->MainOp = S.MainOp;
+        E->AltOp = S.AltOp;
+        VL0 = S.MainOp;
       }
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
 
       LoadInst *LI = cast<LoadInst>(VL0);
       Type *ScalarLoadTy = LI->getType();
@@ -4022,7 +4050,7 @@
       unsigned Alignment = SI->getAlignment();
       unsigned AS = SI->getPointerAddressSpace();
 
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
 
       Value *VecValue = vectorizeTree(E->getOperand(0));
       Value *ScalarPtr = SI->getPointerOperand();
@@ -4049,7 +4077,7 @@
       return V;
     }
     case Instruction::GetElementPtr: {
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
 
       Value *Op0 = vectorizeTree(E->getOperand(0));
 
@@ -4076,7 +4104,7 @@
     }
     case Instruction::Call: {
       CallInst *CI = cast<CallInst>(VL0);
-      setInsertPointAfterBundle(E->Scalars, S);
+      setInsertPointAfterBundle(E);
       Function *FI;
       Intrinsic::ID IID  = Intrinsic::not_intrinsic;
       Value *ScalarArg = nullptr;
@@ -4124,20 +4152,20 @@
       return V;
     }
     case Instruction::ShuffleVector: {
-      assert(S.isAltShuffle() &&
-             ((Instruction::isBinaryOp(S.getOpcode()) &&
-               Instruction::isBinaryOp(S.getAltOpcode())) ||
-              (Instruction::isCast(S.getOpcode()) &&
-               Instruction::isCast(S.getAltOpcode()))) &&
+      assert(E->isAltShuffle() &&
+             ((Instruction::isBinaryOp(E->MainOp->getOpcode()) &&
+               Instruction::isBinaryOp(E->AltOp->getOpcode())) ||
+              (Instruction::isCast(E->MainOp->getOpcode()) &&
+               Instruction::isCast(E->AltOp->getOpcode()))) &&
              "Invalid Shuffle Vector Operand");
 
       Value *LHS, *RHS;
-      if (Instruction::isBinaryOp(S.getOpcode())) {
-        setInsertPointAfterBundle(E->Scalars, S);
+      if (Instruction::isBinaryOp(E->MainOp->getOpcode())) {
+        setInsertPointAfterBundle(E);
         LHS = vectorizeTree(E->getOperand(0));
         RHS = vectorizeTree(E->getOperand(1));
       } else {
-        setInsertPointAfterBundle(E->Scalars, S);
+        setInsertPointAfterBundle(E);
         LHS = vectorizeTree(E->getOperand(0));
       }
 
@@ -4147,16 +4175,20 @@
       }
 
       Value *V0, *V1;
-      if (Instruction::isBinaryOp(S.getOpcode())) {
+      if (Instruction::isBinaryOp(E->MainOp->getOpcode())) {
         V0 = Builder.CreateBinOp(
-          static_cast<Instruction::BinaryOps>(S.getOpcode()), LHS, RHS);
+            static_cast<Instruction::BinaryOps>(E->MainOp->getOpcode()), LHS,
+            RHS);
         V1 = Builder.CreateBinOp(
-          static_cast<Instruction::BinaryOps>(S.getAltOpcode()), LHS, RHS);
+            static_cast<Instruction::BinaryOps>(E->AltOp->getOpcode()), LHS,
+            RHS);
       } else {
         V0 = Builder.CreateCast(
-            static_cast<Instruction::CastOps>(S.getOpcode()), LHS, VecTy);
+            static_cast<Instruction::CastOps>(E->MainOp->getOpcode()), LHS,
+            VecTy);
         V1 = Builder.CreateCast(
-            static_cast<Instruction::CastOps>(S.getAltOpcode()), LHS, VecTy);
+            static_cast<Instruction::CastOps>(E->AltOp->getOpcode()), LHS,
+            VecTy);
       }
 
       // Create shuffle to take alternate operations from the vector.
@@ -4167,8 +4199,8 @@
       SmallVector<Constant *, 8> Mask(e);
       for (unsigned i = 0; i < e; ++i) {
         auto *OpInst = cast<Instruction>(E->Scalars[i]);
-        assert(S.isOpcodeOrAlt(OpInst) && "Unexpected main/alternate opcode");
-        if (OpInst->getOpcode() == S.getAltOpcode()) {
+        assert(E->isOpcodeOrAlt(OpInst) && "Unexpected main/alternate opcode");
+        if (OpInst->getOpcode() == E->AltOp->getOpcode()) {
           Mask[i] = Builder.getInt32(e + i);
           AltScalars.push_back(E->Scalars[i]);
         } else {