vendor/llvm/llvm-trunk-r304222

2 files changed, 26 insertions, 50 deletions

diff --git a/lib/Transforms/Scalar/NewGVN.cpp b/lib/Transforms/Scalar/NewGVN.cpp
index 67abc31169880..5e9f40019ce8c 100644
--- a/lib/Transforms/Scalar/NewGVN.cpp
+++ b/lib/Transforms/Scalar/NewGVN.cpp
@@ -377,7 +377,6 @@ private:
   int StoreCount = 0;
 };
 
-struct HashedExpression;
 namespace llvm {
 template <> struct DenseMapInfo<const Expression *> {
   static const Expression *getEmptyKey() {
@@ -391,41 +390,25 @@ template <> struct DenseMapInfo<const Expression *> {
     return reinterpret_cast<const Expression *>(Val);
   }
   static unsigned getHashValue(const Expression *E) {
-    return static_cast<unsigned>(E->getHashValue());
+    return static_cast<unsigned>(E->getComputedHash());
   }
-  static unsigned getHashValue(const HashedExpression &HE);
-  static bool isEqual(const HashedExpression &LHS, const Expression *RHS);
   static bool isEqual(const Expression *LHS, const Expression *RHS) {
     if (LHS == RHS)
       return true;
     if (LHS == getTombstoneKey() || RHS == getTombstoneKey() ||
         LHS == getEmptyKey() || RHS == getEmptyKey())
       return false;
+    // Compare hashes before equality.  This is *not* what the hashtable does,
+    // since it is computing it modulo the number of buckets, whereas we are
+    // using the full hash keyspace.  Since the hashes are precomputed, this
+    // check is *much* faster than equality.
+    if (LHS->getComputedHash() != RHS->getComputedHash())
+      return false;
     return *LHS == *RHS;
   }
 };
 } // end namespace llvm
 
-// This is just a wrapper around Expression that computes the hash value once at
-// creation time.  Hash values for an Expression can't change once they are
-// inserted into the DenseMap (it breaks DenseMap), so they must be immutable at
-// that point anyway.
-struct HashedExpression {
-  const Expression *E;
-  unsigned HashVal;
-  HashedExpression(const Expression *E)
-      : E(E), HashVal(DenseMapInfo<const Expression *>::getHashValue(E)) {}
-};
-
-unsigned
-DenseMapInfo<const Expression *>::getHashValue(const HashedExpression &HE) {
-  return HE.HashVal;
-}
-bool DenseMapInfo<const Expression *>::isEqual(const HashedExpression &LHS,
-                                               const Expression *RHS) {
-  return isEqual(LHS.E, RHS);
-}
-
 namespace {
 class NewGVN {
   Function &F;
@@ -707,7 +690,7 @@ private:
   void markPredicateUsersTouched(Instruction *);
   void markValueLeaderChangeTouched(CongruenceClass *CC);
   void markMemoryLeaderChangeTouched(CongruenceClass *CC);
-  void markPhiOfOpsChanged(const HashedExpression &HE);
+  void markPhiOfOpsChanged(const Expression *E);
   void addPredicateUsers(const PredicateBase *, Instruction *) const;
   void addMemoryUsers(const MemoryAccess *To, MemoryAccess *U) const;
   void addAdditionalUsers(Value *To, Value *User) const;
@@ -956,8 +939,12 @@ const Expression *NewGVN::checkSimplificationResults(Expression *E,
   if (CC && CC->getDefiningExpr()) {
     // If we simplified to something else, we need to communicate
     // that we're users of the value we simplified to.
-    if (I != V)
-      addAdditionalUsers(V, I);
+    if (I != V) {
+      // Don't add temporary instructions to the user lists.
+      if (!AllTempInstructions.count(I))
+        addAdditionalUsers(V, I);
+    }
+
     if (I)
       DEBUG(dbgs() << "Simplified " << *I << " to "
                    << " expression " << *CC->getDefiningExpr() << "\n");
@@ -2195,8 +2182,8 @@ void NewGVN::moveValueToNewCongruenceClass(Instruction *I, const Expression *E,
 
 // For a given expression, mark the phi of ops instructions that could have
 // changed as a result.
-void NewGVN::markPhiOfOpsChanged(const HashedExpression &HE) {
-  touchAndErase(ExpressionToPhiOfOps, HE);
+void NewGVN::markPhiOfOpsChanged(const Expression *E) {
+  touchAndErase(ExpressionToPhiOfOps, E);
 }
 
 // Perform congruence finding on a given value numbering expression.
@@ -2210,14 +2197,13 @@ void NewGVN::performCongruenceFinding(Instruction *I, const Expression *E) {
   assert(!IClass->isDead() && "Found a dead class");
 
   CongruenceClass *EClass = nullptr;
-  HashedExpression HE(E);
   if (const auto *VE = dyn_cast<VariableExpression>(E)) {
     EClass = ValueToClass.lookup(VE->getVariableValue());
   } else if (isa<DeadExpression>(E)) {
     EClass = TOPClass;
   }
   if (!EClass) {
-    auto lookupResult = ExpressionToClass.insert_as({E, nullptr}, HE);
+    auto lookupResult = ExpressionToClass.insert({E, nullptr});
 
     // If it's not in the value table, create a new congruence class.
     if (lookupResult.second) {
@@ -2268,7 +2254,7 @@ void NewGVN::performCongruenceFinding(Instruction *I, const Expression *E) {
                  << "\n");
     if (ClassChanged) {
       moveValueToNewCongruenceClass(I, E, IClass, EClass);
-      markPhiOfOpsChanged(HE);
+      markPhiOfOpsChanged(E);
     }
 
     markUsersTouched(I);
@@ -2502,9 +2488,8 @@ NewGVN::makePossiblePhiOfOps(Instruction *I, bool HasBackedge,
         // Clone the instruction, create an expression from it, and see if we
         // have a leader.
         Instruction *ValueOp = I->clone();
-        auto Iter = TempToMemory.end();
         if (MemAccess)
-          Iter = TempToMemory.insert({ValueOp, MemAccess}).first;
+          TempToMemory.insert({ValueOp, MemAccess});
 
         for (auto &Op : ValueOp->operands()) {
           Op = Op->DoPHITranslation(PHIBlock, PredBB);
@@ -2523,7 +2508,7 @@ NewGVN::makePossiblePhiOfOps(Instruction *I, bool HasBackedge,
         AllTempInstructions.erase(ValueOp);
         ValueOp->deleteValue();
         if (MemAccess)
-          TempToMemory.erase(Iter);
+          TempToMemory.erase(ValueOp);
         if (!E)
           return nullptr;
         FoundVal = findPhiOfOpsLeader(E, PredBB);
diff --git a/lib/Transforms/Vectorize/LoopVectorize.cpp b/lib/Transforms/Vectorize/LoopVectorize.cpp
index 3b036a6ac430e..2b83b8426d147 100644
--- a/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -7173,7 +7173,7 @@ LoopVectorizationCostModel::getInstructionCost(Instruction *I, unsigned VF) {
   // Note: Even if all instructions are scalarized, return true if any memory
   // accesses appear in the loop to get benefits from address folding etc.
   bool TypeNotScalarized =
-      VF > 1 && VectorTy->isVectorTy() && TTI.getNumberOfParts(VectorTy) < VF;
+      VF > 1 && !VectorTy->isVoidTy() && TTI.getNumberOfParts(VectorTy) < VF;
   return VectorizationCostTy(C, TypeNotScalarized);
 }
 
@@ -7312,7 +7312,7 @@ unsigned LoopVectorizationCostModel::getInstructionCost(Instruction *I,
   Type *RetTy = I->getType();
   if (canTruncateToMinimalBitwidth(I, VF))
     RetTy = IntegerType::get(RetTy->getContext(), MinBWs[I]);
-  VectorTy = isScalarAfterVectorization(I, VF) ? RetTy : ToVectorTy(RetTy, VF);
+  VectorTy = ToVectorTy(RetTy, VF);
   auto SE = PSE.getSE();
 
   // TODO: We need to estimate the cost of intrinsic calls.
@@ -7445,10 +7445,9 @@ unsigned LoopVectorizationCostModel::getInstructionCost(Instruction *I,
     } else if (Legal->isUniform(Op2)) {
       Op2VK = TargetTransformInfo::OK_UniformValue;
     }
-    SmallVector<const Value *, 4> Operands(I->operand_values());
-    unsigned N = isScalarAfterVectorization(I, VF) ? VF : 1;
-    return N * TTI.getArithmeticInstrCost(I->getOpcode(), VectorTy, Op1VK,
-                                          Op2VK, Op1VP, Op2VP, Operands);
+    SmallVector<const Value *, 4> Operands(I->operand_values()); 
+    return TTI.getArithmeticInstrCost(I->getOpcode(), VectorTy, Op1VK,
+                                      Op2VK, Op1VP, Op2VP, Operands);
   }
   case Instruction::Select: {
     SelectInst *SI = cast<SelectInst>(I);
@@ -7471,15 +7470,7 @@ unsigned LoopVectorizationCostModel::getInstructionCost(Instruction *I,
   }
   case Instruction::Store:
   case Instruction::Load: {
-    unsigned Width = VF;
-    if (Width > 1) {
-      InstWidening Decision = getWideningDecision(I, Width);
-      assert(Decision != CM_Unknown &&
-             "CM decision should be taken at this point");
-      if (Decision == CM_Scalarize)
-        Width = 1;
-    }
-    VectorTy = ToVectorTy(getMemInstValueType(I), Width);
+    VectorTy = ToVectorTy(getMemInstValueType(I), VF);
     return getMemoryInstructionCost(I, VF);
   }
   case Instruction::ZExt: