[PredicateInfo] Use SmallVector instead of SmallPtrSet.

We do not need the SmallPtrSet to avoid adding duplicates to
OpsToRename, because we already keep a ValueInfo mapping. If we see an
op for the first time, Infos will be empty and we can also add it to
OpsToRename.

We process operands by visiting BBs depth-first and then iterate over
all instructions & users, so the order should be deterministic.
Therefore we can skip one round of sorting, which we purely needed for
guaranteeing a deterministic order when iterating over the SmallPtrSet.

Reviewers: efriedma, davide

Reviewed By: efriedma

Differential Revision: https://reviews.llvm.org/D64816

llvm-svn: 367028

diff --git a/llvm/include/llvm/Transforms/Utils/PredicateInfo.h b/llvm/include/llvm/Transforms/Utils/PredicateInfo.h
index da4a5dc..7c7a8eb 100644 (file)
--- a/llvm/include/llvm/Transforms/Utils/PredicateInfo.h
+++ b/llvm/include/llvm/Transforms/Utils/PredicateInfo.h
@@ -229,10 +229,10 @@ protected:
 
 private:
   void buildPredicateInfo();
-  void processAssume(IntrinsicInst *, BasicBlock *, SmallPtrSetImpl<Value *> &);
-  void processBranch(BranchInst *, BasicBlock *, SmallPtrSetImpl<Value *> &);
-  void processSwitch(SwitchInst *, BasicBlock *, SmallPtrSetImpl<Value *> &);
-  void renameUses(SmallPtrSetImpl<Value *> &);
+  void processAssume(IntrinsicInst *, BasicBlock *, SmallVectorImpl<Value *> &);
+  void processBranch(BranchInst *, BasicBlock *, SmallVectorImpl<Value *> &);
+  void processSwitch(SwitchInst *, BasicBlock *, SmallVectorImpl<Value *> &);
+  void renameUses(SmallVectorImpl<Value *> &);
   using ValueDFS = PredicateInfoClasses::ValueDFS;
   typedef SmallVectorImpl<ValueDFS> ValueDFSStack;
   void convertUsesToDFSOrdered(Value *, SmallVectorImpl<ValueDFS> &);
@@ -240,7 +240,7 @@ private:
   bool stackIsInScope(const ValueDFSStack &, const ValueDFS &) const;
   void popStackUntilDFSScope(ValueDFSStack &, const ValueDFS &);
   ValueInfo &getOrCreateValueInfo(Value *);
-  void addInfoFor(SmallPtrSetImpl<Value *> &OpsToRename, Value *Op,
+  void addInfoFor(SmallVectorImpl<Value *> &OpsToRename, Value *Op,
                   PredicateBase *PB);
   const ValueInfo &getValueInfo(Value *) const;
   Function &F;

diff --git a/llvm/lib/Transforms/Utils/PredicateInfo.cpp b/llvm/lib/Transforms/Utils/PredicateInfo.cpp
index bdf24d8..79cd42d 100644 (file)
--- a/llvm/lib/Transforms/Utils/PredicateInfo.cpp
+++ b/llvm/lib/Transforms/Utils/PredicateInfo.cpp
@@ -306,10 +306,11 @@ void collectCmpOps(CmpInst *Comparison, SmallVectorImpl<Value *> &CmpOperands) {
 }
 
 // Add Op, PB to the list of value infos for Op, and mark Op to be renamed.
-void PredicateInfo::addInfoFor(SmallPtrSetImpl<Value *> &OpsToRename, Value *Op,
+void PredicateInfo::addInfoFor(SmallVectorImpl<Value *> &OpsToRename, Value *Op,
                                PredicateBase *PB) {
-  OpsToRename.insert(Op);
   auto &OperandInfo = getOrCreateValueInfo(Op);
+  if (OperandInfo.Infos.empty())
+    OpsToRename.push_back(Op);
   AllInfos.push_back(PB);
   OperandInfo.Infos.push_back(PB);
 }
@@ -317,7 +318,7 @@ void PredicateInfo::addInfoFor(SmallPtrSetImpl<Value *> &OpsToRename, Value *Op,
 // Process an assume instruction and place relevant operations we want to rename
 // into OpsToRename.
 void PredicateInfo::processAssume(IntrinsicInst *II, BasicBlock *AssumeBB,
-                                  SmallPtrSetImpl<Value *> &OpsToRename) {
+                                  SmallVectorImpl<Value *> &OpsToRename) {
   // See if we have a comparison we support
   SmallVector<Value *, 8> CmpOperands;
   SmallVector<Value *, 2> ConditionsToProcess;
@@ -357,7 +358,7 @@ void PredicateInfo::processAssume(IntrinsicInst *II, BasicBlock *AssumeBB,
 // Process a block terminating branch, and place relevant operations to be
 // renamed into OpsToRename.
 void PredicateInfo::processBranch(BranchInst *BI, BasicBlock *BranchBB,
-                                  SmallPtrSetImpl<Value *> &OpsToRename) {
+                                  SmallVectorImpl<Value *> &OpsToRename) {
   BasicBlock *FirstBB = BI->getSuccessor(0);
   BasicBlock *SecondBB = BI->getSuccessor(1);
   SmallVector<BasicBlock *, 2> SuccsToProcess;
@@ -427,7 +428,7 @@ void PredicateInfo::processBranch(BranchInst *BI, BasicBlock *BranchBB,
 // Process a block terminating switch, and place relevant operations to be
 // renamed into OpsToRename.
 void PredicateInfo::processSwitch(SwitchInst *SI, BasicBlock *BranchBB,
-                                  SmallPtrSetImpl<Value *> &OpsToRename) {
+                                  SmallVectorImpl<Value *> &OpsToRename) {
   Value *Op = SI->getCondition();
   if ((!isa<Instruction>(Op) && !isa<Argument>(Op)) || Op->hasOneUse())
     return;
@@ -457,7 +458,7 @@ void PredicateInfo::buildPredicateInfo() {
   DT.updateDFSNumbers();
   // Collect operands to rename from all conditional branch terminators, as well
   // as assume statements.
-  SmallPtrSet<Value *, 8> OpsToRename;
+  SmallVector<Value *, 8> OpsToRename;
   for (auto DTN : depth_first(DT.getRootNode())) {
     BasicBlock *BranchBB = DTN->getBlock();
     if (auto *BI = dyn_cast<BranchInst>(BranchBB->getTerminator())) {
@@ -565,13 +566,7 @@ Value *PredicateInfo::materializeStack(unsigned int &Counter,
 //
 // TODO: Use this algorithm to perform fast single-variable renaming in
 // promotememtoreg and memoryssa.
-void PredicateInfo::renameUses(SmallPtrSetImpl<Value *> &OpSet) {
-  // Sort OpsToRename since we are going to iterate it.
-  SmallVector<Value *, 8> OpsToRename(OpSet.begin(), OpSet.end());
-  auto Comparator = [&](const Value *A, const Value *B) {
-    return valueComesBefore(OI, A, B);
-  };
-  llvm::sort(OpsToRename, Comparator);
+void PredicateInfo::renameUses(SmallVectorImpl<Value *> &OpsToRename) {
   ValueDFS_Compare Compare(OI);
   // Compute liveness, and rename in O(uses) per Op.
   for (auto *Op : OpsToRename) {

diff --git a/llvm/test/Transforms/Util/PredicateInfo/condprop.ll b/llvm/test/Transforms/Util/PredicateInfo/condprop.ll
index 0cd6ef3..daf6bb8 100644 (file)
--- a/llvm/test/Transforms/Util/PredicateInfo/condprop.ll
+++ b/llvm/test/Transforms/Util/PredicateInfo/condprop.ll
@@ -98,10 +98,10 @@ define void @test3(i32 %x, i32 %y) {
 ; CHECK-NEXT:    [[XZ:%.*]] = icmp eq i32 [[X:%.*]], 0
 ; CHECK-NEXT:    [[YZ:%.*]] = icmp eq i32 [[Y:%.*]], 0
 ; CHECK-NEXT:    [[Z:%.*]] = and i1 [[XZ]], [[YZ]]
-; CHECK:         [[X_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
-; CHECK:         [[Y_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[Y]])
 ; CHECK:         [[XZ_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[XZ]])
+; CHECK:         [[X_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
 ; CHECK:         [[YZ_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[YZ]])
+; CHECK:         [[Y_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[Y]])
 ; CHECK:         [[Z_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[Z]])
 ; CHECK-NEXT:    br i1 [[Z]], label [[BOTH_ZERO:%.*]], label [[NOPE:%.*]]
 ; CHECK:       both_zero:
@@ -382,8 +382,8 @@ ret:
 define i32 @test10(i32 %j, i32 %i) {
 ; CHECK-LABEL: @test10(
 ; CHECK-NEXT:    [[CMP:%.*]] = icmp eq i32 [[I:%.*]], [[J:%.*]]
-; CHECK:         [[J_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[J]])
 ; CHECK:         [[I_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[I]])
+; CHECK:         [[J_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[J]])
 ; CHECK-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[RET:%.*]]
 ; CHECK:       cond_true:
 ; CHECK-NEXT:    [[DIFF:%.*]] = sub i32 [[I_0]], [[J_0]]

diff --git a/llvm/test/Transforms/Util/PredicateInfo/testandor.ll b/llvm/test/Transforms/Util/PredicateInfo/testandor.ll
index 86d1a0d..c84562c 100644 (file)
--- a/llvm/test/Transforms/Util/PredicateInfo/testandor.ll
+++ b/llvm/test/Transforms/Util/PredicateInfo/testandor.ll
@@ -10,10 +10,10 @@ define void @testor(i32 %x, i32 %y) {
 ; CHECK-NEXT:    [[XZ:%.*]] = icmp eq i32 [[X:%.*]], 0
 ; CHECK-NEXT:    [[YZ:%.*]] = icmp eq i32 [[Y:%.*]], 0
 ; CHECK-NEXT:    [[Z:%.*]] = or i1 [[XZ]], [[YZ]]
-; CHECK:         [[X_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
-; CHECK:         [[Y_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[Y]])
 ; CHECK:         [[XZ_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[XZ]])
+; CHECK:         [[X_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
 ; CHECK:         [[YZ_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[YZ]])
+; CHECK:         [[Y_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[Y]])
 ; CHECK:         [[Z_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[Z]])
 ; CHECK-NEXT:    br i1 [[Z]], label [[ONEOF:%.*]], label [[NEITHER:%.*]]
 ; CHECK:       oneof:
@@ -54,10 +54,10 @@ define void @testand(i32 %x, i32 %y) {
 ; CHECK-NEXT:    [[XZ:%.*]] = icmp eq i32 [[X:%.*]], 0
 ; CHECK-NEXT:    [[YZ:%.*]] = icmp eq i32 [[Y:%.*]], 0
 ; CHECK-NEXT:    [[Z:%.*]] = and i1 [[XZ]], [[YZ]]
-; CHECK:         [[X_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
-; CHECK:         [[Y_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[Y]])
 ; CHECK:         [[XZ_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[XZ]])
+; CHECK:         [[X_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
 ; CHECK:         [[YZ_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[YZ]])
+; CHECK:         [[Y_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[Y]])
 ; CHECK:         [[Z_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[Z]])
 ; CHECK-NEXT:    br i1 [[Z]], label [[BOTH:%.*]], label [[NOPE:%.*]]
 ; CHECK:       both:
@@ -98,9 +98,9 @@ define void @testandsame(i32 %x, i32 %y) {
 ; CHECK-NEXT:    [[XGT:%.*]] = icmp sgt i32 [[X:%.*]], 0
 ; CHECK-NEXT:    [[XLT:%.*]] = icmp slt i32 [[X]], 100
 ; CHECK-NEXT:    [[Z:%.*]] = and i1 [[XGT]], [[XLT]]
+; CHECK:         [[XGT_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[XGT]])
 ; CHECK:         [[X_0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
 ; CHECK:         [[X_0_1:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X_0]])
-; CHECK:         [[XGT_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[XGT]])
 ; CHECK:         [[XLT_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[XLT]])
 ; CHECK:         [[Z_0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[Z]])
 ; CHECK-NEXT:    br i1 [[Z]], label [[BOTH:%.*]], label [[NOPE:%.*]]
@@ -136,23 +136,23 @@ define void @testandassume(i32 %x, i32 %y) {
 ; CHECK-NEXT:    [[XZ:%.*]] = icmp eq i32 [[X:%.*]], 0
 ; CHECK-NEXT:    [[YZ:%.*]] = icmp eq i32 [[Y:%.*]], 0
 ; CHECK-NEXT:    [[Z:%.*]] = and i1 [[XZ]], [[YZ]]
-; CHECK:         [[TMP1:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
-; CHECK:         [[TMP2:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[Y]])
-; CHECK:         [[TMP3:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[XZ]])
-; CHECK:         [[TMP4:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[YZ]])
+; CHECK:         [[TMP1:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[XZ]])
+; CHECK:         [[TMP2:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[X]])
+; CHECK:         [[TMP3:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[YZ]])
+; CHECK:         [[TMP4:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[Y]])
 ; CHECK:         [[TMP5:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[Z]])
 ; CHECK-NEXT:    call void @llvm.assume(i1 [[TMP5]])
-; CHECK:         [[DOT0:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[TMP1]])
+; CHECK:         [[DOT0:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[TMP1]])
 ; CHECK:         [[DOT01:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[TMP2]])
 ; CHECK:         [[DOT02:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[TMP3]])
-; CHECK:         [[DOT03:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[TMP4]])
+; CHECK:         [[DOT03:%.*]] = call i32 @llvm.ssa.copy.{{.+}}(i32 [[TMP4]])
 ; CHECK:         [[DOT04:%.*]] = call i1 @llvm.ssa.copy.{{.+}}(i1 [[TMP5]])
 ; CHECK-NEXT:    br i1 [[TMP5]], label [[BOTH:%.*]], label [[NOPE:%.*]]
 ; CHECK:       both:
+; CHECK-NEXT:    call void @foo(i1 [[DOT0]])
 ; CHECK-NEXT:    call void @foo(i1 [[DOT02]])
-; CHECK-NEXT:    call void @foo(i1 [[DOT03]])
-; CHECK-NEXT:    call void @bar(i32 [[DOT0]])
 ; CHECK-NEXT:    call void @bar(i32 [[DOT01]])
+; CHECK-NEXT:    call void @bar(i32 [[DOT03]])
 ; CHECK-NEXT:    ret void
 ; CHECK:       nope:
 ; CHECK-NEXT:    call void @foo(i1 [[DOT04]])