//===- Reassociate.cpp - Reassociate binary expressions -------------------===//
//
+// The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
// This pass reassociates commutative expressions in an order that is designed
-// to promote better constant propogation, GCSE, LICM, PRE...
+// to promote better constant propagation, GCSE, LICM, PRE...
//
// For example: 4 + (x + 5) -> x + (4 + 5)
//
-// Note that this pass works best if left shifts have been promoted to explicit
-// multiplies before this pass executes.
-//
// In the implementation of this algorithm, constants are assigned rank = 0,
// function arguments are rank = 1, and other values are assigned ranks
// corresponding to the reverse post order traversal of current function
//
//===----------------------------------------------------------------------===//
+#define DEBUG_TYPE "reassociate"
#include "llvm/Transforms/Scalar.h"
+#include "llvm/Constants.h"
+#include "llvm/DerivedTypes.h"
#include "llvm/Function.h"
-#include "llvm/BasicBlock.h"
-#include "llvm/iOperators.h"
-#include "llvm/Type.h"
+#include "llvm/Instructions.h"
+#include "llvm/IntrinsicInst.h"
+#include "llvm/LLVMContext.h"
#include "llvm/Pass.h"
-#include "llvm/Constant.h"
+#include "llvm/Assembly/Writer.h"
#include "llvm/Support/CFG.h"
-#include "Support/PostOrderIterator.h"
-#include "Support/StatisticReporter.h"
+#include "llvm/Support/Compiler.h"
+#include "llvm/Support/Debug.h"
+#include "llvm/Support/ValueHandle.h"
+#include "llvm/ADT/PostOrderIterator.h"
+#include "llvm/ADT/Statistic.h"
+#include <algorithm>
+#include <map>
+using namespace llvm;
+
+STATISTIC(NumLinear , "Number of insts linearized");
+STATISTIC(NumChanged, "Number of insts reassociated");
+STATISTIC(NumAnnihil, "Number of expr tree annihilated");
+STATISTIC(NumFactor , "Number of multiplies factored");
-static Statistic<> NumLinear ("reassociate\t- Number of insts linearized");
-static Statistic<> NumChanged("reassociate\t- Number of insts reassociated");
-static Statistic<> NumSwapped("reassociate\t- Number of insts with operands swapped");
+namespace {
+ struct VISIBILITY_HIDDEN ValueEntry {
+ unsigned Rank;
+ Value *Op;
+ ValueEntry(unsigned R, Value *O) : Rank(R), Op(O) {}
+ };
+ inline bool operator<(const ValueEntry &LHS, const ValueEntry &RHS) {
+ return LHS.Rank > RHS.Rank; // Sort so that highest rank goes to start.
+ }
+}
+#ifndef NDEBUG
+/// PrintOps - Print out the expression identified in the Ops list.
+///
+static void PrintOps(Instruction *I, const std::vector<ValueEntry> &Ops) {
+ Module *M = I->getParent()->getParent()->getParent();
+ cerr << Instruction::getOpcodeName(I->getOpcode()) << " "
+ << *Ops[0].Op->getType();
+ for (unsigned i = 0, e = Ops.size(); i != e; ++i) {
+ WriteAsOperand(*cerr.stream() << " ", Ops[i].Op, false, M);
+ cerr << "," << Ops[i].Rank;
+ }
+}
+#endif
+
namespace {
- class Reassociate : public FunctionPass {
- map<BasicBlock*, unsigned> RankMap;
+ class VISIBILITY_HIDDEN Reassociate : public FunctionPass {
+ std::map<BasicBlock*, unsigned> RankMap;
+ std::map<AssertingVH<>, unsigned> ValueRankMap;
+ bool MadeChange;
public:
- const char *getPassName() const {
- return "Expression Reassociation";
- }
+ static char ID; // Pass identification, replacement for typeid
+ Reassociate() : FunctionPass(&ID) {}
bool runOnFunction(Function &F);
virtual void getAnalysisUsage(AnalysisUsage &AU) const {
- AU.preservesCFG();
+ AU.setPreservesCFG();
}
private:
void BuildRankMap(Function &F);
unsigned getRank(Value *V);
- bool ReassociateExpr(BinaryOperator *I);
- bool ReassociateBB(BasicBlock *BB);
+ void ReassociateExpression(BinaryOperator *I);
+ void RewriteExprTree(BinaryOperator *I, std::vector<ValueEntry> &Ops,
+ unsigned Idx = 0);
+ Value *OptimizeExpression(BinaryOperator *I, std::vector<ValueEntry> &Ops);
+ void LinearizeExprTree(BinaryOperator *I, std::vector<ValueEntry> &Ops);
+ void LinearizeExpr(BinaryOperator *I);
+ Value *RemoveFactorFromExpression(Value *V, Value *Factor);
+ void ReassociateBB(BasicBlock *BB);
+
+ void RemoveDeadBinaryOp(Value *V);
};
}
-Pass *createReassociatePass() { return new Reassociate(); }
+char Reassociate::ID = 0;
+static RegisterPass<Reassociate> X("reassociate", "Reassociate expressions");
+
+// Public interface to the Reassociate pass
+FunctionPass *llvm::createReassociatePass() { return new Reassociate(); }
+
+void Reassociate::RemoveDeadBinaryOp(Value *V) {
+ Instruction *Op = dyn_cast<Instruction>(V);
+ if (!Op || !isa<BinaryOperator>(Op) || !isa<CmpInst>(Op) || !Op->use_empty())
+ return;
+
+ Value *LHS = Op->getOperand(0), *RHS = Op->getOperand(1);
+ RemoveDeadBinaryOp(LHS);
+ RemoveDeadBinaryOp(RHS);
+}
+
+
+static bool isUnmovableInstruction(Instruction *I) {
+ if (I->getOpcode() == Instruction::PHI ||
+ I->getOpcode() == Instruction::Alloca ||
+ I->getOpcode() == Instruction::Load ||
+ I->getOpcode() == Instruction::Malloc ||
+ I->getOpcode() == Instruction::Invoke ||
+ (I->getOpcode() == Instruction::Call &&
+ !isa<DbgInfoIntrinsic>(I)) ||
+ I->getOpcode() == Instruction::UDiv ||
+ I->getOpcode() == Instruction::SDiv ||
+ I->getOpcode() == Instruction::FDiv ||
+ I->getOpcode() == Instruction::URem ||
+ I->getOpcode() == Instruction::SRem ||
+ I->getOpcode() == Instruction::FRem)
+ return true;
+ return false;
+}
void Reassociate::BuildRankMap(Function &F) {
- unsigned i = 1;
+ unsigned i = 2;
+
+ // Assign distinct ranks to function arguments
+ for (Function::arg_iterator I = F.arg_begin(), E = F.arg_end(); I != E; ++I)
+ ValueRankMap[&*I] = ++i;
+
ReversePostOrderTraversal<Function*> RPOT(&F);
for (ReversePostOrderTraversal<Function*>::rpo_iterator I = RPOT.begin(),
- E = RPOT.end(); I != E; ++I)
- RankMap[*I] = ++i;
+ E = RPOT.end(); I != E; ++I) {
+ BasicBlock *BB = *I;
+ unsigned BBRank = RankMap[BB] = ++i << 16;
+
+ // Walk the basic block, adding precomputed ranks for any instructions that
+ // we cannot move. This ensures that the ranks for these instructions are
+ // all different in the block.
+ for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E; ++I)
+ if (isUnmovableInstruction(I))
+ ValueRankMap[&*I] = ++BBRank;
+ }
}
unsigned Reassociate::getRank(Value *V) {
- if (isa<Argument>(V)) return 1; // Function argument...
- if (Instruction *I = dyn_cast<Instruction>(V)) {
- // If this is an expression, return the MAX(rank(LHS), rank(RHS)) so that we
- // can reassociate expressions for code motion! Since we do not recurse for
- // PHI nodes, we cannot have infinite recursion here, because there cannot
- // be loops in the value graph (except for PHI nodes).
- //
- if (I->getOpcode() == Instruction::PHINode ||
- I->getOpcode() == Instruction::Alloca ||
- I->getOpcode() == Instruction::Malloc || isa<TerminatorInst>(I) ||
- I->hasSideEffects())
- return RankMap[I->getParent()];
-
- unsigned Rank = 0, MaxRank = RankMap[I->getParent()];
- for (unsigned i = 0, e = I->getNumOperands();
- i != e && Rank != MaxRank; ++i)
- Rank = std::max(Rank, getRank(I->getOperand(i)));
-
- return Rank;
- }
+ if (isa<Argument>(V)) return ValueRankMap[V]; // Function argument...
+
+ Instruction *I = dyn_cast<Instruction>(V);
+ if (I == 0) return 0; // Otherwise it's a global or constant, rank 0.
- // Otherwise it's a global or constant, rank 0.
+ unsigned &CachedRank = ValueRankMap[I];
+ if (CachedRank) return CachedRank; // Rank already known?
+
+ // If this is an expression, return the 1+MAX(rank(LHS), rank(RHS)) so that
+ // we can reassociate expressions for code motion! Since we do not recurse
+ // for PHI nodes, we cannot have infinite recursion here, because there
+ // cannot be loops in the value graph that do not go through PHI nodes.
+ unsigned Rank = 0, MaxRank = RankMap[I->getParent()];
+ for (unsigned i = 0, e = I->getNumOperands();
+ i != e && Rank != MaxRank; ++i)
+ Rank = std::max(Rank, getRank(I->getOperand(i)));
+
+ // If this is a not or neg instruction, do not count it for rank. This
+ // assures us that X and ~X will have the same rank.
+ if (!I->getType()->isInteger() ||
+ (!BinaryOperator::isNot(I) && !BinaryOperator::isNeg(I)))
+ ++Rank;
+
+ //DOUT << "Calculated Rank[" << V->getName() << "] = "
+ // << Rank << "\n";
+
+ return CachedRank = Rank;
+}
+
+/// isReassociableOp - Return true if V is an instruction of the specified
+/// opcode and if it only has one use.
+static BinaryOperator *isReassociableOp(Value *V, unsigned Opcode) {
+ if ((V->hasOneUse() || V->use_empty()) && isa<Instruction>(V) &&
+ cast<Instruction>(V)->getOpcode() == Opcode)
+ return cast<BinaryOperator>(V);
return 0;
}
+/// LowerNegateToMultiply - Replace 0-X with X*-1.
+///
+static Instruction *LowerNegateToMultiply(Instruction *Neg,
+ std::map<AssertingVH<>, unsigned> &ValueRankMap,
+ LLVMContext *Context) {
+ Constant *Cst = Context->getAllOnesValue(Neg->getType());
+
+ Instruction *Res = BinaryOperator::CreateMul(Neg->getOperand(1), Cst, "",Neg);
+ ValueRankMap.erase(Neg);
+ Res->takeName(Neg);
+ Neg->replaceAllUsesWith(Res);
+ Neg->eraseFromParent();
+ return Res;
+}
-// isCommutativeOperator - Return true if the specified instruction is
-// commutative and associative. If the instruction is not commutative and
-// associative, we can not reorder its operands!
-//
-static inline BinaryOperator *isCommutativeOperator(Instruction *I) {
- // Floating point operations do not commute!
- if (I->getType()->isFloatingPoint()) return 0;
-
- if (I->getOpcode() == Instruction::Add ||
- I->getOpcode() == Instruction::Mul ||
- I->getOpcode() == Instruction::And ||
- I->getOpcode() == Instruction::Or ||
- I->getOpcode() == Instruction::Xor)
- return cast<BinaryOperator>(I);
- return 0;
+// Given an expression of the form '(A+B)+(D+C)', turn it into '(((A+B)+C)+D)'.
+// Note that if D is also part of the expression tree that we recurse to
+// linearize it as well. Besides that case, this does not recurse into A,B, or
+// C.
+void Reassociate::LinearizeExpr(BinaryOperator *I) {
+ BinaryOperator *LHS = cast<BinaryOperator>(I->getOperand(0));
+ BinaryOperator *RHS = cast<BinaryOperator>(I->getOperand(1));
+ assert(isReassociableOp(LHS, I->getOpcode()) &&
+ isReassociableOp(RHS, I->getOpcode()) &&
+ "Not an expression that needs linearization?");
+
+ DOUT << "Linear" << *LHS << *RHS << *I;
+
+ // Move the RHS instruction to live immediately before I, avoiding breaking
+ // dominator properties.
+ RHS->moveBefore(I);
+
+ // Move operands around to do the linearization.
+ I->setOperand(1, RHS->getOperand(0));
+ RHS->setOperand(0, LHS);
+ I->setOperand(0, RHS);
+
+ ++NumLinear;
+ MadeChange = true;
+ DOUT << "Linearized: " << *I;
+
+ // If D is part of this expression tree, tail recurse.
+ if (isReassociableOp(I->getOperand(1), I->getOpcode()))
+ LinearizeExpr(I);
}
-bool Reassociate::ReassociateExpr(BinaryOperator *I) {
- Value *LHS = I->getOperand(0);
- Value *RHS = I->getOperand(1);
- unsigned LHSRank = getRank(LHS);
- unsigned RHSRank = getRank(RHS);
-
- bool Changed = false;
-
- // Make sure the LHS of the operand always has the greater rank...
- if (LHSRank < RHSRank) {
- I->swapOperands();
- std::swap(LHS, RHS);
- std::swap(LHSRank, RHSRank);
- Changed = true;
- ++NumSwapped;
- DEBUG(std::cerr << "Transposed: " << I << " Result BB: " << I->getParent());
+/// LinearizeExprTree - Given an associative binary expression tree, traverse
+/// all of the uses putting it into canonical form. This forces a left-linear
+/// form of the the expression (((a+b)+c)+d), and collects information about the
+/// rank of the non-tree operands.
+///
+/// NOTE: These intentionally destroys the expression tree operands (turning
+/// them into undef values) to reduce #uses of the values. This means that the
+/// caller MUST use something like RewriteExprTree to put the values back in.
+///
+void Reassociate::LinearizeExprTree(BinaryOperator *I,
+ std::vector<ValueEntry> &Ops) {
+ Value *LHS = I->getOperand(0), *RHS = I->getOperand(1);
+ unsigned Opcode = I->getOpcode();
+
+ // First step, linearize the expression if it is in ((A+B)+(C+D)) form.
+ BinaryOperator *LHSBO = isReassociableOp(LHS, Opcode);
+ BinaryOperator *RHSBO = isReassociableOp(RHS, Opcode);
+
+ // If this is a multiply expression tree and it contains internal negations,
+ // transform them into multiplies by -1 so they can be reassociated.
+ if (I->getOpcode() == Instruction::Mul) {
+ if (!LHSBO && LHS->hasOneUse() && BinaryOperator::isNeg(LHS)) {
+ LHS = LowerNegateToMultiply(cast<Instruction>(LHS),
+ ValueRankMap, Context);
+ LHSBO = isReassociableOp(LHS, Opcode);
+ }
+ if (!RHSBO && RHS->hasOneUse() && BinaryOperator::isNeg(RHS)) {
+ RHS = LowerNegateToMultiply(cast<Instruction>(RHS),
+ ValueRankMap, Context);
+ RHSBO = isReassociableOp(RHS, Opcode);
+ }
}
-
- // If the LHS is the same operator as the current one is, and if we are the
- // only expression using it...
- //
- if (BinaryOperator *LHSI = dyn_cast<BinaryOperator>(LHS))
- if (LHSI->getOpcode() == I->getOpcode() && LHSI->use_size() == 1) {
- // If the rank of our current RHS is less than the rank of the LHS's LHS,
- // then we reassociate the two instructions...
- if (RHSRank < getRank(LHSI->getOperand(0))) {
- unsigned TakeOp = 0;
- if (BinaryOperator *IOp = dyn_cast<BinaryOperator>(LHSI->getOperand(0)))
- if (IOp->getOpcode() == LHSI->getOpcode())
- TakeOp = 1; // Hoist out non-tree portion
-
- // Convert ((a + 12) + 10) into (a + (12 + 10))
- I->setOperand(0, LHSI->getOperand(TakeOp));
- LHSI->setOperand(TakeOp, RHS);
- I->setOperand(1, LHSI);
-
- ++NumChanged;
- DEBUG(std::cerr << "Reassociated: " << I << " Result BB: "
- << I->getParent());
-
- // Since we modified the RHS instruction, make sure that we recheck it.
- ReassociateExpr(LHSI);
- return true;
- }
+
+ if (!LHSBO) {
+ if (!RHSBO) {
+ // Neither the LHS or RHS as part of the tree, thus this is a leaf. As
+ // such, just remember these operands and their rank.
+ Ops.push_back(ValueEntry(getRank(LHS), LHS));
+ Ops.push_back(ValueEntry(getRank(RHS), RHS));
+
+ // Clear the leaves out.
+ I->setOperand(0, Context->getUndef(I->getType()));
+ I->setOperand(1, Context->getUndef(I->getType()));
+ return;
+ } else {
+ // Turn X+(Y+Z) -> (Y+Z)+X
+ std::swap(LHSBO, RHSBO);
+ std::swap(LHS, RHS);
+ bool Success = !I->swapOperands();
+ assert(Success && "swapOperands failed");
+ Success = false;
+ MadeChange = true;
}
+ } else if (RHSBO) {
+ // Turn (A+B)+(C+D) -> (((A+B)+C)+D). This guarantees the the RHS is not
+ // part of the expression tree.
+ LinearizeExpr(I);
+ LHS = LHSBO = cast<BinaryOperator>(I->getOperand(0));
+ RHS = I->getOperand(1);
+ RHSBO = 0;
+ }
- return Changed;
+ // Okay, now we know that the LHS is a nested expression and that the RHS is
+ // not. Perform reassociation.
+ assert(!isReassociableOp(RHS, Opcode) && "LinearizeExpr failed!");
+
+ // Move LHS right before I to make sure that the tree expression dominates all
+ // values.
+ LHSBO->moveBefore(I);
+
+ // Linearize the expression tree on the LHS.
+ LinearizeExprTree(LHSBO, Ops);
+
+ // Remember the RHS operand and its rank.
+ Ops.push_back(ValueEntry(getRank(RHS), RHS));
+
+ // Clear the RHS leaf out.
+ I->setOperand(1, Context->getUndef(I->getType()));
+}
+
+// RewriteExprTree - Now that the operands for this expression tree are
+// linearized and optimized, emit them in-order. This function is written to be
+// tail recursive.
+void Reassociate::RewriteExprTree(BinaryOperator *I,
+ std::vector<ValueEntry> &Ops,
+ unsigned i) {
+ if (i+2 == Ops.size()) {
+ if (I->getOperand(0) != Ops[i].Op ||
+ I->getOperand(1) != Ops[i+1].Op) {
+ Value *OldLHS = I->getOperand(0);
+ DOUT << "RA: " << *I;
+ I->setOperand(0, Ops[i].Op);
+ I->setOperand(1, Ops[i+1].Op);
+ DOUT << "TO: " << *I;
+ MadeChange = true;
+ ++NumChanged;
+
+ // If we reassociated a tree to fewer operands (e.g. (1+a+2) -> (a+3)
+ // delete the extra, now dead, nodes.
+ RemoveDeadBinaryOp(OldLHS);
+ }
+ return;
+ }
+ assert(i+2 < Ops.size() && "Ops index out of range!");
+
+ if (I->getOperand(1) != Ops[i].Op) {
+ DOUT << "RA: " << *I;
+ I->setOperand(1, Ops[i].Op);
+ DOUT << "TO: " << *I;
+ MadeChange = true;
+ ++NumChanged;
+ }
+
+ BinaryOperator *LHS = cast<BinaryOperator>(I->getOperand(0));
+ assert(LHS->getOpcode() == I->getOpcode() &&
+ "Improper expression tree!");
+
+ // Compactify the tree instructions together with each other to guarantee
+ // that the expression tree is dominated by all of Ops.
+ LHS->moveBefore(I);
+ RewriteExprTree(LHS, Ops, i+1);
}
+
// NegateValue - Insert instructions before the instruction pointed to by BI,
// that computes the negative version of the value specified. The negative
// version of the value is returned, and BI is left pointing at the instruction
// that should be processed next by the reassociation pass.
//
-static Value *NegateValue(Value *V, BasicBlock *BB, BasicBlock::iterator &BI) {
+static Value *NegateValue(LLVMContext *Context, Value *V, Instruction *BI) {
// We are trying to expose opportunity for reassociation. One of the things
// that we want to do to achieve this is to push a negation as deep into an
// expression chain as possible, to expose the add instructions. In practice,
// X = -(A+12+C+D) into X = -A + -12 + -C + -D = -12 + -A + -C + -D
// so that later, a: Y = 12+X could get reassociated with the -12 to eliminate
// the constants. We assume that instcombine will clean up the mess later if
- // we introduce tons of unneccesary negation instructions...
+ // we introduce tons of unnecessary negation instructions...
//
if (Instruction *I = dyn_cast<Instruction>(V))
- if (I->getOpcode() == Instruction::Add && I->use_size() == 1) {
- Value *RHS = NegateValue(I->getOperand(1), BB, BI);
- Value *LHS = NegateValue(I->getOperand(0), BB, BI);
-
- // We must actually insert a new add instruction here, because the neg
- // instructions do not dominate the old add instruction in general. By
- // adding it now, we are assured that the neg instructions we just
- // inserted dominate the instruction we are about to insert after them.
+ if (I->getOpcode() == Instruction::Add && I->hasOneUse()) {
+ // Push the negates through the add.
+ I->setOperand(0, NegateValue(Context, I->getOperand(0), BI));
+ I->setOperand(1, NegateValue(Context, I->getOperand(1), BI));
+
+ // We must move the add instruction here, because the neg instructions do
+ // not dominate the old add instruction in general. By moving it, we are
+ // assured that the neg instructions we just inserted dominate the
+ // instruction we are about to insert after them.
//
- BasicBlock::iterator NBI = cast<Instruction>(RHS);
-
- Instruction *Add =
- BinaryOperator::create(Instruction::Add, LHS, RHS, I->getName()+".neg");
- BB->getInstList().insert(++NBI, Add); // Add to the basic block...
- return Add;
+ I->moveBefore(BI);
+ I->setName(I->getName()+".neg");
+ return I;
}
// Insert a 'neg' instruction that subtracts the value from zero to get the
// negation.
//
- Instruction *Neg =
- BinaryOperator::create(Instruction::Sub,
- Constant::getNullValue(V->getType()), V,
- V->getName()+".neg");
- BI = BB->getInstList().insert(BI, Neg); // Add to the basic block...
- return Neg;
+ return BinaryOperator::CreateNeg(*Context, V, V->getName() + ".neg", BI);
+}
+
+/// ShouldBreakUpSubtract - Return true if we should break up this subtract of
+/// X-Y into (X + -Y).
+static bool ShouldBreakUpSubtract(LLVMContext *Context, Instruction *Sub) {
+ // If this is a negation, we can't split it up!
+ if (BinaryOperator::isNeg(Sub))
+ return false;
+
+ // Don't bother to break this up unless either the LHS is an associable add or
+ // subtract or if this is only used by one.
+ if (isReassociableOp(Sub->getOperand(0), Instruction::Add) ||
+ isReassociableOp(Sub->getOperand(0), Instruction::Sub))
+ return true;
+ if (isReassociableOp(Sub->getOperand(1), Instruction::Add) ||
+ isReassociableOp(Sub->getOperand(1), Instruction::Sub))
+ return true;
+ if (Sub->hasOneUse() &&
+ (isReassociableOp(Sub->use_back(), Instruction::Add) ||
+ isReassociableOp(Sub->use_back(), Instruction::Sub)))
+ return true;
+
+ return false;
+}
+
+/// BreakUpSubtract - If we have (X-Y), and if either X is an add, or if this is
+/// only used by an add, transform this into (X+(0-Y)) to promote better
+/// reassociation.
+static Instruction *BreakUpSubtract(LLVMContext *Context, Instruction *Sub,
+ std::map<AssertingVH<>, unsigned> &ValueRankMap) {
+ // Convert a subtract into an add and a neg instruction... so that sub
+ // instructions can be commuted with other add instructions...
+ //
+ // Calculate the negative value of Operand 1 of the sub instruction...
+ // and set it as the RHS of the add instruction we just made...
+ //
+ Value *NegVal = NegateValue(Context, Sub->getOperand(1), Sub);
+ Instruction *New =
+ BinaryOperator::CreateAdd(Sub->getOperand(0), NegVal, "", Sub);
+ New->takeName(Sub);
+
+ // Everyone now refers to the add instruction.
+ ValueRankMap.erase(Sub);
+ Sub->replaceAllUsesWith(New);
+ Sub->eraseFromParent();
+
+ DOUT << "Negated: " << *New;
+ return New;
}
+/// ConvertShiftToMul - If this is a shift of a reassociable multiply or is used
+/// by one, change this into a multiply by a constant to assist with further
+/// reassociation.
+static Instruction *ConvertShiftToMul(Instruction *Shl,
+ std::map<AssertingVH<>, unsigned> &ValueRankMap,
+ LLVMContext *Context) {
+ // If an operand of this shift is a reassociable multiply, or if the shift
+ // is used by a reassociable multiply or add, turn into a multiply.
+ if (isReassociableOp(Shl->getOperand(0), Instruction::Mul) ||
+ (Shl->hasOneUse() &&
+ (isReassociableOp(Shl->use_back(), Instruction::Mul) ||
+ isReassociableOp(Shl->use_back(), Instruction::Add)))) {
+ Constant *MulCst = Context->getConstantInt(Shl->getType(), 1);
+ MulCst =
+ Context->getConstantExprShl(MulCst, cast<Constant>(Shl->getOperand(1)));
+
+ Instruction *Mul = BinaryOperator::CreateMul(Shl->getOperand(0), MulCst,
+ "", Shl);
+ ValueRankMap.erase(Shl);
+ Mul->takeName(Shl);
+ Shl->replaceAllUsesWith(Mul);
+ Shl->eraseFromParent();
+ return Mul;
+ }
+ return 0;
+}
-bool Reassociate::ReassociateBB(BasicBlock *BB) {
- bool Changed = false;
- for (BasicBlock::iterator BI = BB->begin(); BI != BB->end(); ++BI) {
-
- // If this instruction is a commutative binary operator, and the ranks of
- // the two operands are sorted incorrectly, fix it now.
- //
- if (BinaryOperator *I = isCommutativeOperator(BI)) {
- if (!I->use_empty()) {
- // Make sure that we don't have a tree-shaped computation. If we do,
- // linearize it. Convert (A+B)+(C+D) into ((A+B)+C)+D
- //
- Instruction *LHSI = dyn_cast<Instruction>(I->getOperand(0));
- Instruction *RHSI = dyn_cast<Instruction>(I->getOperand(1));
- if (LHSI && (int)LHSI->getOpcode() == I->getOpcode() &&
- RHSI && (int)RHSI->getOpcode() == I->getOpcode() &&
- RHSI->use_size() == 1) {
- // Insert a new temporary instruction... (A+B)+C
- BinaryOperator *Tmp = BinaryOperator::create(I->getOpcode(), LHSI,
- RHSI->getOperand(0),
- RHSI->getName()+".ra");
- BI = BB->getInstList().insert(BI, Tmp); // Add to the basic block...
- I->setOperand(0, Tmp);
- I->setOperand(1, RHSI->getOperand(1));
-
- // Process the temporary instruction for reassociation now.
- I = Tmp;
- ++NumLinear;
- Changed = true;
- DEBUG(std::cerr << "Linearized: " << I << " Result BB: " << BB);
+// Scan backwards and forwards among values with the same rank as element i to
+// see if X exists. If X does not exist, return i.
+static unsigned FindInOperandList(std::vector<ValueEntry> &Ops, unsigned i,
+ Value *X) {
+ unsigned XRank = Ops[i].Rank;
+ unsigned e = Ops.size();
+ for (unsigned j = i+1; j != e && Ops[j].Rank == XRank; ++j)
+ if (Ops[j].Op == X)
+ return j;
+ // Scan backwards
+ for (unsigned j = i-1; j != ~0U && Ops[j].Rank == XRank; --j)
+ if (Ops[j].Op == X)
+ return j;
+ return i;
+}
+
+/// EmitAddTreeOfValues - Emit a tree of add instructions, summing Ops together
+/// and returning the result. Insert the tree before I.
+static Value *EmitAddTreeOfValues(Instruction *I, std::vector<Value*> &Ops) {
+ if (Ops.size() == 1) return Ops.back();
+
+ Value *V1 = Ops.back();
+ Ops.pop_back();
+ Value *V2 = EmitAddTreeOfValues(I, Ops);
+ return BinaryOperator::CreateAdd(V2, V1, "tmp", I);
+}
+
+/// RemoveFactorFromExpression - If V is an expression tree that is a
+/// multiplication sequence, and if this sequence contains a multiply by Factor,
+/// remove Factor from the tree and return the new tree.
+Value *Reassociate::RemoveFactorFromExpression(Value *V, Value *Factor) {
+ BinaryOperator *BO = isReassociableOp(V, Instruction::Mul);
+ if (!BO) return 0;
+
+ std::vector<ValueEntry> Factors;
+ LinearizeExprTree(BO, Factors);
+
+ bool FoundFactor = false;
+ for (unsigned i = 0, e = Factors.size(); i != e; ++i)
+ if (Factors[i].Op == Factor) {
+ FoundFactor = true;
+ Factors.erase(Factors.begin()+i);
+ break;
+ }
+ if (!FoundFactor) {
+ // Make sure to restore the operands to the expression tree.
+ RewriteExprTree(BO, Factors);
+ return 0;
+ }
+
+ if (Factors.size() == 1) return Factors[0].Op;
+
+ RewriteExprTree(BO, Factors);
+ return BO;
+}
+
+/// FindSingleUseMultiplyFactors - If V is a single-use multiply, recursively
+/// add its operands as factors, otherwise add V to the list of factors.
+static void FindSingleUseMultiplyFactors(Value *V,
+ std::vector<Value*> &Factors) {
+ BinaryOperator *BO;
+ if ((!V->hasOneUse() && !V->use_empty()) ||
+ !(BO = dyn_cast<BinaryOperator>(V)) ||
+ BO->getOpcode() != Instruction::Mul) {
+ Factors.push_back(V);
+ return;
+ }
+
+ // Otherwise, add the LHS and RHS to the list of factors.
+ FindSingleUseMultiplyFactors(BO->getOperand(1), Factors);
+ FindSingleUseMultiplyFactors(BO->getOperand(0), Factors);
+}
+
+
+
+Value *Reassociate::OptimizeExpression(BinaryOperator *I,
+ std::vector<ValueEntry> &Ops) {
+ // Now that we have the linearized expression tree, try to optimize it.
+ // Start by folding any constants that we found.
+ bool IterateOptimization = false;
+ if (Ops.size() == 1) return Ops[0].Op;
+
+ unsigned Opcode = I->getOpcode();
+
+ if (Constant *V1 = dyn_cast<Constant>(Ops[Ops.size()-2].Op))
+ if (Constant *V2 = dyn_cast<Constant>(Ops.back().Op)) {
+ Ops.pop_back();
+ Ops.back().Op = Context->getConstantExpr(Opcode, V1, V2);
+ return OptimizeExpression(I, Ops);
+ }
+
+ // Check for destructive annihilation due to a constant being used.
+ if (ConstantInt *CstVal = dyn_cast<ConstantInt>(Ops.back().Op))
+ switch (Opcode) {
+ default: break;
+ case Instruction::And:
+ if (CstVal->isZero()) { // ... & 0 -> 0
+ ++NumAnnihil;
+ return CstVal;
+ } else if (CstVal->isAllOnesValue()) { // ... & -1 -> ...
+ Ops.pop_back();
+ }
+ break;
+ case Instruction::Mul:
+ if (CstVal->isZero()) { // ... * 0 -> 0
+ ++NumAnnihil;
+ return CstVal;
+ } else if (cast<ConstantInt>(CstVal)->isOne()) {
+ Ops.pop_back(); // ... * 1 -> ...
+ }
+ break;
+ case Instruction::Or:
+ if (CstVal->isAllOnesValue()) { // ... | -1 -> -1
+ ++NumAnnihil;
+ return CstVal;
+ }
+ // FALLTHROUGH!
+ case Instruction::Add:
+ case Instruction::Xor:
+ if (CstVal->isZero()) // ... [|^+] 0 -> ...
+ Ops.pop_back();
+ break;
+ }
+ if (Ops.size() == 1) return Ops[0].Op;
+
+ // Handle destructive annihilation do to identities between elements in the
+ // argument list here.
+ switch (Opcode) {
+ default: break;
+ case Instruction::And:
+ case Instruction::Or:
+ case Instruction::Xor:
+ // Scan the operand lists looking for X and ~X pairs, along with X,X pairs.
+ // If we find any, we can simplify the expression. X&~X == 0, X|~X == -1.
+ for (unsigned i = 0, e = Ops.size(); i != e; ++i) {
+ // First, check for X and ~X in the operand list.
+ assert(i < Ops.size());
+ if (BinaryOperator::isNot(Ops[i].Op)) { // Cannot occur for ^.
+ Value *X = BinaryOperator::getNotArgument(Ops[i].Op);
+ unsigned FoundX = FindInOperandList(Ops, i, X);
+ if (FoundX != i) {
+ if (Opcode == Instruction::And) { // ...&X&~X = 0
+ ++NumAnnihil;
+ return Context->getNullValue(X->getType());
+ } else if (Opcode == Instruction::Or) { // ...|X|~X = -1
+ ++NumAnnihil;
+ return Context->getAllOnesValue(X->getType());
+ }
+ }
+ }
+
+ // Next, check for duplicate pairs of values, which we assume are next to
+ // each other, due to our sorting criteria.
+ assert(i < Ops.size());
+ if (i+1 != Ops.size() && Ops[i+1].Op == Ops[i].Op) {
+ if (Opcode == Instruction::And || Opcode == Instruction::Or) {
+ // Drop duplicate values.
+ Ops.erase(Ops.begin()+i);
+ --i; --e;
+ IterateOptimization = true;
+ ++NumAnnihil;
+ } else {
+ assert(Opcode == Instruction::Xor);
+ if (e == 2) {
+ ++NumAnnihil;
+ return Context->getNullValue(Ops[0].Op->getType());
+ }
+ // ... X^X -> ...
+ Ops.erase(Ops.begin()+i, Ops.begin()+i+2);
+ i -= 1; e -= 2;
+ IterateOptimization = true;
+ ++NumAnnihil;
+ }
+ }
+ }
+ break;
+
+ case Instruction::Add:
+ // Scan the operand lists looking for X and -X pairs. If we find any, we
+ // can simplify the expression. X+-X == 0.
+ for (unsigned i = 0, e = Ops.size(); i != e; ++i) {
+ assert(i < Ops.size());
+ // Check for X and -X in the operand list.
+ if (BinaryOperator::isNeg(Ops[i].Op)) {
+ Value *X = BinaryOperator::getNegArgument(Ops[i].Op);
+ unsigned FoundX = FindInOperandList(Ops, i, X);
+ if (FoundX != i) {
+ // Remove X and -X from the operand list.
+ if (Ops.size() == 2) {
+ ++NumAnnihil;
+ return Context->getNullValue(X->getType());
+ } else {
+ Ops.erase(Ops.begin()+i);
+ if (i < FoundX)
+ --FoundX;
+ else
+ --i; // Need to back up an extra one.
+ Ops.erase(Ops.begin()+FoundX);
+ IterateOptimization = true;
+ ++NumAnnihil;
+ --i; // Revisit element.
+ e -= 2; // Removed two elements.
+ }
+ }
+ }
+ }
+
+
+ // Scan the operand list, checking to see if there are any common factors
+ // between operands. Consider something like A*A+A*B*C+D. We would like to
+ // reassociate this to A*(A+B*C)+D, which reduces the number of multiplies.
+ // To efficiently find this, we count the number of times a factor occurs
+ // for any ADD operands that are MULs.
+ std::map<Value*, unsigned> FactorOccurrences;
+ unsigned MaxOcc = 0;
+ Value *MaxOccVal = 0;
+ for (unsigned i = 0, e = Ops.size(); i != e; ++i) {
+ if (BinaryOperator *BOp = dyn_cast<BinaryOperator>(Ops[i].Op)) {
+ if (BOp->getOpcode() == Instruction::Mul && BOp->use_empty()) {
+ // Compute all of the factors of this added value.
+ std::vector<Value*> Factors;
+ FindSingleUseMultiplyFactors(BOp, Factors);
+ assert(Factors.size() > 1 && "Bad linearize!");
+
+ // Add one to FactorOccurrences for each unique factor in this op.
+ if (Factors.size() == 2) {
+ unsigned Occ = ++FactorOccurrences[Factors[0]];
+ if (Occ > MaxOcc) { MaxOcc = Occ; MaxOccVal = Factors[0]; }
+ if (Factors[0] != Factors[1]) { // Don't double count A*A.
+ Occ = ++FactorOccurrences[Factors[1]];
+ if (Occ > MaxOcc) { MaxOcc = Occ; MaxOccVal = Factors[1]; }
+ }
+ } else {
+ std::set<Value*> Duplicates;
+ for (unsigned i = 0, e = Factors.size(); i != e; ++i) {
+ if (Duplicates.insert(Factors[i]).second) {
+ unsigned Occ = ++FactorOccurrences[Factors[i]];
+ if (Occ > MaxOcc) { MaxOcc = Occ; MaxOccVal = Factors[i]; }
+ }
+ }
+ }
}
+ }
+ }
- // Make sure that this expression is correctly reassociated with respect
- // to it's used values...
- //
- Changed |= ReassociateExpr(I);
+ // If any factor occurred more than one time, we can pull it out.
+ if (MaxOcc > 1) {
+ DOUT << "\nFACTORING [" << MaxOcc << "]: " << *MaxOccVal << "\n";
+
+ // Create a new instruction that uses the MaxOccVal twice. If we don't do
+ // this, we could otherwise run into situations where removing a factor
+ // from an expression will drop a use of maxocc, and this can cause
+ // RemoveFactorFromExpression on successive values to behave differently.
+ Instruction *DummyInst = BinaryOperator::CreateAdd(MaxOccVal, MaxOccVal);
+ std::vector<Value*> NewMulOps;
+ for (unsigned i = 0, e = Ops.size(); i != e; ++i) {
+ if (Value *V = RemoveFactorFromExpression(Ops[i].Op, MaxOccVal)) {
+ NewMulOps.push_back(V);
+ Ops.erase(Ops.begin()+i);
+ --i; --e;
+ }
}
+
+ // No need for extra uses anymore.
+ delete DummyInst;
+
+ unsigned NumAddedValues = NewMulOps.size();
+ Value *V = EmitAddTreeOfValues(I, NewMulOps);
+ Value *V2 = BinaryOperator::CreateMul(V, MaxOccVal, "tmp", I);
+
+ // Now that we have inserted V and its sole use, optimize it. This allows
+ // us to handle cases that require multiple factoring steps, such as this:
+ // A*A*B + A*A*C --> A*(A*B+A*C) --> A*(A*(B+C))
+ if (NumAddedValues > 1)
+ ReassociateExpression(cast<BinaryOperator>(V));
+
+ ++NumFactor;
+
+ if (Ops.empty())
+ return V2;
+
+ // Add the new value to the list of things being added.
+ Ops.insert(Ops.begin(), ValueEntry(getRank(V2), V2));
+
+ // Rewrite the tree so that there is now a use of V.
+ RewriteExprTree(I, Ops);
+ return OptimizeExpression(I, Ops);
+ }
+ break;
+ //case Instruction::Mul:
+ }
- } else if (BI->getOpcode() == Instruction::Sub &&
- BI->getOperand(0) != Constant::getNullValue(BI->getType())) {
- // Convert a subtract into an add and a neg instruction... so that sub
- // instructions can be commuted with other add instructions...
- //
- Instruction *New = BinaryOperator::create(Instruction::Add,
- BI->getOperand(0),
- BI->getOperand(1),
- BI->getName());
- Value *NegatedValue = BI->getOperand(1);
-
- // Everyone now refers to the add instruction...
- BI->replaceAllUsesWith(New);
-
- // Put the new add in the place of the subtract... deleting the subtract
- BI = BB->getInstList().erase(BI);
- BI = ++BB->getInstList().insert(BI, New);
-
- // Calculate the negative value of Operand 1 of the sub instruction...
- // and set it as the RHS of the add instruction we just made...
- New->setOperand(1, NegateValue(NegatedValue, BB, BI));
- --BI;
- Changed = true;
- DEBUG(std::cerr << "Negated: " << New << " Result BB: " << BB);
+ if (IterateOptimization)
+ return OptimizeExpression(I, Ops);
+ return 0;
+}
+
+
+/// ReassociateBB - Inspect all of the instructions in this basic block,
+/// reassociating them as we go.
+void Reassociate::ReassociateBB(BasicBlock *BB) {
+ for (BasicBlock::iterator BBI = BB->begin(); BBI != BB->end(); ) {
+ Instruction *BI = BBI++;
+ if (BI->getOpcode() == Instruction::Shl &&
+ isa<ConstantInt>(BI->getOperand(1)))
+ if (Instruction *NI = ConvertShiftToMul(BI, ValueRankMap, Context)) {
+ MadeChange = true;
+ BI = NI;
+ }
+
+ // Reject cases where it is pointless to do this.
+ if (!isa<BinaryOperator>(BI) || BI->getType()->isFloatingPoint() ||
+ isa<VectorType>(BI->getType()))
+ continue; // Floating point ops are not associative.
+
+ // If this is a subtract instruction which is not already in negate form,
+ // see if we can convert it to X+-Y.
+ if (BI->getOpcode() == Instruction::Sub) {
+ if (ShouldBreakUpSubtract(Context, BI)) {
+ BI = BreakUpSubtract(Context, BI, ValueRankMap);
+ MadeChange = true;
+ } else if (BinaryOperator::isNeg(BI)) {
+ // Otherwise, this is a negation. See if the operand is a multiply tree
+ // and if this is not an inner node of a multiply tree.
+ if (isReassociableOp(BI->getOperand(1), Instruction::Mul) &&
+ (!BI->hasOneUse() ||
+ !isReassociableOp(BI->use_back(), Instruction::Mul))) {
+ BI = LowerNegateToMultiply(BI, ValueRankMap, Context);
+ MadeChange = true;
+ }
+ }
}
+
+ // If this instruction is a commutative binary operator, process it.
+ if (!BI->isAssociative()) continue;
+ BinaryOperator *I = cast<BinaryOperator>(BI);
+
+ // If this is an interior node of a reassociable tree, ignore it until we
+ // get to the root of the tree, to avoid N^2 analysis.
+ if (I->hasOneUse() && isReassociableOp(I->use_back(), I->getOpcode()))
+ continue;
+
+ // If this is an add tree that is used by a sub instruction, ignore it
+ // until we process the subtract.
+ if (I->hasOneUse() && I->getOpcode() == Instruction::Add &&
+ cast<Instruction>(I->use_back())->getOpcode() == Instruction::Sub)
+ continue;
+
+ ReassociateExpression(I);
}
+}
- return Changed;
+void Reassociate::ReassociateExpression(BinaryOperator *I) {
+
+ // First, walk the expression tree, linearizing the tree, collecting
+ std::vector<ValueEntry> Ops;
+ LinearizeExprTree(I, Ops);
+
+ DOUT << "RAIn:\t"; DEBUG(PrintOps(I, Ops)); DOUT << "\n";
+
+ // Now that we have linearized the tree to a list and have gathered all of
+ // the operands and their ranks, sort the operands by their rank. Use a
+ // stable_sort so that values with equal ranks will have their relative
+ // positions maintained (and so the compiler is deterministic). Note that
+ // this sorts so that the highest ranking values end up at the beginning of
+ // the vector.
+ std::stable_sort(Ops.begin(), Ops.end());
+
+ // OptimizeExpression - Now that we have the expression tree in a convenient
+ // sorted form, optimize it globally if possible.
+ if (Value *V = OptimizeExpression(I, Ops)) {
+ // This expression tree simplified to something that isn't a tree,
+ // eliminate it.
+ DOUT << "Reassoc to scalar: " << *V << "\n";
+ I->replaceAllUsesWith(V);
+ RemoveDeadBinaryOp(I);
+ return;
+ }
+
+ // We want to sink immediates as deeply as possible except in the case where
+ // this is a multiply tree used only by an add, and the immediate is a -1.
+ // In this case we reassociate to put the negation on the outside so that we
+ // can fold the negation into the add: (-X)*Y + Z -> Z-X*Y
+ if (I->getOpcode() == Instruction::Mul && I->hasOneUse() &&
+ cast<Instruction>(I->use_back())->getOpcode() == Instruction::Add &&
+ isa<ConstantInt>(Ops.back().Op) &&
+ cast<ConstantInt>(Ops.back().Op)->isAllOnesValue()) {
+ Ops.insert(Ops.begin(), Ops.back());
+ Ops.pop_back();
+ }
+
+ DOUT << "RAOut:\t"; DEBUG(PrintOps(I, Ops)); DOUT << "\n";
+
+ if (Ops.size() == 1) {
+ // This expression tree simplified to something that isn't a tree,
+ // eliminate it.
+ I->replaceAllUsesWith(Ops[0].Op);
+ RemoveDeadBinaryOp(I);
+ } else {
+ // Now that we ordered and optimized the expressions, splat them back into
+ // the expression tree, removing any unneeded nodes.
+ RewriteExprTree(I, Ops);
+ }
}
// Recalculate the rank map for F
BuildRankMap(F);
- bool Changed = false;
+ MadeChange = false;
for (Function::iterator FI = F.begin(), FE = F.end(); FI != FE; ++FI)
- Changed |= ReassociateBB(FI);
+ ReassociateBB(FI);
// We are done with the rank map...
RankMap.clear();
- return Changed;
+ ValueRankMap.clear();
+ return MadeChange;
}
+