X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTransforms%2FVectorize%2FSLPVectorizer.cpp;h=3aced0b4c4094e8b69ff9e450f85f5b5091816a9;hb=6ca2d8b7c7461910610405e8bde5ef0ca9506d62;hp=e13ba956c398eb9f14490ca256ee90850cfd9a10;hpb=68c7a1cb98399c770af6dc103bec45b1b7ca3c29;p=oota-llvm.git diff --git a/lib/Transforms/Vectorize/SLPVectorizer.cpp b/lib/Transforms/Vectorize/SLPVectorizer.cpp index e13ba956c39..3aced0b4c40 100644 --- a/lib/Transforms/Vectorize/SLPVectorizer.cpp +++ b/lib/Transforms/Vectorize/SLPVectorizer.cpp @@ -19,6 +19,7 @@ #include "llvm/ADT/MapVector.h" #include "llvm/ADT/PostOrderIterator.h" #include "llvm/ADT/SetVector.h" +#include "llvm/ADT/Statistic.h" #include "llvm/Analysis/AliasAnalysis.h" #include "llvm/Analysis/LoopInfo.h" #include "llvm/Analysis/ScalarEvolution.h" @@ -42,12 +43,15 @@ #include "llvm/Transforms/Utils/VectorUtils.h" #include #include +#include using namespace llvm; #define SV_NAME "slp-vectorizer" #define DEBUG_TYPE "SLP" +STATISTIC(NumVectorInstructions, "Number of vector instructions generated"); + static cl::opt SLPCostThreshold("slp-threshold", cl::init(0), cl::Hidden, cl::desc("Only vectorize if you gain more than this " @@ -68,53 +72,6 @@ static const unsigned MinVecRegSize = 128; static const unsigned RecursionMaxDepth = 12; -/// A helper class for numbering instructions in multiple blocks. -/// Numbers start at zero for each basic block. -struct BlockNumbering { - - BlockNumbering(BasicBlock *Bb) : BB(Bb), Valid(false) {} - - void numberInstructions() { - unsigned Loc = 0; - InstrIdx.clear(); - InstrVec.clear(); - // Number the instructions in the block. - for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { - InstrIdx[it] = Loc++; - InstrVec.push_back(it); - assert(InstrVec[InstrIdx[it]] == it && "Invalid allocation"); - } - Valid = true; - } - - int getIndex(Instruction *I) { - assert(I->getParent() == BB && "Invalid instruction"); - if (!Valid) - numberInstructions(); - assert(InstrIdx.count(I) && "Unknown instruction"); - return InstrIdx[I]; - } - - Instruction *getInstruction(unsigned loc) { - if (!Valid) - numberInstructions(); - assert(InstrVec.size() > loc && "Invalid Index"); - return InstrVec[loc]; - } - - void forget() { Valid = false; } - -private: - /// The block we are numbering. - BasicBlock *BB; - /// Is the block numbered. - bool Valid; - /// Maps instructions to numbers and back. - SmallDenseMap InstrIdx; - /// Maps integers to Instructions. - SmallVector InstrVec; -}; - /// \returns the parent basic block if all of the instructions in \p VL /// are in the same block or null otherwise. static BasicBlock *getSameBlock(ArrayRef VL) { @@ -149,6 +106,48 @@ static bool isSplat(ArrayRef VL) { return true; } +///\returns Opcode that can be clubbed with \p Op to create an alternate +/// sequence which can later be merged as a ShuffleVector instruction. +static unsigned getAltOpcode(unsigned Op) { + switch (Op) { + case Instruction::FAdd: + return Instruction::FSub; + case Instruction::FSub: + return Instruction::FAdd; + case Instruction::Add: + return Instruction::Sub; + case Instruction::Sub: + return Instruction::Add; + default: + return 0; + } +} + +///\returns bool representing if Opcode \p Op can be part +/// of an alternate sequence which can later be merged as +/// a ShuffleVector instruction. +static bool canCombineAsAltInst(unsigned Op) { + if (Op == Instruction::FAdd || Op == Instruction::FSub || + Op == Instruction::Sub || Op == Instruction::Add) + return true; + return false; +} + +/// \returns ShuffleVector instruction if intructions in \p VL have +/// alternate fadd,fsub / fsub,fadd/add,sub/sub,add sequence. +/// (i.e. e.g. opcodes of fadd,fsub,fadd,fsub...) +static unsigned isAltInst(ArrayRef VL) { + Instruction *I0 = dyn_cast(VL[0]); + unsigned Opcode = I0->getOpcode(); + unsigned AltOpcode = getAltOpcode(Opcode); + for (int i = 1, e = VL.size(); i < e; i++) { + Instruction *I = dyn_cast(VL[i]); + if (!I || I->getOpcode() != ((i & 1) ? AltOpcode : Opcode)) + return 0; + } + return Instruction::ShuffleVector; +} + /// \returns The opcode if all of the Instructions in \p VL have the same /// opcode, or zero. static unsigned getSameOpcode(ArrayRef VL) { @@ -158,8 +157,11 @@ static unsigned getSameOpcode(ArrayRef VL) { unsigned Opcode = I0->getOpcode(); for (int i = 1, e = VL.size(); i < e; i++) { Instruction *I = dyn_cast(VL[i]); - if (!I || Opcode != I->getOpcode()) + if (!I || Opcode != I->getOpcode()) { + if (canCombineAsAltInst(Opcode) && i == 1) + return isAltInst(VL); return 0; + } } return Opcode; } @@ -185,6 +187,10 @@ static Instruction *propagateMetadata(Instruction *I, ArrayRef VL) { case LLVMContext::MD_tbaa: MD = MDNode::getMostGenericTBAA(MD, IMD); break; + case LLVMContext::MD_alias_scope: + case LLVMContext::MD_noalias: + MD = MDNode::intersect(MD, IMD); + break; case LLVMContext::MD_fpmath: MD = MDNode::getMostGenericFPMath(MD, IMD); break; @@ -347,13 +353,18 @@ public: BoUpSLP(Function *Func, ScalarEvolution *Se, const DataLayout *Dl, TargetTransformInfo *Tti, TargetLibraryInfo *TLi, AliasAnalysis *Aa, LoopInfo *Li, DominatorTree *Dt) - : F(Func), SE(Se), DL(Dl), TTI(Tti), TLI(TLi), AA(Aa), LI(Li), DT(Dt), + : NumLoadsWantToKeepOrder(0), NumLoadsWantToChangeOrder(0), + F(Func), SE(Se), DL(Dl), TTI(Tti), TLI(TLi), AA(Aa), LI(Li), DT(Dt), Builder(Se->getContext()) {} /// \brief Vectorize the tree that starts with the elements in \p VL. /// Returns the vectorized root. Value *vectorizeTree(); + /// \returns the cost incurred by unwanted spills and fills, caused by + /// holding live values over call sites. + int getSpillCost(); + /// \returns the vectorization cost of the subtree that starts at \p VL. /// A negative number means that this is profitable. int getTreeCost(); @@ -369,7 +380,12 @@ public: ScalarToTreeEntry.clear(); MustGather.clear(); ExternalUses.clear(); - MemBarrierIgnoreList.clear(); + NumLoadsWantToKeepOrder = 0; + NumLoadsWantToChangeOrder = 0; + for (auto &Iter : BlocksSchedules) { + BlockScheduling *BS = Iter.second.get(); + BS->clear(); + } } /// \returns true if the memory operations A and B are consecutive. @@ -377,6 +393,12 @@ public: /// \brief Perform LICM and CSE on the newly generated gather sequences. void optimizeGatherSequence(); + + /// \returns true if it is benefitial to reverse the vector order. + bool shouldReorder() const { + return NumLoadsWantToChangeOrder > NumLoadsWantToKeepOrder; + } + private: struct TreeEntry; @@ -413,20 +435,6 @@ private: /// roots. This method calculates the cost of extracting the values. int getGatherCost(ArrayRef VL); - /// \returns the AA location that is being access by the instruction. - AliasAnalysis::Location getLocation(Instruction *I); - - /// \brief Checks if it is possible to sink an instruction from - /// \p Src to \p Dst. - /// \returns the pointer to the barrier instruction if we can't sink. - Value *getSinkBarrier(Instruction *Src, Instruction *Dst); - - /// \returns the index of the last instruction in the BB from \p VL. - int getLastIndex(ArrayRef VL); - - /// \returns the Instruction in the bundle \p VL. - Instruction *getLastInstruction(ArrayRef VL); - /// \brief Set the Builder insert point to one after the last instruction in /// the bundle void setInsertPointAfterBundle(ArrayRef VL); @@ -439,7 +447,7 @@ private: bool isFullyVectorizableTinyTree(); struct TreeEntry { - TreeEntry() : Scalars(), VectorizedValue(nullptr), LastScalarIndex(0), + TreeEntry() : Scalars(), VectorizedValue(nullptr), NeedToGather(0) {} /// \returns true if the scalars in VL are equal to this entry. @@ -454,9 +462,6 @@ private: /// The Scalars are vectorized into this value. It is initialized to Null. Value *VectorizedValue; - /// The index in the basic block of the last scalar. - int LastScalarIndex; - /// Do we need to gather this sequence ? bool NeedToGather; }; @@ -469,18 +474,16 @@ private: Last->Scalars.insert(Last->Scalars.begin(), VL.begin(), VL.end()); Last->NeedToGather = !Vectorized; if (Vectorized) { - Last->LastScalarIndex = getLastIndex(VL); for (int i = 0, e = VL.size(); i != e; ++i) { assert(!ScalarToTreeEntry.count(VL[i]) && "Scalar already in tree!"); ScalarToTreeEntry[VL[i]] = idx; } } else { - Last->LastScalarIndex = 0; MustGather.insert(VL.begin(), VL.end()); } return Last; } - + /// -- Vectorization State -- /// Holds all of the tree entries. std::vector VectorizableTree; @@ -508,28 +511,316 @@ private: /// This list holds pairs of (Internal Scalar : External User). UserList ExternalUses; - /// A list of instructions to ignore while sinking - /// memory instructions. This map must be reset between runs of getCost. - ValueSet MemBarrierIgnoreList; - /// Holds all of the instructions that we gathered. SetVector GatherSeq; /// A list of blocks that we are going to CSE. SetVector CSEBlocks; - /// Numbers instructions in different blocks. - DenseMap BlocksNumbers; + /// Contains all scheduling relevant data for an instruction. + /// A ScheduleData either represents a single instruction or a member of an + /// instruction bundle (= a group of instructions which is combined into a + /// vector instruction). + struct ScheduleData { + + // The initial value for the dependency counters. It means that the + // dependencies are not calculated yet. + enum { InvalidDeps = -1 }; + + ScheduleData() + : Inst(nullptr), FirstInBundle(nullptr), NextInBundle(nullptr), + NextLoadStore(nullptr), SchedulingRegionID(0), SchedulingPriority(0), + Dependencies(InvalidDeps), UnscheduledDeps(InvalidDeps), + UnscheduledDepsInBundle(InvalidDeps), IsScheduled(false) {} + + void init(int BlockSchedulingRegionID) { + FirstInBundle = this; + NextInBundle = nullptr; + NextLoadStore = nullptr; + IsScheduled = false; + SchedulingRegionID = BlockSchedulingRegionID; + UnscheduledDepsInBundle = UnscheduledDeps; + clearDependencies(); + } + + /// Returns true if the dependency information has been calculated. + bool hasValidDependencies() const { return Dependencies != InvalidDeps; } + + /// Returns true for single instructions and for bundle representatives + /// (= the head of a bundle). + bool isSchedulingEntity() const { return FirstInBundle == this; } + + /// Returns true if it represents an instruction bundle and not only a + /// single instruction. + bool isPartOfBundle() const { + return NextInBundle != nullptr || FirstInBundle != this; + } + + /// Returns true if it is ready for scheduling, i.e. it has no more + /// unscheduled depending instructions/bundles. + bool isReady() const { + assert(isSchedulingEntity() && + "can't consider non-scheduling entity for ready list"); + return UnscheduledDepsInBundle == 0 && !IsScheduled; + } + + /// Modifies the number of unscheduled dependencies, also updating it for + /// the whole bundle. + int incrementUnscheduledDeps(int Incr) { + UnscheduledDeps += Incr; + return FirstInBundle->UnscheduledDepsInBundle += Incr; + } + + /// Sets the number of unscheduled dependencies to the number of + /// dependencies. + void resetUnscheduledDeps() { + incrementUnscheduledDeps(Dependencies - UnscheduledDeps); + } + + /// Clears all dependency information. + void clearDependencies() { + Dependencies = InvalidDeps; + resetUnscheduledDeps(); + MemoryDependencies.clear(); + } + + void dump(raw_ostream &os) const { + if (!isSchedulingEntity()) { + os << "/ " << *Inst; + } else if (NextInBundle) { + os << '[' << *Inst; + ScheduleData *SD = NextInBundle; + while (SD) { + os << ';' << *SD->Inst; + SD = SD->NextInBundle; + } + os << ']'; + } else { + os << *Inst; + } + } - /// \brief Get the corresponding instruction numbering list for a given - /// BasicBlock. The list is allocated lazily. - BlockNumbering &getBlockNumbering(BasicBlock *BB) { - auto I = BlocksNumbers.insert(std::make_pair(BB, BlockNumbering(BB))); - return I.first->second; - } + Instruction *Inst; + + /// Points to the head in an instruction bundle (and always to this for + /// single instructions). + ScheduleData *FirstInBundle; + + /// Single linked list of all instructions in a bundle. Null if it is a + /// single instruction. + ScheduleData *NextInBundle; + + /// Single linked list of all memory instructions (e.g. load, store, call) + /// in the block - until the end of the scheduling region. + ScheduleData *NextLoadStore; + + /// The dependent memory instructions. + /// This list is derived on demand in calculateDependencies(). + SmallVector MemoryDependencies; + + /// This ScheduleData is in the current scheduling region if this matches + /// the current SchedulingRegionID of BlockScheduling. + int SchedulingRegionID; + + /// Used for getting a "good" final ordering of instructions. + int SchedulingPriority; + + /// The number of dependencies. Constitutes of the number of users of the + /// instruction plus the number of dependent memory instructions (if any). + /// This value is calculated on demand. + /// If InvalidDeps, the number of dependencies is not calculated yet. + /// + int Dependencies; + + /// The number of dependencies minus the number of dependencies of scheduled + /// instructions. As soon as this is zero, the instruction/bundle gets ready + /// for scheduling. + /// Note that this is negative as long as Dependencies is not calculated. + int UnscheduledDeps; + + /// The sum of UnscheduledDeps in a bundle. Equals to UnscheduledDeps for + /// single instructions. + int UnscheduledDepsInBundle; + + /// True if this instruction is scheduled (or considered as scheduled in the + /// dry-run). + bool IsScheduled; + }; + +#ifndef NDEBUG + friend raw_ostream &operator<<(raw_ostream &os, + const BoUpSLP::ScheduleData &SD); +#endif + + /// Contains all scheduling data for a basic block. + /// + struct BlockScheduling { + + BlockScheduling(BasicBlock *BB) + : BB(BB), ChunkSize(BB->size()), ChunkPos(ChunkSize), + ScheduleStart(nullptr), ScheduleEnd(nullptr), + FirstLoadStoreInRegion(nullptr), LastLoadStoreInRegion(nullptr), + // Make sure that the initial SchedulingRegionID is greater than the + // initial SchedulingRegionID in ScheduleData (which is 0). + SchedulingRegionID(1) {} + + void clear() { + ReadyInsts.clear(); + ScheduleStart = nullptr; + ScheduleEnd = nullptr; + FirstLoadStoreInRegion = nullptr; + LastLoadStoreInRegion = nullptr; + + // Make a new scheduling region, i.e. all existing ScheduleData is not + // in the new region yet. + ++SchedulingRegionID; + } + + ScheduleData *getScheduleData(Value *V) { + ScheduleData *SD = ScheduleDataMap[V]; + if (SD && SD->SchedulingRegionID == SchedulingRegionID) + return SD; + return nullptr; + } + + bool isInSchedulingRegion(ScheduleData *SD) { + return SD->SchedulingRegionID == SchedulingRegionID; + } + + /// Marks an instruction as scheduled and puts all dependent ready + /// instructions into the ready-list. + template + void schedule(ScheduleData *SD, ReadyListType &ReadyList) { + SD->IsScheduled = true; + DEBUG(dbgs() << "SLP: schedule " << *SD << "\n"); + + ScheduleData *BundleMember = SD; + while (BundleMember) { + // Handle the def-use chain dependencies. + for (Use &U : BundleMember->Inst->operands()) { + ScheduleData *OpDef = getScheduleData(U.get()); + if (OpDef && OpDef->hasValidDependencies() && + OpDef->incrementUnscheduledDeps(-1) == 0) { + // There are no more unscheduled dependencies after decrementing, + // so we can put the dependent instruction into the ready list. + ScheduleData *DepBundle = OpDef->FirstInBundle; + assert(!DepBundle->IsScheduled && + "already scheduled bundle gets ready"); + ReadyList.insert(DepBundle); + DEBUG(dbgs() << "SLP: gets ready (def): " << *DepBundle << "\n"); + } + } + // Handle the memory dependencies. + for (ScheduleData *MemoryDepSD : BundleMember->MemoryDependencies) { + if (MemoryDepSD->incrementUnscheduledDeps(-1) == 0) { + // There are no more unscheduled dependencies after decrementing, + // so we can put the dependent instruction into the ready list. + ScheduleData *DepBundle = MemoryDepSD->FirstInBundle; + assert(!DepBundle->IsScheduled && + "already scheduled bundle gets ready"); + ReadyList.insert(DepBundle); + DEBUG(dbgs() << "SLP: gets ready (mem): " << *DepBundle << "\n"); + } + } + BundleMember = BundleMember->NextInBundle; + } + } + + /// Put all instructions into the ReadyList which are ready for scheduling. + template + void initialFillReadyList(ReadyListType &ReadyList) { + for (auto *I = ScheduleStart; I != ScheduleEnd; I = I->getNextNode()) { + ScheduleData *SD = getScheduleData(I); + if (SD->isSchedulingEntity() && SD->isReady()) { + ReadyList.insert(SD); + DEBUG(dbgs() << "SLP: initially in ready list: " << *I << "\n"); + } + } + } + + /// Checks if a bundle of instructions can be scheduled, i.e. has no + /// cyclic dependencies. This is only a dry-run, no instructions are + /// actually moved at this stage. + bool tryScheduleBundle(ArrayRef VL, AliasAnalysis *AA); + + /// Un-bundles a group of instructions. + void cancelScheduling(ArrayRef VL); + + /// Extends the scheduling region so that V is inside the region. + void extendSchedulingRegion(Value *V); + + /// Initialize the ScheduleData structures for new instructions in the + /// scheduling region. + void initScheduleData(Instruction *FromI, Instruction *ToI, + ScheduleData *PrevLoadStore, + ScheduleData *NextLoadStore); + + /// Updates the dependency information of a bundle and of all instructions/ + /// bundles which depend on the original bundle. + void calculateDependencies(ScheduleData *SD, bool InsertInReadyList, + AliasAnalysis *AA); + + /// Sets all instruction in the scheduling region to un-scheduled. + void resetSchedule(); + + BasicBlock *BB; + + /// Simple memory allocation for ScheduleData. + std::vector> ScheduleDataChunks; + + /// The size of a ScheduleData array in ScheduleDataChunks. + int ChunkSize; + + /// The allocator position in the current chunk, which is the last entry + /// of ScheduleDataChunks. + int ChunkPos; + + /// Attaches ScheduleData to Instruction. + /// Note that the mapping survives during all vectorization iterations, i.e. + /// ScheduleData structures are recycled. + DenseMap ScheduleDataMap; + + struct ReadyList : SmallVector { + void insert(ScheduleData *SD) { push_back(SD); } + }; + + /// The ready-list for scheduling (only used for the dry-run). + ReadyList ReadyInsts; + + /// The first instruction of the scheduling region. + Instruction *ScheduleStart; + + /// The first instruction _after_ the scheduling region. + Instruction *ScheduleEnd; + + /// The first memory accessing instruction in the scheduling region + /// (can be null). + ScheduleData *FirstLoadStoreInRegion; + + /// The last memory accessing instruction in the scheduling region + /// (can be null). + ScheduleData *LastLoadStoreInRegion; + + /// The ID of the scheduling region. For a new vectorization iteration this + /// is incremented which "removes" all ScheduleData from the region. + int SchedulingRegionID; + }; + + /// Attaches the BlockScheduling structures to basic blocks. + DenseMap> BlocksSchedules; + + /// Performs the "real" scheduling. Done before vectorization is actually + /// performed in a basic block. + void scheduleBlock(BlockScheduling *BS); /// List of users to ignore during scheduling and that don't need extracting. ArrayRef UserIgnoreList; + // Number of load-bundles, which contain consecutive loads. + int NumLoadsWantToKeepOrder; + + // Number of load-bundles of size 2, which are consecutive loads if reversed. + int NumLoadsWantToChangeOrder; + // Analysis and block reference. Function *F; ScalarEvolution *SE; @@ -543,6 +834,13 @@ private: IRBuilder<> Builder; }; +#ifndef NDEBUG +raw_ostream &operator<<(raw_ostream &os, const BoUpSLP::ScheduleData &SD) { + SD.dump(os); + return os; +} +#endif + void BoUpSLP::buildTree(ArrayRef Roots, ArrayRef UserIgnoreLst) { deleteTree(); @@ -594,6 +892,7 @@ void BoUpSLP::buildTree(ArrayRef Roots, void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { bool SameTy = getSameType(VL); (void)SameTy; + bool isAltShuffle = false; assert(SameTy && "Invalid types!"); if (Depth == RecursionMaxDepth) { @@ -615,10 +914,19 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { newTreeEntry(VL, false); return; } + unsigned Opcode = getSameOpcode(VL); + + // Check that this shuffle vector refers to the alternate + // sequence of opcodes. + if (Opcode == Instruction::ShuffleVector) { + Instruction *I0 = dyn_cast(VL[0]); + unsigned Op = I0->getOpcode(); + if (Op != Instruction::ShuffleVector) + isAltShuffle = true; + } // If all of the operands are identical or constant we have a simple solution. - if (allConstant(VL) || isSplat(VL) || !getSameBlock(VL) || - !getSameOpcode(VL)) { + if (allConstant(VL) || isSplat(VL) || !getSameBlock(VL) || !Opcode) { DEBUG(dbgs() << "SLP: Gathering due to C,S,B,O. \n"); newTreeEntry(VL, false); return; @@ -666,69 +974,16 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { // Check that all of the users of the scalars that we want to vectorize are // schedulable. Instruction *VL0 = cast(VL[0]); - int MyLastIndex = getLastIndex(VL); BasicBlock *BB = cast(VL0)->getParent(); - for (unsigned i = 0, e = VL.size(); i != e; ++i) { - Instruction *Scalar = cast(VL[i]); - DEBUG(dbgs() << "SLP: Checking users of " << *Scalar << ". \n"); - for (User *U : Scalar->users()) { - DEBUG(dbgs() << "SLP: \tUser " << *U << ". \n"); - Instruction *UI = dyn_cast(U); - if (!UI) { - DEBUG(dbgs() << "SLP: Gathering due unknown user. \n"); - newTreeEntry(VL, false); - return; - } - - // We don't care if the user is in a different basic block. - BasicBlock *UserBlock = UI->getParent(); - if (UserBlock != BB) { - DEBUG(dbgs() << "SLP: User from a different basic block " - << *UI << ". \n"); - continue; - } - - // If this is a PHINode within this basic block then we can place the - // extract wherever we want. - if (isa(*UI)) { - DEBUG(dbgs() << "SLP: \tWe can schedule PHIs:" << *UI << ". \n"); - continue; - } - - // Check if this is a safe in-tree user. - if (ScalarToTreeEntry.count(UI)) { - int Idx = ScalarToTreeEntry[UI]; - int VecLocation = VectorizableTree[Idx].LastScalarIndex; - if (VecLocation <= MyLastIndex) { - DEBUG(dbgs() << "SLP: Gathering due to unschedulable vector. \n"); - newTreeEntry(VL, false); - return; - } - DEBUG(dbgs() << "SLP: In-tree user (" << *UI << ") at #" << - VecLocation << " vector value (" << *Scalar << ") at #" - << MyLastIndex << ".\n"); - continue; - } - - // Ignore users in the user ignore list. - if (std::find(UserIgnoreList.begin(), UserIgnoreList.end(), UI) != - UserIgnoreList.end()) - continue; - - // Make sure that we can schedule this unknown user. - BlockNumbering &BN = getBlockNumbering(BB); - int UserIndex = BN.getIndex(UI); - if (UserIndex < MyLastIndex) { - - DEBUG(dbgs() << "SLP: Can't schedule extractelement for " - << *UI << ". \n"); - newTreeEntry(VL, false); - return; - } - } + if (!DT->isReachableFromEntry(BB)) { + // Don't go into unreachable blocks. They may contain instructions with + // dependency cycles which confuse the final scheduling. + DEBUG(dbgs() << "SLP: bundle in unreachable block.\n"); + newTreeEntry(VL, false); + return; } - + // Check that every instructions appears once in this bundle. for (unsigned i = 0, e = VL.size(); i < e; ++i) for (unsigned j = i+1; j < e; ++j) @@ -738,40 +993,19 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { return; } - // Check that instructions in this bundle don't reference other instructions. - // The runtime of this check is O(N * N-1 * uses(N)) and a typical N is 4. - for (unsigned i = 0, e = VL.size(); i < e; ++i) { - for (User *U : VL[i]->users()) { - for (unsigned j = 0; j < e; ++j) { - if (i != j && U == VL[j]) { - DEBUG(dbgs() << "SLP: Intra-bundle dependencies!" << *U << ". \n"); - newTreeEntry(VL, false); - return; - } - } - } + auto &BSRef = BlocksSchedules[BB]; + if (!BSRef) { + BSRef = llvm::make_unique(BB); } + BlockScheduling &BS = *BSRef.get(); - DEBUG(dbgs() << "SLP: We are able to schedule this bundle.\n"); - - unsigned Opcode = getSameOpcode(VL); - - // Check if it is safe to sink the loads or the stores. - if (Opcode == Instruction::Load || Opcode == Instruction::Store) { - Instruction *Last = getLastInstruction(VL); - - for (unsigned i = 0, e = VL.size(); i < e; ++i) { - if (VL[i] == Last) - continue; - Value *Barrier = getSinkBarrier(cast(VL[i]), Last); - if (Barrier) { - DEBUG(dbgs() << "SLP: Can't sink " << *VL[i] << "\n down to " << *Last - << "\n because of " << *Barrier << ". Gathering.\n"); - newTreeEntry(VL, false); - return; - } - } + if (!BS.tryScheduleBundle(VL, AA)) { + DEBUG(dbgs() << "SLP: We are not able to schedule this bundle!\n"); + BS.cancelScheduling(VL); + newTreeEntry(VL, false); + return; } + DEBUG(dbgs() << "SLP: We are able to schedule this bundle.\n"); switch (Opcode) { case Instruction::PHI: { @@ -784,6 +1018,7 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { cast(VL[j])->getIncomingValueForBlock(PH->getIncomingBlock(i))); if (Term) { DEBUG(dbgs() << "SLP: Need to swizzle PHINodes (TerminatorInst use).\n"); + BS.cancelScheduling(VL); newTreeEntry(VL, false); return; } @@ -807,6 +1042,8 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { bool Reuse = CanReuseExtract(VL); if (Reuse) { DEBUG(dbgs() << "SLP: Reusing extract sequence.\n"); + } else { + BS.cancelScheduling(VL); } newTreeEntry(VL, Reuse); return; @@ -815,12 +1052,23 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { // Check if the loads are consecutive or of we need to swizzle them. for (unsigned i = 0, e = VL.size() - 1; i < e; ++i) { LoadInst *L = cast(VL[i]); - if (!L->isSimple() || !isConsecutiveAccess(VL[i], VL[i + 1])) { + if (!L->isSimple()) { + BS.cancelScheduling(VL); newTreeEntry(VL, false); - DEBUG(dbgs() << "SLP: Need to swizzle loads.\n"); + DEBUG(dbgs() << "SLP: Gathering non-simple loads.\n"); + return; + } + if (!isConsecutiveAccess(VL[i], VL[i + 1])) { + if (VL.size() == 2 && isConsecutiveAccess(VL[1], VL[0])) { + ++NumLoadsWantToChangeOrder; + } + BS.cancelScheduling(VL); + newTreeEntry(VL, false); + DEBUG(dbgs() << "SLP: Gathering non-consecutive loads.\n"); return; } } + ++NumLoadsWantToKeepOrder; newTreeEntry(VL, true); DEBUG(dbgs() << "SLP: added a vector of loads.\n"); return; @@ -841,6 +1089,7 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { for (unsigned i = 0; i < VL.size(); ++i) { Type *Ty = cast(VL[i])->getOperand(0)->getType(); if (Ty != SrcTy || Ty->isAggregateType() || Ty->isVectorTy()) { + BS.cancelScheduling(VL); newTreeEntry(VL, false); DEBUG(dbgs() << "SLP: Gathering casts with different src types.\n"); return; @@ -868,6 +1117,7 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { CmpInst *Cmp = cast(VL[i]); if (Cmp->getPredicate() != P0 || Cmp->getOperand(0)->getType() != ComparedTy) { + BS.cancelScheduling(VL); newTreeEntry(VL, false); DEBUG(dbgs() << "SLP: Gathering cmp with different predicate.\n"); return; @@ -929,10 +1179,59 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { } return; } + case Instruction::GetElementPtr: { + // We don't combine GEPs with complicated (nested) indexing. + for (unsigned j = 0; j < VL.size(); ++j) { + if (cast(VL[j])->getNumOperands() != 2) { + DEBUG(dbgs() << "SLP: not-vectorizable GEP (nested indexes).\n"); + BS.cancelScheduling(VL); + newTreeEntry(VL, false); + return; + } + } + + // We can't combine several GEPs into one vector if they operate on + // different types. + Type *Ty0 = cast(VL0)->getOperand(0)->getType(); + for (unsigned j = 0; j < VL.size(); ++j) { + Type *CurTy = cast(VL[j])->getOperand(0)->getType(); + if (Ty0 != CurTy) { + DEBUG(dbgs() << "SLP: not-vectorizable GEP (different types).\n"); + BS.cancelScheduling(VL); + newTreeEntry(VL, false); + return; + } + } + + // We don't combine GEPs with non-constant indexes. + for (unsigned j = 0; j < VL.size(); ++j) { + auto Op = cast(VL[j])->getOperand(1); + if (!isa(Op)) { + DEBUG( + dbgs() << "SLP: not-vectorizable GEP (non-constant indexes).\n"); + BS.cancelScheduling(VL); + newTreeEntry(VL, false); + return; + } + } + + newTreeEntry(VL, true); + DEBUG(dbgs() << "SLP: added a vector of GEPs.\n"); + for (unsigned i = 0, e = 2; i < e; ++i) { + ValueList Operands; + // Prepare the operand vector. + for (unsigned j = 0; j < VL.size(); ++j) + Operands.push_back(cast(VL[j])->getOperand(i)); + + buildTree_rec(Operands, Depth + 1); + } + return; + } case Instruction::Store: { // Check if the stores are consecutive or of we need to swizzle them. for (unsigned i = 0, e = VL.size() - 1; i < e; ++i) if (!isConsecutiveAccess(VL[i], VL[i + 1])) { + BS.cancelScheduling(VL); newTreeEntry(VL, false); DEBUG(dbgs() << "SLP: Non-consecutive store.\n"); return; @@ -945,8 +1244,6 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { for (unsigned j = 0; j < VL.size(); ++j) Operands.push_back(cast(VL[j])->getOperand(0)); - // We can ignore these values because we are sinking them down. - MemBarrierIgnoreList.insert(VL.begin(), VL.end()); buildTree_rec(Operands, Depth + 1); return; } @@ -957,22 +1254,38 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { // represented by an intrinsic call Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI); if (!isTriviallyVectorizable(ID)) { + BS.cancelScheduling(VL); newTreeEntry(VL, false); DEBUG(dbgs() << "SLP: Non-vectorizable call.\n"); return; } - Function *Int = CI->getCalledFunction(); - + Value *A1I = nullptr; + if (hasVectorInstrinsicScalarOpd(ID, 1)) + A1I = CI->getArgOperand(1); for (unsigned i = 1, e = VL.size(); i != e; ++i) { CallInst *CI2 = dyn_cast(VL[i]); if (!CI2 || CI2->getCalledFunction() != Int || getIntrinsicIDForCall(CI2, TLI) != ID) { + BS.cancelScheduling(VL); newTreeEntry(VL, false); DEBUG(dbgs() << "SLP: mismatched calls:" << *CI << "!=" << *VL[i] << "\n"); return; } + // ctlz,cttz and powi are special intrinsics whose second argument + // should be same in order for them to be vectorized. + if (hasVectorInstrinsicScalarOpd(ID, 1)) { + Value *A1J = CI2->getArgOperand(1); + if (A1I != A1J) { + BS.cancelScheduling(VL); + newTreeEntry(VL, false); + DEBUG(dbgs() << "SLP: mismatched arguments in call:" << *CI + << " argument "<< A1I<<"!=" << A1J + << "\n"); + return; + } + } } newTreeEntry(VL, true); @@ -987,7 +1300,29 @@ void BoUpSLP::buildTree_rec(ArrayRef VL, unsigned Depth) { } return; } + case Instruction::ShuffleVector: { + // If this is not an alternate sequence of opcode like add-sub + // then do not vectorize this instruction. + if (!isAltShuffle) { + BS.cancelScheduling(VL); + newTreeEntry(VL, false); + DEBUG(dbgs() << "SLP: ShuffleVector are not vectorized.\n"); + return; + } + newTreeEntry(VL, true); + DEBUG(dbgs() << "SLP: added a ShuffleVector op.\n"); + for (unsigned i = 0, e = VL0->getNumOperands(); i < e; ++i) { + ValueList Operands; + // Prepare the operand vector. + for (unsigned j = 0; j < VL.size(); ++j) + Operands.push_back(cast(VL[j])->getOperand(i)); + + buildTree_rec(Operands, Depth + 1); + } + return; + } default: + BS.cancelScheduling(VL); newTreeEntry(VL, false); DEBUG(dbgs() << "SLP: Gathering unknown instruction.\n"); return; @@ -1010,11 +1345,9 @@ int BoUpSLP::getEntryCost(TreeEntry *E) { } return getGatherCost(E->Scalars); } - - assert(getSameOpcode(VL) && getSameType(VL) && getSameBlock(VL) && - "Invalid VL"); + unsigned Opcode = getSameOpcode(VL); + assert(Opcode && getSameType(VL) && getSameBlock(VL) && "Invalid VL"); Instruction *VL0 = cast(VL[0]); - unsigned Opcode = VL0->getOpcode(); switch (Opcode) { case Instruction::PHI: { return 0; @@ -1121,6 +1454,20 @@ int BoUpSLP::getEntryCost(TreeEntry *E) { } return VecCost - ScalarCost; } + case Instruction::GetElementPtr: { + TargetTransformInfo::OperandValueKind Op1VK = + TargetTransformInfo::OK_AnyValue; + TargetTransformInfo::OperandValueKind Op2VK = + TargetTransformInfo::OK_UniformConstantValue; + + int ScalarCost = + VecTy->getNumElements() * + TTI->getArithmeticInstrCost(Instruction::Add, ScalarTy, Op1VK, Op2VK); + int VecCost = + TTI->getArithmeticInstrCost(Instruction::Add, VecTy, Op1VK, Op2VK); + + return VecCost - ScalarCost; + } case Instruction::Load: { // Cost of wide load - cost of scalar loads. int ScalarLdCost = VecTy->getNumElements() * @@ -1158,6 +1505,32 @@ int BoUpSLP::getEntryCost(TreeEntry *E) { return VecCallCost - ScalarCallCost; } + case Instruction::ShuffleVector: { + TargetTransformInfo::OperandValueKind Op1VK = + TargetTransformInfo::OK_AnyValue; + TargetTransformInfo::OperandValueKind Op2VK = + TargetTransformInfo::OK_AnyValue; + int ScalarCost = 0; + int VecCost = 0; + for (unsigned i = 0; i < VL.size(); ++i) { + Instruction *I = cast(VL[i]); + if (!I) + break; + ScalarCost += + TTI->getArithmeticInstrCost(I->getOpcode(), ScalarTy, Op1VK, Op2VK); + } + // VecCost is equal to sum of the cost of creating 2 vectors + // and the cost of creating shuffle. + Instruction *I0 = cast(VL[0]); + VecCost = + TTI->getArithmeticInstrCost(I0->getOpcode(), VecTy, Op1VK, Op2VK); + Instruction *I1 = cast(VL[1]); + VecCost += + TTI->getArithmeticInstrCost(I1->getOpcode(), VecTy, Op1VK, Op2VK); + VecCost += + TTI->getShuffleCost(TargetTransformInfo::SK_Alternate, VecTy, 0); + return VecCost - ScalarCost; + } default: llvm_unreachable("Unknown instruction"); } @@ -1182,6 +1555,68 @@ bool BoUpSLP::isFullyVectorizableTinyTree() { return true; } +int BoUpSLP::getSpillCost() { + // Walk from the bottom of the tree to the top, tracking which values are + // live. When we see a call instruction that is not part of our tree, + // query TTI to see if there is a cost to keeping values live over it + // (for example, if spills and fills are required). + unsigned BundleWidth = VectorizableTree.front().Scalars.size(); + int Cost = 0; + + SmallPtrSet LiveValues; + Instruction *PrevInst = nullptr; + + for (unsigned N = 0; N < VectorizableTree.size(); ++N) { + Instruction *Inst = dyn_cast(VectorizableTree[N].Scalars[0]); + if (!Inst) + continue; + + if (!PrevInst) { + PrevInst = Inst; + continue; + } + + DEBUG( + dbgs() << "SLP: #LV: " << LiveValues.size(); + for (auto *X : LiveValues) + dbgs() << " " << X->getName(); + dbgs() << ", Looking at "; + Inst->dump(); + ); + + // Update LiveValues. + LiveValues.erase(PrevInst); + for (auto &J : PrevInst->operands()) { + if (isa(&*J) && ScalarToTreeEntry.count(&*J)) + LiveValues.insert(cast(&*J)); + } + + // Now find the sequence of instructions between PrevInst and Inst. + BasicBlock::reverse_iterator InstIt(Inst), PrevInstIt(PrevInst); + --PrevInstIt; + while (InstIt != PrevInstIt) { + if (PrevInstIt == PrevInst->getParent()->rend()) { + PrevInstIt = Inst->getParent()->rbegin(); + continue; + } + + if (isa(&*PrevInstIt) && &*PrevInstIt != PrevInst) { + SmallVector V; + for (auto *II : LiveValues) + V.push_back(VectorType::get(II->getType(), BundleWidth)); + Cost += TTI->getCostOfKeepingLiveOverCall(V); + } + + ++PrevInstIt; + } + + PrevInst = Inst; + } + + DEBUG(dbgs() << "SLP: SpillCost=" << Cost << "\n"); + return Cost; +} + int BoUpSLP::getTreeCost() { int Cost = 0; DEBUG(dbgs() << "SLP: Calculating cost for tree of size " << @@ -1217,6 +1652,8 @@ int BoUpSLP::getTreeCost() { I->Lane); } + Cost += getSpillCost(); + DEBUG(dbgs() << "SLP: Total Cost " << Cost + ExtractCost<< ".\n"); return Cost + ExtractCost; } @@ -1238,14 +1675,6 @@ int BoUpSLP::getGatherCost(ArrayRef VL) { return getGatherCost(VecTy); } -AliasAnalysis::Location BoUpSLP::getLocation(Instruction *I) { - if (StoreInst *SI = dyn_cast(I)) - return AA->getLocation(SI); - if (LoadInst *LI = dyn_cast(I)) - return AA->getLocation(LI); - return AliasAnalysis::Location(); -} - Value *BoUpSLP::getPointerOperand(Value *I) { if (LoadInst *LI = dyn_cast(I)) return LI->getPointerOperand(); @@ -1303,59 +1732,9 @@ bool BoUpSLP::isConsecutiveAccess(Value *A, Value *B) { return X == PtrSCEVB; } -Value *BoUpSLP::getSinkBarrier(Instruction *Src, Instruction *Dst) { - assert(Src->getParent() == Dst->getParent() && "Not the same BB"); - BasicBlock::iterator I = Src, E = Dst; - /// Scan all of the instruction from SRC to DST and check if - /// the source may alias. - for (++I; I != E; ++I) { - // Ignore store instructions that are marked as 'ignore'. - if (MemBarrierIgnoreList.count(I)) - continue; - if (Src->mayWriteToMemory()) /* Write */ { - if (!I->mayReadOrWriteMemory()) - continue; - } else /* Read */ { - if (!I->mayWriteToMemory()) - continue; - } - AliasAnalysis::Location A = getLocation(&*I); - AliasAnalysis::Location B = getLocation(Src); - - if (!A.Ptr || !B.Ptr || AA->alias(A, B)) - return I; - } - return nullptr; -} - -int BoUpSLP::getLastIndex(ArrayRef VL) { - BasicBlock *BB = cast(VL[0])->getParent(); - assert(BB == getSameBlock(VL) && "Invalid block"); - BlockNumbering &BN = getBlockNumbering(BB); - - int MaxIdx = BN.getIndex(BB->getFirstNonPHI()); - for (unsigned i = 0, e = VL.size(); i < e; ++i) - MaxIdx = std::max(MaxIdx, BN.getIndex(cast(VL[i]))); - return MaxIdx; -} - -Instruction *BoUpSLP::getLastInstruction(ArrayRef VL) { - BasicBlock *BB = cast(VL[0])->getParent(); - assert(BB == getSameBlock(VL) && "Invalid block"); - BlockNumbering &BN = getBlockNumbering(BB); - - int MaxIdx = BN.getIndex(cast(VL[0])); - for (unsigned i = 1, e = VL.size(); i < e; ++i) - MaxIdx = std::max(MaxIdx, BN.getIndex(cast(VL[i]))); - Instruction *I = BN.getInstruction(MaxIdx); - assert(I && "bad location"); - return I; -} - void BoUpSLP::setInsertPointAfterBundle(ArrayRef VL) { Instruction *VL0 = cast(VL[0]); - Instruction *LastInst = getLastInstruction(VL); - BasicBlock::iterator NextInst = LastInst; + BasicBlock::iterator NextInst = VL0; ++NextInst; Builder.SetInsertPoint(VL0->getParent(), NextInst); Builder.SetCurrentDebugLocation(VL0->getDebugLoc()); @@ -1439,8 +1818,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { return Gather(E->Scalars, VecTy); } - unsigned Opcode = VL0->getOpcode(); - assert(Opcode == getSameOpcode(E->Scalars) && "Invalid opcode"); + unsigned Opcode = getSameOpcode(E->Scalars); switch (Opcode) { case Instruction::PHI: { @@ -1513,6 +1891,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { CastInst *CI = dyn_cast(VL0); Value *V = Builder.CreateCast(CI->getOpcode(), InVec, VecTy); E->VectorizedValue = V; + ++NumVectorInstructions; return V; } case Instruction::FCmp: @@ -1539,6 +1918,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { V = Builder.CreateICmp(P0, L, R); E->VectorizedValue = V; + ++NumVectorInstructions; return V; } case Instruction::Select: { @@ -1560,6 +1940,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { Value *V = Builder.CreateSelect(Cond, True, False); E->VectorizedValue = V; + ++NumVectorInstructions; return V; } case Instruction::Add: @@ -1604,6 +1985,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { BinaryOperator *BinOp = cast(VL0); Value *V = Builder.CreateBinOp(BinOp->getOpcode(), LHS, RHS); E->VectorizedValue = V; + ++NumVectorInstructions; if (Instruction *I = dyn_cast(V)) return propagateMetadata(I, E->Scalars); @@ -1616,6 +1998,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { setInsertPointAfterBundle(E->Scalars); LoadInst *LI = cast(VL0); + Type *ScalarLoadTy = LI->getType(); unsigned AS = LI->getPointerAddressSpace(); Value *VecPtr = Builder.CreateBitCast(LI->getPointerOperand(), @@ -1623,9 +2006,10 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { unsigned Alignment = LI->getAlignment(); LI = Builder.CreateLoad(VecPtr); if (!Alignment) - Alignment = DL->getABITypeAlignment(LI->getPointerOperand()->getType()); + Alignment = DL->getABITypeAlignment(ScalarLoadTy); LI->setAlignment(Alignment); E->VectorizedValue = LI; + ++NumVectorInstructions; return propagateMetadata(LI, E->Scalars); } case Instruction::Store: { @@ -1644,17 +2028,59 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { VecTy->getPointerTo(AS)); StoreInst *S = Builder.CreateStore(VecValue, VecPtr); if (!Alignment) - Alignment = DL->getABITypeAlignment(SI->getPointerOperand()->getType()); + Alignment = DL->getABITypeAlignment(SI->getValueOperand()->getType()); S->setAlignment(Alignment); E->VectorizedValue = S; + ++NumVectorInstructions; return propagateMetadata(S, E->Scalars); } + case Instruction::GetElementPtr: { + setInsertPointAfterBundle(E->Scalars); + + ValueList Op0VL; + for (int i = 0, e = E->Scalars.size(); i < e; ++i) + Op0VL.push_back(cast(E->Scalars[i])->getOperand(0)); + + Value *Op0 = vectorizeTree(Op0VL); + + std::vector OpVecs; + for (int j = 1, e = cast(VL0)->getNumOperands(); j < e; + ++j) { + ValueList OpVL; + for (int i = 0, e = E->Scalars.size(); i < e; ++i) + OpVL.push_back(cast(E->Scalars[i])->getOperand(j)); + + Value *OpVec = vectorizeTree(OpVL); + OpVecs.push_back(OpVec); + } + + Value *V = Builder.CreateGEP(Op0, OpVecs); + E->VectorizedValue = V; + ++NumVectorInstructions; + + if (Instruction *I = dyn_cast(V)) + return propagateMetadata(I, E->Scalars); + + return V; + } case Instruction::Call: { CallInst *CI = cast(VL0); setInsertPointAfterBundle(E->Scalars); + Function *FI; + Intrinsic::ID IID = Intrinsic::not_intrinsic; + if (CI && (FI = CI->getCalledFunction())) { + IID = (Intrinsic::ID) FI->getIntrinsicID(); + } std::vector OpVecs; for (int j = 0, e = CI->getNumArgOperands(); j < e; ++j) { ValueList OpVL; + // ctlz,cttz and powi are special intrinsics whose second argument is + // a scalar. This argument should not be vectorized. + if (hasVectorInstrinsicScalarOpd(IID, 1) && j == 1) { + CallInst *CEI = cast(E->Scalars[0]); + OpVecs.push_back(CEI->getArgOperand(j)); + continue; + } for (int i = 0, e = E->Scalars.size(); i < e; ++i) { CallInst *CEI = cast(E->Scalars[i]); OpVL.push_back(CEI->getArgOperand(j)); @@ -1671,6 +2097,51 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { Function *CF = Intrinsic::getDeclaration(M, ID, Tys); Value *V = Builder.CreateCall(CF, OpVecs); E->VectorizedValue = V; + ++NumVectorInstructions; + return V; + } + case Instruction::ShuffleVector: { + ValueList LHSVL, RHSVL; + for (int i = 0, e = E->Scalars.size(); i < e; ++i) { + LHSVL.push_back(cast(E->Scalars[i])->getOperand(0)); + RHSVL.push_back(cast(E->Scalars[i])->getOperand(1)); + } + setInsertPointAfterBundle(E->Scalars); + + Value *LHS = vectorizeTree(LHSVL); + Value *RHS = vectorizeTree(RHSVL); + + if (Value *V = alreadyVectorized(E->Scalars)) + return V; + + // Create a vector of LHS op1 RHS + BinaryOperator *BinOp0 = cast(VL0); + Value *V0 = Builder.CreateBinOp(BinOp0->getOpcode(), LHS, RHS); + + // Create a vector of LHS op2 RHS + Instruction *VL1 = cast(E->Scalars[1]); + BinaryOperator *BinOp1 = cast(VL1); + Value *V1 = Builder.CreateBinOp(BinOp1->getOpcode(), LHS, RHS); + + // Create appropriate shuffle to take alternative operations from + // the vector. + std::vector Mask(E->Scalars.size()); + unsigned e = E->Scalars.size(); + for (unsigned i = 0; i < e; ++i) { + if (i & 1) + Mask[i] = Builder.getInt32(e + i); + else + Mask[i] = Builder.getInt32(i); + } + + Value *ShuffleMask = ConstantVector::get(Mask); + + Value *V = Builder.CreateShuffleVector(V0, V1, ShuffleMask); + E->VectorizedValue = V; + ++NumVectorInstructions; + if (Instruction *I = dyn_cast(V)) + return propagateMetadata(I, E->Scalars); + return V; } default: @@ -1680,6 +2151,12 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) { } Value *BoUpSLP::vectorizeTree() { + + // All blocks must be scheduled before any instructions are inserted. + for (auto &BSIter : BlocksSchedules) { + scheduleBlock(BSIter.second.get()); + } + Builder.SetInsertPoint(F->getEntryBlock().begin()); vectorizeTree(&VectorizableTree[0]); @@ -1741,7 +2218,6 @@ Value *BoUpSLP::vectorizeTree() { // For each lane: for (int Lane = 0, LE = Entry->Scalars.size(); Lane != LE; ++Lane) { Value *Scalar = Entry->Scalars[Lane]; - // No need to handle users of gathered values. if (Entry->NeedToGather) continue; @@ -1769,9 +2245,6 @@ Value *BoUpSLP::vectorizeTree() { } } - for (auto &BN : BlocksNumbers) - BN.second.forget(); - Builder.ClearInsertionPoint(); return VectorizableTree[0].VectorizedValue; @@ -1865,6 +2338,363 @@ void BoUpSLP::optimizeGatherSequence() { GatherSeq.clear(); } +// Groups the instructions to a bundle (which is then a single scheduling entity) +// and schedules instructions until the bundle gets ready. +bool BoUpSLP::BlockScheduling::tryScheduleBundle(ArrayRef VL, + AliasAnalysis *AA) { + if (isa(VL[0])) + return true; + + // Initialize the instruction bundle. + Instruction *OldScheduleEnd = ScheduleEnd; + ScheduleData *PrevInBundle = nullptr; + ScheduleData *Bundle = nullptr; + bool ReSchedule = false; + DEBUG(dbgs() << "SLP: bundle: " << *VL[0] << "\n"); + for (Value *V : VL) { + extendSchedulingRegion(V); + ScheduleData *BundleMember = getScheduleData(V); + assert(BundleMember && + "no ScheduleData for bundle member (maybe not in same basic block)"); + if (BundleMember->IsScheduled) { + // A bundle member was scheduled as single instruction before and now + // needs to be scheduled as part of the bundle. We just get rid of the + // existing schedule. + DEBUG(dbgs() << "SLP: reset schedule because " << *BundleMember + << " was already scheduled\n"); + ReSchedule = true; + } + assert(BundleMember->isSchedulingEntity() && + "bundle member already part of other bundle"); + if (PrevInBundle) { + PrevInBundle->NextInBundle = BundleMember; + } else { + Bundle = BundleMember; + } + BundleMember->UnscheduledDepsInBundle = 0; + Bundle->UnscheduledDepsInBundle += BundleMember->UnscheduledDeps; + + // Group the instructions to a bundle. + BundleMember->FirstInBundle = Bundle; + PrevInBundle = BundleMember; + } + if (ScheduleEnd != OldScheduleEnd) { + // The scheduling region got new instructions at the lower end (or it is a + // new region for the first bundle). This makes it necessary to + // recalculate all dependencies. + // It is seldom that this needs to be done a second time after adding the + // initial bundle to the region. + for (auto *I = ScheduleStart; I != ScheduleEnd; I = I->getNextNode()) { + ScheduleData *SD = getScheduleData(I); + SD->clearDependencies(); + } + ReSchedule = true; + } + if (ReSchedule) { + resetSchedule(); + initialFillReadyList(ReadyInsts); + } + + DEBUG(dbgs() << "SLP: try schedule bundle " << *Bundle << " in block " + << BB->getName() << "\n"); + + calculateDependencies(Bundle, true, AA); + + // Now try to schedule the new bundle. As soon as the bundle is "ready" it + // means that there are no cyclic dependencies and we can schedule it. + // Note that's important that we don't "schedule" the bundle yet (see + // cancelScheduling). + while (!Bundle->isReady() && !ReadyInsts.empty()) { + + ScheduleData *pickedSD = ReadyInsts.back(); + ReadyInsts.pop_back(); + + if (pickedSD->isSchedulingEntity() && pickedSD->isReady()) { + schedule(pickedSD, ReadyInsts); + } + } + return Bundle->isReady(); +} + +void BoUpSLP::BlockScheduling::cancelScheduling(ArrayRef VL) { + if (isa(VL[0])) + return; + + ScheduleData *Bundle = getScheduleData(VL[0]); + DEBUG(dbgs() << "SLP: cancel scheduling of " << *Bundle << "\n"); + assert(!Bundle->IsScheduled && + "Can't cancel bundle which is already scheduled"); + assert(Bundle->isSchedulingEntity() && Bundle->isPartOfBundle() && + "tried to unbundle something which is not a bundle"); + + // Un-bundle: make single instructions out of the bundle. + ScheduleData *BundleMember = Bundle; + while (BundleMember) { + assert(BundleMember->FirstInBundle == Bundle && "corrupt bundle links"); + BundleMember->FirstInBundle = BundleMember; + ScheduleData *Next = BundleMember->NextInBundle; + BundleMember->NextInBundle = nullptr; + BundleMember->UnscheduledDepsInBundle = BundleMember->UnscheduledDeps; + if (BundleMember->UnscheduledDepsInBundle == 0) { + ReadyInsts.insert(BundleMember); + } + BundleMember = Next; + } +} + +void BoUpSLP::BlockScheduling::extendSchedulingRegion(Value *V) { + if (getScheduleData(V)) + return; + Instruction *I = dyn_cast(V); + assert(I && "bundle member must be an instruction"); + assert(!isa(I) && "phi nodes don't need to be scheduled"); + if (!ScheduleStart) { + // It's the first instruction in the new region. + initScheduleData(I, I->getNextNode(), nullptr, nullptr); + ScheduleStart = I; + ScheduleEnd = I->getNextNode(); + assert(ScheduleEnd && "tried to vectorize a TerminatorInst?"); + DEBUG(dbgs() << "SLP: initialize schedule region to " << *I << "\n"); + return; + } + // Search up and down at the same time, because we don't know if the new + // instruction is above or below the existing scheduling region. + BasicBlock::reverse_iterator UpIter(ScheduleStart); + BasicBlock::reverse_iterator UpperEnd = BB->rend(); + BasicBlock::iterator DownIter(ScheduleEnd); + BasicBlock::iterator LowerEnd = BB->end(); + for (;;) { + if (UpIter != UpperEnd) { + if (&*UpIter == I) { + initScheduleData(I, ScheduleStart, nullptr, FirstLoadStoreInRegion); + ScheduleStart = I; + DEBUG(dbgs() << "SLP: extend schedule region start to " << *I << "\n"); + return; + } + UpIter++; + } + if (DownIter != LowerEnd) { + if (&*DownIter == I) { + initScheduleData(ScheduleEnd, I->getNextNode(), LastLoadStoreInRegion, + nullptr); + ScheduleEnd = I->getNextNode(); + assert(ScheduleEnd && "tried to vectorize a TerminatorInst?"); + DEBUG(dbgs() << "SLP: extend schedule region end to " << *I << "\n"); + return; + } + DownIter++; + } + assert((UpIter != UpperEnd || DownIter != LowerEnd) && + "instruction not found in block"); + } +} + +void BoUpSLP::BlockScheduling::initScheduleData(Instruction *FromI, + Instruction *ToI, + ScheduleData *PrevLoadStore, + ScheduleData *NextLoadStore) { + ScheduleData *CurrentLoadStore = PrevLoadStore; + for (Instruction *I = FromI; I != ToI; I = I->getNextNode()) { + ScheduleData *SD = ScheduleDataMap[I]; + if (!SD) { + // Allocate a new ScheduleData for the instruction. + if (ChunkPos >= ChunkSize) { + ScheduleDataChunks.push_back( + llvm::make_unique(ChunkSize)); + ChunkPos = 0; + } + SD = &(ScheduleDataChunks.back()[ChunkPos++]); + ScheduleDataMap[I] = SD; + SD->Inst = I; + } + assert(!isInSchedulingRegion(SD) && + "new ScheduleData already in scheduling region"); + SD->init(SchedulingRegionID); + + if (I->mayReadOrWriteMemory()) { + // Update the linked list of memory accessing instructions. + if (CurrentLoadStore) { + CurrentLoadStore->NextLoadStore = SD; + } else { + FirstLoadStoreInRegion = SD; + } + CurrentLoadStore = SD; + } + } + if (NextLoadStore) { + if (CurrentLoadStore) + CurrentLoadStore->NextLoadStore = NextLoadStore; + } else { + LastLoadStoreInRegion = CurrentLoadStore; + } +} + +/// \returns the AA location that is being access by the instruction. +static AliasAnalysis::Location getLocation(Instruction *I, AliasAnalysis *AA) { + if (StoreInst *SI = dyn_cast(I)) + return AA->getLocation(SI); + if (LoadInst *LI = dyn_cast(I)) + return AA->getLocation(LI); + return AliasAnalysis::Location(); +} + +void BoUpSLP::BlockScheduling::calculateDependencies(ScheduleData *SD, + bool InsertInReadyList, + AliasAnalysis *AA) { + assert(SD->isSchedulingEntity()); + + SmallVector WorkList; + WorkList.push_back(SD); + + while (!WorkList.empty()) { + ScheduleData *SD = WorkList.back(); + WorkList.pop_back(); + + ScheduleData *BundleMember = SD; + while (BundleMember) { + assert(isInSchedulingRegion(BundleMember)); + if (!BundleMember->hasValidDependencies()) { + + DEBUG(dbgs() << "SLP: update deps of " << *BundleMember << "\n"); + BundleMember->Dependencies = 0; + BundleMember->resetUnscheduledDeps(); + + // Handle def-use chain dependencies. + for (User *U : BundleMember->Inst->users()) { + if (isa(U)) { + ScheduleData *UseSD = getScheduleData(U); + if (UseSD && isInSchedulingRegion(UseSD->FirstInBundle)) { + BundleMember->Dependencies++; + ScheduleData *DestBundle = UseSD->FirstInBundle; + if (!DestBundle->IsScheduled) { + BundleMember->incrementUnscheduledDeps(1); + } + if (!DestBundle->hasValidDependencies()) { + WorkList.push_back(DestBundle); + } + } + } else { + // I'm not sure if this can ever happen. But we need to be safe. + // This lets the instruction/bundle never be scheduled and eventally + // disable vectorization. + BundleMember->Dependencies++; + BundleMember->incrementUnscheduledDeps(1); + } + } + + // Handle the memory dependencies. + ScheduleData *DepDest = BundleMember->NextLoadStore; + if (DepDest) { + AliasAnalysis::Location SrcLoc = getLocation(BundleMember->Inst, AA); + bool SrcMayWrite = BundleMember->Inst->mayWriteToMemory(); + + while (DepDest) { + assert(isInSchedulingRegion(DepDest)); + if (SrcMayWrite || DepDest->Inst->mayWriteToMemory()) { + AliasAnalysis::Location DstLoc = getLocation(DepDest->Inst, AA); + if (!SrcLoc.Ptr || !DstLoc.Ptr || AA->alias(SrcLoc, DstLoc)) { + DepDest->MemoryDependencies.push_back(BundleMember); + BundleMember->Dependencies++; + ScheduleData *DestBundle = DepDest->FirstInBundle; + if (!DestBundle->IsScheduled) { + BundleMember->incrementUnscheduledDeps(1); + } + if (!DestBundle->hasValidDependencies()) { + WorkList.push_back(DestBundle); + } + } + } + DepDest = DepDest->NextLoadStore; + } + } + } + BundleMember = BundleMember->NextInBundle; + } + if (InsertInReadyList && SD->isReady()) { + ReadyInsts.push_back(SD); + DEBUG(dbgs() << "SLP: gets ready on update: " << *SD->Inst << "\n"); + } + } +} + +void BoUpSLP::BlockScheduling::resetSchedule() { + assert(ScheduleStart && + "tried to reset schedule on block which has not been scheduled"); + for (Instruction *I = ScheduleStart; I != ScheduleEnd; I = I->getNextNode()) { + ScheduleData *SD = getScheduleData(I); + assert(isInSchedulingRegion(SD)); + SD->IsScheduled = false; + SD->resetUnscheduledDeps(); + } + ReadyInsts.clear(); +} + +void BoUpSLP::scheduleBlock(BlockScheduling *BS) { + + if (!BS->ScheduleStart) + return; + + DEBUG(dbgs() << "SLP: schedule block " << BS->BB->getName() << "\n"); + + BS->resetSchedule(); + + // For the real scheduling we use a more sophisticated ready-list: it is + // sorted by the original instruction location. This lets the final schedule + // be as close as possible to the original instruction order. + struct ScheduleDataCompare { + bool operator()(ScheduleData *SD1, ScheduleData *SD2) { + return SD2->SchedulingPriority < SD1->SchedulingPriority; + } + }; + std::set ReadyInsts; + + // Ensure that all depencency data is updated and fill the ready-list with + // initial instructions. + int Idx = 0; + int NumToSchedule = 0; + for (auto *I = BS->ScheduleStart; I != BS->ScheduleEnd; + I = I->getNextNode()) { + ScheduleData *SD = BS->getScheduleData(I); + assert( + SD->isPartOfBundle() == (ScalarToTreeEntry.count(SD->Inst) != 0) && + "scheduler and vectorizer have different opinion on what is a bundle"); + SD->FirstInBundle->SchedulingPriority = Idx++; + if (SD->isSchedulingEntity()) { + BS->calculateDependencies(SD, false, AA); + NumToSchedule++; + } + } + BS->initialFillReadyList(ReadyInsts); + + Instruction *LastScheduledInst = BS->ScheduleEnd; + + // Do the "real" scheduling. + while (!ReadyInsts.empty()) { + ScheduleData *picked = *ReadyInsts.begin(); + ReadyInsts.erase(ReadyInsts.begin()); + + // Move the scheduled instruction(s) to their dedicated places, if not + // there yet. + ScheduleData *BundleMember = picked; + while (BundleMember) { + Instruction *pickedInst = BundleMember->Inst; + if (LastScheduledInst->getNextNode() != pickedInst) { + BS->BB->getInstList().remove(pickedInst); + BS->BB->getInstList().insert(LastScheduledInst, pickedInst); + } + LastScheduledInst = pickedInst; + BundleMember = BundleMember->NextInBundle; + } + + BS->schedule(picked, ReadyInsts); + NumToSchedule--; + } + assert(NumToSchedule == 0 && "could not schedule all instructions"); + + // Avoid duplicate scheduling of the block. + BS->ScheduleStart = nullptr; +} + /// The SLPVectorizer Pass. struct SLPVectorizer : public FunctionPass { typedef SmallVector StoreList; @@ -1925,7 +2755,6 @@ struct SLPVectorizer : public FunctionPass { for (po_iterator it = po_begin(&F.getEntryBlock()), e = po_end(&F.getEntryBlock()); it != e; ++it) { BasicBlock *BB = *it; - // Vectorize trees that end at stores. if (unsigned count = collectStores(BB, R)) { (void)count; @@ -1973,7 +2802,8 @@ private: /// scheduling and that don't need extracting. /// \returns true if a value was vectorized. bool tryToVectorizeList(ArrayRef VL, BoUpSLP &R, - ArrayRef BuildVector = None); + ArrayRef BuildVector = None, + bool allowReorder = false); /// \brief Try to vectorize a chain that may start at the operands of \V; bool tryToVectorize(BinaryOperator *V, BoUpSLP &R); @@ -2143,11 +2973,12 @@ bool SLPVectorizer::tryToVectorizePair(Value *A, Value *B, BoUpSLP &R) { if (!A || !B) return false; Value *VL[] = { A, B }; - return tryToVectorizeList(VL, R); + return tryToVectorizeList(VL, R, None, true); } bool SLPVectorizer::tryToVectorizeList(ArrayRef VL, BoUpSLP &R, - ArrayRef BuildVector) { + ArrayRef BuildVector, + bool allowReorder) { if (VL.size() < 2) return false; @@ -2202,6 +3033,14 @@ bool SLPVectorizer::tryToVectorizeList(ArrayRef VL, BoUpSLP &R, BuildVectorSlice = BuildVector.slice(i, OpsWidth); R.buildTree(Ops, BuildVectorSlice); + // TODO: check if we can allow reordering also for other cases than + // tryToVectorizePair() + if (allowReorder && R.shouldReorder()) { + assert(Ops.size() == 2); + assert(BuildVectorSlice.empty()); + Value *ReorderedOps[] = { Ops[1], Ops[0] }; + R.buildTree(ReorderedOps, None); + } int Cost = R.getTreeCost(); if (Cost < -SLPCostThreshold) { @@ -2753,15 +3592,15 @@ bool SLPVectorizer::vectorizeChainsInBlock(BasicBlock *BB, BoUpSLP &R) { } for (int i = 0; i < 2; ++i) { - if (BinaryOperator *BI = dyn_cast(CI->getOperand(i))) { - if (tryToVectorizePair(BI->getOperand(0), BI->getOperand(1), R)) { - Changed = true; - // We would like to start over since some instructions are deleted - // and the iterator may become invalid value. - it = BB->begin(); - e = BB->end(); - } - } + if (BinaryOperator *BI = dyn_cast(CI->getOperand(i))) { + if (tryToVectorizePair(BI->getOperand(0), BI->getOperand(1), R)) { + Changed = true; + // We would like to start over since some instructions are deleted + // and the iterator may become invalid value. + it = BB->begin(); + e = BB->end(); + } + } } continue; }