HELLO·Android
系统源代码
IT资讯
技术文章
我的收藏
注册
登录
-
我收藏的文章
创建代码块
我的代码块
我的账号
Lollipop MR1
|
5.1.0_r3
下载
查看原文件
收藏
根目录
external
llvm
lib
Transforms
Vectorize
LoopVectorize.cpp
//===- LoopVectorize.cpp - A Loop Vectorizer ------------------------------===// // // The LLVM Compiler Infrastructure // // This file is distributed under the University of Illinois Open Source // License. See LICENSE.TXT for details. // //===----------------------------------------------------------------------===// // // This is the LLVM loop vectorizer. This pass modifies 'vectorizable' loops // and generates target-independent LLVM-IR. // The vectorizer uses the TargetTransformInfo analysis to estimate the costs // of instructions in order to estimate the profitability of vectorization. // // The loop vectorizer combines consecutive loop iterations into a single // 'wide' iteration. After this transformation the index is incremented // by the SIMD vector width, and not by one. // // This pass has three parts: // 1. The main loop pass that drives the different parts. // 2. LoopVectorizationLegality - A unit that checks for the legality // of the vectorization. // 3. InnerLoopVectorizer - A unit that performs the actual // widening of instructions. // 4. LoopVectorizationCostModel - A unit that checks for the profitability // of vectorization. It decides on the optimal vector width, which // can be one, if vectorization is not profitable. // //===----------------------------------------------------------------------===// // // The reduction-variable vectorization is based on the paper: // D. Nuzman and R. Henderson. Multi-platform Auto-vectorization. // // Variable uniformity checks are inspired by: // Karrenberg, R. and Hack, S. Whole Function Vectorization. // // Other ideas/concepts are from: // A. Zaks and D. Nuzman. Autovectorization in GCC-two years later. // // S. Maleki, Y. Gao, M. Garzaran, T. Wong and D. Padua. An Evaluation of // Vectorizing Compilers. // //===----------------------------------------------------------------------===// #include "llvm/Transforms/Vectorize.h" #include "llvm/ADT/DenseMap.h" #include "llvm/ADT/EquivalenceClasses.h" #include "llvm/ADT/Hashing.h" #include "llvm/ADT/MapVector.h" #include "llvm/ADT/SetVector.h" #include "llvm/ADT/SmallPtrSet.h" #include "llvm/ADT/SmallSet.h" #include "llvm/ADT/SmallVector.h" #include "llvm/ADT/Statistic.h" #include "llvm/ADT/StringExtras.h" #include "llvm/Analysis/AliasAnalysis.h" #include "llvm/Analysis/BlockFrequencyInfo.h" #include "llvm/Analysis/LoopInfo.h" #include "llvm/Analysis/LoopIterator.h" #include "llvm/Analysis/LoopPass.h" #include "llvm/Analysis/ScalarEvolution.h" #include "llvm/Analysis/ScalarEvolutionExpander.h" #include "llvm/Analysis/ScalarEvolutionExpressions.h" #include "llvm/Analysis/TargetTransformInfo.h" #include "llvm/Analysis/ValueTracking.h" #include "llvm/IR/Constants.h" #include "llvm/IR/DataLayout.h" #include "llvm/IR/DebugInfo.h" #include "llvm/IR/DerivedTypes.h" #include "llvm/IR/DiagnosticInfo.h" #include "llvm/IR/Dominators.h" #include "llvm/IR/Function.h" #include "llvm/IR/IRBuilder.h" #include "llvm/IR/Instructions.h" #include "llvm/IR/IntrinsicInst.h" #include "llvm/IR/LLVMContext.h" #include "llvm/IR/Module.h" #include "llvm/IR/PatternMatch.h" #include "llvm/IR/Type.h" #include "llvm/IR/Value.h" #include "llvm/IR/ValueHandle.h" #include "llvm/IR/Verifier.h" #include "llvm/Pass.h" #include "llvm/Support/BranchProbability.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/Debug.h" #include "llvm/Support/raw_ostream.h" #include "llvm/Transforms/Scalar.h" #include "llvm/Transforms/Utils/BasicBlockUtils.h" #include "llvm/Transforms/Utils/Local.h" #include "llvm/Transforms/Utils/VectorUtils.h" #include
#include
#include
using namespace llvm; using namespace llvm::PatternMatch; #define LV_NAME "loop-vectorize" #define DEBUG_TYPE LV_NAME STATISTIC(LoopsVectorized, "Number of loops vectorized"); STATISTIC(LoopsAnalyzed, "Number of loops analyzed for vectorization"); static cl::opt
VectorizationFactor("force-vector-width", cl::init(0), cl::Hidden, cl::desc("Sets the SIMD width. Zero is autoselect.")); static cl::opt
VectorizationUnroll("force-vector-unroll", cl::init(0), cl::Hidden, cl::desc("Sets the vectorization unroll count. " "Zero is autoselect.")); static cl::opt
EnableIfConversion("enable-if-conversion", cl::init(true), cl::Hidden, cl::desc("Enable if-conversion during vectorization.")); /// We don't vectorize loops with a known constant trip count below this number. static cl::opt
TinyTripCountVectorThreshold("vectorizer-min-trip-count", cl::init(16), cl::Hidden, cl::desc("Don't vectorize loops with a constant " "trip count that is smaller than this " "value.")); /// This enables versioning on the strides of symbolically striding memory /// accesses in code like the following. /// for (i = 0; i < N; ++i) /// A[i * Stride1] += B[i * Stride2] ... /// /// Will be roughly translated to /// if (Stride1 == 1 && Stride2 == 1) { /// for (i = 0; i < N; i+=4) /// A[i:i+3] += ... /// } else /// ... static cl::opt
EnableMemAccessVersioning( "enable-mem-access-versioning", cl::init(true), cl::Hidden, cl::desc("Enable symblic stride memory access versioning")); /// We don't unroll loops with a known constant trip count below this number. static const unsigned TinyTripCountUnrollThreshold = 128; /// When performing memory disambiguation checks at runtime do not make more /// than this number of comparisons. static const unsigned RuntimeMemoryCheckThreshold = 8; /// Maximum simd width. static const unsigned MaxVectorWidth = 64; static cl::opt
ForceTargetNumScalarRegs( "force-target-num-scalar-regs", cl::init(0), cl::Hidden, cl::desc("A flag that overrides the target's number of scalar registers.")); static cl::opt
ForceTargetNumVectorRegs( "force-target-num-vector-regs", cl::init(0), cl::Hidden, cl::desc("A flag that overrides the target's number of vector registers.")); /// Maximum vectorization unroll count. static const unsigned MaxUnrollFactor = 16; static cl::opt
ForceTargetMaxScalarUnrollFactor( "force-target-max-scalar-unroll", cl::init(0), cl::Hidden, cl::desc("A flag that overrides the target's max unroll factor for scalar " "loops.")); static cl::opt
ForceTargetMaxVectorUnrollFactor( "force-target-max-vector-unroll", cl::init(0), cl::Hidden, cl::desc("A flag that overrides the target's max unroll factor for " "vectorized loops.")); static cl::opt
ForceTargetInstructionCost( "force-target-instruction-cost", cl::init(0), cl::Hidden, cl::desc("A flag that overrides the target's expected cost for " "an instruction to a single constant value. Mostly " "useful for getting consistent testing.")); static cl::opt
SmallLoopCost( "small-loop-cost", cl::init(20), cl::Hidden, cl::desc("The cost of a loop that is considered 'small' by the unroller.")); static cl::opt
LoopVectorizeWithBlockFrequency( "loop-vectorize-with-block-frequency", cl::init(false), cl::Hidden, cl::desc("Enable the use of the block frequency analysis to access PGO " "heuristics minimizing code growth in cold regions and being more " "aggressive in hot regions.")); // Runtime unroll loops for load/store throughput. static cl::opt
EnableLoadStoreRuntimeUnroll( "enable-loadstore-runtime-unroll", cl::init(true), cl::Hidden, cl::desc("Enable runtime unrolling until load/store ports are saturated")); /// The number of stores in a loop that are allowed to need predication. static cl::opt
NumberOfStoresToPredicate( "vectorize-num-stores-pred", cl::init(1), cl::Hidden, cl::desc("Max number of stores to be predicated behind an if.")); static cl::opt
EnableIndVarRegisterHeur( "enable-ind-var-reg-heur", cl::init(true), cl::Hidden, cl::desc("Count the induction variable only once when unrolling")); static cl::opt
EnableCondStoresVectorization( "enable-cond-stores-vec", cl::init(false), cl::Hidden, cl::desc("Enable if predication of stores during vectorization.")); namespace { // Forward declarations. class LoopVectorizationLegality; class LoopVectorizationCostModel; /// Optimization analysis message produced during vectorization. Messages inform /// the user why vectorization did not occur. class Report { std::string Message; raw_string_ostream Out; Instruction *Instr; public: Report(Instruction *I = nullptr) : Out(Message), Instr(I) { Out << "loop not vectorized: "; } template
Report &operator<<(const A &Value) { Out << Value; return *this; } Instruction *getInstr() { return Instr; } std::string &str() { return Out.str(); } operator Twine() { return Out.str(); } }; /// InnerLoopVectorizer vectorizes loops which contain only one basic /// block to a specified vectorization factor (VF). /// This class performs the widening of scalars into vectors, or multiple /// scalars. This class also implements the following features: /// * It inserts an epilogue loop for handling loops that don't have iteration /// counts that are known to be a multiple of the vectorization factor. /// * It handles the code generation for reduction variables. /// * Scalarization (implementation using scalars) of un-vectorizable /// instructions. /// InnerLoopVectorizer does not perform any vectorization-legality /// checks, and relies on the caller to check for the different legality /// aspects. The InnerLoopVectorizer relies on the /// LoopVectorizationLegality class to provide information about the induction /// and reduction variables that were found to a given vectorization factor. class InnerLoopVectorizer { public: InnerLoopVectorizer(Loop *OrigLoop, ScalarEvolution *SE, LoopInfo *LI, DominatorTree *DT, const DataLayout *DL, const TargetLibraryInfo *TLI, unsigned VecWidth, unsigned UnrollFactor) : OrigLoop(OrigLoop), SE(SE), LI(LI), DT(DT), DL(DL), TLI(TLI), VF(VecWidth), UF(UnrollFactor), Builder(SE->getContext()), Induction(nullptr), OldInduction(nullptr), WidenMap(UnrollFactor), Legal(nullptr) {} // Perform the actual loop widening (vectorization). void vectorize(LoopVectorizationLegality *L) { Legal = L; // Create a new empty loop. Unlink the old loop and connect the new one. createEmptyLoop(); // Widen each instruction in the old loop to a new one in the new loop. // Use the Legality module to find the induction and reduction variables. vectorizeLoop(); // Register the new loop and update the analysis passes. updateAnalysis(); } virtual ~InnerLoopVectorizer() {} protected: /// A small list of PHINodes. typedef SmallVector
PhiVector; /// When we unroll loops we have multiple vector values for each scalar. /// This data structure holds the unrolled and vectorized values that /// originated from one scalar instruction. typedef SmallVector
VectorParts; // When we if-convert we need create edge masks. We have to cache values so // that we don't end up with exponential recursion/IR. typedef DenseMap
, VectorParts> EdgeMaskCache; /// \brief Add code that checks at runtime if the accessed arrays overlap. /// /// Returns a pair of instructions where the first element is the first /// instruction generated in possibly a sequence of instructions and the /// second value is the final comparator value or NULL if no check is needed. std::pair
addRuntimeCheck(Instruction *Loc); /// \brief Add checks for strides that where assumed to be 1. /// /// Returns the last check instruction and the first check instruction in the /// pair as (first, last). std::pair
addStrideCheck(Instruction *Loc); /// Create an empty loop, based on the loop ranges of the old loop. void createEmptyLoop(); /// Copy and widen the instructions from the old loop. virtual void vectorizeLoop(); /// \brief The Loop exit block may have single value PHI nodes where the /// incoming value is 'Undef'. While vectorizing we only handled real values /// that were defined inside the loop. Here we fix the 'undef case'. /// See PR14725. void fixLCSSAPHIs(); /// A helper function that computes the predicate of the block BB, assuming /// that the header block of the loop is set to True. It returns the *entry* /// mask for the block BB. VectorParts createBlockInMask(BasicBlock *BB); /// A helper function that computes the predicate of the edge between SRC /// and DST. VectorParts createEdgeMask(BasicBlock *Src, BasicBlock *Dst); /// A helper function to vectorize a single BB within the innermost loop. void vectorizeBlockInLoop(BasicBlock *BB, PhiVector *PV); /// Vectorize a single PHINode in a block. This method handles the induction /// variable canonicalization. It supports both VF = 1 for unrolled loops and /// arbitrary length vectors. void widenPHIInstruction(Instruction *PN, VectorParts &Entry, unsigned UF, unsigned VF, PhiVector *PV); /// Insert the new loop to the loop hierarchy and pass manager /// and update the analysis passes. void updateAnalysis(); /// This instruction is un-vectorizable. Implement it as a sequence /// of scalars. If \p IfPredicateStore is true we need to 'hide' each /// scalarized instruction behind an if block predicated on the control /// dependence of the instruction. virtual void scalarizeInstruction(Instruction *Instr, bool IfPredicateStore=false); /// Vectorize Load and Store instructions, virtual void vectorizeMemoryInstruction(Instruction *Instr); /// Create a broadcast instruction. This method generates a broadcast /// instruction (shuffle) for loop invariant values and for the induction /// value. If this is the induction variable then we extend it to N, N+1, ... /// this is needed because each iteration in the loop corresponds to a SIMD /// element. virtual Value *getBroadcastInstrs(Value *V); /// This function adds 0, 1, 2 ... to each vector element, starting at zero. /// If Negate is set then negative numbers are added e.g. (0, -1, -2, ...). /// The sequence starts at StartIndex. virtual Value *getConsecutiveVector(Value* Val, int StartIdx, bool Negate); /// When we go over instructions in the basic block we rely on previous /// values within the current basic block or on loop invariant values. /// When we widen (vectorize) values we place them in the map. If the values /// are not within the map, they have to be loop invariant, so we simply /// broadcast them into a vector. VectorParts &getVectorValue(Value *V); /// Generate a shuffle sequence that will reverse the vector Vec. virtual Value *reverseVector(Value *Vec); /// This is a helper class that holds the vectorizer state. It maps scalar /// instructions to vector instructions. When the code is 'unrolled' then /// then a single scalar value is mapped to multiple vector parts. The parts /// are stored in the VectorPart type. struct ValueMap { /// C'tor. UnrollFactor controls the number of vectors ('parts') that /// are mapped. ValueMap(unsigned UnrollFactor) : UF(UnrollFactor) {} /// \return True if 'Key' is saved in the Value Map. bool has(Value *Key) const { return MapStorage.count(Key); } /// Initializes a new entry in the map. Sets all of the vector parts to the /// save value in 'Val'. /// \return A reference to a vector with splat values. VectorParts &splat(Value *Key, Value *Val) { VectorParts &Entry = MapStorage[Key]; Entry.assign(UF, Val); return Entry; } ///\return A reference to the value that is stored at 'Key'. VectorParts &get(Value *Key) { VectorParts &Entry = MapStorage[Key]; if (Entry.empty()) Entry.resize(UF); assert(Entry.size() == UF); return Entry; } private: /// The unroll factor. Each entry in the map stores this number of vector /// elements. unsigned UF; /// Map storage. We use std::map and not DenseMap because insertions to a /// dense map invalidates its iterators. std::map
MapStorage; }; /// The original loop. Loop *OrigLoop; /// Scev analysis to use. ScalarEvolution *SE; /// Loop Info. LoopInfo *LI; /// Dominator Tree. DominatorTree *DT; /// Data Layout. const DataLayout *DL; /// Target Library Info. const TargetLibraryInfo *TLI; /// The vectorization SIMD factor to use. Each vector will have this many /// vector elements. unsigned VF; protected: /// The vectorization unroll factor to use. Each scalar is vectorized to this /// many different vector instructions. unsigned UF; /// The builder that we use IRBuilder<> Builder; // --- Vectorization state --- /// The vector-loop preheader. BasicBlock *LoopVectorPreHeader; /// The scalar-loop preheader. BasicBlock *LoopScalarPreHeader; /// Middle Block between the vector and the scalar. BasicBlock *LoopMiddleBlock; ///The ExitBlock of the scalar loop. BasicBlock *LoopExitBlock; ///The vector loop body. SmallVector
LoopVectorBody; ///The scalar loop body. BasicBlock *LoopScalarBody; /// A list of all bypass blocks. The first block is the entry of the loop. SmallVector
LoopBypassBlocks; /// The new Induction variable which was added to the new block. PHINode *Induction; /// The induction variable of the old basic block. PHINode *OldInduction; /// Holds the extended (to the widest induction type) start index. Value *ExtendedIdx; /// Maps scalars to widened vectors. ValueMap WidenMap; EdgeMaskCache MaskCache; LoopVectorizationLegality *Legal; }; class InnerLoopUnroller : public InnerLoopVectorizer { public: InnerLoopUnroller(Loop *OrigLoop, ScalarEvolution *SE, LoopInfo *LI, DominatorTree *DT, const DataLayout *DL, const TargetLibraryInfo *TLI, unsigned UnrollFactor) : InnerLoopVectorizer(OrigLoop, SE, LI, DT, DL, TLI, 1, UnrollFactor) { } private: void scalarizeInstruction(Instruction *Instr, bool IfPredicateStore = false) override; void vectorizeMemoryInstruction(Instruction *Instr) override; Value *getBroadcastInstrs(Value *V) override; Value *getConsecutiveVector(Value* Val, int StartIdx, bool Negate) override; Value *reverseVector(Value *Vec) override; }; /// \brief Look for a meaningful debug location on the instruction or it's /// operands. static Instruction *getDebugLocFromInstOrOperands(Instruction *I) { if (!I) return I; DebugLoc Empty; if (I->getDebugLoc() != Empty) return I; for (User::op_iterator OI = I->op_begin(), OE = I->op_end(); OI != OE; ++OI) { if (Instruction *OpInst = dyn_cast
(*OI)) if (OpInst->getDebugLoc() != Empty) return OpInst; } return I; } /// \brief Set the debug location in the builder using the debug location in the /// instruction. static void setDebugLocFromInst(IRBuilder<> &B, const Value *Ptr) { if (const Instruction *Inst = dyn_cast_or_null
(Ptr)) B.SetCurrentDebugLocation(Inst->getDebugLoc()); else B.SetCurrentDebugLocation(DebugLoc()); } #ifndef NDEBUG /// \return string containing a file name and a line # for the given loop. static std::string getDebugLocString(const Loop *L) { std::string Result; if (L) { raw_string_ostream OS(Result); const DebugLoc LoopDbgLoc = L->getStartLoc(); if (!LoopDbgLoc.isUnknown()) LoopDbgLoc.print(L->getHeader()->getContext(), OS); else // Just print the module name. OS << L->getHeader()->getParent()->getParent()->getModuleIdentifier(); OS.flush(); } return Result; } #endif /// LoopVectorizationLegality checks if it is legal to vectorize a loop, and /// to what vectorization factor. /// This class does not look at the profitability of vectorization, only the /// legality. This class has two main kinds of checks: /// * Memory checks - The code in canVectorizeMemory checks if vectorization /// will change the order of memory accesses in a way that will change the /// correctness of the program. /// * Scalars checks - The code in canVectorizeInstrs and canVectorizeMemory /// checks for a number of different conditions, such as the availability of a /// single induction variable, that all types are supported and vectorize-able, /// etc. This code reflects the capabilities of InnerLoopVectorizer. /// This class is also used by InnerLoopVectorizer for identifying /// induction variable and the different reduction variables. class LoopVectorizationLegality { public: unsigned NumLoads; unsigned NumStores; unsigned NumPredStores; LoopVectorizationLegality(Loop *L, ScalarEvolution *SE, const DataLayout *DL, DominatorTree *DT, TargetLibraryInfo *TLI, Function *F) : NumLoads(0), NumStores(0), NumPredStores(0), TheLoop(L), SE(SE), DL(DL), DT(DT), TLI(TLI), TheFunction(F), Induction(nullptr), WidestIndTy(nullptr), HasFunNoNaNAttr(false), MaxSafeDepDistBytes(-1U) { } /// This enum represents the kinds of reductions that we support. enum ReductionKind { RK_NoReduction, ///< Not a reduction. RK_IntegerAdd, ///< Sum of integers. RK_IntegerMult, ///< Product of integers. RK_IntegerOr, ///< Bitwise or logical OR of numbers. RK_IntegerAnd, ///< Bitwise or logical AND of numbers. RK_IntegerXor, ///< Bitwise or logical XOR of numbers. RK_IntegerMinMax, ///< Min/max implemented in terms of select(cmp()). RK_FloatAdd, ///< Sum of floats. RK_FloatMult, ///< Product of floats. RK_FloatMinMax ///< Min/max implemented in terms of select(cmp()). }; /// This enum represents the kinds of inductions that we support. enum InductionKind { IK_NoInduction, ///< Not an induction variable. IK_IntInduction, ///< Integer induction variable. Step = 1. IK_ReverseIntInduction, ///< Reverse int induction variable. Step = -1. IK_PtrInduction, ///< Pointer induction var. Step = sizeof(elem). IK_ReversePtrInduction ///< Reverse ptr indvar. Step = - sizeof(elem). }; // This enum represents the kind of minmax reduction. enum MinMaxReductionKind { MRK_Invalid, MRK_UIntMin, MRK_UIntMax, MRK_SIntMin, MRK_SIntMax, MRK_FloatMin, MRK_FloatMax }; /// This struct holds information about reduction variables. struct ReductionDescriptor { ReductionDescriptor() : StartValue(nullptr), LoopExitInstr(nullptr), Kind(RK_NoReduction), MinMaxKind(MRK_Invalid) {} ReductionDescriptor(Value *Start, Instruction *Exit, ReductionKind K, MinMaxReductionKind MK) : StartValue(Start), LoopExitInstr(Exit), Kind(K), MinMaxKind(MK) {} // The starting value of the reduction. // It does not have to be zero! TrackingVH
StartValue; // The instruction who's value is used outside the loop. Instruction *LoopExitInstr; // The kind of the reduction. ReductionKind Kind; // If this a min/max reduction the kind of reduction. MinMaxReductionKind MinMaxKind; }; /// This POD struct holds information about a potential reduction operation. struct ReductionInstDesc { ReductionInstDesc(bool IsRedux, Instruction *I) : IsReduction(IsRedux), PatternLastInst(I), MinMaxKind(MRK_Invalid) {} ReductionInstDesc(Instruction *I, MinMaxReductionKind K) : IsReduction(true), PatternLastInst(I), MinMaxKind(K) {} // Is this instruction a reduction candidate. bool IsReduction; // The last instruction in a min/max pattern (select of the select(icmp()) // pattern), or the current reduction instruction otherwise. Instruction *PatternLastInst; // If this is a min/max pattern the comparison predicate. MinMaxReductionKind MinMaxKind; }; /// This struct holds information about the memory runtime legality /// check that a group of pointers do not overlap. struct RuntimePointerCheck { RuntimePointerCheck() : Need(false) {} /// Reset the state of the pointer runtime information. void reset() { Need = false; Pointers.clear(); Starts.clear(); Ends.clear(); IsWritePtr.clear(); DependencySetId.clear(); } /// Insert a pointer and calculate the start and end SCEVs. void insert(ScalarEvolution *SE, Loop *Lp, Value *Ptr, bool WritePtr, unsigned DepSetId, ValueToValueMap &Strides); /// This flag indicates if we need to add the runtime check. bool Need; /// Holds the pointers that we need to check. SmallVector
, 2> Pointers; /// Holds the pointer value at the beginning of the loop. SmallVector
Starts; /// Holds the pointer value at the end of the loop. SmallVector
Ends; /// Holds the information if this pointer is used for writing to memory. SmallVector
IsWritePtr; /// Holds the id of the set of pointers that could be dependent because of a /// shared underlying object. SmallVector
DependencySetId; }; /// A struct for saving information about induction variables. struct InductionInfo { InductionInfo(Value *Start, InductionKind K) : StartValue(Start), IK(K) {} InductionInfo() : StartValue(nullptr), IK(IK_NoInduction) {} /// Start value. TrackingVH
StartValue; /// Induction kind. InductionKind IK; }; /// ReductionList contains the reduction descriptors for all /// of the reductions that were found in the loop. typedef DenseMap
ReductionList; /// InductionList saves induction variables and maps them to the /// induction descriptor. typedef MapVector
InductionList; /// Returns true if it is legal to vectorize this loop. /// This does not mean that it is profitable to vectorize this /// loop, only that it is legal to do so. bool canVectorize(); /// Returns the Induction variable. PHINode *getInduction() { return Induction; } /// Returns the reduction variables found in the loop. ReductionList *getReductionVars() { return &Reductions; } /// Returns the induction variables found in the loop. InductionList *getInductionVars() { return &Inductions; } /// Returns the widest induction type. Type *getWidestInductionType() { return WidestIndTy; } /// Returns True if V is an induction variable in this loop. bool isInductionVariable(const Value *V); /// Return true if the block BB needs to be predicated in order for the loop /// to be vectorized. bool blockNeedsPredication(BasicBlock *BB); /// Check if this pointer is consecutive when vectorizing. This happens /// when the last index of the GEP is the induction variable, or that the /// pointer itself is an induction variable. /// This check allows us to vectorize A[idx] into a wide load/store. /// Returns: /// 0 - Stride is unknown or non-consecutive. /// 1 - Address is consecutive. /// -1 - Address is consecutive, and decreasing. int isConsecutivePtr(Value *Ptr); /// Returns true if the value V is uniform within the loop. bool isUniform(Value *V); /// Returns true if this instruction will remain scalar after vectorization. bool isUniformAfterVectorization(Instruction* I) { return Uniforms.count(I); } /// Returns the information that we collected about runtime memory check. RuntimePointerCheck *getRuntimePointerCheck() { return &PtrRtCheck; } /// This function returns the identity element (or neutral element) for /// the operation K. static Constant *getReductionIdentity(ReductionKind K, Type *Tp); unsigned getMaxSafeDepDistBytes() { return MaxSafeDepDistBytes; } bool hasStride(Value *V) { return StrideSet.count(V); } bool mustCheckStrides() { return !StrideSet.empty(); } SmallPtrSet
::iterator strides_begin() { return StrideSet.begin(); } SmallPtrSet
::iterator strides_end() { return StrideSet.end(); } private: /// Check if a single basic block loop is vectorizable. /// At this point we know that this is a loop with a constant trip count /// and we only need to check individual instructions. bool canVectorizeInstrs(); /// When we vectorize loops we may change the order in which /// we read and write from memory. This method checks if it is /// legal to vectorize the code, considering only memory constrains. /// Returns true if the loop is vectorizable bool canVectorizeMemory(); /// Return true if we can vectorize this loop using the IF-conversion /// transformation. bool canVectorizeWithIfConvert(); /// Collect the variables that need to stay uniform after vectorization. void collectLoopUniforms(); /// Return true if all of the instructions in the block can be speculatively /// executed. \p SafePtrs is a list of addresses that are known to be legal /// and we know that we can read from them without segfault. bool blockCanBePredicated(BasicBlock *BB, SmallPtrSet
& SafePtrs); /// Returns True, if 'Phi' is the kind of reduction variable for type /// 'Kind'. If this is a reduction variable, it adds it to ReductionList. bool AddReductionVar(PHINode *Phi, ReductionKind Kind); /// Returns a struct describing if the instruction 'I' can be a reduction /// variable of type 'Kind'. If the reduction is a min/max pattern of /// select(icmp()) this function advances the instruction pointer 'I' from the /// compare instruction to the select instruction and stores this pointer in /// 'PatternLastInst' member of the returned struct. ReductionInstDesc isReductionInstr(Instruction *I, ReductionKind Kind, ReductionInstDesc &Desc); /// Returns true if the instruction is a Select(ICmp(X, Y), X, Y) instruction /// pattern corresponding to a min(X, Y) or max(X, Y). static ReductionInstDesc isMinMaxSelectCmpPattern(Instruction *I, ReductionInstDesc &Prev); /// Returns the induction kind of Phi. This function may return NoInduction /// if the PHI is not an induction variable. InductionKind isInductionVariable(PHINode *Phi); /// \brief Collect memory access with loop invariant strides. /// /// Looks for accesses like "a[i * StrideA]" where "StrideA" is loop /// invariant. void collectStridedAcccess(Value *LoadOrStoreInst); /// Report an analysis message to assist the user in diagnosing loops that are /// not vectorized. void emitAnalysis(Report &Message) { DebugLoc DL = TheLoop->getStartLoc(); if (Instruction *I = Message.getInstr()) DL = I->getDebugLoc(); emitOptimizationRemarkAnalysis(TheFunction->getContext(), DEBUG_TYPE, *TheFunction, DL, Message.str()); } /// The loop that we evaluate. Loop *TheLoop; /// Scev analysis. ScalarEvolution *SE; /// DataLayout analysis. const DataLayout *DL; /// Dominators. DominatorTree *DT; /// Target Library Info. TargetLibraryInfo *TLI; /// Parent function Function *TheFunction; // --- vectorization state --- // /// Holds the integer induction variable. This is the counter of the /// loop. PHINode *Induction; /// Holds the reduction variables. ReductionList Reductions; /// Holds all of the induction variables that we found in the loop. /// Notice that inductions don't need to start at zero and that induction /// variables can be pointers. InductionList Inductions; /// Holds the widest induction type encountered. Type *WidestIndTy; /// Allowed outside users. This holds the reduction /// vars which can be accessed from outside the loop. SmallPtrSet
AllowedExit; /// This set holds the variables which are known to be uniform after /// vectorization. SmallPtrSet
Uniforms; /// We need to check that all of the pointers in this list are disjoint /// at runtime. RuntimePointerCheck PtrRtCheck; /// Can we assume the absence of NaNs. bool HasFunNoNaNAttr; unsigned MaxSafeDepDistBytes; ValueToValueMap Strides; SmallPtrSet
StrideSet; }; /// LoopVectorizationCostModel - estimates the expected speedups due to /// vectorization. /// In many cases vectorization is not profitable. This can happen because of /// a number of reasons. In this class we mainly attempt to predict the /// expected speedup/slowdowns due to the supported instruction set. We use the /// TargetTransformInfo to query the different backends for the cost of /// different operations. class LoopVectorizationCostModel { public: LoopVectorizationCostModel(Loop *L, ScalarEvolution *SE, LoopInfo *LI, LoopVectorizationLegality *Legal, const TargetTransformInfo &TTI, const DataLayout *DL, const TargetLibraryInfo *TLI) : TheLoop(L), SE(SE), LI(LI), Legal(Legal), TTI(TTI), DL(DL), TLI(TLI) {} /// Information about vectorization costs struct VectorizationFactor { unsigned Width; // Vector width with best cost unsigned Cost; // Cost of the loop with that width }; /// \return The most profitable vectorization factor and the cost of that VF. /// This method checks every power of two up to VF. If UserVF is not ZERO /// then this vectorization factor will be selected if vectorization is /// possible. VectorizationFactor selectVectorizationFactor(bool OptForSize, unsigned UserVF, bool ForceVectorization); /// \return The size (in bits) of the widest type in the code that /// needs to be vectorized. We ignore values that remain scalar such as /// 64 bit loop indices. unsigned getWidestType(); /// \return The most profitable unroll factor. /// If UserUF is non-zero then this method finds the best unroll-factor /// based on register pressure and other parameters. /// VF and LoopCost are the selected vectorization factor and the cost of the /// selected VF. unsigned selectUnrollFactor(bool OptForSize, unsigned UserUF, unsigned VF, unsigned LoopCost); /// \brief A struct that represents some properties of the register usage /// of a loop. struct RegisterUsage { /// Holds the number of loop invariant values that are used in the loop. unsigned LoopInvariantRegs; /// Holds the maximum number of concurrent live intervals in the loop. unsigned MaxLocalUsers; /// Holds the number of instructions in the loop. unsigned NumInstructions; }; /// \return information about the register usage of the loop. RegisterUsage calculateRegisterUsage(); private: /// Returns the expected execution cost. The unit of the cost does /// not matter because we use the 'cost' units to compare different /// vector widths. The cost that is returned is *not* normalized by /// the factor width. unsigned expectedCost(unsigned VF); /// Returns the execution time cost of an instruction for a given vector /// width. Vector width of one means scalar. unsigned getInstructionCost(Instruction *I, unsigned VF); /// A helper function for converting Scalar types to vector types. /// If the incoming type is void, we return void. If the VF is 1, we return /// the scalar type. static Type* ToVectorTy(Type *Scalar, unsigned VF); /// Returns whether the instruction is a load or store and will be a emitted /// as a vector operation. bool isConsecutiveLoadOrStore(Instruction *I); /// The loop that we evaluate. Loop *TheLoop; /// Scev analysis. ScalarEvolution *SE; /// Loop Info analysis. LoopInfo *LI; /// Vectorization legality. LoopVectorizationLegality *Legal; /// Vector target information. const TargetTransformInfo &TTI; /// Target data layout information. const DataLayout *DL; /// Target Library Info. const TargetLibraryInfo *TLI; }; /// Utility class for getting and setting loop vectorizer hints in the form /// of loop metadata. class LoopVectorizeHints { public: enum ForceKind { FK_Undefined = -1, ///< Not selected. FK_Disabled = 0, ///< Forcing disabled. FK_Enabled = 1, ///< Forcing enabled. }; LoopVectorizeHints(const Loop *L, bool DisableUnrolling) : Width(VectorizationFactor), Unroll(DisableUnrolling), Force(FK_Undefined), LoopID(L->getLoopID()) { getHints(L); // force-vector-unroll overrides DisableUnrolling. if (VectorizationUnroll.getNumOccurrences() > 0) Unroll = VectorizationUnroll; DEBUG(if (DisableUnrolling && Unroll == 1) dbgs() << "LV: Unrolling disabled by the pass manager\n"); } /// Return the loop vectorizer metadata prefix. static StringRef Prefix() { return "llvm.loop.vectorize."; } MDNode *createHint(LLVMContext &Context, StringRef Name, unsigned V) const { SmallVector
Vals; Vals.push_back(MDString::get(Context, Name)); Vals.push_back(ConstantInt::get(Type::getInt32Ty(Context), V)); return MDNode::get(Context, Vals); } /// Mark the loop L as already vectorized by setting the width to 1. void setAlreadyVectorized(Loop *L) { LLVMContext &Context = L->getHeader()->getContext(); Width = 1; // Create a new loop id with one more operand for the already_vectorized // hint. If the loop already has a loop id then copy the existing operands. SmallVector
Vals(1); if (LoopID) for (unsigned i = 1, ie = LoopID->getNumOperands(); i < ie; ++i) Vals.push_back(LoopID->getOperand(i)); Vals.push_back(createHint(Context, Twine(Prefix(), "width").str(), Width)); Vals.push_back(createHint(Context, Twine(Prefix(), "unroll").str(), 1)); MDNode *NewLoopID = MDNode::get(Context, Vals); // Set operand 0 to refer to the loop id itself. NewLoopID->replaceOperandWith(0, NewLoopID); L->setLoopID(NewLoopID); if (LoopID) LoopID->replaceAllUsesWith(NewLoopID); LoopID = NewLoopID; } std::string emitRemark() const { Report R; R << "vectorization "; switch (Force) { case LoopVectorizeHints::FK_Disabled: R << "is explicitly disabled"; break; case LoopVectorizeHints::FK_Enabled: R << "is explicitly enabled"; if (Width != 0 && Unroll != 0) R << " with width " << Width << " and interleave count " << Unroll; else if (Width != 0) R << " with width " << Width; else if (Unroll != 0) R << " with interleave count " << Unroll; break; case LoopVectorizeHints::FK_Undefined: R << "was not specified"; break; } return R.str(); } unsigned getWidth() const { return Width; } unsigned getUnroll() const { return Unroll; } enum ForceKind getForce() const { return Force; } MDNode *getLoopID() const { return LoopID; } private: /// Find hints specified in the loop metadata. void getHints(const Loop *L) { if (!LoopID) return; // First operand should refer to the loop id itself. assert(LoopID->getNumOperands() > 0 && "requires at least one operand"); assert(LoopID->getOperand(0) == LoopID && "invalid loop id"); for (unsigned i = 1, ie = LoopID->getNumOperands(); i < ie; ++i) { const MDString *S = nullptr; SmallVector
Args; // The expected hint is either a MDString or a MDNode with the first // operand a MDString. if (const MDNode *MD = dyn_cast
(LoopID->getOperand(i))) { if (!MD || MD->getNumOperands() == 0) continue; S = dyn_cast
(MD->getOperand(0)); for (unsigned i = 1, ie = MD->getNumOperands(); i < ie; ++i) Args.push_back(MD->getOperand(i)); } else { S = dyn_cast
(LoopID->getOperand(i)); assert(Args.size() == 0 && "too many arguments for MDString"); } if (!S) continue; // Check if the hint starts with the vectorizer prefix. StringRef Hint = S->getString(); if (!Hint.startswith(Prefix())) continue; // Remove the prefix. Hint = Hint.substr(Prefix().size(), StringRef::npos); if (Args.size() == 1) getHint(Hint, Args[0]); } } // Check string hint with one operand. void getHint(StringRef Hint, Value *Arg) { const ConstantInt *C = dyn_cast
(Arg); if (!C) return; unsigned Val = C->getZExtValue(); if (Hint == "width") { if (isPowerOf2_32(Val) && Val <= MaxVectorWidth) Width = Val; else DEBUG(dbgs() << "LV: ignoring invalid width hint metadata\n"); } else if (Hint == "unroll") { if (isPowerOf2_32(Val) && Val <= MaxUnrollFactor) Unroll = Val; else DEBUG(dbgs() << "LV: ignoring invalid unroll hint metadata\n"); } else if (Hint == "enable") { if (C->getBitWidth() == 1) Force = Val == 1 ? LoopVectorizeHints::FK_Enabled : LoopVectorizeHints::FK_Disabled; else DEBUG(dbgs() << "LV: ignoring invalid enable hint metadata\n"); } else { DEBUG(dbgs() << "LV: ignoring unknown hint " << Hint << '\n'); } } /// Vectorization width. unsigned Width; /// Vectorization unroll factor. unsigned Unroll; /// Vectorization forced enum ForceKind Force; MDNode *LoopID; }; static void addInnerLoop(Loop &L, SmallVectorImpl
&V) { if (L.empty()) return V.push_back(&L); for (Loop *InnerL : L) addInnerLoop(*InnerL, V); } /// The LoopVectorize Pass. struct LoopVectorize : public FunctionPass { /// Pass identification, replacement for typeid static char ID; explicit LoopVectorize(bool NoUnrolling = false, bool AlwaysVectorize = true) : FunctionPass(ID), DisableUnrolling(NoUnrolling), AlwaysVectorize(AlwaysVectorize) { initializeLoopVectorizePass(*PassRegistry::getPassRegistry()); } ScalarEvolution *SE; const DataLayout *DL; LoopInfo *LI; TargetTransformInfo *TTI; DominatorTree *DT; BlockFrequencyInfo *BFI; TargetLibraryInfo *TLI; bool DisableUnrolling; bool AlwaysVectorize; BlockFrequency ColdEntryFreq; bool runOnFunction(Function &F) override { SE = &getAnalysis
(); DataLayoutPass *DLP = getAnalysisIfAvailable
(); DL = DLP ? &DLP->getDataLayout() : nullptr; LI = &getAnalysis
(); TTI = &getAnalysis
(); DT = &getAnalysis
().getDomTree(); BFI = &getAnalysis
(); TLI = getAnalysisIfAvailable
(); // Compute some weights outside of the loop over the loops. Compute this // using a BranchProbability to re-use its scaling math. const BranchProbability ColdProb(1, 5); // 20% ColdEntryFreq = BlockFrequency(BFI->getEntryFreq()) * ColdProb; // If the target claims to have no vector registers don't attempt // vectorization. if (!TTI->getNumberOfRegisters(true)) return false; if (!DL) { DEBUG(dbgs() << "\nLV: Not vectorizing " << F.getName() << ": Missing data layout\n"); return false; } // Build up a worklist of inner-loops to vectorize. This is necessary as // the act of vectorizing or partially unrolling a loop creates new loops // and can invalidate iterators across the loops. SmallVector
Worklist; for (Loop *L : *LI) addInnerLoop(*L, Worklist); LoopsAnalyzed += Worklist.size(); // Now walk the identified inner loops. bool Changed = false; while (!Worklist.empty()) Changed |= processLoop(Worklist.pop_back_val()); // Process each loop nest in the function. return Changed; } bool processLoop(Loop *L) { assert(L->empty() && "Only process inner loops."); #ifndef NDEBUG const std::string DebugLocStr = getDebugLocString(L); #endif /* NDEBUG */ DEBUG(dbgs() << "\nLV: Checking a loop in \"" << L->getHeader()->getParent()->getName() << "\" from " << DebugLocStr << "\n"); LoopVectorizeHints Hints(L, DisableUnrolling); DEBUG(dbgs() << "LV: Loop hints:" << " force=" << (Hints.getForce() == LoopVectorizeHints::FK_Disabled ? "disabled" : (Hints.getForce() == LoopVectorizeHints::FK_Enabled ? "enabled" : "?")) << " width=" << Hints.getWidth() << " unroll=" << Hints.getUnroll() << "\n"); // Function containing loop Function *F = L->getHeader()->getParent(); // Looking at the diagnostic output is the only way to determine if a loop // was vectorized (other than looking at the IR or machine code), so it // is important to generate an optimization remark for each loop. Most of // these messages are generated by emitOptimizationRemarkAnalysis. Remarks // generated by emitOptimizationRemark and emitOptimizationRemarkMissed are // less verbose reporting vectorized loops and unvectorized loops that may // benefit from vectorization, respectively. if (Hints.getForce() == LoopVectorizeHints::FK_Disabled) { DEBUG(dbgs() << "LV: Not vectorizing: #pragma vectorize disable.\n"); emitOptimizationRemarkAnalysis(F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), Hints.emitRemark()); return false; } if (!AlwaysVectorize && Hints.getForce() != LoopVectorizeHints::FK_Enabled) { DEBUG(dbgs() << "LV: Not vectorizing: No #pragma vectorize enable.\n"); emitOptimizationRemarkAnalysis(F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), Hints.emitRemark()); return false; } if (Hints.getWidth() == 1 && Hints.getUnroll() == 1) { DEBUG(dbgs() << "LV: Not vectorizing: Disabled/already vectorized.\n"); emitOptimizationRemarkAnalysis( F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), "loop not vectorized: vector width and interleave count are " "explicitly set to 1"); return false; } // Check the loop for a trip count threshold: // do not vectorize loops with a tiny trip count. BasicBlock *Latch = L->getLoopLatch(); const unsigned TC = SE->getSmallConstantTripCount(L, Latch); if (TC > 0u && TC < TinyTripCountVectorThreshold) { DEBUG(dbgs() << "LV: Found a loop with a very small trip count. " << "This loop is not worth vectorizing."); if (Hints.getForce() == LoopVectorizeHints::FK_Enabled) DEBUG(dbgs() << " But vectorizing was explicitly forced.\n"); else { DEBUG(dbgs() << "\n"); emitOptimizationRemarkAnalysis( F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), "vectorization is not beneficial and is not explicitly forced"); return false; } } // Check if it is legal to vectorize the loop. LoopVectorizationLegality LVL(L, SE, DL, DT, TLI, F); if (!LVL.canVectorize()) { DEBUG(dbgs() << "LV: Not vectorizing: Cannot prove legality.\n"); emitOptimizationRemarkMissed(F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), Hints.emitRemark()); return false; } // Use the cost model. LoopVectorizationCostModel CM(L, SE, LI, &LVL, *TTI, DL, TLI); // Check the function attributes to find out if this function should be // optimized for size. bool OptForSize = Hints.getForce() != LoopVectorizeHints::FK_Enabled && F->hasFnAttribute(Attribute::OptimizeForSize); // Compute the weighted frequency of this loop being executed and see if it // is less than 20% of the function entry baseline frequency. Note that we // always have a canonical loop here because we think we *can* vectoriez. // FIXME: This is hidden behind a flag due to pervasive problems with // exactly what block frequency models. if (LoopVectorizeWithBlockFrequency) { BlockFrequency LoopEntryFreq = BFI->getBlockFreq(L->getLoopPreheader()); if (Hints.getForce() != LoopVectorizeHints::FK_Enabled && LoopEntryFreq < ColdEntryFreq) OptForSize = true; } // Check the function attributes to see if implicit floats are allowed.a // FIXME: This check doesn't seem possibly correct -- what if the loop is // an integer loop and the vector instructions selected are purely integer // vector instructions? if (F->hasFnAttribute(Attribute::NoImplicitFloat)) { DEBUG(dbgs() << "LV: Can't vectorize when the NoImplicitFloat" "attribute is used.\n"); emitOptimizationRemarkAnalysis( F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), "loop not vectorized due to NoImplicitFloat attribute"); emitOptimizationRemarkMissed(F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), Hints.emitRemark()); return false; } // Select the optimal vectorization factor. const LoopVectorizationCostModel::VectorizationFactor VF = CM.selectVectorizationFactor(OptForSize, Hints.getWidth(), Hints.getForce() == LoopVectorizeHints::FK_Enabled); // Select the unroll factor. const unsigned UF = CM.selectUnrollFactor(OptForSize, Hints.getUnroll(), VF.Width, VF.Cost); DEBUG(dbgs() << "LV: Found a vectorizable loop (" << VF.Width << ") in " << DebugLocStr << '\n'); DEBUG(dbgs() << "LV: Unroll Factor is " << UF << '\n'); if (VF.Width == 1) { DEBUG(dbgs() << "LV: Vectorization is possible but not beneficial\n"); if (UF == 1) { emitOptimizationRemarkAnalysis( F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), "not beneficial to vectorize and user disabled interleaving"); return false; } DEBUG(dbgs() << "LV: Trying to at least unroll the loops.\n"); // Report the unrolling decision. emitOptimizationRemark(F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), Twine("unrolled with interleaving factor " + Twine(UF) + " (vectorization not beneficial)")); // We decided not to vectorize, but we may want to unroll. InnerLoopUnroller Unroller(L, SE, LI, DT, DL, TLI, UF); Unroller.vectorize(&LVL); } else { // If we decided that it is *legal* to vectorize the loop then do it. InnerLoopVectorizer LB(L, SE, LI, DT, DL, TLI, VF.Width, UF); LB.vectorize(&LVL); ++LoopsVectorized; // Report the vectorization decision. emitOptimizationRemark( F->getContext(), DEBUG_TYPE, *F, L->getStartLoc(), Twine("vectorized loop (vectorization factor: ") + Twine(VF.Width) + ", unrolling interleave factor: " + Twine(UF) + ")"); } // Mark the loop as already vectorized to avoid vectorizing again. Hints.setAlreadyVectorized(L); DEBUG(verifyFunction(*L->getHeader()->getParent())); return true; } void getAnalysisUsage(AnalysisUsage &AU) const override { AU.addRequiredID(LoopSimplifyID); AU.addRequiredID(LCSSAID); AU.addRequired
(); AU.addRequired
(); AU.addRequired
(); AU.addRequired
(); AU.addRequired
(); AU.addPreserved
(); AU.addPreserved
(); } }; } // end anonymous namespace //===----------------------------------------------------------------------===// // Implementation of LoopVectorizationLegality, InnerLoopVectorizer and // LoopVectorizationCostModel. //===----------------------------------------------------------------------===// static Value *stripIntegerCast(Value *V) { if (CastInst *CI = dyn_cast
(V)) if (CI->getOperand(0)->getType()->isIntegerTy()) return CI->getOperand(0); return V; } ///\brief Replaces the symbolic stride in a pointer SCEV expression by one. /// /// If \p OrigPtr is not null, use it to look up the stride value instead of /// \p Ptr. static const SCEV *replaceSymbolicStrideSCEV(ScalarEvolution *SE, ValueToValueMap &PtrToStride, Value *Ptr, Value *OrigPtr = nullptr) { const SCEV *OrigSCEV = SE->getSCEV(Ptr); // If there is an entry in the map return the SCEV of the pointer with the // symbolic stride replaced by one. ValueToValueMap::iterator SI = PtrToStride.find(OrigPtr ? OrigPtr : Ptr); if (SI != PtrToStride.end()) { Value *StrideVal = SI->second; // Strip casts. StrideVal = stripIntegerCast(StrideVal); // Replace symbolic stride by one. Value *One = ConstantInt::get(StrideVal->getType(), 1); ValueToValueMap RewriteMap; RewriteMap[StrideVal] = One; const SCEV *ByOne = SCEVParameterRewriter::rewrite(OrigSCEV, *SE, RewriteMap, true); DEBUG(dbgs() << "LV: Replacing SCEV: " << *OrigSCEV << " by: " << *ByOne << "\n"); return ByOne; } // Otherwise, just return the SCEV of the original pointer. return SE->getSCEV(Ptr); } void LoopVectorizationLegality::RuntimePointerCheck::insert( ScalarEvolution *SE, Loop *Lp, Value *Ptr, bool WritePtr, unsigned DepSetId, ValueToValueMap &Strides) { // Get the stride replaced scev. const SCEV *Sc = replaceSymbolicStrideSCEV(SE, Strides, Ptr); const SCEVAddRecExpr *AR = dyn_cast
(Sc); assert(AR && "Invalid addrec expression"); const SCEV *Ex = SE->getBackedgeTakenCount(Lp); const SCEV *ScEnd = AR->evaluateAtIteration(Ex, *SE); Pointers.push_back(Ptr); Starts.push_back(AR->getStart()); Ends.push_back(ScEnd); IsWritePtr.push_back(WritePtr); DependencySetId.push_back(DepSetId); } Value *InnerLoopVectorizer::getBroadcastInstrs(Value *V) { // We need to place the broadcast of invariant variables outside the loop. Instruction *Instr = dyn_cast
(V); bool NewInstr = (Instr && std::find(LoopVectorBody.begin(), LoopVectorBody.end(), Instr->getParent()) != LoopVectorBody.end()); bool Invariant = OrigLoop->isLoopInvariant(V) && !NewInstr; // Place the code for broadcasting invariant variables in the new preheader. IRBuilder<>::InsertPointGuard Guard(Builder); if (Invariant) Builder.SetInsertPoint(LoopVectorPreHeader->getTerminator()); // Broadcast the scalar into all locations in the vector. Value *Shuf = Builder.CreateVectorSplat(VF, V, "broadcast"); return Shuf; } Value *InnerLoopVectorizer::getConsecutiveVector(Value* Val, int StartIdx, bool Negate) { assert(Val->getType()->isVectorTy() && "Must be a vector"); assert(Val->getType()->getScalarType()->isIntegerTy() && "Elem must be an integer"); // Create the types. Type *ITy = Val->getType()->getScalarType(); VectorType *Ty = cast
(Val->getType()); int VLen = Ty->getNumElements(); SmallVector
Indices; // Create a vector of consecutive numbers from zero to VF. for (int i = 0; i < VLen; ++i) { int64_t Idx = Negate ? (-i) : i; Indices.push_back(ConstantInt::get(ITy, StartIdx + Idx, Negate)); } // Add the consecutive indices to the vector value. Constant *Cv = ConstantVector::get(Indices); assert(Cv->getType() == Val->getType() && "Invalid consecutive vec"); return Builder.CreateAdd(Val, Cv, "induction"); } /// \brief Find the operand of the GEP that should be checked for consecutive /// stores. This ignores trailing indices that have no effect on the final /// pointer. static unsigned getGEPInductionOperand(const DataLayout *DL, const GetElementPtrInst *Gep) { unsigned LastOperand = Gep->getNumOperands() - 1; unsigned GEPAllocSize = DL->getTypeAllocSize( cast
(Gep->getType()->getScalarType())->getElementType()); // Walk backwards and try to peel off zeros. while (LastOperand > 1 && match(Gep->getOperand(LastOperand), m_Zero())) { // Find the type we're currently indexing into. gep_type_iterator GEPTI = gep_type_begin(Gep); std::advance(GEPTI, LastOperand - 1); // If it's a type with the same allocation size as the result of the GEP we // can peel off the zero index. if (DL->getTypeAllocSize(*GEPTI) != GEPAllocSize) break; --LastOperand; } return LastOperand; } int LoopVectorizationLegality::isConsecutivePtr(Value *Ptr) { assert(Ptr->getType()->isPointerTy() && "Unexpected non-ptr"); // Make sure that the pointer does not point to structs. if (Ptr->getType()->getPointerElementType()->isAggregateType()) return 0; // If this value is a pointer induction variable we know it is consecutive. PHINode *Phi = dyn_cast_or_null
(Ptr); if (Phi && Inductions.count(Phi)) { InductionInfo II = Inductions[Phi]; if (IK_PtrInduction == II.IK) return 1; else if (IK_ReversePtrInduction == II.IK) return -1; } GetElementPtrInst *Gep = dyn_cast_or_null
(Ptr); if (!Gep) return 0; unsigned NumOperands = Gep->getNumOperands(); Value *GpPtr = Gep->getPointerOperand(); // If this GEP value is a consecutive pointer induction variable and all of // the indices are constant then we know it is consecutive. We can Phi = dyn_cast
(GpPtr); if (Phi && Inductions.count(Phi)) { // Make sure that the pointer does not point to structs. PointerType *GepPtrType = cast
(GpPtr->getType()); if (GepPtrType->getElementType()->isAggregateType()) return 0; // Make sure that all of the index operands are loop invariant. for (unsigned i = 1; i < NumOperands; ++i) if (!SE->isLoopInvariant(SE->getSCEV(Gep->getOperand(i)), TheLoop)) return 0; InductionInfo II = Inductions[Phi]; if (IK_PtrInduction == II.IK) return 1; else if (IK_ReversePtrInduction == II.IK) return -1; } unsigned InductionOperand = getGEPInductionOperand(DL, Gep); // Check that all of the gep indices are uniform except for our induction // operand. for (unsigned i = 0; i != NumOperands; ++i) if (i != InductionOperand && !SE->isLoopInvariant(SE->getSCEV(Gep->getOperand(i)), TheLoop)) return 0; // We can emit wide load/stores only if the last non-zero index is the // induction variable. const SCEV *Last = nullptr; if (!Strides.count(Gep)) Last = SE->getSCEV(Gep->getOperand(InductionOperand)); else { // Because of the multiplication by a stride we can have a s/zext cast. // We are going to replace this stride by 1 so the cast is safe to ignore. // // %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ] // %0 = trunc i64 %indvars.iv to i32 // %mul = mul i32 %0, %Stride1 // %idxprom = zext i32 %mul to i64 << Safe cast. // %arrayidx = getelementptr inbounds i32* %B, i64 %idxprom // Last = replaceSymbolicStrideSCEV(SE, Strides, Gep->getOperand(InductionOperand), Gep); if (const SCEVCastExpr *C = dyn_cast
(Last)) Last = (C->getSCEVType() == scSignExtend || C->getSCEVType() == scZeroExtend) ? C->getOperand() : Last; } if (const SCEVAddRecExpr *AR = dyn_cast
(Last)) { const SCEV *Step = AR->getStepRecurrence(*SE); // The memory is consecutive because the last index is consecutive // and all other indices are loop invariant. if (Step->isOne()) return 1; if (Step->isAllOnesValue()) return -1; } return 0; } bool LoopVectorizationLegality::isUniform(Value *V) { return (SE->isLoopInvariant(SE->getSCEV(V), TheLoop)); } InnerLoopVectorizer::VectorParts& InnerLoopVectorizer::getVectorValue(Value *V) { assert(V != Induction && "The new induction variable should not be used."); assert(!V->getType()->isVectorTy() && "Can't widen a vector"); // If we have a stride that is replaced by one, do it here. if (Legal->hasStride(V)) V = ConstantInt::get(V->getType(), 1); // If we have this scalar in the map, return it. if (WidenMap.has(V)) return WidenMap.get(V); // If this scalar is unknown, assume that it is a constant or that it is // loop invariant. Broadcast V and save the value for future uses. Value *B = getBroadcastInstrs(V); return WidenMap.splat(V, B); } Value *InnerLoopVectorizer::reverseVector(Value *Vec) { assert(Vec->getType()->isVectorTy() && "Invalid type"); SmallVector
ShuffleMask; for (unsigned i = 0; i < VF; ++i) ShuffleMask.push_back(Builder.getInt32(VF - i - 1)); return Builder.CreateShuffleVector(Vec, UndefValue::get(Vec->getType()), ConstantVector::get(ShuffleMask), "reverse"); } void InnerLoopVectorizer::vectorizeMemoryInstruction(Instruction *Instr) { // Attempt to issue a wide load. LoadInst *LI = dyn_cast
(Instr); StoreInst *SI = dyn_cast
(Instr); assert((LI || SI) && "Invalid Load/Store instruction"); Type *ScalarDataTy = LI ? LI->getType() : SI->getValueOperand()->getType(); Type *DataTy = VectorType::get(ScalarDataTy, VF); Value *Ptr = LI ? LI->getPointerOperand() : SI->getPointerOperand(); unsigned Alignment = LI ? LI->getAlignment() : SI->getAlignment(); // An alignment of 0 means target abi alignment. We need to use the scalar's // target abi alignment in such a case. if (!Alignment) Alignment = DL->getABITypeAlignment(ScalarDataTy); unsigned AddressSpace = Ptr->getType()->getPointerAddressSpace(); unsigned ScalarAllocatedSize = DL->getTypeAllocSize(ScalarDataTy); unsigned VectorElementSize = DL->getTypeStoreSize(DataTy)/VF; if (SI && Legal->blockNeedsPredication(SI->getParent())) return scalarizeInstruction(Instr, true); if (ScalarAllocatedSize != VectorElementSize) return scalarizeInstruction(Instr); // If the pointer is loop invariant or if it is non-consecutive, // scalarize the load. int ConsecutiveStride = Legal->isConsecutivePtr(Ptr); bool Reverse = ConsecutiveStride < 0; bool UniformLoad = LI && Legal->isUniform(Ptr); if (!ConsecutiveStride || UniformLoad) return scalarizeInstruction(Instr); Constant *Zero = Builder.getInt32(0); VectorParts &Entry = WidenMap.get(Instr); // Handle consecutive loads/stores. GetElementPtrInst *Gep = dyn_cast
(Ptr); if (Gep && Legal->isInductionVariable(Gep->getPointerOperand())) { setDebugLocFromInst(Builder, Gep); Value *PtrOperand = Gep->getPointerOperand(); Value *FirstBasePtr = getVectorValue(PtrOperand)[0]; FirstBasePtr = Builder.CreateExtractElement(FirstBasePtr, Zero); // Create the new GEP with the new induction variable. GetElementPtrInst *Gep2 = cast
(Gep->clone()); Gep2->setOperand(0, FirstBasePtr); Gep2->setName("gep.indvar.base"); Ptr = Builder.Insert(Gep2); } else if (Gep) { setDebugLocFromInst(Builder, Gep); assert(SE->isLoopInvariant(SE->getSCEV(Gep->getPointerOperand()), OrigLoop) && "Base ptr must be invariant"); // The last index does not have to be the induction. It can be // consecutive and be a function of the index. For example A[I+1]; unsigned NumOperands = Gep->getNumOperands(); unsigned InductionOperand = getGEPInductionOperand(DL, Gep); // Create the new GEP with the new induction variable. GetElementPtrInst *Gep2 = cast
(Gep->clone()); for (unsigned i = 0; i < NumOperands; ++i) { Value *GepOperand = Gep->getOperand(i); Instruction *GepOperandInst = dyn_cast
(GepOperand); // Update last index or loop invariant instruction anchored in loop. if (i == InductionOperand || (GepOperandInst && OrigLoop->contains(GepOperandInst))) { assert((i == InductionOperand || SE->isLoopInvariant(SE->getSCEV(GepOperandInst), OrigLoop)) && "Must be last index or loop invariant"); VectorParts &GEPParts = getVectorValue(GepOperand); Value *Index = GEPParts[0]; Index = Builder.CreateExtractElement(Index, Zero); Gep2->setOperand(i, Index); Gep2->setName("gep.indvar.idx"); } } Ptr = Builder.Insert(Gep2); } else { // Use the induction element ptr. assert(isa
(Ptr) && "Invalid induction ptr"); setDebugLocFromInst(Builder, Ptr); VectorParts &PtrVal = getVectorValue(Ptr); Ptr = Builder.CreateExtractElement(PtrVal[0], Zero); } // Handle Stores: if (SI) { assert(!Legal->isUniform(SI->getPointerOperand()) && "We do not allow storing to uniform addresses"); setDebugLocFromInst(Builder, SI); // We don't want to update the value in the map as it might be used in // another expression. So don't use a reference type for "StoredVal". VectorParts StoredVal = getVectorValue(SI->getValueOperand()); for (unsigned Part = 0; Part < UF; ++Part) { // Calculate the pointer for the specific unroll-part. Value *PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(Part * VF)); if (Reverse) { // If we store to reverse consecutive memory locations then we need // to reverse the order of elements in the stored value. StoredVal[Part] = reverseVector(StoredVal[Part]); // If the address is consecutive but reversed, then the // wide store needs to start at the last vector element. PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(-Part * VF)); PartPtr = Builder.CreateGEP(PartPtr, Builder.getInt32(1 - VF)); } Value *VecPtr = Builder.CreateBitCast(PartPtr, DataTy->getPointerTo(AddressSpace)); Builder.CreateStore(StoredVal[Part], VecPtr)->setAlignment(Alignment); } return; } // Handle loads. assert(LI && "Must have a load instruction"); setDebugLocFromInst(Builder, LI); for (unsigned Part = 0; Part < UF; ++Part) { // Calculate the pointer for the specific unroll-part. Value *PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(Part * VF)); if (Reverse) { // If the address is consecutive but reversed, then the // wide store needs to start at the last vector element. PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(-Part * VF)); PartPtr = Builder.CreateGEP(PartPtr, Builder.getInt32(1 - VF)); } Value *VecPtr = Builder.CreateBitCast(PartPtr, DataTy->getPointerTo(AddressSpace)); Value *LI = Builder.CreateLoad(VecPtr, "wide.load"); cast
(LI)->setAlignment(Alignment); Entry[Part] = Reverse ? reverseVector(LI) : LI; } } void InnerLoopVectorizer::scalarizeInstruction(Instruction *Instr, bool IfPredicateStore) { assert(!Instr->getType()->isAggregateType() && "Can't handle vectors"); // Holds vector parameters or scalars, in case of uniform vals. SmallVector
Params; setDebugLocFromInst(Builder, Instr); // Find all of the vectorized parameters. for (unsigned op = 0, e = Instr->getNumOperands(); op != e; ++op) { Value *SrcOp = Instr->getOperand(op); // If we are accessing the old induction variable, use the new one. if (SrcOp == OldInduction) { Params.push_back(getVectorValue(SrcOp)); continue; } // Try using previously calculated values. Instruction *SrcInst = dyn_cast
(SrcOp); // If the src is an instruction that appeared earlier in the basic block // then it should already be vectorized. if (SrcInst && OrigLoop->contains(SrcInst)) { assert(WidenMap.has(SrcInst) && "Source operand is unavailable"); // The parameter is a vector value from earlier. Params.push_back(WidenMap.get(SrcInst)); } else { // The parameter is a scalar from outside the loop. Maybe even a constant. VectorParts Scalars; Scalars.append(UF, SrcOp); Params.push_back(Scalars); } } assert(Params.size() == Instr->getNumOperands() && "Invalid number of operands"); // Does this instruction return a value ? bool IsVoidRetTy = Instr->getType()->isVoidTy(); Value *UndefVec = IsVoidRetTy ? nullptr : UndefValue::get(VectorType::get(Instr->getType(), VF)); // Create a new entry in the WidenMap and initialize it to Undef or Null. VectorParts &VecResults = WidenMap.splat(Instr, UndefVec); Instruction *InsertPt = Builder.GetInsertPoint(); BasicBlock *IfBlock = Builder.GetInsertBlock(); BasicBlock *CondBlock = nullptr; VectorParts Cond; Loop *VectorLp = nullptr; if (IfPredicateStore) { assert(Instr->getParent()->getSinglePredecessor() && "Only support single predecessor blocks"); Cond = createEdgeMask(Instr->getParent()->getSinglePredecessor(), Instr->getParent()); VectorLp = LI->getLoopFor(IfBlock); assert(VectorLp && "Must have a loop for this block"); } // For each vector unroll 'part': for (unsigned Part = 0; Part < UF; ++Part) { // For each scalar that we create: for (unsigned Width = 0; Width < VF; ++Width) { // Start if-block. Value *Cmp = nullptr; if (IfPredicateStore) { Cmp = Builder.CreateExtractElement(Cond[Part], Builder.getInt32(Width)); Cmp = Builder.CreateICmp(ICmpInst::ICMP_EQ, Cmp, ConstantInt::get(Cmp->getType(), 1)); CondBlock = IfBlock->splitBasicBlock(InsertPt, "cond.store"); LoopVectorBody.push_back(CondBlock); VectorLp->addBasicBlockToLoop(CondBlock, LI->getBase()); // Update Builder with newly created basic block. Builder.SetInsertPoint(InsertPt); } Instruction *Cloned = Instr->clone(); if (!IsVoidRetTy) Cloned->setName(Instr->getName() + ".cloned"); // Replace the operands of the cloned instructions with extracted scalars. for (unsigned op = 0, e = Instr->getNumOperands(); op != e; ++op) { Value *Op = Params[op][Part]; // Param is a vector. Need to extract the right lane. if (Op->getType()->isVectorTy()) Op = Builder.CreateExtractElement(Op, Builder.getInt32(Width)); Cloned->setOperand(op, Op); } // Place the cloned scalar in the new loop. Builder.Insert(Cloned); // If the original scalar returns a value we need to place it in a vector // so that future users will be able to use it. if (!IsVoidRetTy) VecResults[Part] = Builder.CreateInsertElement(VecResults[Part], Cloned, Builder.getInt32(Width)); // End if-block. if (IfPredicateStore) { BasicBlock *NewIfBlock = CondBlock->splitBasicBlock(InsertPt, "else"); LoopVectorBody.push_back(NewIfBlock); VectorLp->addBasicBlockToLoop(NewIfBlock, LI->getBase()); Builder.SetInsertPoint(InsertPt); Instruction *OldBr = IfBlock->getTerminator(); BranchInst::Create(CondBlock, NewIfBlock, Cmp, OldBr); OldBr->eraseFromParent(); IfBlock = NewIfBlock; } } } } static Instruction *getFirstInst(Instruction *FirstInst, Value *V, Instruction *Loc) { if (FirstInst) return FirstInst; if (Instruction *I = dyn_cast
(V)) return I->getParent() == Loc->getParent() ? I : nullptr; return nullptr; } std::pair
InnerLoopVectorizer::addStrideCheck(Instruction *Loc) { Instruction *tnullptr = nullptr; if (!Legal->mustCheckStrides()) return std::pair
(tnullptr, tnullptr); IRBuilder<> ChkBuilder(Loc); // Emit checks. Value *Check = nullptr; Instruction *FirstInst = nullptr; for (SmallPtrSet
::iterator SI = Legal->strides_begin(), SE = Legal->strides_end(); SI != SE; ++SI) { Value *Ptr = stripIntegerCast(*SI); Value *C = ChkBuilder.CreateICmpNE(Ptr, ConstantInt::get(Ptr->getType(), 1), "stride.chk"); // Store the first instruction we create. FirstInst = getFirstInst(FirstInst, C, Loc); if (Check) Check = ChkBuilder.CreateOr(Check, C); else Check = C; } // We have to do this trickery because the IRBuilder might fold the check to a // constant expression in which case there is no Instruction anchored in a // the block. LLVMContext &Ctx = Loc->getContext(); Instruction *TheCheck = BinaryOperator::CreateAnd(Check, ConstantInt::getTrue(Ctx)); ChkBuilder.Insert(TheCheck, "stride.not.one"); FirstInst = getFirstInst(FirstInst, TheCheck, Loc); return std::make_pair(FirstInst, TheCheck); } std::pair
InnerLoopVectorizer::addRuntimeCheck(Instruction *Loc) { LoopVectorizationLegality::RuntimePointerCheck *PtrRtCheck = Legal->getRuntimePointerCheck(); Instruction *tnullptr = nullptr; if (!PtrRtCheck->Need) return std::pair
(tnullptr, tnullptr); unsigned NumPointers = PtrRtCheck->Pointers.size(); SmallVector
, 2> Starts; SmallVector
, 2> Ends; LLVMContext &Ctx = Loc->getContext(); SCEVExpander Exp(*SE, "induction"); Instruction *FirstInst = nullptr; for (unsigned i = 0; i < NumPointers; ++i) { Value *Ptr = PtrRtCheck->Pointers[i]; const SCEV *Sc = SE->getSCEV(Ptr); if (SE->isLoopInvariant(Sc, OrigLoop)) { DEBUG(dbgs() << "LV: Adding RT check for a loop invariant ptr:" << *Ptr <<"\n"); Starts.push_back(Ptr); Ends.push_back(Ptr); } else { DEBUG(dbgs() << "LV: Adding RT check for range:" << *Ptr << '\n'); unsigned AS = Ptr->getType()->getPointerAddressSpace(); // Use this type for pointer arithmetic. Type *PtrArithTy = Type::getInt8PtrTy(Ctx, AS); Value *Start = Exp.expandCodeFor(PtrRtCheck->Starts[i], PtrArithTy, Loc); Value *End = Exp.expandCodeFor(PtrRtCheck->Ends[i], PtrArithTy, Loc); Starts.push_back(Start); Ends.push_back(End); } } IRBuilder<> ChkBuilder(Loc); // Our instructions might fold to a constant. Value *MemoryRuntimeCheck = nullptr; for (unsigned i = 0; i < NumPointers; ++i) { for (unsigned j = i+1; j < NumPointers; ++j) { // No need to check if two readonly pointers intersect. if (!PtrRtCheck->IsWritePtr[i] && !PtrRtCheck->IsWritePtr[j]) continue; // Only need to check pointers between two different dependency sets. if (PtrRtCheck->DependencySetId[i] == PtrRtCheck->DependencySetId[j]) continue; unsigned AS0 = Starts[i]->getType()->getPointerAddressSpace(); unsigned AS1 = Starts[j]->getType()->getPointerAddressSpace(); assert((AS0 == Ends[j]->getType()->getPointerAddressSpace()) && (AS1 == Ends[i]->getType()->getPointerAddressSpace()) && "Trying to bounds check pointers with different address spaces"); Type *PtrArithTy0 = Type::getInt8PtrTy(Ctx, AS0); Type *PtrArithTy1 = Type::getInt8PtrTy(Ctx, AS1); Value *Start0 = ChkBuilder.CreateBitCast(Starts[i], PtrArithTy0, "bc"); Value *Start1 = ChkBuilder.CreateBitCast(Starts[j], PtrArithTy1, "bc"); Value *End0 = ChkBuilder.CreateBitCast(Ends[i], PtrArithTy1, "bc"); Value *End1 = ChkBuilder.CreateBitCast(Ends[j], PtrArithTy0, "bc"); Value *Cmp0 = ChkBuilder.CreateICmpULE(Start0, End1, "bound0"); FirstInst = getFirstInst(FirstInst, Cmp0, Loc); Value *Cmp1 = ChkBuilder.CreateICmpULE(Start1, End0, "bound1"); FirstInst = getFirstInst(FirstInst, Cmp1, Loc); Value *IsConflict = ChkBuilder.CreateAnd(Cmp0, Cmp1, "found.conflict"); FirstInst = getFirstInst(FirstInst, IsConflict, Loc); if (MemoryRuntimeCheck) { IsConflict = ChkBuilder.CreateOr(MemoryRuntimeCheck, IsConflict, "conflict.rdx"); FirstInst = getFirstInst(FirstInst, IsConflict, Loc); } MemoryRuntimeCheck = IsConflict; } } // We have to do this trickery because the IRBuilder might fold the check to a // constant expression in which case there is no Instruction anchored in a // the block. Instruction *Check = BinaryOperator::CreateAnd(MemoryRuntimeCheck, ConstantInt::getTrue(Ctx)); ChkBuilder.Insert(Check, "memcheck.conflict"); FirstInst = getFirstInst(FirstInst, Check, Loc); return std::make_pair(FirstInst, Check); } void InnerLoopVectorizer::createEmptyLoop() { /* In this function we generate a new loop. The new loop will contain the vectorized instructions while the old loop will continue to run the scalar remainder. [ ] <-- Back-edge taken count overflow check. / | / v | [ ] <-- vector loop bypass (may consist of multiple blocks). | / | | / v || [ ] <-- vector pre header. || | || v || [ ] \ || [ ]_| <-- vector loop. || | | \ v | >[ ] <--- middle-block. | / | | / v -|- >[ ] <--- new preheader. | | | v | [ ] \ | [ ]_| <-- old scalar loop to handle remainder. \ | \ v >[ ] <-- exit block. ... */ BasicBlock *OldBasicBlock = OrigLoop->getHeader(); BasicBlock *BypassBlock = OrigLoop->getLoopPreheader(); BasicBlock *ExitBlock = OrigLoop->getExitBlock(); assert(BypassBlock && "Invalid loop structure"); assert(ExitBlock && "Must have an exit block"); // Some loops have a single integer induction variable, while other loops // don't. One example is c++ iterators that often have multiple pointer // induction variables. In the code below we also support a case where we // don't have a single induction variable. OldInduction = Legal->getInduction(); Type *IdxTy = Legal->getWidestInductionType(); // Find the loop boundaries. const SCEV *ExitCount = SE->getBackedgeTakenCount(OrigLoop); assert(ExitCount != SE->getCouldNotCompute() && "Invalid loop count"); // The exit count might have the type of i64 while the phi is i32. This can // happen if we have an induction variable that is sign extended before the // compare. The only way that we get a backedge taken count is that the // induction variable was signed and as such will not overflow. In such a case // truncation is legal. if (ExitCount->getType()->getPrimitiveSizeInBits() > IdxTy->getPrimitiveSizeInBits()) ExitCount = SE->getTruncateOrNoop(ExitCount, IdxTy); const SCEV *BackedgeTakeCount = SE->getNoopOrZeroExtend(ExitCount, IdxTy); // Get the total trip count from the count by adding 1. ExitCount = SE->getAddExpr(BackedgeTakeCount, SE->getConstant(BackedgeTakeCount->getType(), 1)); // Expand the trip count and place the new instructions in the preheader. // Notice that the pre-header does not change, only the loop body. SCEVExpander Exp(*SE, "induction"); // We need to test whether the backedge-taken count is uint##_max. Adding one // to it will cause overflow and an incorrect loop trip count in the vector // body. In case of overflow we want to directly jump to the scalar remainder // loop. Value *BackedgeCount = Exp.expandCodeFor(BackedgeTakeCount, BackedgeTakeCount->getType(), BypassBlock->getTerminator()); if (BackedgeCount->getType()->isPointerTy()) BackedgeCount = CastInst::CreatePointerCast(BackedgeCount, IdxTy, "backedge.ptrcnt.to.int", BypassBlock->getTerminator()); Instruction *CheckBCOverflow = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_EQ, BackedgeCount, Constant::getAllOnesValue(BackedgeCount->getType()), "backedge.overflow", BypassBlock->getTerminator()); // The loop index does not have to start at Zero. Find the original start // value from the induction PHI node. If we don't have an induction variable // then we know that it starts at zero. Builder.SetInsertPoint(BypassBlock->getTerminator()); Value *StartIdx = ExtendedIdx = OldInduction ? Builder.CreateZExt(OldInduction->getIncomingValueForBlock(BypassBlock), IdxTy): ConstantInt::get(IdxTy, 0); // We need an instruction to anchor the overflow check on. StartIdx needs to // be defined before the overflow check branch. Because the scalar preheader // is going to merge the start index and so the overflow branch block needs to // contain a definition of the start index. Instruction *OverflowCheckAnchor = BinaryOperator::CreateAdd( StartIdx, ConstantInt::get(IdxTy, 0), "overflow.check.anchor", BypassBlock->getTerminator()); // Count holds the overall loop count (N). Value *Count = Exp.expandCodeFor(ExitCount, ExitCount->getType(), BypassBlock->getTerminator()); LoopBypassBlocks.push_back(BypassBlock); // Split the single block loop into the two loop structure described above. BasicBlock *VectorPH = BypassBlock->splitBasicBlock(BypassBlock->getTerminator(), "vector.ph"); BasicBlock *VecBody = VectorPH->splitBasicBlock(VectorPH->getTerminator(), "vector.body"); BasicBlock *MiddleBlock = VecBody->splitBasicBlock(VecBody->getTerminator(), "middle.block"); BasicBlock *ScalarPH = MiddleBlock->splitBasicBlock(MiddleBlock->getTerminator(), "scalar.ph"); // Create and register the new vector loop. Loop* Lp = new Loop(); Loop *ParentLoop = OrigLoop->getParentLoop(); // Insert the new loop into the loop nest and register the new basic blocks // before calling any utilities such as SCEV that require valid LoopInfo. if (ParentLoop) { ParentLoop->addChildLoop(Lp); ParentLoop->addBasicBlockToLoop(ScalarPH, LI->getBase()); ParentLoop->addBasicBlockToLoop(VectorPH, LI->getBase()); ParentLoop->addBasicBlockToLoop(MiddleBlock, LI->getBase()); } else { LI->addTopLevelLoop(Lp); } Lp->addBasicBlockToLoop(VecBody, LI->getBase()); // Use this IR builder to create the loop instructions (Phi, Br, Cmp) // inside the loop. Builder.SetInsertPoint(VecBody->getFirstNonPHI()); // Generate the induction variable. setDebugLocFromInst(Builder, getDebugLocFromInstOrOperands(OldInduction)); Induction = Builder.CreatePHI(IdxTy, 2, "index"); // The loop step is equal to the vectorization factor (num of SIMD elements) // times the unroll factor (num of SIMD instructions). Constant *Step = ConstantInt::get(IdxTy, VF * UF); // This is the IR builder that we use to add all of the logic for bypassing // the new vector loop. IRBuilder<> BypassBuilder(BypassBlock->getTerminator()); setDebugLocFromInst(BypassBuilder, getDebugLocFromInstOrOperands(OldInduction)); // We may need to extend the index in case there is a type mismatch. // We know that the count starts at zero and does not overflow. if (Count->getType() != IdxTy) { // The exit count can be of pointer type. Convert it to the correct // integer type. if (ExitCount->getType()->isPointerTy()) Count = BypassBuilder.CreatePointerCast(Count, IdxTy, "ptrcnt.to.int"); else Count = BypassBuilder.CreateZExtOrTrunc(Count, IdxTy, "cnt.cast"); } // Add the start index to the loop count to get the new end index. Value *IdxEnd = BypassBuilder.CreateAdd(Count, StartIdx, "end.idx"); // Now we need to generate the expression for N - (N % VF), which is // the part that the vectorized body will execute. Value *R = BypassBuilder.CreateURem(Count, Step, "n.mod.vf"); Value *CountRoundDown = BypassBuilder.CreateSub(Count, R, "n.vec"); Value *IdxEndRoundDown = BypassBuilder.CreateAdd(CountRoundDown, StartIdx, "end.idx.rnd.down"); // Now, compare the new count to zero. If it is zero skip the vector loop and // jump to the scalar loop. Value *Cmp = BypassBuilder.CreateICmpEQ(IdxEndRoundDown, StartIdx, "cmp.zero"); BasicBlock *LastBypassBlock = BypassBlock; // Generate code to check that the loops trip count that we computed by adding // one to the backedge-taken count will not overflow. { auto PastOverflowCheck = std::next(BasicBlock::iterator(OverflowCheckAnchor)); BasicBlock *CheckBlock = LastBypassBlock->splitBasicBlock(PastOverflowCheck, "overflow.checked"); if (ParentLoop) ParentLoop->addBasicBlockToLoop(CheckBlock, LI->getBase()); LoopBypassBlocks.push_back(CheckBlock); Instruction *OldTerm = LastBypassBlock->getTerminator(); BranchInst::Create(ScalarPH, CheckBlock, CheckBCOverflow, OldTerm); OldTerm->eraseFromParent(); LastBypassBlock = CheckBlock; } // Generate the code to check that the strides we assumed to be one are really // one. We want the new basic block to start at the first instruction in a // sequence of instructions that form a check. Instruction *StrideCheck; Instruction *FirstCheckInst; std::tie(FirstCheckInst, StrideCheck) = addStrideCheck(LastBypassBlock->getTerminator()); if (StrideCheck) { // Create a new block containing the stride check. BasicBlock *CheckBlock = LastBypassBlock->splitBasicBlock(FirstCheckInst, "vector.stridecheck"); if (ParentLoop) ParentLoop->addBasicBlockToLoop(CheckBlock, LI->getBase()); LoopBypassBlocks.push_back(CheckBlock); // Replace the branch into the memory check block with a conditional branch // for the "few elements case". Instruction *OldTerm = LastBypassBlock->getTerminator(); BranchInst::Create(MiddleBlock, CheckBlock, Cmp, OldTerm); OldTerm->eraseFromParent(); Cmp = StrideCheck; LastBypassBlock = CheckBlock; } // Generate the code that checks in runtime if arrays overlap. We put the // checks into a separate block to make the more common case of few elements // faster. Instruction *MemRuntimeCheck; std::tie(FirstCheckInst, MemRuntimeCheck) = addRuntimeCheck(LastBypassBlock->getTerminator()); if (MemRuntimeCheck) { // Create a new block containing the memory check. BasicBlock *CheckBlock = LastBypassBlock->splitBasicBlock(MemRuntimeCheck, "vector.memcheck"); if (ParentLoop) ParentLoop->addBasicBlockToLoop(CheckBlock, LI->getBase()); LoopBypassBlocks.push_back(CheckBlock); // Replace the branch into the memory check block with a conditional branch // for the "few elements case". Instruction *OldTerm = LastBypassBlock->getTerminator(); BranchInst::Create(MiddleBlock, CheckBlock, Cmp, OldTerm); OldTerm->eraseFromParent(); Cmp = MemRuntimeCheck; LastBypassBlock = CheckBlock; } LastBypassBlock->getTerminator()->eraseFromParent(); BranchInst::Create(MiddleBlock, VectorPH, Cmp, LastBypassBlock); // We are going to resume the execution of the scalar loop. // Go over all of the induction variables that we found and fix the // PHIs that are left in the scalar version of the loop. // The starting values of PHI nodes depend on the counter of the last // iteration in the vectorized loop. // If we come from a bypass edge then we need to start from the original // start value. // This variable saves the new starting index for the scalar loop. PHINode *ResumeIndex = nullptr; LoopVectorizationLegality::InductionList::iterator I, E; LoopVectorizationLegality::InductionList *List = Legal->getInductionVars(); // Set builder to point to last bypass block. BypassBuilder.SetInsertPoint(LoopBypassBlocks.back()->getTerminator()); for (I = List->begin(), E = List->end(); I != E; ++I) { PHINode *OrigPhi = I->first; LoopVectorizationLegality::InductionInfo II = I->second; Type *ResumeValTy = (OrigPhi == OldInduction) ? IdxTy : OrigPhi->getType(); PHINode *ResumeVal = PHINode::Create(ResumeValTy, 2, "resume.val", MiddleBlock->getTerminator()); // We might have extended the type of the induction variable but we need a // truncated version for the scalar loop. PHINode *TruncResumeVal = (OrigPhi == OldInduction) ? PHINode::Create(OrigPhi->getType(), 2, "trunc.resume.val", MiddleBlock->getTerminator()) : nullptr; // Create phi nodes to merge from the backedge-taken check block. PHINode *BCResumeVal = PHINode::Create(ResumeValTy, 3, "bc.resume.val", ScalarPH->getTerminator()); BCResumeVal->addIncoming(ResumeVal, MiddleBlock); PHINode *BCTruncResumeVal = nullptr; if (OrigPhi == OldInduction) { BCTruncResumeVal = PHINode::Create(OrigPhi->getType(), 2, "bc.trunc.resume.val", ScalarPH->getTerminator()); BCTruncResumeVal->addIncoming(TruncResumeVal, MiddleBlock); } Value *EndValue = nullptr; switch (II.IK) { case LoopVectorizationLegality::IK_NoInduction: llvm_unreachable("Unknown induction"); case LoopVectorizationLegality::IK_IntInduction: { // Handle the integer induction counter. assert(OrigPhi->getType()->isIntegerTy() && "Invalid type"); // We have the canonical induction variable. if (OrigPhi == OldInduction) { // Create a truncated version of the resume value for the scalar loop, // we might have promoted the type to a larger width. EndValue = BypassBuilder.CreateTrunc(IdxEndRoundDown, OrigPhi->getType()); // The new PHI merges the original incoming value, in case of a bypass, // or the value at the end of the vectorized loop. for (unsigned I = 1, E = LoopBypassBlocks.size(); I != E; ++I) TruncResumeVal->addIncoming(II.StartValue, LoopBypassBlocks[I]); TruncResumeVal->addIncoming(EndValue, VecBody); BCTruncResumeVal->addIncoming(II.StartValue, LoopBypassBlocks[0]); // We know what the end value is. EndValue = IdxEndRoundDown; // We also know which PHI node holds it. ResumeIndex = ResumeVal; break; } // Not the canonical induction variable - add the vector loop count to the // start value. Value *CRD = BypassBuilder.CreateSExtOrTrunc(CountRoundDown, II.StartValue->getType(), "cast.crd"); EndValue = BypassBuilder.CreateAdd(CRD, II.StartValue , "ind.end"); break; } case LoopVectorizationLegality::IK_ReverseIntInduction: { // Convert the CountRoundDown variable to the PHI size. Value *CRD = BypassBuilder.CreateSExtOrTrunc(CountRoundDown, II.StartValue->getType(), "cast.crd"); // Handle reverse integer induction counter. EndValue = BypassBuilder.CreateSub(II.StartValue, CRD, "rev.ind.end"); break; } case LoopVectorizationLegality::IK_PtrInduction: { // For pointer induction variables, calculate the offset using // the end index. EndValue = BypassBuilder.CreateGEP(II.StartValue, CountRoundDown, "ptr.ind.end"); break; } case LoopVectorizationLegality::IK_ReversePtrInduction: { // The value at the end of the loop for the reverse pointer is calculated // by creating a GEP with a negative index starting from the start value. Value *Zero = ConstantInt::get(CountRoundDown->getType(), 0); Value *NegIdx = BypassBuilder.CreateSub(Zero, CountRoundDown, "rev.ind.end"); EndValue = BypassBuilder.CreateGEP(II.StartValue, NegIdx, "rev.ptr.ind.end"); break; } }// end of case // The new PHI merges the original incoming value, in case of a bypass, // or the value at the end of the vectorized loop. for (unsigned I = 1, E = LoopBypassBlocks.size(); I != E; ++I) { if (OrigPhi == OldInduction) ResumeVal->addIncoming(StartIdx, LoopBypassBlocks[I]); else ResumeVal->addIncoming(II.StartValue, LoopBypassBlocks[I]); } ResumeVal->addIncoming(EndValue, VecBody); // Fix the scalar body counter (PHI node). unsigned BlockIdx = OrigPhi->getBasicBlockIndex(ScalarPH); // The old induction's phi node in the scalar body needs the truncated // value. if (OrigPhi == OldInduction) { BCResumeVal->addIncoming(StartIdx, LoopBypassBlocks[0]); OrigPhi->setIncomingValue(BlockIdx, BCTruncResumeVal); } else { BCResumeVal->addIncoming(II.StartValue, LoopBypassBlocks[0]); OrigPhi->setIncomingValue(BlockIdx, BCResumeVal); } } // If we are generating a new induction variable then we also need to // generate the code that calculates the exit value. This value is not // simply the end of the counter because we may skip the vectorized body // in case of a runtime check. if (!OldInduction){ assert(!ResumeIndex && "Unexpected resume value found"); ResumeIndex = PHINode::Create(IdxTy, 2, "new.indc.resume.val", MiddleBlock->getTerminator()); for (unsigned I = 1, E = LoopBypassBlocks.size(); I != E; ++I) ResumeIndex->addIncoming(StartIdx, LoopBypassBlocks[I]); ResumeIndex->addIncoming(IdxEndRoundDown, VecBody); } // Make sure that we found the index where scalar loop needs to continue. assert(ResumeIndex && ResumeIndex->getType()->isIntegerTy() && "Invalid resume Index"); // Add a check in the middle block to see if we have completed // all of the iterations in the first vector loop. // If (N - N%VF) == N, then we *don't* need to run the remainder. Value *CmpN = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_EQ, IdxEnd, ResumeIndex, "cmp.n", MiddleBlock->getTerminator()); BranchInst::Create(ExitBlock, ScalarPH, CmpN, MiddleBlock->getTerminator()); // Remove the old terminator. MiddleBlock->getTerminator()->eraseFromParent(); // Create i+1 and fill the PHINode. Value *NextIdx = Builder.CreateAdd(Induction, Step, "index.next"); Induction->addIncoming(StartIdx, VectorPH); Induction->addIncoming(NextIdx, VecBody); // Create the compare. Value *ICmp = Builder.CreateICmpEQ(NextIdx, IdxEndRoundDown); Builder.CreateCondBr(ICmp, MiddleBlock, VecBody); // Now we have two terminators. Remove the old one from the block. VecBody->getTerminator()->eraseFromParent(); // Get ready to start creating new instructions into the vectorized body. Builder.SetInsertPoint(VecBody->getFirstInsertionPt()); // Save the state. LoopVectorPreHeader = VectorPH; LoopScalarPreHeader = ScalarPH; LoopMiddleBlock = MiddleBlock; LoopExitBlock = ExitBlock; LoopVectorBody.push_back(VecBody); LoopScalarBody = OldBasicBlock; LoopVectorizeHints Hints(Lp, true); Hints.setAlreadyVectorized(Lp); } /// This function returns the identity element (or neutral element) for /// the operation K. Constant* LoopVectorizationLegality::getReductionIdentity(ReductionKind K, Type *Tp) { switch (K) { case RK_IntegerXor: case RK_IntegerAdd: case RK_IntegerOr: // Adding, Xoring, Oring zero to a number does not change it. return ConstantInt::get(Tp, 0); case RK_IntegerMult: // Multiplying a number by 1 does not change it. return ConstantInt::get(Tp, 1); case RK_IntegerAnd: // AND-ing a number with an all-1 value does not change it. return ConstantInt::get(Tp, -1, true); case RK_FloatMult: // Multiplying a number by 1 does not change it. return ConstantFP::get(Tp, 1.0L); case RK_FloatAdd: // Adding zero to a number does not change it. return ConstantFP::get(Tp, 0.0L); default: llvm_unreachable("Unknown reduction kind"); } } /// This function translates the reduction kind to an LLVM binary operator. static unsigned getReductionBinOp(LoopVectorizationLegality::ReductionKind Kind) { switch (Kind) { case LoopVectorizationLegality::RK_IntegerAdd: return Instruction::Add; case LoopVectorizationLegality::RK_IntegerMult: return Instruction::Mul; case LoopVectorizationLegality::RK_IntegerOr: return Instruction::Or; case LoopVectorizationLegality::RK_IntegerAnd: return Instruction::And; case LoopVectorizationLegality::RK_IntegerXor: return Instruction::Xor; case LoopVectorizationLegality::RK_FloatMult: return Instruction::FMul; case LoopVectorizationLegality::RK_FloatAdd: return Instruction::FAdd; case LoopVectorizationLegality::RK_IntegerMinMax: return Instruction::ICmp; case LoopVectorizationLegality::RK_FloatMinMax: return Instruction::FCmp; default: llvm_unreachable("Unknown reduction operation"); } } Value *createMinMaxOp(IRBuilder<> &Builder, LoopVectorizationLegality::MinMaxReductionKind RK, Value *Left, Value *Right) { CmpInst::Predicate P = CmpInst::ICMP_NE; switch (RK) { default: llvm_unreachable("Unknown min/max reduction kind"); case LoopVectorizationLegality::MRK_UIntMin: P = CmpInst::ICMP_ULT; break; case LoopVectorizationLegality::MRK_UIntMax: P = CmpInst::ICMP_UGT; break; case LoopVectorizationLegality::MRK_SIntMin: P = CmpInst::ICMP_SLT; break; case LoopVectorizationLegality::MRK_SIntMax: P = CmpInst::ICMP_SGT; break; case LoopVectorizationLegality::MRK_FloatMin: P = CmpInst::FCMP_OLT; break; case LoopVectorizationLegality::MRK_FloatMax: P = CmpInst::FCMP_OGT; break; } Value *Cmp; if (RK == LoopVectorizationLegality::MRK_FloatMin || RK == LoopVectorizationLegality::MRK_FloatMax) Cmp = Builder.CreateFCmp(P, Left, Right, "rdx.minmax.cmp"); else Cmp = Builder.CreateICmp(P, Left, Right, "rdx.minmax.cmp"); Value *Select = Builder.CreateSelect(Cmp, Left, Right, "rdx.minmax.select"); return Select; } namespace { struct CSEDenseMapInfo { static bool canHandle(Instruction *I) { return isa
(I) || isa
(I) || isa
(I) || isa
(I); } static inline Instruction *getEmptyKey() { return DenseMapInfo
::getEmptyKey(); } static inline Instruction *getTombstoneKey() { return DenseMapInfo
::getTombstoneKey(); } static unsigned getHashValue(Instruction *I) { assert(canHandle(I) && "Unknown instruction!"); return hash_combine(I->getOpcode(), hash_combine_range(I->value_op_begin(), I->value_op_end())); } static bool isEqual(Instruction *LHS, Instruction *RHS) { if (LHS == getEmptyKey() || RHS == getEmptyKey() || LHS == getTombstoneKey() || RHS == getTombstoneKey()) return LHS == RHS; return LHS->isIdenticalTo(RHS); } }; } /// \brief Check whether this block is a predicated block. /// Due to if predication of stores we might create a sequence of "if(pred) a[i] /// = ...; " blocks. We start with one vectorized basic block. For every /// conditional block we split this vectorized block. Therefore, every second /// block will be a predicated one. static bool isPredicatedBlock(unsigned BlockNum) { return BlockNum % 2; } ///\brief Perform cse of induction variable instructions. static void cse(SmallVector
&BBs) { // Perform simple cse. SmallDenseMap
CSEMap; for (unsigned i = 0, e = BBs.size(); i != e; ++i) { BasicBlock *BB = BBs[i]; for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E;) { Instruction *In = I++; if (!CSEDenseMapInfo::canHandle(In)) continue; // Check if we can replace this instruction with any of the // visited instructions. if (Instruction *V = CSEMap.lookup(In)) { In->replaceAllUsesWith(V); In->eraseFromParent(); continue; } // Ignore instructions in conditional blocks. We create "if (pred) a[i] = // ...;" blocks for predicated stores. Every second block is a predicated // block. if (isPredicatedBlock(i)) continue; CSEMap[In] = In; } } } /// \brief Adds a 'fast' flag to floating point operations. static Value *addFastMathFlag(Value *V) { if (isa
(V)){ FastMathFlags Flags; Flags.setUnsafeAlgebra(); cast
(V)->setFastMathFlags(Flags); } return V; } void InnerLoopVectorizer::vectorizeLoop() { //===------------------------------------------------===// // // Notice: any optimization or new instruction that go // into the code below should be also be implemented in // the cost-model. // //===------------------------------------------------===// Constant *Zero = Builder.getInt32(0); // In order to support reduction variables we need to be able to vectorize // Phi nodes. Phi nodes have cycles, so we need to vectorize them in two // stages. First, we create a new vector PHI node with no incoming edges. // We use this value when we vectorize all of the instructions that use the // PHI. Next, after all of the instructions in the block are complete we // add the new incoming edges to the PHI. At this point all of the // instructions in the basic block are vectorized, so we can use them to // construct the PHI. PhiVector RdxPHIsToFix; // Scan the loop in a topological order to ensure that defs are vectorized // before users. LoopBlocksDFS DFS(OrigLoop); DFS.perform(LI); // Vectorize all of the blocks in the original loop. for (LoopBlocksDFS::RPOIterator bb = DFS.beginRPO(), be = DFS.endRPO(); bb != be; ++bb) vectorizeBlockInLoop(*bb, &RdxPHIsToFix); // At this point every instruction in the original loop is widened to // a vector form. We are almost done. Now, we need to fix the PHI nodes // that we vectorized. The PHI nodes are currently empty because we did // not want to introduce cycles. Notice that the remaining PHI nodes // that we need to fix are reduction variables. // Create the 'reduced' values for each of the induction vars. // The reduced values are the vector values that we scalarize and combine // after the loop is finished. for (PhiVector::iterator it = RdxPHIsToFix.begin(), e = RdxPHIsToFix.end(); it != e; ++it) { PHINode *RdxPhi = *it; assert(RdxPhi && "Unable to recover vectorized PHI"); // Find the reduction variable descriptor. assert(Legal->getReductionVars()->count(RdxPhi) && "Unable to find the reduction variable"); LoopVectorizationLegality::ReductionDescriptor RdxDesc = (*Legal->getReductionVars())[RdxPhi]; setDebugLocFromInst(Builder, RdxDesc.StartValue); // We need to generate a reduction vector from the incoming scalar. // To do so, we need to generate the 'identity' vector and override // one of the elements with the incoming scalar reduction. We need // to do it in the vector-loop preheader. Builder.SetInsertPoint(LoopBypassBlocks[1]->getTerminator()); // This is the vector-clone of the value that leaves the loop. VectorParts &VectorExit = getVectorValue(RdxDesc.LoopExitInstr); Type *VecTy = VectorExit[0]->getType(); // Find the reduction identity variable. Zero for addition, or, xor, // one for multiplication, -1 for And. Value *Identity; Value *VectorStart; if (RdxDesc.Kind == LoopVectorizationLegality::RK_IntegerMinMax || RdxDesc.Kind == LoopVectorizationLegality::RK_FloatMinMax) { // MinMax reduction have the start value as their identify. if (VF == 1) { VectorStart = Identity = RdxDesc.StartValue; } else { VectorStart = Identity = Builder.CreateVectorSplat(VF, RdxDesc.StartValue, "minmax.ident"); } } else { // Handle other reduction kinds: Constant *Iden = LoopVectorizationLegality::getReductionIdentity(RdxDesc.Kind, VecTy->getScalarType()); if (VF == 1) { Identity = Iden; // This vector is the Identity vector where the first element is the // incoming scalar reduction. VectorStart = RdxDesc.StartValue; } else { Identity = ConstantVector::getSplat(VF, Iden); // This vector is the Identity vector where the first element is the // incoming scalar reduction. VectorStart = Builder.CreateInsertElement(Identity, RdxDesc.StartValue, Zero); } } // Fix the vector-loop phi. // We created the induction variable so we know that the // preheader is the first entry. BasicBlock *VecPreheader = Induction->getIncomingBlock(0); // Reductions do not have to start at zero. They can start with // any loop invariant values. VectorParts &VecRdxPhi = WidenMap.get(RdxPhi); BasicBlock *Latch = OrigLoop->getLoopLatch(); Value *LoopVal = RdxPhi->getIncomingValueForBlock(Latch); VectorParts &Val = getVectorValue(LoopVal); for (unsigned part = 0; part < UF; ++part) { // Make sure to add the reduction stat value only to the // first unroll part. Value *StartVal = (part == 0) ? VectorStart : Identity; cast
(VecRdxPhi[part])->addIncoming(StartVal, VecPreheader); cast
(VecRdxPhi[part])->addIncoming(Val[part], LoopVectorBody.back()); } // Before each round, move the insertion point right between // the PHIs and the values we are going to write. // This allows us to write both PHINodes and the extractelement // instructions. Builder.SetInsertPoint(LoopMiddleBlock->getFirstInsertionPt()); VectorParts RdxParts; setDebugLocFromInst(Builder, RdxDesc.LoopExitInstr); for (unsigned part = 0; part < UF; ++part) { // This PHINode contains the vectorized reduction variable, or // the initial value vector, if we bypass the vector loop. VectorParts &RdxExitVal = getVectorValue(RdxDesc.LoopExitInstr); PHINode *NewPhi = Builder.CreatePHI(VecTy, 2, "rdx.vec.exit.phi"); Value *StartVal = (part == 0) ? VectorStart : Identity; for (unsigned I = 1, E = LoopBypassBlocks.size(); I != E; ++I) NewPhi->addIncoming(StartVal, LoopBypassBlocks[I]); NewPhi->addIncoming(RdxExitVal[part], LoopVectorBody.back()); RdxParts.push_back(NewPhi); } // Reduce all of the unrolled parts into a single vector. Value *ReducedPartRdx = RdxParts[0]; unsigned Op = getReductionBinOp(RdxDesc.Kind); setDebugLocFromInst(Builder, ReducedPartRdx); for (unsigned part = 1; part < UF; ++part) { if (Op != Instruction::ICmp && Op != Instruction::FCmp) // Floating point operations had to be 'fast' to enable the reduction. ReducedPartRdx = addFastMathFlag( Builder.CreateBinOp((Instruction::BinaryOps)Op, RdxParts[part], ReducedPartRdx, "bin.rdx")); else ReducedPartRdx = createMinMaxOp(Builder, RdxDesc.MinMaxKind, ReducedPartRdx, RdxParts[part]); } if (VF > 1) { // VF is a power of 2 so we can emit the reduction using log2(VF) shuffles // and vector ops, reducing the set of values being computed by half each // round. assert(isPowerOf2_32(VF) && "Reduction emission only supported for pow2 vectors!"); Value *TmpVec = ReducedPartRdx; SmallVector
ShuffleMask(VF, nullptr); for (unsigned i = VF; i != 1; i >>= 1) { // Move the upper half of the vector to the lower half. for (unsigned j = 0; j != i/2; ++j) ShuffleMask[j] = Builder.getInt32(i/2 + j); // Fill the rest of the mask with undef. std::fill(&ShuffleMask[i/2], ShuffleMask.end(), UndefValue::get(Builder.getInt32Ty())); Value *Shuf = Builder.CreateShuffleVector(TmpVec, UndefValue::get(TmpVec->getType()), ConstantVector::get(ShuffleMask), "rdx.shuf"); if (Op != Instruction::ICmp && Op != Instruction::FCmp) // Floating point operations had to be 'fast' to enable the reduction. TmpVec = addFastMathFlag(Builder.CreateBinOp( (Instruction::BinaryOps)Op, TmpVec, Shuf, "bin.rdx")); else TmpVec = createMinMaxOp(Builder, RdxDesc.MinMaxKind, TmpVec, Shuf); } // The result is in the first element of the vector. ReducedPartRdx = Builder.CreateExtractElement(TmpVec, Builder.getInt32(0)); } // Create a phi node that merges control-flow from the backedge-taken check // block and the middle block. PHINode *BCBlockPhi = PHINode::Create(RdxPhi->getType(), 2, "bc.merge.rdx", LoopScalarPreHeader->getTerminator()); BCBlockPhi->addIncoming(RdxDesc.StartValue, LoopBypassBlocks[0]); BCBlockPhi->addIncoming(ReducedPartRdx, LoopMiddleBlock); // Now, we need to fix the users of the reduction variable // inside and outside of the scalar remainder loop. // We know that the loop is in LCSSA form. We need to update the // PHI nodes in the exit blocks. for (BasicBlock::iterator LEI = LoopExitBlock->begin(), LEE = LoopExitBlock->end(); LEI != LEE; ++LEI) { PHINode *LCSSAPhi = dyn_cast
(LEI); if (!LCSSAPhi) break; // All PHINodes need to have a single entry edge, or two if // we already fixed them. assert(LCSSAPhi->getNumIncomingValues() < 3 && "Invalid LCSSA PHI"); // We found our reduction value exit-PHI. Update it with the // incoming bypass edge. if (LCSSAPhi->getIncomingValue(0) == RdxDesc.LoopExitInstr) { // Add an edge coming from the bypass. LCSSAPhi->addIncoming(ReducedPartRdx, LoopMiddleBlock); break; } }// end of the LCSSA phi scan. // Fix the scalar loop reduction variable with the incoming reduction sum // from the vector body and from the backedge value. int IncomingEdgeBlockIdx = (RdxPhi)->getBasicBlockIndex(OrigLoop->getLoopLatch()); assert(IncomingEdgeBlockIdx >= 0 && "Invalid block index"); // Pick the other block. int SelfEdgeBlockIdx = (IncomingEdgeBlockIdx ? 0 : 1); (RdxPhi)->setIncomingValue(SelfEdgeBlockIdx, BCBlockPhi); (RdxPhi)->setIncomingValue(IncomingEdgeBlockIdx, RdxDesc.LoopExitInstr); }// end of for each redux variable. fixLCSSAPHIs(); // Remove redundant induction instructions. cse(LoopVectorBody); } void InnerLoopVectorizer::fixLCSSAPHIs() { for (BasicBlock::iterator LEI = LoopExitBlock->begin(), LEE = LoopExitBlock->end(); LEI != LEE; ++LEI) { PHINode *LCSSAPhi = dyn_cast
(LEI); if (!LCSSAPhi) break; if (LCSSAPhi->getNumIncomingValues() == 1) LCSSAPhi->addIncoming(UndefValue::get(LCSSAPhi->getType()), LoopMiddleBlock); } } InnerLoopVectorizer::VectorParts InnerLoopVectorizer::createEdgeMask(BasicBlock *Src, BasicBlock *Dst) { assert(std::find(pred_begin(Dst), pred_end(Dst), Src) != pred_end(Dst) && "Invalid edge"); // Look for cached value. std::pair
Edge(Src, Dst); EdgeMaskCache::iterator ECEntryIt = MaskCache.find(Edge); if (ECEntryIt != MaskCache.end()) return ECEntryIt->second; VectorParts SrcMask = createBlockInMask(Src); // The terminator has to be a branch inst! BranchInst *BI = dyn_cast
(Src->getTerminator()); assert(BI && "Unexpected terminator found"); if (BI->isConditional()) { VectorParts EdgeMask = getVectorValue(BI->getCondition()); if (BI->getSuccessor(0) != Dst) for (unsigned part = 0; part < UF; ++part) EdgeMask[part] = Builder.CreateNot(EdgeMask[part]); for (unsigned part = 0; part < UF; ++part) EdgeMask[part] = Builder.CreateAnd(EdgeMask[part], SrcMask[part]); MaskCache[Edge] = EdgeMask; return EdgeMask; } MaskCache[Edge] = SrcMask; return SrcMask; } InnerLoopVectorizer::VectorParts InnerLoopVectorizer::createBlockInMask(BasicBlock *BB) { assert(OrigLoop->contains(BB) && "Block is not a part of a loop"); // Loop incoming mask is all-one. if (OrigLoop->getHeader() == BB) { Value *C = ConstantInt::get(IntegerType::getInt1Ty(BB->getContext()), 1); return getVectorValue(C); } // This is the block mask. We OR all incoming edges, and with zero. Value *Zero = ConstantInt::get(IntegerType::getInt1Ty(BB->getContext()), 0); VectorParts BlockMask = getVectorValue(Zero); // For each pred: for (pred_iterator it = pred_begin(BB), e = pred_end(BB); it != e; ++it) { VectorParts EM = createEdgeMask(*it, BB); for (unsigned part = 0; part < UF; ++part) BlockMask[part] = Builder.CreateOr(BlockMask[part], EM[part]); } return BlockMask; } void InnerLoopVectorizer::widenPHIInstruction(Instruction *PN, InnerLoopVectorizer::VectorParts &Entry, unsigned UF, unsigned VF, PhiVector *PV) { PHINode* P = cast
(PN); // Handle reduction variables: if (Legal->getReductionVars()->count(P)) { for (unsigned part = 0; part < UF; ++part) { // This is phase one of vectorizing PHIs. Type *VecTy = (VF == 1) ? PN->getType() : VectorType::get(PN->getType(), VF); Entry[part] = PHINode::Create(VecTy, 2, "vec.phi", LoopVectorBody.back()-> getFirstInsertionPt()); } PV->push_back(P); return; } setDebugLocFromInst(Builder, P); // Check for PHI nodes that are lowered to vector selects. if (P->getParent() != OrigLoop->getHeader()) { // We know that all PHIs in non-header blocks are converted into // selects, so we don't have to worry about the insertion order and we // can just use the builder. // At this point we generate the predication tree. There may be // duplications since this is a simple recursive scan, but future // optimizations will clean it up. unsigned NumIncoming = P->getNumIncomingValues(); // Generate a sequence of selects of the form: // SELECT(Mask3, In3, // SELECT(Mask2, In2, // ( ...))) for (unsigned In = 0; In < NumIncoming; In++) { VectorParts Cond = createEdgeMask(P->getIncomingBlock(In), P->getParent()); VectorParts &In0 = getVectorValue(P->getIncomingValue(In)); for (unsigned part = 0; part < UF; ++part) { // We might have single edge PHIs (blocks) - use an identity // 'select' for the first PHI operand. if (In == 0) Entry[part] = Builder.CreateSelect(Cond[part], In0[part], In0[part]); else // Select between the current value and the previous incoming edge // based on the incoming mask. Entry[part] = Builder.CreateSelect(Cond[part], In0[part], Entry[part], "predphi"); } } return; } // This PHINode must be an induction variable. // Make sure that we know about it. assert(Legal->getInductionVars()->count(P) && "Not an induction variable"); LoopVectorizationLegality::InductionInfo II = Legal->getInductionVars()->lookup(P); switch (II.IK) { case LoopVectorizationLegality::IK_NoInduction: llvm_unreachable("Unknown induction"); case LoopVectorizationLegality::IK_IntInduction: { assert(P->getType() == II.StartValue->getType() && "Types must match"); Type *PhiTy = P->getType(); Value *Broadcasted; if (P == OldInduction) { // Handle the canonical induction variable. We might have had to // extend the type. Broadcasted = Builder.CreateTrunc(Induction, PhiTy); } else { // Handle other induction variables that are now based on the // canonical one. Value *NormalizedIdx = Builder.CreateSub(Induction, ExtendedIdx, "normalized.idx"); NormalizedIdx = Builder.CreateSExtOrTrunc(NormalizedIdx, PhiTy); Broadcasted = Builder.CreateAdd(II.StartValue, NormalizedIdx, "offset.idx"); } Broadcasted = getBroadcastInstrs(Broadcasted); // After broadcasting the induction variable we need to make the vector // consecutive by adding 0, 1, 2, etc. for (unsigned part = 0; part < UF; ++part) Entry[part] = getConsecutiveVector(Broadcasted, VF * part, false); return; } case LoopVectorizationLegality::IK_ReverseIntInduction: case LoopVectorizationLegality::IK_PtrInduction: case LoopVectorizationLegality::IK_ReversePtrInduction: // Handle reverse integer and pointer inductions. Value *StartIdx = ExtendedIdx; // This is the normalized GEP that starts counting at zero. Value *NormalizedIdx = Builder.CreateSub(Induction, StartIdx, "normalized.idx"); // Handle the reverse integer induction variable case. if (LoopVectorizationLegality::IK_ReverseIntInduction == II.IK) { IntegerType *DstTy = cast
(II.StartValue->getType()); Value *CNI = Builder.CreateSExtOrTrunc(NormalizedIdx, DstTy, "resize.norm.idx"); Value *ReverseInd = Builder.CreateSub(II.StartValue, CNI, "reverse.idx"); // This is a new value so do not hoist it out. Value *Broadcasted = getBroadcastInstrs(ReverseInd); // After broadcasting the induction variable we need to make the // vector consecutive by adding ... -3, -2, -1, 0. for (unsigned part = 0; part < UF; ++part) Entry[part] = getConsecutiveVector(Broadcasted, -(int)VF * part, true); return; } // Handle the pointer induction variable case. assert(P->getType()->isPointerTy() && "Unexpected type."); // Is this a reverse induction ptr or a consecutive induction ptr. bool Reverse = (LoopVectorizationLegality::IK_ReversePtrInduction == II.IK); // This is the vector of results. Notice that we don't generate // vector geps because scalar geps result in better code. for (unsigned part = 0; part < UF; ++part) { if (VF == 1) { int EltIndex = (part) * (Reverse ? -1 : 1); Constant *Idx = ConstantInt::get(Induction->getType(), EltIndex); Value *GlobalIdx; if (Reverse) GlobalIdx = Builder.CreateSub(Idx, NormalizedIdx, "gep.ridx"); else GlobalIdx = Builder.CreateAdd(NormalizedIdx, Idx, "gep.idx"); Value *SclrGep = Builder.CreateGEP(II.StartValue, GlobalIdx, "next.gep"); Entry[part] = SclrGep; continue; } Value *VecVal = UndefValue::get(VectorType::get(P->getType(), VF)); for (unsigned int i = 0; i < VF; ++i) { int EltIndex = (i + part * VF) * (Reverse ? -1 : 1); Constant *Idx = ConstantInt::get(Induction->getType(), EltIndex); Value *GlobalIdx; if (!Reverse) GlobalIdx = Builder.CreateAdd(NormalizedIdx, Idx, "gep.idx"); else GlobalIdx = Builder.CreateSub(Idx, NormalizedIdx, "gep.ridx"); Value *SclrGep = Builder.CreateGEP(II.StartValue, GlobalIdx, "next.gep"); VecVal = Builder.CreateInsertElement(VecVal, SclrGep, Builder.getInt32(i), "insert.gep"); } Entry[part] = VecVal; } return; } } void InnerLoopVectorizer::vectorizeBlockInLoop(BasicBlock *BB, PhiVector *PV) { // For each instruction in the old loop. for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { VectorParts &Entry = WidenMap.get(it); switch (it->getOpcode()) { case Instruction::Br: // Nothing to do for PHIs and BR, since we already took care of the // loop control flow instructions. continue; case Instruction::PHI:{ // Vectorize PHINodes. widenPHIInstruction(it, Entry, UF, VF, PV); continue; }// End of PHI. case Instruction::Add: case Instruction::FAdd: case Instruction::Sub: case Instruction::FSub: case Instruction::Mul: case Instruction::FMul: case Instruction::UDiv: case Instruction::SDiv: case Instruction::FDiv: case Instruction::URem: case Instruction::SRem: case Instruction::FRem: case Instruction::Shl: case Instruction::LShr: case Instruction::AShr: case Instruction::And: case Instruction::Or: case Instruction::Xor: { // Just widen binops. BinaryOperator *BinOp = dyn_cast
(it); setDebugLocFromInst(Builder, BinOp); VectorParts &A = getVectorValue(it->getOperand(0)); VectorParts &B = getVectorValue(it->getOperand(1)); // Use this vector value for all users of the original instruction. for (unsigned Part = 0; Part < UF; ++Part) { Value *V = Builder.CreateBinOp(BinOp->getOpcode(), A[Part], B[Part]); // Update the NSW, NUW and Exact flags. Notice: V can be an Undef. BinaryOperator *VecOp = dyn_cast
(V); if (VecOp && isa
(BinOp)) { VecOp->setHasNoSignedWrap(BinOp->hasNoSignedWrap()); VecOp->setHasNoUnsignedWrap(BinOp->hasNoUnsignedWrap()); } if (VecOp && isa
(VecOp)) VecOp->setIsExact(BinOp->isExact()); // Copy the fast-math flags. if (VecOp && isa
(V)) VecOp->setFastMathFlags(it->getFastMathFlags()); Entry[Part] = V; } break; } case Instruction::Select: { // Widen selects. // If the selector is loop invariant we can create a select // instruction with a scalar condition. Otherwise, use vector-select. bool InvariantCond = SE->isLoopInvariant(SE->getSCEV(it->getOperand(0)), OrigLoop); setDebugLocFromInst(Builder, it); // The condition can be loop invariant but still defined inside the // loop. This means that we can't just use the original 'cond' value. // We have to take the 'vectorized' value and pick the first lane. // Instcombine will make this a no-op. VectorParts &Cond = getVectorValue(it->getOperand(0)); VectorParts &Op0 = getVectorValue(it->getOperand(1)); VectorParts &Op1 = getVectorValue(it->getOperand(2)); Value *ScalarCond = (VF == 1) ? Cond[0] : Builder.CreateExtractElement(Cond[0], Builder.getInt32(0)); for (unsigned Part = 0; Part < UF; ++Part) { Entry[Part] = Builder.CreateSelect( InvariantCond ? ScalarCond : Cond[Part], Op0[Part], Op1[Part]); } break; } case Instruction::ICmp: case Instruction::FCmp: { // Widen compares. Generate vector compares. bool FCmp = (it->getOpcode() == Instruction::FCmp); CmpInst *Cmp = dyn_cast
(it); setDebugLocFromInst(Builder, it); VectorParts &A = getVectorValue(it->getOperand(0)); VectorParts &B = getVectorValue(it->getOperand(1)); for (unsigned Part = 0; Part < UF; ++Part) { Value *C = nullptr; if (FCmp) C = Builder.CreateFCmp(Cmp->getPredicate(), A[Part], B[Part]); else C = Builder.CreateICmp(Cmp->getPredicate(), A[Part], B[Part]); Entry[Part] = C; } break; } case Instruction::Store: case Instruction::Load: vectorizeMemoryInstruction(it); break; case Instruction::ZExt: case Instruction::SExt: case Instruction::FPToUI: case Instruction::FPToSI: case Instruction::FPExt: case Instruction::PtrToInt: case Instruction::IntToPtr: case Instruction::SIToFP: case Instruction::UIToFP: case Instruction::Trunc: case Instruction::FPTrunc: case Instruction::BitCast: { CastInst *CI = dyn_cast
(it); setDebugLocFromInst(Builder, it); /// Optimize the special case where the source is the induction /// variable. Notice that we can only optimize the 'trunc' case /// because: a. FP conversions lose precision, b. sext/zext may wrap, /// c. other casts depend on pointer size. if (CI->getOperand(0) == OldInduction && it->getOpcode() == Instruction::Trunc) { Value *ScalarCast = Builder.CreateCast(CI->getOpcode(), Induction, CI->getType()); Value *Broadcasted = getBroadcastInstrs(ScalarCast); for (unsigned Part = 0; Part < UF; ++Part) Entry[Part] = getConsecutiveVector(Broadcasted, VF * Part, false); break; } /// Vectorize casts. Type *DestTy = (VF == 1) ? CI->getType() : VectorType::get(CI->getType(), VF); VectorParts &A = getVectorValue(it->getOperand(0)); for (unsigned Part = 0; Part < UF; ++Part) Entry[Part] = Builder.CreateCast(CI->getOpcode(), A[Part], DestTy); break; } case Instruction::Call: { // Ignore dbg intrinsics. if (isa
(it)) break; setDebugLocFromInst(Builder, it); Module *M = BB->getParent()->getParent(); CallInst *CI = cast
(it); Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI); assert(ID && "Not an intrinsic call!"); switch (ID) { case Intrinsic::lifetime_end: case Intrinsic::lifetime_start: scalarizeInstruction(it); break; default: bool HasScalarOpd = hasVectorInstrinsicScalarOpd(ID, 1); for (unsigned Part = 0; Part < UF; ++Part) { SmallVector
Args; for (unsigned i = 0, ie = CI->getNumArgOperands(); i != ie; ++i) { if (HasScalarOpd && i == 1) { Args.push_back(CI->getArgOperand(i)); continue; } VectorParts &Arg = getVectorValue(CI->getArgOperand(i)); Args.push_back(Arg[Part]); } Type *Tys[] = {CI->getType()}; if (VF > 1) Tys[0] = VectorType::get(CI->getType()->getScalarType(), VF); Function *F = Intrinsic::getDeclaration(M, ID, Tys); Entry[Part] = Builder.CreateCall(F, Args); } break; } break; } default: // All other instructions are unsupported. Scalarize them. scalarizeInstruction(it); break; }// end of switch. }// end of for_each instr. } void InnerLoopVectorizer::updateAnalysis() { // Forget the original basic block. SE->forgetLoop(OrigLoop); // Update the dominator tree information. assert(DT->properlyDominates(LoopBypassBlocks.front(), LoopExitBlock) && "Entry does not dominate exit."); for (unsigned I = 1, E = LoopBypassBlocks.size(); I != E; ++I) DT->addNewBlock(LoopBypassBlocks[I], LoopBypassBlocks[I-1]); DT->addNewBlock(LoopVectorPreHeader, LoopBypassBlocks.back()); // Due to if predication of stores we might create a sequence of "if(pred) // a[i] = ...; " blocks. for (unsigned i = 0, e = LoopVectorBody.size(); i != e; ++i) { if (i == 0) DT->addNewBlock(LoopVectorBody[0], LoopVectorPreHeader); else if (isPredicatedBlock(i)) { DT->addNewBlock(LoopVectorBody[i], LoopVectorBody[i-1]); } else { DT->addNewBlock(LoopVectorBody[i], LoopVectorBody[i-2]); } } DT->addNewBlock(LoopMiddleBlock, LoopBypassBlocks[1]); DT->addNewBlock(LoopScalarPreHeader, LoopBypassBlocks[0]); DT->changeImmediateDominator(LoopScalarBody, LoopScalarPreHeader); DT->changeImmediateDominator(LoopExitBlock, LoopMiddleBlock); DEBUG(DT->verifyDomTree()); } /// \brief Check whether it is safe to if-convert this phi node. /// /// Phi nodes with constant expressions that can trap are not safe to if /// convert. static bool canIfConvertPHINodes(BasicBlock *BB) { for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E; ++I) { PHINode *Phi = dyn_cast
(I); if (!Phi) return true; for (unsigned p = 0, e = Phi->getNumIncomingValues(); p != e; ++p) if (Constant *C = dyn_cast
(Phi->getIncomingValue(p))) if (C->canTrap()) return false; } return true; } bool LoopVectorizationLegality::canVectorizeWithIfConvert() { if (!EnableIfConversion) { emitAnalysis(Report() << "if-conversion is disabled"); return false; } assert(TheLoop->getNumBlocks() > 1 && "Single block loops are vectorizable"); // A list of pointers that we can safely read and write to. SmallPtrSet
SafePointes; // Collect safe addresses. for (Loop::block_iterator BI = TheLoop->block_begin(), BE = TheLoop->block_end(); BI != BE; ++BI) { BasicBlock *BB = *BI; if (blockNeedsPredication(BB)) continue; for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E; ++I) { if (LoadInst *LI = dyn_cast
(I)) SafePointes.insert(LI->getPointerOperand()); else if (StoreInst *SI = dyn_cast
(I)) SafePointes.insert(SI->getPointerOperand()); } } // Collect the blocks that need predication. BasicBlock *Header = TheLoop->getHeader(); for (Loop::block_iterator BI = TheLoop->block_begin(), BE = TheLoop->block_end(); BI != BE; ++BI) { BasicBlock *BB = *BI; // We don't support switch statements inside loops. if (!isa
(BB->getTerminator())) { emitAnalysis(Report(BB->getTerminator()) << "loop contains a switch statement"); return false; } // We must be able to predicate all blocks that need to be predicated. if (blockNeedsPredication(BB)) { if (!blockCanBePredicated(BB, SafePointes)) { emitAnalysis(Report(BB->getTerminator()) << "control flow cannot be substituted for a select"); return false; } } else if (BB != Header && !canIfConvertPHINodes(BB)) { emitAnalysis(Report(BB->getTerminator()) << "control flow cannot be substituted for a select"); return false; } } // We can if-convert this loop. return true; } bool LoopVectorizationLegality::canVectorize() { // We must have a loop in canonical form. Loops with indirectbr in them cannot // be canonicalized. if (!TheLoop->getLoopPreheader()) { emitAnalysis( Report() << "loop control flow is not understood by vectorizer"); return false; } // We can only vectorize innermost loops. if (TheLoop->getSubLoopsVector().size()) { emitAnalysis(Report() << "loop is not the innermost loop"); return false; } // We must have a single backedge. if (TheLoop->getNumBackEdges() != 1) { emitAnalysis( Report() << "loop control flow is not understood by vectorizer"); return false; } // We must have a single exiting block. if (!TheLoop->getExitingBlock()) { emitAnalysis( Report() << "loop control flow is not understood by vectorizer"); return false; } // We need to have a loop header. DEBUG(dbgs() << "LV: Found a loop: " << TheLoop->getHeader()->getName() << '\n'); // Check if we can if-convert non-single-bb loops. unsigned NumBlocks = TheLoop->getNumBlocks(); if (NumBlocks != 1 && !canVectorizeWithIfConvert()) { DEBUG(dbgs() << "LV: Can't if-convert the loop.\n"); return false; } // ScalarEvolution needs to be able to find the exit count. const SCEV *ExitCount = SE->getBackedgeTakenCount(TheLoop); if (ExitCount == SE->getCouldNotCompute()) { emitAnalysis(Report() << "could not determine number of loop iterations"); DEBUG(dbgs() << "LV: SCEV could not compute the loop exit count.\n"); return false; } // Check if we can vectorize the instructions and CFG in this loop. if (!canVectorizeInstrs()) { DEBUG(dbgs() << "LV: Can't vectorize the instructions or CFG\n"); return false; } // Go over each instruction and look at memory deps. if (!canVectorizeMemory()) { DEBUG(dbgs() << "LV: Can't vectorize due to memory conflicts\n"); return false; } // Collect all of the variables that remain uniform after vectorization. collectLoopUniforms(); DEBUG(dbgs() << "LV: We can vectorize this loop" << (PtrRtCheck.Need ? " (with a runtime bound check)" : "") <<"!\n"); // Okay! We can vectorize. At this point we don't have any other mem analysis // which may limit our maximum vectorization factor, so just return true with // no restrictions. return true; } static Type *convertPointerToIntegerType(const DataLayout &DL, Type *Ty) { if (Ty->isPointerTy()) return DL.getIntPtrType(Ty); // It is possible that char's or short's overflow when we ask for the loop's // trip count, work around this by changing the type size. if (Ty->getScalarSizeInBits() < 32) return Type::getInt32Ty(Ty->getContext()); return Ty; } static Type* getWiderType(const DataLayout &DL, Type *Ty0, Type *Ty1) { Ty0 = convertPointerToIntegerType(DL, Ty0); Ty1 = convertPointerToIntegerType(DL, Ty1); if (Ty0->getScalarSizeInBits() > Ty1->getScalarSizeInBits()) return Ty0; return Ty1; } /// \brief Check that the instruction has outside loop users and is not an /// identified reduction variable. static bool hasOutsideLoopUser(const Loop *TheLoop, Instruction *Inst, SmallPtrSet
&Reductions) { // Reduction instructions are allowed to have exit users. All other // instructions must not have external users. if (!Reductions.count(Inst)) //Check that all of the users of the loop are inside the BB. for (User *U : Inst->users()) { Instruction *UI = cast
(U); // This user may be a reduction exit value. if (!TheLoop->contains(UI)) { DEBUG(dbgs() << "LV: Found an outside user for : " << *UI << '\n'); return true; } } return false; } bool LoopVectorizationLegality::canVectorizeInstrs() { BasicBlock *PreHeader = TheLoop->getLoopPreheader(); BasicBlock *Header = TheLoop->getHeader(); // Look for the attribute signaling the absence of NaNs. Function &F = *Header->getParent(); if (F.hasFnAttribute("no-nans-fp-math")) HasFunNoNaNAttr = F.getAttributes().getAttribute( AttributeSet::FunctionIndex, "no-nans-fp-math").getValueAsString() == "true"; // For each block in the loop. for (Loop::block_iterator bb = TheLoop->block_begin(), be = TheLoop->block_end(); bb != be; ++bb) { // Scan the instructions in the block and look for hazards. for (BasicBlock::iterator it = (*bb)->begin(), e = (*bb)->end(); it != e; ++it) { if (PHINode *Phi = dyn_cast
(it)) { Type *PhiTy = Phi->getType(); // Check that this PHI type is allowed. if (!PhiTy->isIntegerTy() && !PhiTy->isFloatingPointTy() && !PhiTy->isPointerTy()) { emitAnalysis(Report(it) << "loop control flow is not understood by vectorizer"); DEBUG(dbgs() << "LV: Found an non-int non-pointer PHI.\n"); return false; } // If this PHINode is not in the header block, then we know that we // can convert it to select during if-conversion. No need to check if // the PHIs in this block are induction or reduction variables. if (*bb != Header) { // Check that this instruction has no outside users or is an // identified reduction value with an outside user. if (!hasOutsideLoopUser(TheLoop, it, AllowedExit)) continue; emitAnalysis(Report(it) << "value that could not be identified as " "reduction is used outside the loop"); return false; } // We only allow if-converted PHIs with more than two incoming values. if (Phi->getNumIncomingValues() != 2) { emitAnalysis(Report(it) << "control flow not understood by vectorizer"); DEBUG(dbgs() << "LV: Found an invalid PHI.\n"); return false; } // This is the value coming from the preheader. Value *StartValue = Phi->getIncomingValueForBlock(PreHeader); // Check if this is an induction variable. InductionKind IK = isInductionVariable(Phi); if (IK_NoInduction != IK) { // Get the widest type. if (!WidestIndTy) WidestIndTy = convertPointerToIntegerType(*DL, PhiTy); else WidestIndTy = getWiderType(*DL, PhiTy, WidestIndTy); // Int inductions are special because we only allow one IV. if (IK == IK_IntInduction) { // Use the phi node with the widest type as induction. Use the last // one if there are multiple (no good reason for doing this other // than it is expedient). if (!Induction || PhiTy == WidestIndTy) Induction = Phi; } DEBUG(dbgs() << "LV: Found an induction variable.\n"); Inductions[Phi] = InductionInfo(StartValue, IK); // Until we explicitly handle the case of an induction variable with // an outside loop user we have to give up vectorizing this loop. if (hasOutsideLoopUser(TheLoop, it, AllowedExit)) { emitAnalysis(Report(it) << "use of induction value outside of the " "loop is not handled by vectorizer"); return false; } continue; } if (AddReductionVar(Phi, RK_IntegerAdd)) { DEBUG(dbgs() << "LV: Found an ADD reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerMult)) { DEBUG(dbgs() << "LV: Found a MUL reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerOr)) { DEBUG(dbgs() << "LV: Found an OR reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerAnd)) { DEBUG(dbgs() << "LV: Found an AND reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerXor)) { DEBUG(dbgs() << "LV: Found a XOR reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerMinMax)) { DEBUG(dbgs() << "LV: Found a MINMAX reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_FloatMult)) { DEBUG(dbgs() << "LV: Found an FMult reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_FloatAdd)) { DEBUG(dbgs() << "LV: Found an FAdd reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_FloatMinMax)) { DEBUG(dbgs() << "LV: Found an float MINMAX reduction PHI."<< *Phi << "\n"); continue; } emitAnalysis(Report(it) << "unvectorizable operation"); DEBUG(dbgs() << "LV: Found an unidentified PHI."<< *Phi <<"\n"); return false; }// end of PHI handling // We still don't handle functions. However, we can ignore dbg intrinsic // calls and we do handle certain intrinsic and libm functions. CallInst *CI = dyn_cast
(it); if (CI && !getIntrinsicIDForCall(CI, TLI) && !isa
(CI)) { emitAnalysis(Report(it) << "call instruction cannot be vectorized"); DEBUG(dbgs() << "LV: Found a call site.\n"); return false; } // Intrinsics such as powi,cttz and ctlz are legal to vectorize if the // second argument is the same (i.e. loop invariant) if (CI && hasVectorInstrinsicScalarOpd(getIntrinsicIDForCall(CI, TLI), 1)) { if (!SE->isLoopInvariant(SE->getSCEV(CI->getOperand(1)), TheLoop)) { emitAnalysis(Report(it) << "intrinsic instruction cannot be vectorized"); DEBUG(dbgs() << "LV: Found unvectorizable intrinsic " << *CI << "\n"); return false; } } // Check that the instruction return type is vectorizable. // Also, we can't vectorize extractelement instructions. if ((!VectorType::isValidElementType(it->getType()) && !it->getType()->isVoidTy()) || isa
(it)) { emitAnalysis(Report(it) << "instruction return type cannot be vectorized"); DEBUG(dbgs() << "LV: Found unvectorizable type.\n"); return false; } // Check that the stored type is vectorizable. if (StoreInst *ST = dyn_cast
(it)) { Type *T = ST->getValueOperand()->getType(); if (!VectorType::isValidElementType(T)) { emitAnalysis(Report(ST) << "store instruction cannot be vectorized"); return false; } if (EnableMemAccessVersioning) collectStridedAcccess(ST); } if (EnableMemAccessVersioning) if (LoadInst *LI = dyn_cast
(it)) collectStridedAcccess(LI); // Reduction instructions are allowed to have exit users. // All other instructions must not have external users. if (hasOutsideLoopUser(TheLoop, it, AllowedExit)) { emitAnalysis(Report(it) << "value cannot be used outside the loop"); return false; } } // next instr. } if (!Induction) { DEBUG(dbgs() << "LV: Did not find one integer induction var.\n"); if (Inductions.empty()) { emitAnalysis(Report() << "loop induction variable could not be identified"); return false; } } return true; } ///\brief Remove GEPs whose indices but the last one are loop invariant and /// return the induction operand of the gep pointer. static Value *stripGetElementPtr(Value *Ptr, ScalarEvolution *SE, const DataLayout *DL, Loop *Lp) { GetElementPtrInst *GEP = dyn_cast
(Ptr); if (!GEP) return Ptr; unsigned InductionOperand = getGEPInductionOperand(DL, GEP); // Check that all of the gep indices are uniform except for our induction // operand. for (unsigned i = 0, e = GEP->getNumOperands(); i != e; ++i) if (i != InductionOperand && !SE->isLoopInvariant(SE->getSCEV(GEP->getOperand(i)), Lp)) return Ptr; return GEP->getOperand(InductionOperand); } ///\brief Look for a cast use of the passed value. static Value *getUniqueCastUse(Value *Ptr, Loop *Lp, Type *Ty) { Value *UniqueCast = nullptr; for (User *U : Ptr->users()) { CastInst *CI = dyn_cast
(U); if (CI && CI->getType() == Ty) { if (!UniqueCast) UniqueCast = CI; else return nullptr; } } return UniqueCast; } ///\brief Get the stride of a pointer access in a loop. /// Looks for symbolic strides "a[i*stride]". Returns the symbolic stride as a /// pointer to the Value, or null otherwise. static Value *getStrideFromPointer(Value *Ptr, ScalarEvolution *SE, const DataLayout *DL, Loop *Lp) { const PointerType *PtrTy = dyn_cast
(Ptr->getType()); if (!PtrTy || PtrTy->isAggregateType()) return nullptr; // Try to remove a gep instruction to make the pointer (actually index at this // point) easier analyzable. If OrigPtr is equal to Ptr we are analzying the // pointer, otherwise, we are analyzing the index. Value *OrigPtr = Ptr; // The size of the pointer access. int64_t PtrAccessSize = 1; Ptr = stripGetElementPtr(Ptr, SE, DL, Lp); const SCEV *V = SE->getSCEV(Ptr); if (Ptr != OrigPtr) // Strip off casts. while (const SCEVCastExpr *C = dyn_cast
(V)) V = C->getOperand(); const SCEVAddRecExpr *S = dyn_cast
(V); if (!S) return nullptr; V = S->getStepRecurrence(*SE); if (!V) return nullptr; // Strip off the size of access multiplication if we are still analyzing the // pointer. if (OrigPtr == Ptr) { DL->getTypeAllocSize(PtrTy->getElementType()); if (const SCEVMulExpr *M = dyn_cast
(V)) { if (M->getOperand(0)->getSCEVType() != scConstant) return nullptr; const APInt &APStepVal = cast
(M->getOperand(0))->getValue()->getValue(); // Huge step value - give up. if (APStepVal.getBitWidth() > 64) return nullptr; int64_t StepVal = APStepVal.getSExtValue(); if (PtrAccessSize != StepVal) return nullptr; V = M->getOperand(1); } } // Strip off casts. Type *StripedOffRecurrenceCast = nullptr; if (const SCEVCastExpr *C = dyn_cast
(V)) { StripedOffRecurrenceCast = C->getType(); V = C->getOperand(); } // Look for the loop invariant symbolic value. const SCEVUnknown *U = dyn_cast
(V); if (!U) return nullptr; Value *Stride = U->getValue(); if (!Lp->isLoopInvariant(Stride)) return nullptr; // If we have stripped off the recurrence cast we have to make sure that we // return the value that is used in this loop so that we can replace it later. if (StripedOffRecurrenceCast) Stride = getUniqueCastUse(Stride, Lp, StripedOffRecurrenceCast); return Stride; } void LoopVectorizationLegality::collectStridedAcccess(Value *MemAccess) { Value *Ptr = nullptr; if (LoadInst *LI = dyn_cast
(MemAccess)) Ptr = LI->getPointerOperand(); else if (StoreInst *SI = dyn_cast
(MemAccess)) Ptr = SI->getPointerOperand(); else return; Value *Stride = getStrideFromPointer(Ptr, SE, DL, TheLoop); if (!Stride) return; DEBUG(dbgs() << "LV: Found a strided access that we can version"); DEBUG(dbgs() << " Ptr: " << *Ptr << " Stride: " << *Stride << "\n"); Strides[Ptr] = Stride; StrideSet.insert(Stride); } void LoopVectorizationLegality::collectLoopUniforms() { // We now know that the loop is vectorizable! // Collect variables that will remain uniform after vectorization. std::vector
Worklist; BasicBlock *Latch = TheLoop->getLoopLatch(); // Start with the conditional branch and walk up the block. Worklist.push_back(Latch->getTerminator()->getOperand(0)); // Also add all consecutive pointer values; these values will be uniform // after vectorization (and subsequent cleanup) and, until revectorization is // supported, all dependencies must also be uniform. for (Loop::block_iterator B = TheLoop->block_begin(), BE = TheLoop->block_end(); B != BE; ++B) for (BasicBlock::iterator I = (*B)->begin(), IE = (*B)->end(); I != IE; ++I) if (I->getType()->isPointerTy() && isConsecutivePtr(I)) Worklist.insert(Worklist.end(), I->op_begin(), I->op_end()); while (Worklist.size()) { Instruction *I = dyn_cast
(Worklist.back()); Worklist.pop_back(); // Look at instructions inside this loop. // Stop when reaching PHI nodes. // TODO: we need to follow values all over the loop, not only in this block. if (!I || !TheLoop->contains(I) || isa
(I)) continue; // This is a known uniform. Uniforms.insert(I); // Insert all operands. Worklist.insert(Worklist.end(), I->op_begin(), I->op_end()); } } namespace { /// \brief Analyses memory accesses in a loop. /// /// Checks whether run time pointer checks are needed and builds sets for data /// dependence checking. class AccessAnalysis { public: /// \brief Read or write access location. typedef PointerIntPair
MemAccessInfo; typedef SmallPtrSet
MemAccessInfoSet; /// \brief Set of potential dependent memory accesses. typedef EquivalenceClasses
DepCandidates; AccessAnalysis(const DataLayout *Dl, DepCandidates &DA) : DL(Dl), DepCands(DA), AreAllWritesIdentified(true), AreAllReadsIdentified(true), IsRTCheckNeeded(false) {} /// \brief Register a load and whether it is only read from. void addLoad(Value *Ptr, bool IsReadOnly) { Accesses.insert(MemAccessInfo(Ptr, false)); if (IsReadOnly) ReadOnlyPtr.insert(Ptr); } /// \brief Register a store. void addStore(Value *Ptr) { Accesses.insert(MemAccessInfo(Ptr, true)); } /// \brief Check whether we can check the pointers at runtime for /// non-intersection. bool canCheckPtrAtRT(LoopVectorizationLegality::RuntimePointerCheck &RtCheck, unsigned &NumComparisons, ScalarEvolution *SE, Loop *TheLoop, ValueToValueMap &Strides, bool ShouldCheckStride = false); /// \brief Goes over all memory accesses, checks whether a RT check is needed /// and builds sets of dependent accesses. void buildDependenceSets() { // Process read-write pointers first. processMemAccesses(false); // Next, process read pointers. processMemAccesses(true); } bool isRTCheckNeeded() { return IsRTCheckNeeded; } bool isDependencyCheckNeeded() { return !CheckDeps.empty(); } void resetDepChecks() { CheckDeps.clear(); } MemAccessInfoSet &getDependenciesToCheck() { return CheckDeps; } private: typedef SetVector
PtrAccessSet; typedef DenseMap
UnderlyingObjToAccessMap; /// \brief Go over all memory access or only the deferred ones if /// \p UseDeferred is true and check whether runtime pointer checks are needed /// and build sets of dependency check candidates. void processMemAccesses(bool UseDeferred); /// Set of all accesses. PtrAccessSet Accesses; /// Set of access to check after all writes have been processed. PtrAccessSet DeferredAccesses; /// Map of pointers to last access encountered. UnderlyingObjToAccessMap ObjToLastAccess; /// Set of accesses that need a further dependence check. MemAccessInfoSet CheckDeps; /// Set of pointers that are read only. SmallPtrSet
ReadOnlyPtr; /// Set of underlying objects already written to. SmallPtrSet
WriteObjects; const DataLayout *DL; /// Sets of potentially dependent accesses - members of one set share an /// underlying pointer. The set "CheckDeps" identfies which sets really need a /// dependence check. DepCandidates &DepCands; bool AreAllWritesIdentified; bool AreAllReadsIdentified; bool IsRTCheckNeeded; }; } // end anonymous namespace /// \brief Check whether a pointer can participate in a runtime bounds check. static bool hasComputableBounds(ScalarEvolution *SE, ValueToValueMap &Strides, Value *Ptr) { const SCEV *PtrScev = replaceSymbolicStrideSCEV(SE, Strides, Ptr); const SCEVAddRecExpr *AR = dyn_cast
(PtrScev); if (!AR) return false; return AR->isAffine(); } /// \brief Check the stride of the pointer and ensure that it does not wrap in /// the address space. static int isStridedPtr(ScalarEvolution *SE, const DataLayout *DL, Value *Ptr, const Loop *Lp, ValueToValueMap &StridesMap); bool AccessAnalysis::canCheckPtrAtRT( LoopVectorizationLegality::RuntimePointerCheck &RtCheck, unsigned &NumComparisons, ScalarEvolution *SE, Loop *TheLoop, ValueToValueMap &StridesMap, bool ShouldCheckStride) { // Find pointers with computable bounds. We are going to use this information // to place a runtime bound check. unsigned NumReadPtrChecks = 0; unsigned NumWritePtrChecks = 0; bool CanDoRT = true; bool IsDepCheckNeeded = isDependencyCheckNeeded(); // We assign consecutive id to access from different dependence sets. // Accesses within the same set don't need a runtime check. unsigned RunningDepId = 1; DenseMap
DepSetId; for (PtrAccessSet::iterator AI = Accesses.begin(), AE = Accesses.end(); AI != AE; ++AI) { const MemAccessInfo &Access = *AI; Value *Ptr = Access.getPointer(); bool IsWrite = Access.getInt(); // Just add write checks if we have both. if (!IsWrite && Accesses.count(MemAccessInfo(Ptr, true))) continue; if (IsWrite) ++NumWritePtrChecks; else ++NumReadPtrChecks; if (hasComputableBounds(SE, StridesMap, Ptr) && // When we run after a failing dependency check we have to make sure we // don't have wrapping pointers. (!ShouldCheckStride || isStridedPtr(SE, DL, Ptr, TheLoop, StridesMap) == 1)) { // The id of the dependence set. unsigned DepId; if (IsDepCheckNeeded) { Value *Leader = DepCands.getLeaderValue(Access).getPointer(); unsigned &LeaderId = DepSetId[Leader]; if (!LeaderId) LeaderId = RunningDepId++; DepId = LeaderId; } else // Each access has its own dependence set. DepId = RunningDepId++; RtCheck.insert(SE, TheLoop, Ptr, IsWrite, DepId, StridesMap); DEBUG(dbgs() << "LV: Found a runtime check ptr:" << *Ptr << '\n'); } else { CanDoRT = false; } } if (IsDepCheckNeeded && CanDoRT && RunningDepId == 2) NumComparisons = 0; // Only one dependence set. else { NumComparisons = (NumWritePtrChecks * (NumReadPtrChecks + NumWritePtrChecks - 1)); } // If the pointers that we would use for the bounds comparison have different // address spaces, assume the values aren't directly comparable, so we can't // use them for the runtime check. We also have to assume they could // overlap. In the future there should be metadata for whether address spaces // are disjoint. unsigned NumPointers = RtCheck.Pointers.size(); for (unsigned i = 0; i < NumPointers; ++i) { for (unsigned j = i + 1; j < NumPointers; ++j) { // Only need to check pointers between two different dependency sets. if (RtCheck.DependencySetId[i] == RtCheck.DependencySetId[j]) continue; Value *PtrI = RtCheck.Pointers[i]; Value *PtrJ = RtCheck.Pointers[j]; unsigned ASi = PtrI->getType()->getPointerAddressSpace(); unsigned ASj = PtrJ->getType()->getPointerAddressSpace(); if (ASi != ASj) { DEBUG(dbgs() << "LV: Runtime check would require comparison between" " different address spaces\n"); return false; } } } return CanDoRT; } static bool isFunctionScopeIdentifiedObject(Value *Ptr) { return isNoAliasArgument(Ptr) || isNoAliasCall(Ptr) || isa
(Ptr); } void AccessAnalysis::processMemAccesses(bool UseDeferred) { // We process the set twice: first we process read-write pointers, last we // process read-only pointers. This allows us to skip dependence tests for // read-only pointers. PtrAccessSet &S = UseDeferred ? DeferredAccesses : Accesses; for (PtrAccessSet::iterator AI = S.begin(), AE = S.end(); AI != AE; ++AI) { const MemAccessInfo &Access = *AI; Value *Ptr = Access.getPointer(); bool IsWrite = Access.getInt(); DepCands.insert(Access); // Memorize read-only pointers for later processing and skip them in the // first round (they need to be checked after we have seen all write // pointers). Note: we also mark pointer that are not consecutive as // "read-only" pointers (so that we check "a[b[i]] +="). Hence, we need the // second check for "!IsWrite". bool IsReadOnlyPtr = ReadOnlyPtr.count(Ptr) && !IsWrite; if (!UseDeferred && IsReadOnlyPtr) { DeferredAccesses.insert(Access); continue; } bool NeedDepCheck = false; // Check whether there is the possibility of dependency because of // underlying objects being the same. typedef SmallVector
ValueVector; ValueVector TempObjects; GetUnderlyingObjects(Ptr, TempObjects, DL); for (ValueVector::iterator UI = TempObjects.begin(), UE = TempObjects.end(); UI != UE; ++UI) { Value *UnderlyingObj = *UI; // If this is a write then it needs to be an identified object. If this a // read and all writes (so far) are identified function scope objects we // don't need an identified underlying object but only an Argument (the // next write is going to invalidate this assumption if it is // unidentified). // This is a micro-optimization for the case where all writes are // identified and we have one argument pointer. // Otherwise, we do need a runtime check. if ((IsWrite && !isFunctionScopeIdentifiedObject(UnderlyingObj)) || (!IsWrite && (!AreAllWritesIdentified || !isa
(UnderlyingObj)) && !isIdentifiedObject(UnderlyingObj))) { DEBUG(dbgs() << "LV: Found an unidentified " << (IsWrite ? "write" : "read" ) << " ptr: " << *UnderlyingObj << "\n"); IsRTCheckNeeded = (IsRTCheckNeeded || !isIdentifiedObject(UnderlyingObj) || !AreAllReadsIdentified); if (IsWrite) AreAllWritesIdentified = false; if (!IsWrite) AreAllReadsIdentified = false; } // If this is a write - check other reads and writes for conflicts. If // this is a read only check other writes for conflicts (but only if there // is no other write to the ptr - this is an optimization to catch "a[i] = // a[i] + " without having to do a dependence check). if ((IsWrite || IsReadOnlyPtr) && WriteObjects.count(UnderlyingObj)) NeedDepCheck = true; if (IsWrite) WriteObjects.insert(UnderlyingObj); // Create sets of pointers connected by shared underlying objects. UnderlyingObjToAccessMap::iterator Prev = ObjToLastAccess.find(UnderlyingObj); if (Prev != ObjToLastAccess.end()) DepCands.unionSets(Access, Prev->second); ObjToLastAccess[UnderlyingObj] = Access; } if (NeedDepCheck) CheckDeps.insert(Access); } } namespace { /// \brief Checks memory dependences among accesses to the same underlying /// object to determine whether there vectorization is legal or not (and at /// which vectorization factor). /// /// This class works under the assumption that we already checked that memory /// locations with different underlying pointers are "must-not alias". /// We use the ScalarEvolution framework to symbolically evalutate access /// functions pairs. Since we currently don't restructure the loop we can rely /// on the program order of memory accesses to determine their safety. /// At the moment we will only deem accesses as safe for: /// * A negative constant distance assuming program order. /// /// Safe: tmp = a[i + 1]; OR a[i + 1] = x; /// a[i] = tmp; y = a[i]; /// /// The latter case is safe because later checks guarantuee that there can't /// be a cycle through a phi node (that is, we check that "x" and "y" is not /// the same variable: a header phi can only be an induction or a reduction, a /// reduction can't have a memory sink, an induction can't have a memory /// source). This is important and must not be violated (or we have to /// resort to checking for cycles through memory). /// /// * A positive constant distance assuming program order that is bigger /// than the biggest memory access. /// /// tmp = a[i] OR b[i] = x /// a[i+2] = tmp y = b[i+2]; /// /// Safe distance: 2 x sizeof(a[0]), and 2 x sizeof(b[0]), respectively. /// /// * Zero distances and all accesses have the same size. /// class MemoryDepChecker { public: typedef PointerIntPair
MemAccessInfo; typedef SmallPtrSet
MemAccessInfoSet; MemoryDepChecker(ScalarEvolution *Se, const DataLayout *Dl, const Loop *L) : SE(Se), DL(Dl), InnermostLoop(L), AccessIdx(0), ShouldRetryWithRuntimeCheck(false) {} /// \brief Register the location (instructions are given increasing numbers) /// of a write access. void addAccess(StoreInst *SI) { Value *Ptr = SI->getPointerOperand(); Accesses[MemAccessInfo(Ptr, true)].push_back(AccessIdx); InstMap.push_back(SI); ++AccessIdx; } /// \brief Register the location (instructions are given increasing numbers) /// of a write access. void addAccess(LoadInst *LI) { Value *Ptr = LI->getPointerOperand(); Accesses[MemAccessInfo(Ptr, false)].push_back(AccessIdx); InstMap.push_back(LI); ++AccessIdx; } /// \brief Check whether the dependencies between the accesses are safe. /// /// Only checks sets with elements in \p CheckDeps. bool areDepsSafe(AccessAnalysis::DepCandidates &AccessSets, MemAccessInfoSet &CheckDeps, ValueToValueMap &Strides); /// \brief The maximum number of bytes of a vector register we can vectorize /// the accesses safely with. unsigned getMaxSafeDepDistBytes() { return MaxSafeDepDistBytes; } /// \brief In same cases when the dependency check fails we can still /// vectorize the loop with a dynamic array access check. bool shouldRetryWithRuntimeCheck() { return ShouldRetryWithRuntimeCheck; } private: ScalarEvolution *SE; const DataLayout *DL; const Loop *InnermostLoop; /// \brief Maps access locations (ptr, read/write) to program order. DenseMap
> Accesses; /// \brief Memory access instructions in program order. SmallVector
InstMap; /// \brief The program order index to be used for the next instruction. unsigned AccessIdx; // We can access this many bytes in parallel safely. unsigned MaxSafeDepDistBytes; /// \brief If we see a non-constant dependence distance we can still try to /// vectorize this loop with runtime checks. bool ShouldRetryWithRuntimeCheck; /// \brief Check whether there is a plausible dependence between the two /// accesses. /// /// Access \p A must happen before \p B in program order. The two indices /// identify the index into the program order map. /// /// This function checks whether there is a plausible dependence (or the /// absence of such can't be proved) between the two accesses. If there is a /// plausible dependence but the dependence distance is bigger than one /// element access it records this distance in \p MaxSafeDepDistBytes (if this /// distance is smaller than any other distance encountered so far). /// Otherwise, this function returns true signaling a possible dependence. bool isDependent(const MemAccessInfo &A, unsigned AIdx, const MemAccessInfo &B, unsigned BIdx, ValueToValueMap &Strides); /// \brief Check whether the data dependence could prevent store-load /// forwarding. bool couldPreventStoreLoadForward(unsigned Distance, unsigned TypeByteSize); }; } // end anonymous namespace static bool isInBoundsGep(Value *Ptr) { if (GetElementPtrInst *GEP = dyn_cast
(Ptr)) return GEP->isInBounds(); return false; } /// \brief Check whether the access through \p Ptr has a constant stride. static int isStridedPtr(ScalarEvolution *SE, const DataLayout *DL, Value *Ptr, const Loop *Lp, ValueToValueMap &StridesMap) { const Type *Ty = Ptr->getType(); assert(Ty->isPointerTy() && "Unexpected non-ptr"); // Make sure that the pointer does not point to aggregate types. const PointerType *PtrTy = cast
(Ty); if (PtrTy->getElementType()->isAggregateType()) { DEBUG(dbgs() << "LV: Bad stride - Not a pointer to a scalar type" << *Ptr << "\n"); return 0; } const SCEV *PtrScev = replaceSymbolicStrideSCEV(SE, StridesMap, Ptr); const SCEVAddRecExpr *AR = dyn_cast
(PtrScev); if (!AR) { DEBUG(dbgs() << "LV: Bad stride - Not an AddRecExpr pointer " << *Ptr << " SCEV: " << *PtrScev << "\n"); return 0; } // The accesss function must stride over the innermost loop. if (Lp != AR->getLoop()) { DEBUG(dbgs() << "LV: Bad stride - Not striding over innermost loop " << *Ptr << " SCEV: " << *PtrScev << "\n"); } // The address calculation must not wrap. Otherwise, a dependence could be // inverted. // An inbounds getelementptr that is a AddRec with a unit stride // cannot wrap per definition. The unit stride requirement is checked later. // An getelementptr without an inbounds attribute and unit stride would have // to access the pointer value "0" which is undefined behavior in address // space 0, therefore we can also vectorize this case. bool IsInBoundsGEP = isInBoundsGep(Ptr); bool IsNoWrapAddRec = AR->getNoWrapFlags(SCEV::NoWrapMask); bool IsInAddressSpaceZero = PtrTy->getAddressSpace() == 0; if (!IsNoWrapAddRec && !IsInBoundsGEP && !IsInAddressSpaceZero) { DEBUG(dbgs() << "LV: Bad stride - Pointer may wrap in the address space " << *Ptr << " SCEV: " << *PtrScev << "\n"); return 0; } // Check the step is constant. const SCEV *Step = AR->getStepRecurrence(*SE); // Calculate the pointer stride and check if it is consecutive. const SCEVConstant *C = dyn_cast
(Step); if (!C) { DEBUG(dbgs() << "LV: Bad stride - Not a constant strided " << *Ptr << " SCEV: " << *PtrScev << "\n"); return 0; } int64_t Size = DL->getTypeAllocSize(PtrTy->getElementType()); const APInt &APStepVal = C->getValue()->getValue(); // Huge step value - give up. if (APStepVal.getBitWidth() > 64) return 0; int64_t StepVal = APStepVal.getSExtValue(); // Strided access. int64_t Stride = StepVal / Size; int64_t Rem = StepVal % Size; if (Rem) return 0; // If the SCEV could wrap but we have an inbounds gep with a unit stride we // know we can't "wrap around the address space". In case of address space // zero we know that this won't happen without triggering undefined behavior. if (!IsNoWrapAddRec && (IsInBoundsGEP || IsInAddressSpaceZero) && Stride != 1 && Stride != -1) return 0; return Stride; } bool MemoryDepChecker::couldPreventStoreLoadForward(unsigned Distance, unsigned TypeByteSize) { // If loads occur at a distance that is not a multiple of a feasible vector // factor store-load forwarding does not take place. // Positive dependences might cause troubles because vectorizing them might // prevent store-load forwarding making vectorized code run a lot slower. // a[i] = a[i-3] ^ a[i-8]; // The stores to a[i:i+1] don't align with the stores to a[i-3:i-2] and // hence on your typical architecture store-load forwarding does not take // place. Vectorizing in such cases does not make sense. // Store-load forwarding distance. const unsigned NumCyclesForStoreLoadThroughMemory = 8*TypeByteSize; // Maximum vector factor. unsigned MaxVFWithoutSLForwardIssues = MaxVectorWidth*TypeByteSize; if(MaxSafeDepDistBytes < MaxVFWithoutSLForwardIssues) MaxVFWithoutSLForwardIssues = MaxSafeDepDistBytes; for (unsigned vf = 2*TypeByteSize; vf <= MaxVFWithoutSLForwardIssues; vf *= 2) { if (Distance % vf && Distance / vf < NumCyclesForStoreLoadThroughMemory) { MaxVFWithoutSLForwardIssues = (vf >>=1); break; } } if (MaxVFWithoutSLForwardIssues< 2*TypeByteSize) { DEBUG(dbgs() << "LV: Distance " << Distance << " that could cause a store-load forwarding conflict\n"); return true; } if (MaxVFWithoutSLForwardIssues < MaxSafeDepDistBytes && MaxVFWithoutSLForwardIssues != MaxVectorWidth*TypeByteSize) MaxSafeDepDistBytes = MaxVFWithoutSLForwardIssues; return false; } bool MemoryDepChecker::isDependent(const MemAccessInfo &A, unsigned AIdx, const MemAccessInfo &B, unsigned BIdx, ValueToValueMap &Strides) { assert (AIdx < BIdx && "Must pass arguments in program order"); Value *APtr = A.getPointer(); Value *BPtr = B.getPointer(); bool AIsWrite = A.getInt(); bool BIsWrite = B.getInt(); // Two reads are independent. if (!AIsWrite && !BIsWrite) return false; const SCEV *AScev = replaceSymbolicStrideSCEV(SE, Strides, APtr); const SCEV *BScev = replaceSymbolicStrideSCEV(SE, Strides, BPtr); int StrideAPtr = isStridedPtr(SE, DL, APtr, InnermostLoop, Strides); int StrideBPtr = isStridedPtr(SE, DL, BPtr, InnermostLoop, Strides); const SCEV *Src = AScev; const SCEV *Sink = BScev; // If the induction step is negative we have to invert source and sink of the // dependence. if (StrideAPtr < 0) { //Src = BScev; //Sink = AScev; std::swap(APtr, BPtr); std::swap(Src, Sink); std::swap(AIsWrite, BIsWrite); std::swap(AIdx, BIdx); std::swap(StrideAPtr, StrideBPtr); } const SCEV *Dist = SE->getMinusSCEV(Sink, Src); DEBUG(dbgs() << "LV: Src Scev: " << *Src << "Sink Scev: " << *Sink << "(Induction step: " << StrideAPtr << ")\n"); DEBUG(dbgs() << "LV: Distance for " << *InstMap[AIdx] << " to " << *InstMap[BIdx] << ": " << *Dist << "\n"); // Need consecutive accesses. We don't want to vectorize // "A[B[i]] += ..." and similar code or pointer arithmetic that could wrap in // the address space. if (!StrideAPtr || !StrideBPtr || StrideAPtr != StrideBPtr){ DEBUG(dbgs() << "Non-consecutive pointer access\n"); return true; } const SCEVConstant *C = dyn_cast
(Dist); if (!C) { DEBUG(dbgs() << "LV: Dependence because of non-constant distance\n"); ShouldRetryWithRuntimeCheck = true; return true; } Type *ATy = APtr->getType()->getPointerElementType(); Type *BTy = BPtr->getType()->getPointerElementType(); unsigned TypeByteSize = DL->getTypeAllocSize(ATy); // Negative distances are not plausible dependencies. const APInt &Val = C->getValue()->getValue(); if (Val.isNegative()) { bool IsTrueDataDependence = (AIsWrite && !BIsWrite); if (IsTrueDataDependence && (couldPreventStoreLoadForward(Val.abs().getZExtValue(), TypeByteSize) || ATy != BTy)) return true; DEBUG(dbgs() << "LV: Dependence is negative: NoDep\n"); return false; } // Write to the same location with the same size. // Could be improved to assert type sizes are the same (i32 == float, etc). if (Val == 0) { if (ATy == BTy) return false; DEBUG(dbgs() << "LV: Zero dependence difference but different types\n"); return true; } assert(Val.isStrictlyPositive() && "Expect a positive value"); // Positive distance bigger than max vectorization factor. if (ATy != BTy) { DEBUG(dbgs() << "LV: ReadWrite-Write positive dependency with different types\n"); return false; } unsigned Distance = (unsigned) Val.getZExtValue(); // Bail out early if passed-in parameters make vectorization not feasible. unsigned ForcedFactor = VectorizationFactor ? VectorizationFactor : 1; unsigned ForcedUnroll = VectorizationUnroll ? VectorizationUnroll : 1; // The distance must be bigger than the size needed for a vectorized version // of the operation and the size of the vectorized operation must not be // bigger than the currrent maximum size. if (Distance < 2*TypeByteSize || 2*TypeByteSize > MaxSafeDepDistBytes || Distance < TypeByteSize * ForcedUnroll * ForcedFactor) { DEBUG(dbgs() << "LV: Failure because of Positive distance " << Val.getSExtValue() << '\n'); return true; } MaxSafeDepDistBytes = Distance < MaxSafeDepDistBytes ? Distance : MaxSafeDepDistBytes; bool IsTrueDataDependence = (!AIsWrite && BIsWrite); if (IsTrueDataDependence && couldPreventStoreLoadForward(Distance, TypeByteSize)) return true; DEBUG(dbgs() << "LV: Positive distance " << Val.getSExtValue() << " with max VF = " << MaxSafeDepDistBytes / TypeByteSize << '\n'); return false; } bool MemoryDepChecker::areDepsSafe(AccessAnalysis::DepCandidates &AccessSets, MemAccessInfoSet &CheckDeps, ValueToValueMap &Strides) { MaxSafeDepDistBytes = -1U; while (!CheckDeps.empty()) { MemAccessInfo CurAccess = *CheckDeps.begin(); // Get the relevant memory access set. EquivalenceClasses
::iterator I = AccessSets.findValue(AccessSets.getLeaderValue(CurAccess)); // Check accesses within this set. EquivalenceClasses
::member_iterator AI, AE; AI = AccessSets.member_begin(I), AE = AccessSets.member_end(); // Check every access pair. while (AI != AE) { CheckDeps.erase(*AI); EquivalenceClasses
::member_iterator OI = std::next(AI); while (OI != AE) { // Check every accessing instruction pair in program order. for (std::vector
::iterator I1 = Accesses[*AI].begin(), I1E = Accesses[*AI].end(); I1 != I1E; ++I1) for (std::vector
::iterator I2 = Accesses[*OI].begin(), I2E = Accesses[*OI].end(); I2 != I2E; ++I2) { if (*I1 < *I2 && isDependent(*AI, *I1, *OI, *I2, Strides)) return false; if (*I2 < *I1 && isDependent(*OI, *I2, *AI, *I1, Strides)) return false; } ++OI; } AI++; } } return true; } bool LoopVectorizationLegality::canVectorizeMemory() { typedef SmallVector
ValueVector; typedef SmallPtrSet
ValueSet; // Holds the Load and Store *instructions*. ValueVector Loads; ValueVector Stores; // Holds all the different accesses in the loop. unsigned NumReads = 0; unsigned NumReadWrites = 0; PtrRtCheck.Pointers.clear(); PtrRtCheck.Need = false; const bool IsAnnotatedParallel = TheLoop->isAnnotatedParallel(); MemoryDepChecker DepChecker(SE, DL, TheLoop); // For each block. for (Loop::block_iterator bb = TheLoop->block_begin(), be = TheLoop->block_end(); bb != be; ++bb) { // Scan the BB and collect legal loads and stores. for (BasicBlock::iterator it = (*bb)->begin(), e = (*bb)->end(); it != e; ++it) { // If this is a load, save it. If this instruction can read from memory // but is not a load, then we quit. Notice that we don't handle function // calls that read or write. if (it->mayReadFromMemory()) { // Many math library functions read the rounding mode. We will only // vectorize a loop if it contains known function calls that don't set // the flag. Therefore, it is safe to ignore this read from memory. CallInst *Call = dyn_cast
(it); if (Call && getIntrinsicIDForCall(Call, TLI)) continue; LoadInst *Ld = dyn_cast
(it); if (!Ld || (!Ld->isSimple() && !IsAnnotatedParallel)) { emitAnalysis(Report(Ld) << "read with atomic ordering or volatile read"); DEBUG(dbgs() << "LV: Found a non-simple load.\n"); return false; } NumLoads++; Loads.push_back(Ld); DepChecker.addAccess(Ld); continue; } // Save 'store' instructions. Abort if other instructions write to memory. if (it->mayWriteToMemory()) { StoreInst *St = dyn_cast
(it); if (!St) { emitAnalysis(Report(it) << "instruction cannot be vectorized"); return false; } if (!St->isSimple() && !IsAnnotatedParallel) { emitAnalysis(Report(St) << "write with atomic ordering or volatile write"); DEBUG(dbgs() << "LV: Found a non-simple store.\n"); return false; } NumStores++; Stores.push_back(St); DepChecker.addAccess(St); } } // Next instr. } // Next block. // Now we have two lists that hold the loads and the stores. // Next, we find the pointers that they use. // Check if we see any stores. If there are no stores, then we don't // care if the pointers are *restrict*. if (!Stores.size()) { DEBUG(dbgs() << "LV: Found a read-only loop!\n"); return true; } AccessAnalysis::DepCandidates DependentAccesses; AccessAnalysis Accesses(DL, DependentAccesses); // Holds the analyzed pointers. We don't want to call GetUnderlyingObjects // multiple times on the same object. If the ptr is accessed twice, once // for read and once for write, it will only appear once (on the write // list). This is okay, since we are going to check for conflicts between // writes and between reads and writes, but not between reads and reads. ValueSet Seen; ValueVector::iterator I, IE; for (I = Stores.begin(), IE = Stores.end(); I != IE; ++I) { StoreInst *ST = cast
(*I); Value* Ptr = ST->getPointerOperand(); if (isUniform(Ptr)) { emitAnalysis( Report(ST) << "write to a loop invariant address could not be vectorized"); DEBUG(dbgs() << "LV: We don't allow storing to uniform addresses\n"); return false; } // If we did *not* see this pointer before, insert it to the read-write // list. At this phase it is only a 'write' list. if (Seen.insert(Ptr)) { ++NumReadWrites; Accesses.addStore(Ptr); } } if (IsAnnotatedParallel) { DEBUG(dbgs() << "LV: A loop annotated parallel, ignore memory dependency " << "checks.\n"); return true; } for (I = Loads.begin(), IE = Loads.end(); I != IE; ++I) { LoadInst *LD = cast
(*I); Value* Ptr = LD->getPointerOperand(); // If we did *not* see this pointer before, insert it to the // read list. If we *did* see it before, then it is already in // the read-write list. This allows us to vectorize expressions // such as A[i] += x; Because the address of A[i] is a read-write // pointer. This only works if the index of A[i] is consecutive. // If the address of i is unknown (for example A[B[i]]) then we may // read a few words, modify, and write a few words, and some of the // words may be written to the same address. bool IsReadOnlyPtr = false; if (Seen.insert(Ptr) || !isStridedPtr(SE, DL, Ptr, TheLoop, Strides)) { ++NumReads; IsReadOnlyPtr = true; } Accesses.addLoad(Ptr, IsReadOnlyPtr); } // If we write (or read-write) to a single destination and there are no // other reads in this loop then is it safe to vectorize. if (NumReadWrites == 1 && NumReads == 0) { DEBUG(dbgs() << "LV: Found a write-only loop!\n"); return true; } // Build dependence sets and check whether we need a runtime pointer bounds // check. Accesses.buildDependenceSets(); bool NeedRTCheck = Accesses.isRTCheckNeeded(); // Find pointers with computable bounds. We are going to use this information // to place a runtime bound check. unsigned NumComparisons = 0; bool CanDoRT = false; if (NeedRTCheck) CanDoRT = Accesses.canCheckPtrAtRT(PtrRtCheck, NumComparisons, SE, TheLoop, Strides); DEBUG(dbgs() << "LV: We need to do " << NumComparisons << " pointer comparisons.\n"); // If we only have one set of dependences to check pointers among we don't // need a runtime check. if (NumComparisons == 0 && NeedRTCheck) NeedRTCheck = false; // Check that we did not collect too many pointers or found an unsizeable // pointer. if (!CanDoRT || NumComparisons > RuntimeMemoryCheckThreshold) { PtrRtCheck.reset(); CanDoRT = false; } if (CanDoRT) { DEBUG(dbgs() << "LV: We can perform a memory runtime check if needed.\n"); } if (NeedRTCheck && !CanDoRT) { emitAnalysis(Report() << "cannot identify array bounds"); DEBUG(dbgs() << "LV: We can't vectorize because we can't find " << "the array bounds.\n"); PtrRtCheck.reset(); return false; } PtrRtCheck.Need = NeedRTCheck; bool CanVecMem = true; if (Accesses.isDependencyCheckNeeded()) { DEBUG(dbgs() << "LV: Checking memory dependencies\n"); CanVecMem = DepChecker.areDepsSafe( DependentAccesses, Accesses.getDependenciesToCheck(), Strides); MaxSafeDepDistBytes = DepChecker.getMaxSafeDepDistBytes(); if (!CanVecMem && DepChecker.shouldRetryWithRuntimeCheck()) { DEBUG(dbgs() << "LV: Retrying with memory checks\n"); NeedRTCheck = true; // Clear the dependency checks. We assume they are not needed. Accesses.resetDepChecks(); PtrRtCheck.reset(); PtrRtCheck.Need = true; CanDoRT = Accesses.canCheckPtrAtRT(PtrRtCheck, NumComparisons, SE, TheLoop, Strides, true); // Check that we did not collect too many pointers or found an unsizeable // pointer. if (!CanDoRT || NumComparisons > RuntimeMemoryCheckThreshold) { if (!CanDoRT && NumComparisons > 0) emitAnalysis(Report() << "cannot check memory dependencies at runtime"); else emitAnalysis(Report() << NumComparisons << " exceeds limit of " << RuntimeMemoryCheckThreshold << " dependent memory operations checked at runtime"); DEBUG(dbgs() << "LV: Can't vectorize with memory checks\n"); PtrRtCheck.reset(); return false; } CanVecMem = true; } } if (!CanVecMem) emitAnalysis(Report() << "unsafe dependent memory operations in loop"); DEBUG(dbgs() << "LV: We" << (NeedRTCheck ? "" : " don't") << " need a runtime memory check.\n"); return CanVecMem; } static bool hasMultipleUsesOf(Instruction *I, SmallPtrSet
&Insts) { unsigned NumUses = 0; for(User::op_iterator Use = I->op_begin(), E = I->op_end(); Use != E; ++Use) { if (Insts.count(dyn_cast
(*Use))) ++NumUses; if (NumUses > 1) return true; } return false; } static bool areAllUsesIn(Instruction *I, SmallPtrSet
&Set) { for(User::op_iterator Use = I->op_begin(), E = I->op_end(); Use != E; ++Use) if (!Set.count(dyn_cast
(*Use))) return false; return true; } bool LoopVectorizationLegality::AddReductionVar(PHINode *Phi, ReductionKind Kind) { if (Phi->getNumIncomingValues() != 2) return false; // Reduction variables are only found in the loop header block. if (Phi->getParent() != TheLoop->getHeader()) return false; // Obtain the reduction start value from the value that comes from the loop // preheader. Value *RdxStart = Phi->getIncomingValueForBlock(TheLoop->getLoopPreheader()); // ExitInstruction is the single value which is used outside the loop. // We only allow for a single reduction value to be used outside the loop. // This includes users of the reduction, variables (which form a cycle // which ends in the phi node). Instruction *ExitInstruction = nullptr; // Indicates that we found a reduction operation in our scan. bool FoundReduxOp = false; // We start with the PHI node and scan for all of the users of this // instruction. All users must be instructions that can be used as reduction // variables (such as ADD). We must have a single out-of-block user. The cycle // must include the original PHI. bool FoundStartPHI = false; // To recognize min/max patterns formed by a icmp select sequence, we store // the number of instruction we saw from the recognized min/max pattern, // to make sure we only see exactly the two instructions. unsigned NumCmpSelectPatternInst = 0; ReductionInstDesc ReduxDesc(false, nullptr); SmallPtrSet
VisitedInsts; SmallVector
Worklist; Worklist.push_back(Phi); VisitedInsts.insert(Phi); // A value in the reduction can be used: // - By the reduction: // - Reduction operation: // - One use of reduction value (safe). // - Multiple use of reduction value (not safe). // - PHI: // - All uses of the PHI must be the reduction (safe). // - Otherwise, not safe. // - By one instruction outside of the loop (safe). // - By further instructions outside of the loop (not safe). // - By an instruction that is not part of the reduction (not safe). // This is either: // * An instruction type other than PHI or the reduction operation. // * A PHI in the header other than the initial PHI. while (!Worklist.empty()) { Instruction *Cur = Worklist.back(); Worklist.pop_back(); // No Users. // If the instruction has no users then this is a broken chain and can't be // a reduction variable. if (Cur->use_empty()) return false; bool IsAPhi = isa
(Cur); // A header PHI use other than the original PHI. if (Cur != Phi && IsAPhi && Cur->getParent() == Phi->getParent()) return false; // Reductions of instructions such as Div, and Sub is only possible if the // LHS is the reduction variable. if (!Cur->isCommutative() && !IsAPhi && !isa
(Cur) && !isa
(Cur) && !isa
(Cur) && !VisitedInsts.count(dyn_cast
(Cur->getOperand(0)))) return false; // Any reduction instruction must be of one of the allowed kinds. ReduxDesc = isReductionInstr(Cur, Kind, ReduxDesc); if (!ReduxDesc.IsReduction) return false; // A reduction operation must only have one use of the reduction value. if (!IsAPhi && Kind != RK_IntegerMinMax && Kind != RK_FloatMinMax && hasMultipleUsesOf(Cur, VisitedInsts)) return false; // All inputs to a PHI node must be a reduction value. if(IsAPhi && Cur != Phi && !areAllUsesIn(Cur, VisitedInsts)) return false; if (Kind == RK_IntegerMinMax && (isa
(Cur) || isa
(Cur))) ++NumCmpSelectPatternInst; if (Kind == RK_FloatMinMax && (isa
(Cur) || isa
(Cur))) ++NumCmpSelectPatternInst; // Check whether we found a reduction operator. FoundReduxOp |= !IsAPhi; // Process users of current instruction. Push non-PHI nodes after PHI nodes // onto the stack. This way we are going to have seen all inputs to PHI // nodes once we get to them. SmallVector
NonPHIs; SmallVector
PHIs; for (User *U : Cur->users()) { Instruction *UI = cast
(U); // Check if we found the exit user. BasicBlock *Parent = UI->getParent(); if (!TheLoop->contains(Parent)) { // Exit if you find multiple outside users or if the header phi node is // being used. In this case the user uses the value of the previous // iteration, in which case we would loose "VF-1" iterations of the // reduction operation if we vectorize. if (ExitInstruction != nullptr || Cur == Phi) return false; // The instruction used by an outside user must be the last instruction // before we feed back to the reduction phi. Otherwise, we loose VF-1 // operations on the value. if (std::find(Phi->op_begin(), Phi->op_end(), Cur) == Phi->op_end()) return false; ExitInstruction = Cur; continue; } // Process instructions only once (termination). Each reduction cycle // value must only be used once, except by phi nodes and min/max // reductions which are represented as a cmp followed by a select. ReductionInstDesc IgnoredVal(false, nullptr); if (VisitedInsts.insert(UI)) { if (isa
(UI)) PHIs.push_back(UI); else NonPHIs.push_back(UI); } else if (!isa
(UI) && ((!isa
(UI) && !isa
(UI) && !isa
(UI)) || !isMinMaxSelectCmpPattern(UI, IgnoredVal).IsReduction)) return false; // Remember that we completed the cycle. if (UI == Phi) FoundStartPHI = true; } Worklist.append(PHIs.begin(), PHIs.end()); Worklist.append(NonPHIs.begin(), NonPHIs.end()); } // This means we have seen one but not the other instruction of the // pattern or more than just a select and cmp. if ((Kind == RK_IntegerMinMax || Kind == RK_FloatMinMax) && NumCmpSelectPatternInst != 2) return false; if (!FoundStartPHI || !FoundReduxOp || !ExitInstruction) return false; // We found a reduction var if we have reached the original phi node and we // only have a single instruction with out-of-loop users. // This instruction is allowed to have out-of-loop users. AllowedExit.insert(ExitInstruction); // Save the description of this reduction variable. ReductionDescriptor RD(RdxStart, ExitInstruction, Kind, ReduxDesc.MinMaxKind); Reductions[Phi] = RD; // We've ended the cycle. This is a reduction variable if we have an // outside user and it has a binary op. return true; } /// Returns true if the instruction is a Select(ICmp(X, Y), X, Y) instruction /// pattern corresponding to a min(X, Y) or max(X, Y). LoopVectorizationLegality::ReductionInstDesc LoopVectorizationLegality::isMinMaxSelectCmpPattern(Instruction *I, ReductionInstDesc &Prev) { assert((isa
(I) || isa
(I) || isa
(I)) && "Expect a select instruction"); Instruction *Cmp = nullptr; SelectInst *Select = nullptr; // We must handle the select(cmp()) as a single instruction. Advance to the // select. if ((Cmp = dyn_cast
(I)) || (Cmp = dyn_cast
(I))) { if (!Cmp->hasOneUse() || !(Select = dyn_cast
(*I->user_begin()))) return ReductionInstDesc(false, I); return ReductionInstDesc(Select, Prev.MinMaxKind); } // Only handle single use cases for now. if (!(Select = dyn_cast
(I))) return ReductionInstDesc(false, I); if (!(Cmp = dyn_cast
(I->getOperand(0))) && !(Cmp = dyn_cast
(I->getOperand(0)))) return ReductionInstDesc(false, I); if (!Cmp->hasOneUse()) return ReductionInstDesc(false, I); Value *CmpLeft; Value *CmpRight; // Look for a min/max pattern. if (m_UMin(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_UIntMin); else if (m_UMax(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_UIntMax); else if (m_SMax(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_SIntMax); else if (m_SMin(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_SIntMin); else if (m_OrdFMin(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_FloatMin); else if (m_OrdFMax(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_FloatMax); else if (m_UnordFMin(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_FloatMin); else if (m_UnordFMax(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_FloatMax); return ReductionInstDesc(false, I); } LoopVectorizationLegality::ReductionInstDesc LoopVectorizationLegality::isReductionInstr(Instruction *I, ReductionKind Kind, ReductionInstDesc &Prev) { bool FP = I->getType()->isFloatingPointTy(); bool FastMath = (FP && I->isCommutative() && I->isAssociative()); switch (I->getOpcode()) { default: return ReductionInstDesc(false, I); case Instruction::PHI: if (FP && (Kind != RK_FloatMult && Kind != RK_FloatAdd && Kind != RK_FloatMinMax)) return ReductionInstDesc(false, I); return ReductionInstDesc(I, Prev.MinMaxKind); case Instruction::Sub: case Instruction::Add: return ReductionInstDesc(Kind == RK_IntegerAdd, I); case Instruction::Mul: return ReductionInstDesc(Kind == RK_IntegerMult, I); case Instruction::And: return ReductionInstDesc(Kind == RK_IntegerAnd, I); case Instruction::Or: return ReductionInstDesc(Kind == RK_IntegerOr, I); case Instruction::Xor: return ReductionInstDesc(Kind == RK_IntegerXor, I); case Instruction::FMul: return ReductionInstDesc(Kind == RK_FloatMult && FastMath, I); case Instruction::FAdd: return ReductionInstDesc(Kind == RK_FloatAdd && FastMath, I); case Instruction::FCmp: case Instruction::ICmp: case Instruction::Select: if (Kind != RK_IntegerMinMax && (!HasFunNoNaNAttr || Kind != RK_FloatMinMax)) return ReductionInstDesc(false, I); return isMinMaxSelectCmpPattern(I, Prev); } } LoopVectorizationLegality::InductionKind LoopVectorizationLegality::isInductionVariable(PHINode *Phi) { Type *PhiTy = Phi->getType(); // We only handle integer and pointer inductions variables. if (!PhiTy->isIntegerTy() && !PhiTy->isPointerTy()) return IK_NoInduction; // Check that the PHI is consecutive. const SCEV *PhiScev = SE->getSCEV(Phi); const SCEVAddRecExpr *AR = dyn_cast
(PhiScev); if (!AR) { DEBUG(dbgs() << "LV: PHI is not a poly recurrence.\n"); return IK_NoInduction; } const SCEV *Step = AR->getStepRecurrence(*SE); // Integer inductions need to have a stride of one. if (PhiTy->isIntegerTy()) { if (Step->isOne()) return IK_IntInduction; if (Step->isAllOnesValue()) return IK_ReverseIntInduction; return IK_NoInduction; } // Calculate the pointer stride and check if it is consecutive. const SCEVConstant *C = dyn_cast
(Step); if (!C) return IK_NoInduction; assert(PhiTy->isPointerTy() && "The PHI must be a pointer"); uint64_t Size = DL->getTypeAllocSize(PhiTy->getPointerElementType()); if (C->getValue()->equalsInt(Size)) return IK_PtrInduction; else if (C->getValue()->equalsInt(0 - Size)) return IK_ReversePtrInduction; return IK_NoInduction; } bool LoopVectorizationLegality::isInductionVariable(const Value *V) { Value *In0 = const_cast
(V); PHINode *PN = dyn_cast_or_null
(In0); if (!PN) return false; return Inductions.count(PN); } bool LoopVectorizationLegality::blockNeedsPredication(BasicBlock *BB) { assert(TheLoop->contains(BB) && "Unknown block used"); // Blocks that do not dominate the latch need predication. BasicBlock* Latch = TheLoop->getLoopLatch(); return !DT->dominates(BB, Latch); } bool LoopVectorizationLegality::blockCanBePredicated(BasicBlock *BB, SmallPtrSet
& SafePtrs) { for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { // We might be able to hoist the load. if (it->mayReadFromMemory()) { LoadInst *LI = dyn_cast
(it); if (!LI || !SafePtrs.count(LI->getPointerOperand())) return false; } // We don't predicate stores at the moment. if (it->mayWriteToMemory()) { StoreInst *SI = dyn_cast
(it); // We only support predication of stores in basic blocks with one // predecessor. if (!SI || ++NumPredStores > NumberOfStoresToPredicate || !SafePtrs.count(SI->getPointerOperand()) || !SI->getParent()->getSinglePredecessor()) return false; } if (it->mayThrow()) return false; // Check that we don't have a constant expression that can trap as operand. for (Instruction::op_iterator OI = it->op_begin(), OE = it->op_end(); OI != OE; ++OI) { if (Constant *C = dyn_cast
(*OI)) if (C->canTrap()) return false; } // The instructions below can trap. switch (it->getOpcode()) { default: continue; case Instruction::UDiv: case Instruction::SDiv: case Instruction::URem: case Instruction::SRem: return false; } } return true; } LoopVectorizationCostModel::VectorizationFactor LoopVectorizationCostModel::selectVectorizationFactor(bool OptForSize, unsigned UserVF, bool ForceVectorization) { // Width 1 means no vectorize VectorizationFactor Factor = { 1U, 0U }; if (OptForSize && Legal->getRuntimePointerCheck()->Need) { DEBUG(dbgs() << "LV: Aborting. Runtime ptr check is required in Os.\n"); return Factor; } if (!EnableCondStoresVectorization && Legal->NumPredStores) { DEBUG(dbgs() << "LV: No vectorization. There are conditional stores.\n"); return Factor; } // Find the trip count. unsigned TC = SE->getSmallConstantTripCount(TheLoop, TheLoop->getLoopLatch()); DEBUG(dbgs() << "LV: Found trip count: " << TC << '\n'); unsigned WidestType = getWidestType(); unsigned WidestRegister = TTI.getRegisterBitWidth(true); unsigned MaxSafeDepDist = -1U; if (Legal->getMaxSafeDepDistBytes() != -1U) MaxSafeDepDist = Legal->getMaxSafeDepDistBytes() * 8; WidestRegister = ((WidestRegister < MaxSafeDepDist) ? WidestRegister : MaxSafeDepDist); unsigned MaxVectorSize = WidestRegister / WidestType; DEBUG(dbgs() << "LV: The Widest type: " << WidestType << " bits.\n"); DEBUG(dbgs() << "LV: The Widest register is: " << WidestRegister << " bits.\n"); if (MaxVectorSize == 0) { DEBUG(dbgs() << "LV: The target has no vector registers.\n"); MaxVectorSize = 1; } assert(MaxVectorSize <= 32 && "Did not expect to pack so many elements" " into one vector!"); unsigned VF = MaxVectorSize; // If we optimize the program for size, avoid creating the tail loop. if (OptForSize) { // If we are unable to calculate the trip count then don't try to vectorize. if (TC < 2) { DEBUG(dbgs() << "LV: Aborting. A tail loop is required in Os.\n"); return Factor; } // Find the maximum SIMD width that can fit within the trip count. VF = TC % MaxVectorSize; if (VF == 0) VF = MaxVectorSize; // If the trip count that we found modulo the vectorization factor is not // zero then we require a tail. if (VF < 2) { DEBUG(dbgs() << "LV: Aborting. A tail loop is required in Os.\n"); return Factor; } } if (UserVF != 0) { assert(isPowerOf2_32(UserVF) && "VF needs to be a power of two"); DEBUG(dbgs() << "LV: Using user VF " << UserVF << ".\n"); Factor.Width = UserVF; return Factor; } float Cost = expectedCost(1); #ifndef NDEBUG const float ScalarCost = Cost; #endif /* NDEBUG */ unsigned Width = 1; DEBUG(dbgs() << "LV: Scalar loop costs: " << (int)ScalarCost << ".\n"); // Ignore scalar width, because the user explicitly wants vectorization. if (ForceVectorization && VF > 1) { Width = 2; Cost = expectedCost(Width) / (float)Width; } for (unsigned i=2; i <= VF; i*=2) { // Notice that the vector loop needs to be executed less times, so // we need to divide the cost of the vector loops by the width of // the vector elements. float VectorCost = expectedCost(i) / (float)i; DEBUG(dbgs() << "LV: Vector loop of width " << i << " costs: " << (int)VectorCost << ".\n"); if (VectorCost < Cost) { Cost = VectorCost; Width = i; } } DEBUG(if (ForceVectorization && Width > 1 && Cost >= ScalarCost) dbgs() << "LV: Vectorization seems to be not beneficial, " << "but was forced by a user.\n"); DEBUG(dbgs() << "LV: Selecting VF: "<< Width << ".\n"); Factor.Width = Width; Factor.Cost = Width * Cost; return Factor; } unsigned LoopVectorizationCostModel::getWidestType() { unsigned MaxWidth = 8; // For each block. for (Loop::block_iterator bb = TheLoop->block_begin(), be = TheLoop->block_end(); bb != be; ++bb) { BasicBlock *BB = *bb; // For each instruction in the loop. for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { Type *T = it->getType(); // Only examine Loads, Stores and PHINodes. if (!isa
(it) && !isa
(it) && !isa
(it)) continue; // Examine PHI nodes that are reduction variables. if (PHINode *PN = dyn_cast
(it)) if (!Legal->getReductionVars()->count(PN)) continue; // Examine the stored values. if (StoreInst *ST = dyn_cast
(it)) T = ST->getValueOperand()->getType(); // Ignore loaded pointer types and stored pointer types that are not // consecutive. However, we do want to take consecutive stores/loads of // pointer vectors into account. if (T->isPointerTy() && !isConsecutiveLoadOrStore(it)) continue; MaxWidth = std::max(MaxWidth, (unsigned)DL->getTypeSizeInBits(T->getScalarType())); } } return MaxWidth; } unsigned LoopVectorizationCostModel::selectUnrollFactor(bool OptForSize, unsigned UserUF, unsigned VF, unsigned LoopCost) { // -- The unroll heuristics -- // We unroll the loop in order to expose ILP and reduce the loop overhead. // There are many micro-architectural considerations that we can't predict // at this level. For example frontend pressure (on decode or fetch) due to // code size, or the number and capabilities of the execution ports. // // We use the following heuristics to select the unroll factor: // 1. If the code has reductions the we unroll in order to break the cross // iteration dependency. // 2. If the loop is really small then we unroll in order to reduce the loop // overhead. // 3. We don't unroll if we think that we will spill registers to memory due // to the increased register pressure. // Use the user preference, unless 'auto' is selected. if (UserUF != 0) return UserUF; // When we optimize for size we don't unroll. if (OptForSize) return 1; // We used the distance for the unroll factor. if (Legal->getMaxSafeDepDistBytes() != -1U) return 1; // Do not unroll loops with a relatively small trip count. unsigned TC = SE->getSmallConstantTripCount(TheLoop, TheLoop->getLoopLatch()); if (TC > 1 && TC < TinyTripCountUnrollThreshold) return 1; unsigned TargetNumRegisters = TTI.getNumberOfRegisters(VF > 1); DEBUG(dbgs() << "LV: The target has " << TargetNumRegisters << " registers\n"); if (VF == 1) { if (ForceTargetNumScalarRegs.getNumOccurrences() > 0) TargetNumRegisters = ForceTargetNumScalarRegs; } else { if (ForceTargetNumVectorRegs.getNumOccurrences() > 0) TargetNumRegisters = ForceTargetNumVectorRegs; } LoopVectorizationCostModel::RegisterUsage R = calculateRegisterUsage(); // We divide by these constants so assume that we have at least one // instruction that uses at least one register. R.MaxLocalUsers = std::max(R.MaxLocalUsers, 1U); R.NumInstructions = std::max(R.NumInstructions, 1U); // We calculate the unroll factor using the following formula. // Subtract the number of loop invariants from the number of available // registers. These registers are used by all of the unrolled instances. // Next, divide the remaining registers by the number of registers that is // required by the loop, in order to estimate how many parallel instances // fit without causing spills. All of this is rounded down if necessary to be // a power of two. We want power of two unroll factors to simplify any // addressing operations or alignment considerations. unsigned UF = PowerOf2Floor((TargetNumRegisters - R.LoopInvariantRegs) / R.MaxLocalUsers); // Don't count the induction variable as unrolled. if (EnableIndVarRegisterHeur) UF = PowerOf2Floor((TargetNumRegisters - R.LoopInvariantRegs - 1) / std::max(1U, (R.MaxLocalUsers - 1))); // Clamp the unroll factor ranges to reasonable factors. unsigned MaxUnrollSize = TTI.getMaximumUnrollFactor(); // Check if the user has overridden the unroll max. if (VF == 1) { if (ForceTargetMaxScalarUnrollFactor.getNumOccurrences() > 0) MaxUnrollSize = ForceTargetMaxScalarUnrollFactor; } else { if (ForceTargetMaxVectorUnrollFactor.getNumOccurrences() > 0) MaxUnrollSize = ForceTargetMaxVectorUnrollFactor; } // If we did not calculate the cost for VF (because the user selected the VF) // then we calculate the cost of VF here. if (LoopCost == 0) LoopCost = expectedCost(VF); // Clamp the calculated UF to be between the 1 and the max unroll factor // that the target allows. if (UF > MaxUnrollSize) UF = MaxUnrollSize; else if (UF < 1) UF = 1; // Unroll if we vectorized this loop and there is a reduction that could // benefit from unrolling. if (VF > 1 && Legal->getReductionVars()->size()) { DEBUG(dbgs() << "LV: Unrolling because of reductions.\n"); return UF; } // Note that if we've already vectorized the loop we will have done the // runtime check and so unrolling won't require further checks. bool UnrollingRequiresRuntimePointerCheck = (VF == 1 && Legal->getRuntimePointerCheck()->Need); // We want to unroll small loops in order to reduce the loop overhead and // potentially expose ILP opportunities. DEBUG(dbgs() << "LV: Loop cost is " << LoopCost << '\n'); if (!UnrollingRequiresRuntimePointerCheck && LoopCost < SmallLoopCost) { // We assume that the cost overhead is 1 and we use the cost model // to estimate the cost of the loop and unroll until the cost of the // loop overhead is about 5% of the cost of the loop. unsigned SmallUF = std::min(UF, (unsigned)PowerOf2Floor(SmallLoopCost / LoopCost)); // Unroll until store/load ports (estimated by max unroll factor) are // saturated. unsigned StoresUF = UF / (Legal->NumStores ? Legal->NumStores : 1); unsigned LoadsUF = UF / (Legal->NumLoads ? Legal->NumLoads : 1); if (EnableLoadStoreRuntimeUnroll && std::max(StoresUF, LoadsUF) > SmallUF) { DEBUG(dbgs() << "LV: Unrolling to saturate store or load ports.\n"); return std::max(StoresUF, LoadsUF); } DEBUG(dbgs() << "LV: Unrolling to reduce branch cost.\n"); return SmallUF; } DEBUG(dbgs() << "LV: Not Unrolling.\n"); return 1; } LoopVectorizationCostModel::RegisterUsage LoopVectorizationCostModel::calculateRegisterUsage() { // This function calculates the register usage by measuring the highest number // of values that are alive at a single location. Obviously, this is a very // rough estimation. We scan the loop in a topological order in order and // assign a number to each instruction. We use RPO to ensure that defs are // met before their users. We assume that each instruction that has in-loop // users starts an interval. We record every time that an in-loop value is // used, so we have a list of the first and last occurrences of each // instruction. Next, we transpose this data structure into a multi map that // holds the list of intervals that *end* at a specific location. This multi // map allows us to perform a linear search. We scan the instructions linearly // and record each time that a new interval starts, by placing it in a set. // If we find this value in the multi-map then we remove it from the set. // The max register usage is the maximum size of the set. // We also search for instructions that are defined outside the loop, but are // used inside the loop. We need this number separately from the max-interval // usage number because when we unroll, loop-invariant values do not take // more register. LoopBlocksDFS DFS(TheLoop); DFS.perform(LI); RegisterUsage R; R.NumInstructions = 0; // Each 'key' in the map opens a new interval. The values // of the map are the index of the 'last seen' usage of the // instruction that is the key. typedef DenseMap
IntervalMap; // Maps instruction to its index. DenseMap
IdxToInstr; // Marks the end of each interval. IntervalMap EndPoint; // Saves the list of instruction indices that are used in the loop. SmallSet
Ends; // Saves the list of values that are used in the loop but are // defined outside the loop, such as arguments and constants. SmallPtrSet
LoopInvariants; unsigned Index = 0; for (LoopBlocksDFS::RPOIterator bb = DFS.beginRPO(), be = DFS.endRPO(); bb != be; ++bb) { R.NumInstructions += (*bb)->size(); for (BasicBlock::iterator it = (*bb)->begin(), e = (*bb)->end(); it != e; ++it) { Instruction *I = it; IdxToInstr[Index++] = I; // Save the end location of each USE. for (unsigned i = 0; i < I->getNumOperands(); ++i) { Value *U = I->getOperand(i); Instruction *Instr = dyn_cast
(U); // Ignore non-instruction values such as arguments, constants, etc. if (!Instr) continue; // If this instruction is outside the loop then record it and continue. if (!TheLoop->contains(Instr)) { LoopInvariants.insert(Instr); continue; } // Overwrite previous end points. EndPoint[Instr] = Index; Ends.insert(Instr); } } } // Saves the list of intervals that end with the index in 'key'. typedef SmallVector
InstrList; DenseMap
TransposeEnds; // Transpose the EndPoints to a list of values that end at each index. for (IntervalMap::iterator it = EndPoint.begin(), e = EndPoint.end(); it != e; ++it) TransposeEnds[it->second].push_back(it->first); SmallSet
OpenIntervals; unsigned MaxUsage = 0; DEBUG(dbgs() << "LV(REG): Calculating max register usage:\n"); for (unsigned int i = 0; i < Index; ++i) { Instruction *I = IdxToInstr[i]; // Ignore instructions that are never used within the loop. if (!Ends.count(I)) continue; // Remove all of the instructions that end at this location. InstrList &List = TransposeEnds[i]; for (unsigned int j=0, e = List.size(); j < e; ++j) OpenIntervals.erase(List[j]); // Count the number of live interals. MaxUsage = std::max(MaxUsage, OpenIntervals.size()); DEBUG(dbgs() << "LV(REG): At #" << i << " Interval # " << OpenIntervals.size() << '\n'); // Add the current instruction to the list of open intervals. OpenIntervals.insert(I); } unsigned Invariant = LoopInvariants.size(); DEBUG(dbgs() << "LV(REG): Found max usage: " << MaxUsage << '\n'); DEBUG(dbgs() << "LV(REG): Found invariant usage: " << Invariant << '\n'); DEBUG(dbgs() << "LV(REG): LoopSize: " << R.NumInstructions << '\n'); R.LoopInvariantRegs = Invariant; R.MaxLocalUsers = MaxUsage; return R; } unsigned LoopVectorizationCostModel::expectedCost(unsigned VF) { unsigned Cost = 0; // For each block. for (Loop::block_iterator bb = TheLoop->block_begin(), be = TheLoop->block_end(); bb != be; ++bb) { unsigned BlockCost = 0; BasicBlock *BB = *bb; // For each instruction in the old loop. for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { // Skip dbg intrinsics. if (isa
(it)) continue; unsigned C = getInstructionCost(it, VF); // Check if we should override the cost. if (ForceTargetInstructionCost.getNumOccurrences() > 0) C = ForceTargetInstructionCost; BlockCost += C; DEBUG(dbgs() << "LV: Found an estimated cost of " << C << " for VF " << VF << " For instruction: " << *it << '\n'); } // We assume that if-converted blocks have a 50% chance of being executed. // When the code is scalar then some of the blocks are avoided due to CF. // When the code is vectorized we execute all code paths. if (VF == 1 && Legal->blockNeedsPredication(*bb)) BlockCost /= 2; Cost += BlockCost; } return Cost; } /// \brief Check whether the address computation for a non-consecutive memory /// access looks like an unlikely candidate for being merged into the indexing /// mode. /// /// We look for a GEP which has one index that is an induction variable and all /// other indices are loop invariant. If the stride of this access is also /// within a small bound we decide that this address computation can likely be /// merged into the addressing mode. /// In all other cases, we identify the address computation as complex. static bool isLikelyComplexAddressComputation(Value *Ptr, LoopVectorizationLegality *Legal, ScalarEvolution *SE, const Loop *TheLoop) { GetElementPtrInst *Gep = dyn_cast
(Ptr); if (!Gep) return true; // We are looking for a gep with all loop invariant indices except for one // which should be an induction variable. unsigned NumOperands = Gep->getNumOperands(); for (unsigned i = 1; i < NumOperands; ++i) { Value *Opd = Gep->getOperand(i); if (!SE->isLoopInvariant(SE->getSCEV(Opd), TheLoop) && !Legal->isInductionVariable(Opd)) return true; } // Now we know we have a GEP ptr, %inv, %ind, %inv. Make sure that the step // can likely be merged into the address computation. unsigned MaxMergeDistance = 64; const SCEVAddRecExpr *AddRec = dyn_cast
(SE->getSCEV(Ptr)); if (!AddRec) return true; // Check the step is constant. const SCEV *Step = AddRec->getStepRecurrence(*SE); // Calculate the pointer stride and check if it is consecutive. const SCEVConstant *C = dyn_cast
(Step); if (!C) return true; const APInt &APStepVal = C->getValue()->getValue(); // Huge step value - give up. if (APStepVal.getBitWidth() > 64) return true; int64_t StepVal = APStepVal.getSExtValue(); return StepVal > MaxMergeDistance; } static bool isStrideMul(Instruction *I, LoopVectorizationLegality *Legal) { if (Legal->hasStride(I->getOperand(0)) || Legal->hasStride(I->getOperand(1))) return true; return false; } unsigned LoopVectorizationCostModel::getInstructionCost(Instruction *I, unsigned VF) { // If we know that this instruction will remain uniform, check the cost of // the scalar version. if (Legal->isUniformAfterVectorization(I)) VF = 1; Type *RetTy = I->getType(); Type *VectorTy = ToVectorTy(RetTy, VF); // TODO: We need to estimate the cost of intrinsic calls. switch (I->getOpcode()) { case Instruction::GetElementPtr: // We mark this instruction as zero-cost because the cost of GEPs in // vectorized code depends on whether the corresponding memory instruction // is scalarized or not. Therefore, we handle GEPs with the memory // instruction cost. return 0; case Instruction::Br: { return TTI.getCFInstrCost(I->getOpcode()); } case Instruction::PHI: //TODO: IF-converted IFs become selects. return 0; case Instruction::Add: case Instruction::FAdd: case Instruction::Sub: case Instruction::FSub: case Instruction::Mul: case Instruction::FMul: case Instruction::UDiv: case Instruction::SDiv: case Instruction::FDiv: case Instruction::URem: case Instruction::SRem: case Instruction::FRem: case Instruction::Shl: case Instruction::LShr: case Instruction::AShr: case Instruction::And: case Instruction::Or: case Instruction::Xor: { // Since we will replace the stride by 1 the multiplication should go away. if (I->getOpcode() == Instruction::Mul && isStrideMul(I, Legal)) return 0; // Certain instructions can be cheaper to vectorize if they have a constant // second vector operand. One example of this are shifts on x86. TargetTransformInfo::OperandValueKind Op1VK = TargetTransformInfo::OK_AnyValue; TargetTransformInfo::OperandValueKind Op2VK = TargetTransformInfo::OK_AnyValue; Value *Op2 = I->getOperand(1); // Check for a splat of a constant or for a non uniform vector of constants. if (isa
(Op2)) Op2VK = TargetTransformInfo::OK_UniformConstantValue; else if (isa
(Op2) || isa
(Op2)) { Op2VK = TargetTransformInfo::OK_NonUniformConstantValue; if (cast
(Op2)->getSplatValue() != nullptr) Op2VK = TargetTransformInfo::OK_UniformConstantValue; } return TTI.getArithmeticInstrCost(I->getOpcode(), VectorTy, Op1VK, Op2VK); } case Instruction::Select: { SelectInst *SI = cast
(I); const SCEV *CondSCEV = SE->getSCEV(SI->getCondition()); bool ScalarCond = (SE->isLoopInvariant(CondSCEV, TheLoop)); Type *CondTy = SI->getCondition()->getType(); if (!ScalarCond) CondTy = VectorType::get(CondTy, VF); return TTI.getCmpSelInstrCost(I->getOpcode(), VectorTy, CondTy); } case Instruction::ICmp: case Instruction::FCmp: { Type *ValTy = I->getOperand(0)->getType(); VectorTy = ToVectorTy(ValTy, VF); return TTI.getCmpSelInstrCost(I->getOpcode(), VectorTy); } case Instruction::Store: case Instruction::Load: { StoreInst *SI = dyn_cast
(I); LoadInst *LI = dyn_cast
(I); Type *ValTy = (SI ? SI->getValueOperand()->getType() : LI->getType()); VectorTy = ToVectorTy(ValTy, VF); unsigned Alignment = SI ? SI->getAlignment() : LI->getAlignment(); unsigned AS = SI ? SI->getPointerAddressSpace() : LI->getPointerAddressSpace(); Value *Ptr = SI ? SI->getPointerOperand() : LI->getPointerOperand(); // We add the cost of address computation here instead of with the gep // instruction because only here we know whether the operation is // scalarized. if (VF == 1) return TTI.getAddressComputationCost(VectorTy) + TTI.getMemoryOpCost(I->getOpcode(), VectorTy, Alignment, AS); // Scalarized loads/stores. int ConsecutiveStride = Legal->isConsecutivePtr(Ptr); bool Reverse = ConsecutiveStride < 0; unsigned ScalarAllocatedSize = DL->getTypeAllocSize(ValTy); unsigned VectorElementSize = DL->getTypeStoreSize(VectorTy)/VF; if (!ConsecutiveStride || ScalarAllocatedSize != VectorElementSize) { bool IsComplexComputation = isLikelyComplexAddressComputation(Ptr, Legal, SE, TheLoop); unsigned Cost = 0; // The cost of extracting from the value vector and pointer vector. Type *PtrTy = ToVectorTy(Ptr->getType(), VF); for (unsigned i = 0; i < VF; ++i) { // The cost of extracting the pointer operand. Cost += TTI.getVectorInstrCost(Instruction::ExtractElement, PtrTy, i); // In case of STORE, the cost of ExtractElement from the vector. // In case of LOAD, the cost of InsertElement into the returned // vector. Cost += TTI.getVectorInstrCost(SI ? Instruction::ExtractElement : Instruction::InsertElement, VectorTy, i); } // The cost of the scalar loads/stores. Cost += VF * TTI.getAddressComputationCost(PtrTy, IsComplexComputation); Cost += VF * TTI.getMemoryOpCost(I->getOpcode(), ValTy->getScalarType(), Alignment, AS); return Cost; } // Wide load/stores. unsigned Cost = TTI.getAddressComputationCost(VectorTy); Cost += TTI.getMemoryOpCost(I->getOpcode(), VectorTy, Alignment, AS); if (Reverse) Cost += TTI.getShuffleCost(TargetTransformInfo::SK_Reverse, VectorTy, 0); return Cost; } case Instruction::ZExt: case Instruction::SExt: case Instruction::FPToUI: case Instruction::FPToSI: case Instruction::FPExt: case Instruction::PtrToInt: case Instruction::IntToPtr: case Instruction::SIToFP: case Instruction::UIToFP: case Instruction::Trunc: case Instruction::FPTrunc: case Instruction::BitCast: { // We optimize the truncation of induction variable. // The cost of these is the same as the scalar operation. if (I->getOpcode() == Instruction::Trunc && Legal->isInductionVariable(I->getOperand(0))) return TTI.getCastInstrCost(I->getOpcode(), I->getType(), I->getOperand(0)->getType()); Type *SrcVecTy = ToVectorTy(I->getOperand(0)->getType(), VF); return TTI.getCastInstrCost(I->getOpcode(), VectorTy, SrcVecTy); } case Instruction::Call: { CallInst *CI = cast