maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

4571 lines
161 KiB
C++
Raw Blame History

//===- SLPVectorizer.cpp - A bottom up SLP Vectorizer ---------------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
// This pass implements the Bottom Up SLP vectorizer. It detects consecutive
// stores that can be put together into vector-stores. Next, it attempts to
// construct vectorizable tree using the use-def chains. If a profitable tree
// was found, the SLP vectorizer performs vectorization on the tree.
//
// The pass is inspired by the work described in the paper:
// "Loop-Aware SLP in GCC" by Ira Rosen, Dorit Nuzman, Ayal Zaks.
//
//===----------------------------------------------------------------------===//
#include "llvm/ADT/MapVector.h"
#include "llvm/ADT/Optional.h"
#include "llvm/ADT/PostOrderIterator.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/CodeMetrics.h"
#include "llvm/Analysis/DemandedBits.h"
#include "llvm/Analysis/GlobalsModRef.h"
#include "llvm/Analysis/LoopAccessAnalysis.h"
#include "llvm/Analysis/LoopAccessAnalysis.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/ScalarEvolution.h"
#include "llvm/Analysis/ScalarEvolutionExpressions.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/Analysis/VectorUtils.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/NoFolder.h"
#include "llvm/IR/Type.h"
#include "llvm/IR/Value.h"
#include "llvm/IR/Verifier.h"
#include "llvm/Pass.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Vectorize.h"
#include <algorithm>
#include <map>
#include <memory>
using namespace llvm;
#define SV_NAME "slp-vectorizer"
#define DEBUG_TYPE "SLP"
STATISTIC(NumVectorInstructions, "Number of vector instructions generated");
static cl::opt<int>
SLPCostThreshold("slp-threshold", cl::init(0), cl::Hidden,
cl::desc("Only vectorize if you gain more than this "
"number "));
static cl::opt<bool>
ShouldVectorizeHor("slp-vectorize-hor", cl::init(true), cl::Hidden,
cl::desc("Attempt to vectorize horizontal reductions"));
static cl::opt<bool> ShouldStartVectorizeHorAtStore(
"slp-vectorize-hor-store", cl::init(false), cl::Hidden,
cl::desc(
"Attempt to vectorize horizontal reductions feeding into a store"));
static cl::opt<int>
MaxVectorRegSizeOption("slp-max-reg-size", cl::init(128), cl::Hidden,
cl::desc("Attempt to vectorize for this register size in bits"));
/// Limits the size of scheduling regions in a block.
/// It avoid long compile times for _very_ large blocks where vector
/// instructions are spread over a wide range.
/// This limit is way higher than needed by real-world functions.
static cl::opt<int>
ScheduleRegionSizeBudget("slp-schedule-budget", cl::init(100000), cl::Hidden,
cl::desc("Limit the size of the SLP scheduling region per block"));
namespace {
// FIXME: Set this via cl::opt to allow overriding.
static const unsigned MinVecRegSize = 128;
static const unsigned RecursionMaxDepth = 12;
// Limit the number of alias checks. The limit is chosen so that
// it has no negative effect on the llvm benchmarks.
static const unsigned AliasedCheckLimit = 10;
// Another limit for the alias checks: The maximum distance between load/store
// instructions where alias checks are done.
// This limit is useful for very large basic blocks.
static const unsigned MaxMemDepDistance = 160;
/// If the ScheduleRegionSizeBudget is exhausted, we allow small scheduling
/// regions to be handled.
static const int MinScheduleRegionSize = 16;
/// \brief Predicate for the element types that the SLP vectorizer supports.
///
/// The most important thing to filter here are types which are invalid in LLVM
/// vectors. We also filter target specific types which have absolutely no
/// meaningful vectorization path such as x86_fp80 and ppc_f128. This just
/// avoids spending time checking the cost model and realizing that they will
/// be inevitably scalarized.
static bool isValidElementType(Type *Ty) {
return VectorType::isValidElementType(Ty) && !Ty->isX86_FP80Ty() &&
!Ty->isPPC_FP128Ty();
}
/// \returns the parent basic block if all of the instructions in \p VL
/// are in the same block or null otherwise.
static BasicBlock *getSameBlock(ArrayRef<Value *> VL) {
Instruction *I0 = dyn_cast<Instruction>(VL[0]);
if (!I0)
return nullptr;
BasicBlock *BB = I0->getParent();
for (int i = 1, e = VL.size(); i < e; i++) {
Instruction *I = dyn_cast<Instruction>(VL[i]);
if (!I)
return nullptr;
if (BB != I->getParent())
return nullptr;
}
return BB;
}
/// \returns True if all of the values in \p VL are constants.
static bool allConstant(ArrayRef<Value *> VL) {
for (unsigned i = 0, e = VL.size(); i < e; ++i)
if (!isa<Constant>(VL[i]))
return false;
return true;
}
/// \returns True if all of the values in \p VL are identical.
static bool isSplat(ArrayRef<Value *> VL) {
for (unsigned i = 1, e = VL.size(); i < e; ++i)
if (VL[i] != VL[0])
return false;
return true;
}
///\returns Opcode that can be clubbed with \p Op to create an alternate
/// sequence which can later be merged as a ShuffleVector instruction.
static unsigned getAltOpcode(unsigned Op) {
switch (Op) {
case Instruction::FAdd:
return Instruction::FSub;
case Instruction::FSub:
return Instruction::FAdd;
case Instruction::Add:
return Instruction::Sub;
case Instruction::Sub:
return Instruction::Add;
default:
return 0;
}
}
///\returns bool representing if Opcode \p Op can be part
/// of an alternate sequence which can later be merged as
/// a ShuffleVector instruction.
static bool canCombineAsAltInst(unsigned Op) {
return Op == Instruction::FAdd || Op == Instruction::FSub ||
Op == Instruction::Sub || Op == Instruction::Add;
}
/// \returns ShuffleVector instruction if instructions in \p VL have
/// alternate fadd,fsub / fsub,fadd/add,sub/sub,add sequence.
/// (i.e. e.g. opcodes of fadd,fsub,fadd,fsub...)
static unsigned isAltInst(ArrayRef<Value *> VL) {
Instruction *I0 = dyn_cast<Instruction>(VL[0]);
unsigned Opcode = I0->getOpcode();
unsigned AltOpcode = getAltOpcode(Opcode);
for (int i = 1, e = VL.size(); i < e; i++) {
Instruction *I = dyn_cast<Instruction>(VL[i]);
if (!I || I->getOpcode() != ((i & 1) ? AltOpcode : Opcode))
return 0;
}
return Instruction::ShuffleVector;
}
/// \returns The opcode if all of the Instructions in \p VL have the same
/// opcode, or zero.
static unsigned getSameOpcode(ArrayRef<Value *> VL) {
Instruction *I0 = dyn_cast<Instruction>(VL[0]);
if (!I0)
return 0;
unsigned Opcode = I0->getOpcode();
for (int i = 1, e = VL.size(); i < e; i++) {
Instruction *I = dyn_cast<Instruction>(VL[i]);
if (!I || Opcode != I->getOpcode()) {
if (canCombineAsAltInst(Opcode) && i == 1)
return isAltInst(VL);
return 0;
}
}
return Opcode;
}
/// Get the intersection (logical and) of all of the potential IR flags
/// of each scalar operation (VL) that will be converted into a vector (I).
/// Flag set: NSW, NUW, exact, and all of fast-math.
static void propagateIRFlags(Value *I, ArrayRef<Value *> VL) {
if (auto *VecOp = dyn_cast<BinaryOperator>(I)) {
if (auto *Intersection = dyn_cast<BinaryOperator>(VL[0])) {
// Intersection is initialized to the 0th scalar,
// so start counting from index '1'.
for (int i = 1, e = VL.size(); i < e; ++i) {
if (auto *Scalar = dyn_cast<BinaryOperator>(VL[i]))
Intersection->andIRFlags(Scalar);
}
VecOp->copyIRFlags(Intersection);
}
}
}
/// \returns \p I after propagating metadata from \p VL.
static Instruction *propagateMetadata(Instruction *I, ArrayRef<Value *> VL) {
Instruction *I0 = cast<Instruction>(VL[0]);
SmallVector<std::pair<unsigned, MDNode *>, 4> Metadata;
I0->getAllMetadataOtherThanDebugLoc(Metadata);
for (unsigned i = 0, n = Metadata.size(); i != n; ++i) {
unsigned Kind = Metadata[i].first;
MDNode *MD = Metadata[i].second;
for (int i = 1, e = VL.size(); MD && i != e; i++) {
Instruction *I = cast<Instruction>(VL[i]);
MDNode *IMD = I->getMetadata(Kind);
switch (Kind) {
default:
MD = nullptr; // Remove unknown metadata
break;
case LLVMContext::MD_tbaa:
MD = MDNode::getMostGenericTBAA(MD, IMD);
break;
case LLVMContext::MD_alias_scope:
MD = MDNode::getMostGenericAliasScope(MD, IMD);
break;
case LLVMContext::MD_noalias:
MD = MDNode::intersect(MD, IMD);
break;
case LLVMContext::MD_fpmath:
MD = MDNode::getMostGenericFPMath(MD, IMD);
break;
case LLVMContext::MD_nontemporal:
MD = MDNode::intersect(MD, IMD);
break;
}
}
I->setMetadata(Kind, MD);
}
return I;
}
/// \returns The type that all of the values in \p VL have or null if there
/// are different types.
static Type* getSameType(ArrayRef<Value *> VL) {
Type *Ty = VL[0]->getType();
for (int i = 1, e = VL.size(); i < e; i++)
if (VL[i]->getType() != Ty)
return nullptr;
return Ty;
}
/// \returns True if the ExtractElement instructions in VL can be vectorized
/// to use the original vector.
static bool CanReuseExtract(ArrayRef<Value *> VL) {
assert(Instruction::ExtractElement == getSameOpcode(VL) && "Invalid opcode");
// Check if all of the extracts come from the same vector and from the
// correct offset.
Value *VL0 = VL[0];
ExtractElementInst *E0 = cast<ExtractElementInst>(VL0);
Value *Vec = E0->getOperand(0);
// We have to extract from the same vector type.
unsigned NElts = Vec->getType()->getVectorNumElements();
if (NElts != VL.size())
return false;
// Check that all of the indices extract from the correct offset.
ConstantInt *CI = dyn_cast<ConstantInt>(E0->getOperand(1));
if (!CI || CI->getZExtValue())
return false;
for (unsigned i = 1, e = VL.size(); i < e; ++i) {
ExtractElementInst *E = cast<ExtractElementInst>(VL[i]);
ConstantInt *CI = dyn_cast<ConstantInt>(E->getOperand(1));
if (!CI || CI->getZExtValue() != i || E->getOperand(0) != Vec)
return false;
}
return true;
}
/// \returns True if in-tree use also needs extract. This refers to
/// possible scalar operand in vectorized instruction.
static bool InTreeUserNeedToExtract(Value *Scalar, Instruction *UserInst,
TargetLibraryInfo *TLI) {
unsigned Opcode = UserInst->getOpcode();
switch (Opcode) {
case Instruction::Load: {
LoadInst *LI = cast<LoadInst>(UserInst);
return (LI->getPointerOperand() == Scalar);
}
case Instruction::Store: {
StoreInst *SI = cast<StoreInst>(UserInst);
return (SI->getPointerOperand() == Scalar);
}
case Instruction::Call: {
CallInst *CI = cast<CallInst>(UserInst);
Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI);
if (hasVectorInstrinsicScalarOpd(ID, 1)) {
return (CI->getArgOperand(1) == Scalar);
}
}
default:
return false;
}
}
/// \returns the AA location that is being access by the instruction.
static MemoryLocation getLocation(Instruction *I, AliasAnalysis *AA) {
if (StoreInst *SI = dyn_cast<StoreInst>(I))
return MemoryLocation::get(SI);
if (LoadInst *LI = dyn_cast<LoadInst>(I))
return MemoryLocation::get(LI);
return MemoryLocation();
}
/// \returns True if the instruction is not a volatile or atomic load/store.
static bool isSimple(Instruction *I) {
if (LoadInst *LI = dyn_cast<LoadInst>(I))
return LI->isSimple();
if (StoreInst *SI = dyn_cast<StoreInst>(I))
return SI->isSimple();
if (MemIntrinsic *MI = dyn_cast<MemIntrinsic>(I))
return !MI->isVolatile();
return true;
}
/// Bottom Up SLP Vectorizer.
class BoUpSLP {
public:
typedef SmallVector<Value *, 8> ValueList;
typedef SmallVector<Instruction *, 16> InstrList;
typedef SmallPtrSet<Value *, 16> ValueSet;
typedef SmallVector<StoreInst *, 8> StoreList;
BoUpSLP(Function *Func, ScalarEvolution *Se, TargetTransformInfo *Tti,
TargetLibraryInfo *TLi, AliasAnalysis *Aa, LoopInfo *Li,
DominatorTree *Dt, AssumptionCache *AC, DemandedBits *DB)
: NumLoadsWantToKeepOrder(0), NumLoadsWantToChangeOrder(0), F(Func),
SE(Se), TTI(Tti), TLI(TLi), AA(Aa), LI(Li), DT(Dt), AC(AC), DB(DB),
Builder(Se->getContext()) {
CodeMetrics::collectEphemeralValues(F, AC, EphValues);
}
/// \brief Vectorize the tree that starts with the elements in \p VL.
/// Returns the vectorized root.
Value *vectorizeTree();
/// \returns the cost incurred by unwanted spills and fills, caused by
/// holding live values over call sites.
int getSpillCost();
/// \returns the vectorization cost of the subtree that starts at \p VL.
/// A negative number means that this is profitable.
int getTreeCost();
/// Construct a vectorizable tree that starts at \p Roots, ignoring users for
/// the purpose of scheduling and extraction in the \p UserIgnoreLst.
void buildTree(ArrayRef<Value *> Roots,
ArrayRef<Value *> UserIgnoreLst = None);
/// Clear the internal data structures that are created by 'buildTree'.
void deleteTree() {
VectorizableTree.clear();
ScalarToTreeEntry.clear();
MustGather.clear();
ExternalUses.clear();
NumLoadsWantToKeepOrder = 0;
NumLoadsWantToChangeOrder = 0;
for (auto &Iter : BlocksSchedules) {
BlockScheduling *BS = Iter.second.get();
BS->clear();
}
MinBWs.clear();
}
/// \brief Perform LICM and CSE on the newly generated gather sequences.
void optimizeGatherSequence();
/// \returns true if it is beneficial to reverse the vector order.
bool shouldReorder() const {
return NumLoadsWantToChangeOrder > NumLoadsWantToKeepOrder;
}
/// \return The vector element size in bits to use when vectorizing the
/// expression tree ending at \p V. If V is a store, the size is the width of
/// the stored value. Otherwise, the size is the width of the largest loaded
/// value reaching V. This method is used by the vectorizer to calculate
/// vectorization factors.
unsigned getVectorElementSize(Value *V);
/// Compute the minimum type sizes required to represent the entries in a
/// vectorizable tree.
void computeMinimumValueSizes();
private:
struct TreeEntry;
/// \returns the cost of the vectorizable entry.
int getEntryCost(TreeEntry *E);
/// This is the recursive part of buildTree.
void buildTree_rec(ArrayRef<Value *> Roots, unsigned Depth);
/// Vectorize a single entry in the tree.
Value *vectorizeTree(TreeEntry *E);
/// Vectorize a single entry in the tree, starting in \p VL.
Value *vectorizeTree(ArrayRef<Value *> VL);
/// \returns the pointer to the vectorized value if \p VL is already
/// vectorized, or NULL. They may happen in cycles.
Value *alreadyVectorized(ArrayRef<Value *> VL) const;
/// \returns the scalarization cost for this type. Scalarization in this
/// context means the creation of vectors from a group of scalars.
int getGatherCost(Type *Ty);
/// \returns the scalarization cost for this list of values. Assuming that
/// this subtree gets vectorized, we may need to extract the values from the
/// roots. This method calculates the cost of extracting the values.
int getGatherCost(ArrayRef<Value *> VL);
/// \brief Set the Builder insert point to one after the last instruction in
/// the bundle
void setInsertPointAfterBundle(ArrayRef<Value *> VL);
/// \returns a vector from a collection of scalars in \p VL.
Value *Gather(ArrayRef<Value *> VL, VectorType *Ty);
/// \returns whether the VectorizableTree is fully vectorizable and will
/// be beneficial even the tree height is tiny.
bool isFullyVectorizableTinyTree();
/// \reorder commutative operands in alt shuffle if they result in
/// vectorized code.
void reorderAltShuffleOperands(ArrayRef<Value *> VL,
SmallVectorImpl<Value *> &Left,
SmallVectorImpl<Value *> &Right);
/// \reorder commutative operands to get better probability of
/// generating vectorized code.
void reorderInputsAccordingToOpcode(ArrayRef<Value *> VL,
SmallVectorImpl<Value *> &Left,
SmallVectorImpl<Value *> &Right);
struct TreeEntry {
TreeEntry() : Scalars(), VectorizedValue(nullptr),
NeedToGather(0) {}
/// \returns true if the scalars in VL are equal to this entry.
bool isSame(ArrayRef<Value *> VL) const {
assert(VL.size() == Scalars.size() && "Invalid size");
return std::equal(VL.begin(), VL.end(), Scalars.begin());
}
/// A vector of scalars.
ValueList Scalars;
/// The Scalars are vectorized into this value. It is initialized to Null.
Value *VectorizedValue;
/// Do we need to gather this sequence ?
bool NeedToGather;
};
/// Create a new VectorizableTree entry.
TreeEntry *newTreeEntry(ArrayRef<Value *> VL, bool Vectorized) {
VectorizableTree.emplace_back();
int idx = VectorizableTree.size() - 1;
TreeEntry *Last = &VectorizableTree[idx];
Last->Scalars.insert(Last->Scalars.begin(), VL.begin(), VL.end());
Last->NeedToGather = !Vectorized;
if (Vectorized) {
for (int i = 0, e = VL.size(); i != e; ++i) {
assert(!ScalarToTreeEntry.count(VL[i]) && "Scalar already in tree!");
ScalarToTreeEntry[VL[i]] = idx;
}
} else {
MustGather.insert(VL.begin(), VL.end());
}
return Last;
}
/// -- Vectorization State --
/// Holds all of the tree entries.
std::vector<TreeEntry> VectorizableTree;
/// Maps a specific scalar to its tree entry.
SmallDenseMap<Value*, int> ScalarToTreeEntry;
/// A list of scalars that we found that we need to keep as scalars.
ValueSet MustGather;
/// This POD struct describes one external user in the vectorized tree.
struct ExternalUser {
ExternalUser (Value *S, llvm::User *U, int L) :
Scalar(S), User(U), Lane(L){}
// Which scalar in our function.
Value *Scalar;
// Which user that uses the scalar.
llvm::User *User;
// Which lane does the scalar belong to.
int Lane;
};
typedef SmallVector<ExternalUser, 16> UserList;
/// Checks if two instructions may access the same memory.
///
/// \p Loc1 is the location of \p Inst1. It is passed explicitly because it
/// is invariant in the calling loop.
bool isAliased(const MemoryLocation &Loc1, Instruction *Inst1,
Instruction *Inst2) {
// First check if the result is already in the cache.
AliasCacheKey key = std::make_pair(Inst1, Inst2);
Optional<bool> &result = AliasCache[key];
if (result.hasValue()) {
return result.getValue();
}
MemoryLocation Loc2 = getLocation(Inst2, AA);
bool aliased = true;
if (Loc1.Ptr && Loc2.Ptr && isSimple(Inst1) && isSimple(Inst2)) {
// Do the alias check.
aliased = AA->alias(Loc1, Loc2);
}
// Store the result in the cache.
result = aliased;
return aliased;
}
typedef std::pair<Instruction *, Instruction *> AliasCacheKey;
/// Cache for alias results.
/// TODO: consider moving this to the AliasAnalysis itself.
DenseMap<AliasCacheKey, Optional<bool>> AliasCache;
/// Removes an instruction from its block and eventually deletes it.
/// It's like Instruction::eraseFromParent() except that the actual deletion
/// is delayed until BoUpSLP is destructed.
/// This is required to ensure that there are no incorrect collisions in the
/// AliasCache, which can happen if a new instruction is allocated at the
/// same address as a previously deleted instruction.
void eraseInstruction(Instruction *I) {
I->removeFromParent();
I->dropAllReferences();
DeletedInstructions.push_back(std::unique_ptr<Instruction>(I));
}
/// Temporary store for deleted instructions. Instructions will be deleted
/// eventually when the BoUpSLP is destructed.
SmallVector<std::unique_ptr<Instruction>, 8> DeletedInstructions;
/// A list of values that need to extracted out of the tree.
/// This list holds pairs of (Internal Scalar : External User).
UserList ExternalUses;
/// Values used only by @llvm.assume calls.
SmallPtrSet<const Value *, 32> EphValues;
/// Holds all of the instructions that we gathered.
SetVector<Instruction *> GatherSeq;
/// A list of blocks that we are going to CSE.
SetVector<BasicBlock *> CSEBlocks;
/// Contains all scheduling relevant data for an instruction.
/// A ScheduleData either represents a single instruction or a member of an
/// instruction bundle (= a group of instructions which is combined into a
/// vector instruction).
struct ScheduleData {
// The initial value for the dependency counters. It means that the
// dependencies are not calculated yet.
enum { InvalidDeps = -1 };
ScheduleData()
: Inst(nullptr), FirstInBundle(nullptr), NextInBundle(nullptr),
NextLoadStore(nullptr), SchedulingRegionID(0), SchedulingPriority(0),
Dependencies(InvalidDeps), UnscheduledDeps(InvalidDeps),
UnscheduledDepsInBundle(InvalidDeps), IsScheduled(false) {}
void init(int BlockSchedulingRegionID) {
FirstInBundle = this;
NextInBundle = nullptr;
NextLoadStore = nullptr;
IsScheduled = false;
SchedulingRegionID = BlockSchedulingRegionID;
UnscheduledDepsInBundle = UnscheduledDeps;
clearDependencies();
}
/// Returns true if the dependency information has been calculated.
bool hasValidDependencies() const { return Dependencies != InvalidDeps; }
/// Returns true for single instructions and for bundle representatives
/// (= the head of a bundle).
bool isSchedulingEntity() const { return FirstInBundle == this; }
/// Returns true if it represents an instruction bundle and not only a
/// single instruction.
bool isPartOfBundle() const {
return NextInBundle != nullptr || FirstInBundle != this;
}
/// Returns true if it is ready for scheduling, i.e. it has no more
/// unscheduled depending instructions/bundles.
bool isReady() const {
assert(isSchedulingEntity() &&
"can't consider non-scheduling entity for ready list");
return UnscheduledDepsInBundle == 0 && !IsScheduled;
}
/// Modifies the number of unscheduled dependencies, also updating it for
/// the whole bundle.
int incrementUnscheduledDeps(int Incr) {
UnscheduledDeps += Incr;
return FirstInBundle->UnscheduledDepsInBundle += Incr;
}
/// Sets the number of unscheduled dependencies to the number of
/// dependencies.
void resetUnscheduledDeps() {
incrementUnscheduledDeps(Dependencies - UnscheduledDeps);
}
/// Clears all dependency information.
void clearDependencies() {
Dependencies = InvalidDeps;
resetUnscheduledDeps();
MemoryDependencies.clear();
}
void dump(raw_ostream &os) const {
if (!isSchedulingEntity()) {
os << "/ " << *Inst;
} else if (NextInBundle) {
os << '[' << *Inst;
ScheduleData *SD = NextInBundle;
while (SD) {
os << ';' << *SD->Inst;
SD = SD->NextInBundle;
}
os << ']';
} else {
os << *Inst;
}
}
Instruction *Inst;
/// Points to the head in an instruction bundle (and always to this for
/// single instructions).
ScheduleData *FirstInBundle;
/// Single linked list of all instructions in a bundle. Null if it is a
/// single instruction.
ScheduleData *NextInBundle;
/// Single linked list of all memory instructions (e.g. load, store, call)
/// in the block - until the end of the scheduling region.
ScheduleData *NextLoadStore;
/// The dependent memory instructions.
/// This list is derived on demand in calculateDependencies().
SmallVector<ScheduleData *, 4> MemoryDependencies;
/// This ScheduleData is in the current scheduling region if this matches
/// the current SchedulingRegionID of BlockScheduling.
int SchedulingRegionID;
/// Used for getting a "good" final ordering of instructions.
int SchedulingPriority;
/// The number of dependencies. Constitutes of the number of users of the
/// instruction plus the number of dependent memory instructions (if any).
/// This value is calculated on demand.
/// If InvalidDeps, the number of dependencies is not calculated yet.
///
int Dependencies;
/// The number of dependencies minus the number of dependencies of scheduled
/// instructions. As soon as this is zero, the instruction/bundle gets ready
/// for scheduling.
/// Note that this is negative as long as Dependencies is not calculated.
int UnscheduledDeps;
/// The sum of UnscheduledDeps in a bundle. Equals to UnscheduledDeps for
/// single instructions.
int UnscheduledDepsInBundle;
/// True if this instruction is scheduled (or considered as scheduled in the
/// dry-run).
bool IsScheduled;
};
#ifndef NDEBUG
friend raw_ostream &operator<<(raw_ostream &os,
const BoUpSLP::ScheduleData &SD);
#endif
/// Contains all scheduling data for a basic block.
///
struct BlockScheduling {
BlockScheduling(BasicBlock *BB)
: BB(BB), ChunkSize(BB->size()), ChunkPos(ChunkSize),
ScheduleStart(nullptr), ScheduleEnd(nullptr),
FirstLoadStoreInRegion(nullptr), LastLoadStoreInRegion(nullptr),
ScheduleRegionSize(0),
ScheduleRegionSizeLimit(ScheduleRegionSizeBudget),
// Make sure that the initial SchedulingRegionID is greater than the
// initial SchedulingRegionID in ScheduleData (which is 0).
SchedulingRegionID(1) {}
void clear() {
ReadyInsts.clear();
ScheduleStart = nullptr;
ScheduleEnd = nullptr;
FirstLoadStoreInRegion = nullptr;
LastLoadStoreInRegion = nullptr;
// Reduce the maximum schedule region size by the size of the
// previous scheduling run.
ScheduleRegionSizeLimit -= ScheduleRegionSize;
if (ScheduleRegionSizeLimit < MinScheduleRegionSize)
ScheduleRegionSizeLimit = MinScheduleRegionSize;
ScheduleRegionSize = 0;
// Make a new scheduling region, i.e. all existing ScheduleData is not
// in the new region yet.
++SchedulingRegionID;
}
ScheduleData *getScheduleData(Value *V) {
ScheduleData *SD = ScheduleDataMap[V];
if (SD && SD->SchedulingRegionID == SchedulingRegionID)
return SD;
return nullptr;
}
bool isInSchedulingRegion(ScheduleData *SD) {
return SD->SchedulingRegionID == SchedulingRegionID;
}
/// Marks an instruction as scheduled and puts all dependent ready
/// instructions into the ready-list.
template <typename ReadyListType>
void schedule(ScheduleData *SD, ReadyListType &ReadyList) {
SD->IsScheduled = true;
DEBUG(dbgs() << "SLP: schedule " << *SD << "\n");
ScheduleData *BundleMember = SD;
while (BundleMember) {
// Handle the def-use chain dependencies.
for (Use &U : BundleMember->Inst->operands()) {
ScheduleData *OpDef = getScheduleData(U.get());
if (OpDef && OpDef->hasValidDependencies() &&
OpDef->incrementUnscheduledDeps(-1) == 0) {
// There are no more unscheduled dependencies after decrementing,
// so we can put the dependent instruction into the ready list.
ScheduleData *DepBundle = OpDef->FirstInBundle;
assert(!DepBundle->IsScheduled &&
"already scheduled bundle gets ready");
ReadyList.insert(DepBundle);
DEBUG(dbgs() << "SLP: gets ready (def): " << *DepBundle << "\n");
}
}
// Handle the memory dependencies.
for (ScheduleData *MemoryDepSD : BundleMember->MemoryDependencies) {
if (MemoryDepSD->incrementUnscheduledDeps(-1) == 0) {
// There are no more unscheduled dependencies after decrementing,
// so we can put the dependent instruction into the ready list.
ScheduleData *DepBundle = MemoryDepSD->FirstInBundle;
assert(!DepBundle->IsScheduled &&
"already scheduled bundle gets ready");
ReadyList.insert(DepBundle);
DEBUG(dbgs() << "SLP: gets ready (mem): " << *DepBundle << "\n");
}
}
BundleMember = BundleMember->NextInBundle;
}
}
/// Put all instructions into the ReadyList which are ready for scheduling.
template <typename ReadyListType>
void initialFillReadyList(ReadyListType &ReadyList) {
for (auto *I = ScheduleStart; I != ScheduleEnd; I = I->getNextNode()) {
ScheduleData *SD = getScheduleData(I);
if (SD->isSchedulingEntity() && SD->isReady()) {
ReadyList.insert(SD);
DEBUG(dbgs() << "SLP: initially in ready list: " << *I << "\n");
}
}
}
/// Checks if a bundle of instructions can be scheduled, i.e. has no
/// cyclic dependencies. This is only a dry-run, no instructions are
/// actually moved at this stage.
bool tryScheduleBundle(ArrayRef<Value *> VL, BoUpSLP *SLP);
/// Un-bundles a group of instructions.
void cancelScheduling(ArrayRef<Value *> VL);
/// Extends the scheduling region so that V is inside the region.
/// \returns true if the region size is within the limit.
bool extendSchedulingRegion(Value *V);
/// Initialize the ScheduleData structures for new instructions in the
/// scheduling region.
void initScheduleData(Instruction *FromI, Instruction *ToI,
ScheduleData *PrevLoadStore,
ScheduleData *NextLoadStore);
/// Updates the dependency information of a bundle and of all instructions/
/// bundles which depend on the original bundle.
void calculateDependencies(ScheduleData *SD, bool InsertInReadyList,
BoUpSLP *SLP);
/// Sets all instruction in the scheduling region to un-scheduled.
void resetSchedule();
BasicBlock *BB;
/// Simple memory allocation for ScheduleData.
std::vector<std::unique_ptr<ScheduleData[]>> ScheduleDataChunks;
/// The size of a ScheduleData array in ScheduleDataChunks.
int ChunkSize;
/// The allocator position in the current chunk, which is the last entry
/// of ScheduleDataChunks.
int ChunkPos;
/// Attaches ScheduleData to Instruction.
/// Note that the mapping survives during all vectorization iterations, i.e.
/// ScheduleData structures are recycled.
DenseMap<Value *, ScheduleData *> ScheduleDataMap;
struct ReadyList : SmallVector<ScheduleData *, 8> {
void insert(ScheduleData *SD) { push_back(SD); }
};
/// The ready-list for scheduling (only used for the dry-run).
ReadyList ReadyInsts;
/// The first instruction of the scheduling region.
Instruction *ScheduleStart;
/// The first instruction _after_ the scheduling region.
Instruction *ScheduleEnd;
/// The first memory accessing instruction in the scheduling region
/// (can be null).
ScheduleData *FirstLoadStoreInRegion;
/// The last memory accessing instruction in the scheduling region
/// (can be null).
ScheduleData *LastLoadStoreInRegion;
/// The current size of the scheduling region.
int ScheduleRegionSize;
/// The maximum size allowed for the scheduling region.
int ScheduleRegionSizeLimit;
/// The ID of the scheduling region. For a new vectorization iteration this
/// is incremented which "removes" all ScheduleData from the region.
int SchedulingRegionID;
};
/// Attaches the BlockScheduling structures to basic blocks.
MapVector<BasicBlock *, std::unique_ptr<BlockScheduling>> BlocksSchedules;
/// Performs the "real" scheduling. Done before vectorization is actually
/// performed in a basic block.
void scheduleBlock(BlockScheduling *BS);
/// List of users to ignore during scheduling and that don't need extracting.
ArrayRef<Value *> UserIgnoreList;
// Number of load-bundles, which contain consecutive loads.
int NumLoadsWantToKeepOrder;
// Number of load-bundles of size 2, which are consecutive loads if reversed.
int NumLoadsWantToChangeOrder;
// Analysis and block reference.
Function *F;
ScalarEvolution *SE;
TargetTransformInfo *TTI;
TargetLibraryInfo *TLI;
AliasAnalysis *AA;
LoopInfo *LI;
DominatorTree *DT;
AssumptionCache *AC;
DemandedBits *DB;
/// Instruction builder to construct the vectorized tree.
IRBuilder<> Builder;
/// A map of scalar integer values to the smallest bit width with which they
/// can legally be represented.
MapVector<Value *, uint64_t> MinBWs;
};
#ifndef NDEBUG
raw_ostream &operator<<(raw_ostream &os, const BoUpSLP::ScheduleData &SD) {
SD.dump(os);
return os;
}
#endif
void BoUpSLP::buildTree(ArrayRef<Value *> Roots,
ArrayRef<Value *> UserIgnoreLst) {
deleteTree();
UserIgnoreList = UserIgnoreLst;
if (!getSameType(Roots))
return;
buildTree_rec(Roots, 0);
// Collect the values that we need to extract from the tree.
for (int EIdx = 0, EE = VectorizableTree.size(); EIdx < EE; ++EIdx) {
TreeEntry *Entry = &VectorizableTree[EIdx];
// For each lane:
for (int Lane = 0, LE = Entry->Scalars.size(); Lane != LE; ++Lane) {
Value *Scalar = Entry->Scalars[Lane];
// No need to handle users of gathered values.
if (Entry->NeedToGather)
continue;
for (User *U : Scalar->users()) {
DEBUG(dbgs() << "SLP: Checking user:" << *U << ".\n");
Instruction *UserInst = dyn_cast<Instruction>(U);
if (!UserInst)
continue;
// Skip in-tree scalars that become vectors
if (ScalarToTreeEntry.count(U)) {
int Idx = ScalarToTreeEntry[U];
TreeEntry *UseEntry = &VectorizableTree[Idx];
Value *UseScalar = UseEntry->Scalars[0];
// Some in-tree scalars will remain as scalar in vectorized
// instructions. If that is the case, the one in Lane 0 will
// be used.
if (UseScalar != U ||
!InTreeUserNeedToExtract(Scalar, UserInst, TLI)) {
DEBUG(dbgs() << "SLP: \tInternal user will be removed:" << *U
<< ".\n");
assert(!VectorizableTree[Idx].NeedToGather && "Bad state");
continue;
}
}
// Ignore users in the user ignore list.
if (std::find(UserIgnoreList.begin(), UserIgnoreList.end(), UserInst) !=
UserIgnoreList.end())
continue;
DEBUG(dbgs() << "SLP: Need to extract:" << *U << " from lane " <<
Lane << " from " << *Scalar << ".\n");
ExternalUses.push_back(ExternalUser(Scalar, U, Lane));
}
}
}
}
void BoUpSLP::buildTree_rec(ArrayRef<Value *> VL, unsigned Depth) {
bool SameTy = getSameType(VL); (void)SameTy;
bool isAltShuffle = false;
assert(SameTy && "Invalid types!");
if (Depth == RecursionMaxDepth) {
DEBUG(dbgs() << "SLP: Gathering due to max recursion depth.\n");
newTreeEntry(VL, false);
return;
}
// Don't handle vectors.
if (VL[0]->getType()->isVectorTy()) {
DEBUG(dbgs() << "SLP: Gathering due to vector type.\n");
newTreeEntry(VL, false);
return;
}
if (StoreInst *SI = dyn_cast<StoreInst>(VL[0]))
if (SI->getValueOperand()->getType()->isVectorTy()) {
DEBUG(dbgs() << "SLP: Gathering due to store vector type.\n");
newTreeEntry(VL, false);
return;
}
unsigned Opcode = getSameOpcode(VL);
// Check that this shuffle vector refers to the alternate
// sequence of opcodes.
if (Opcode == Instruction::ShuffleVector) {
Instruction *I0 = dyn_cast<Instruction>(VL[0]);
unsigned Op = I0->getOpcode();
if (Op != Instruction::ShuffleVector)
isAltShuffle = true;
}
// If all of the operands are identical or constant we have a simple solution.
if (allConstant(VL) || isSplat(VL) || !getSameBlock(VL) || !Opcode) {
DEBUG(dbgs() << "SLP: Gathering due to C,S,B,O. \n");
newTreeEntry(VL, false);
return;
}
// We now know that this is a vector of instructions of the same type from
// the same block.
// Don't vectorize ephemeral values.
for (unsigned i = 0, e = VL.size(); i != e; ++i) {
if (EphValues.count(VL[i])) {
DEBUG(dbgs() << "SLP: The instruction (" << *VL[i] <<
") is ephemeral.\n");
newTreeEntry(VL, false);
return;
}
}
// Check if this is a duplicate of another entry.
if (ScalarToTreeEntry.count(VL[0])) {
int Idx = ScalarToTreeEntry[VL[0]];
TreeEntry *E = &VectorizableTree[Idx];
for (unsigned i = 0, e = VL.size(); i != e; ++i) {
DEBUG(dbgs() << "SLP: \tChecking bundle: " << *VL[i] << ".\n");
if (E->Scalars[i] != VL[i]) {
DEBUG(dbgs() << "SLP: Gathering due to partial overlap.\n");
newTreeEntry(VL, false);
return;
}
}
DEBUG(dbgs() << "SLP: Perfect diamond merge at " << *VL[0] << ".\n");
return;
}
// Check that none of the instructions in the bundle are already in the tree.
for (unsigned i = 0, e = VL.size(); i != e; ++i) {
if (ScalarToTreeEntry.count(VL[i])) {
DEBUG(dbgs() << "SLP: The instruction (" << *VL[i] <<
") is already in tree.\n");
newTreeEntry(VL, false);
return;
}
}
// If any of the scalars is marked as a value that needs to stay scalar then
// we need to gather the scalars.
for (unsigned i = 0, e = VL.size(); i != e; ++i) {
if (MustGather.count(VL[i])) {
DEBUG(dbgs() << "SLP: Gathering due to gathered scalar.\n");
newTreeEntry(VL, false);
return;
}
}
// Check that all of the users of the scalars that we want to vectorize are
// schedulable.
Instruction *VL0 = cast<Instruction>(VL[0]);
BasicBlock *BB = cast<Instruction>(VL0)->getParent();
if (!DT->isReachableFromEntry(BB)) {
// Don't go into unreachable blocks. They may contain instructions with
// dependency cycles which confuse the final scheduling.
DEBUG(dbgs() << "SLP: bundle in unreachable block.\n");
newTreeEntry(VL, false);
return;
}
// Check that every instructions appears once in this bundle.
for (unsigned i = 0, e = VL.size(); i < e; ++i)
for (unsigned j = i+1; j < e; ++j)
if (VL[i] == VL[j]) {
DEBUG(dbgs() << "SLP: Scalar used twice in bundle.\n");
newTreeEntry(VL, false);
return;
}
auto &BSRef = BlocksSchedules[BB];
if (!BSRef) {
BSRef = llvm::make_unique<BlockScheduling>(BB);
}
BlockScheduling &BS = *BSRef.get();
if (!BS.tryScheduleBundle(VL, this)) {
DEBUG(dbgs() << "SLP: We are not able to schedule this bundle!\n");
assert((!BS.getScheduleData(VL[0]) ||
!BS.getScheduleData(VL[0])->isPartOfBundle()) &&
"tryScheduleBundle should cancelScheduling on failure");
newTreeEntry(VL, false);
return;
}
DEBUG(dbgs() << "SLP: We are able to schedule this bundle.\n");
switch (Opcode) {
case Instruction::PHI: {
PHINode *PH = dyn_cast<PHINode>(VL0);
// Check for terminator values (e.g. invoke).
for (unsigned j = 0; j < VL.size(); ++j)
for (unsigned i = 0, e = PH->getNumIncomingValues(); i < e; ++i) {
TerminatorInst *Term = dyn_cast<TerminatorInst>(
cast<PHINode>(VL[j])->getIncomingValueForBlock(PH->getIncomingBlock(i)));
if (Term) {
DEBUG(dbgs() << "SLP: Need to swizzle PHINodes (TerminatorInst use).\n");
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
return;
}
}
newTreeEntry(VL, true);
DEBUG(dbgs() << "SLP: added a vector of PHINodes.\n");
for (unsigned i = 0, e = PH->getNumIncomingValues(); i < e; ++i) {
ValueList Operands;
// Prepare the operand vector.
for (unsigned j = 0; j < VL.size(); ++j)
Operands.push_back(cast<PHINode>(VL[j])->getIncomingValueForBlock(
PH->getIncomingBlock(i)));
buildTree_rec(Operands, Depth + 1);
}
return;
}
case Instruction::ExtractElement: {
bool Reuse = CanReuseExtract(VL);
if (Reuse) {
DEBUG(dbgs() << "SLP: Reusing extract sequence.\n");
} else {
BS.cancelScheduling(VL);
}
newTreeEntry(VL, Reuse);
return;
}
case Instruction::Load: {
// Check that a vectorized load would load the same memory as a scalar
// load.
// For example we don't want vectorize loads that are smaller than 8 bit.
// Even though we have a packed struct {<i2, i2, i2, i2>} LLVM treats
// loading/storing it as an i8 struct. If we vectorize loads/stores from
// such a struct we read/write packed bits disagreeing with the
// unvectorized version.
const DataLayout &DL = F->getParent()->getDataLayout();
Type *ScalarTy = VL[0]->getType();
if (DL.getTypeSizeInBits(ScalarTy) !=
DL.getTypeAllocSizeInBits(ScalarTy)) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: Gathering loads of non-packed type.\n");
return;
}
// Check if the loads are consecutive or of we need to swizzle them.
for (unsigned i = 0, e = VL.size() - 1; i < e; ++i) {
LoadInst *L = cast<LoadInst>(VL[i]);
if (!L->isSimple()) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: Gathering non-simple loads.\n");
return;
}
if (!isConsecutiveAccess(VL[i], VL[i + 1], DL, *SE)) {
if (VL.size() == 2 && isConsecutiveAccess(VL[1], VL[0], DL, *SE)) {
++NumLoadsWantToChangeOrder;
}
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: Gathering non-consecutive loads.\n");
return;
}
}
++NumLoadsWantToKeepOrder;
newTreeEntry(VL, true);
DEBUG(dbgs() << "SLP: added a vector of loads.\n");
return;
}
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: {
Type *SrcTy = VL0->getOperand(0)->getType();
for (unsigned i = 0; i < VL.size(); ++i) {
Type *Ty = cast<Instruction>(VL[i])->getOperand(0)->getType();
if (Ty != SrcTy || !isValidElementType(Ty)) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: Gathering casts with different src types.\n");
return;
}
}
newTreeEntry(VL, true);
DEBUG(dbgs() << "SLP: added a vector of casts.\n");
for (unsigned i = 0, e = VL0->getNumOperands(); i < e; ++i) {
ValueList Operands;
// Prepare the operand vector.
for (unsigned j = 0; j < VL.size(); ++j)
Operands.push_back(cast<Instruction>(VL[j])->getOperand(i));
buildTree_rec(Operands, Depth+1);
}
return;
}
case Instruction::ICmp:
case Instruction::FCmp: {
// Check that all of the compares have the same predicate.
CmpInst::Predicate P0 = cast<CmpInst>(VL0)->getPredicate();
Type *ComparedTy = cast<Instruction>(VL[0])->getOperand(0)->getType();
for (unsigned i = 1, e = VL.size(); i < e; ++i) {
CmpInst *Cmp = cast<CmpInst>(VL[i]);
if (Cmp->getPredicate() != P0 ||
Cmp->getOperand(0)->getType() != ComparedTy) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: Gathering cmp with different predicate.\n");
return;
}
}
newTreeEntry(VL, true);
DEBUG(dbgs() << "SLP: added a vector of compares.\n");
for (unsigned i = 0, e = VL0->getNumOperands(); i < e; ++i) {
ValueList Operands;
// Prepare the operand vector.
for (unsigned j = 0; j < VL.size(); ++j)
Operands.push_back(cast<Instruction>(VL[j])->getOperand(i));
buildTree_rec(Operands, Depth+1);
}
return;
}
case Instruction::Select:
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: {
newTreeEntry(VL, true);
DEBUG(dbgs() << "SLP: added a vector of bin op.\n");
// Sort operands of the instructions so that each side is more likely to
// have the same opcode.
if (isa<BinaryOperator>(VL0) && VL0->isCommutative()) {
ValueList Left, Right;
reorderInputsAccordingToOpcode(VL, Left, Right);
buildTree_rec(Left, Depth + 1);
buildTree_rec(Right, Depth + 1);
return;
}
for (unsigned i = 0, e = VL0->getNumOperands(); i < e; ++i) {
ValueList Operands;
// Prepare the operand vector.
for (unsigned j = 0; j < VL.size(); ++j)
Operands.push_back(cast<Instruction>(VL[j])->getOperand(i));
buildTree_rec(Operands, Depth+1);
}
return;
}
case Instruction::GetElementPtr: {
// We don't combine GEPs with complicated (nested) indexing.
for (unsigned j = 0; j < VL.size(); ++j) {
if (cast<Instruction>(VL[j])->getNumOperands() != 2) {
DEBUG(dbgs() << "SLP: not-vectorizable GEP (nested indexes).\n");
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
return;
}
}
// We can't combine several GEPs into one vector if they operate on
// different types.
Type *Ty0 = cast<Instruction>(VL0)->getOperand(0)->getType();
for (unsigned j = 0; j < VL.size(); ++j) {
Type *CurTy = cast<Instruction>(VL[j])->getOperand(0)->getType();
if (Ty0 != CurTy) {
DEBUG(dbgs() << "SLP: not-vectorizable GEP (different types).\n");
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
return;
}
}
// We don't combine GEPs with non-constant indexes.
for (unsigned j = 0; j < VL.size(); ++j) {
auto Op = cast<Instruction>(VL[j])->getOperand(1);
if (!isa<ConstantInt>(Op)) {
DEBUG(
dbgs() << "SLP: not-vectorizable GEP (non-constant indexes).\n");
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
return;
}
}
newTreeEntry(VL, true);
DEBUG(dbgs() << "SLP: added a vector of GEPs.\n");
for (unsigned i = 0, e = 2; i < e; ++i) {
ValueList Operands;
// Prepare the operand vector.
for (unsigned j = 0; j < VL.size(); ++j)
Operands.push_back(cast<Instruction>(VL[j])->getOperand(i));
buildTree_rec(Operands, Depth + 1);
}
return;
}
case Instruction::Store: {
const DataLayout &DL = F->getParent()->getDataLayout();
// Check if the stores are consecutive or of we need to swizzle them.
for (unsigned i = 0, e = VL.size() - 1; i < e; ++i)
if (!isConsecutiveAccess(VL[i], VL[i + 1], DL, *SE)) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: Non-consecutive store.\n");
return;
}
newTreeEntry(VL, true);
DEBUG(dbgs() << "SLP: added a vector of stores.\n");
ValueList Operands;
for (unsigned j = 0; j < VL.size(); ++j)
Operands.push_back(cast<Instruction>(VL[j])->getOperand(0));
buildTree_rec(Operands, Depth + 1);
return;
}
case Instruction::Call: {
// Check if the calls are all to the same vectorizable intrinsic.
CallInst *CI = cast<CallInst>(VL[0]);
// Check if this is an Intrinsic call or something that can be
// represented by an intrinsic call
Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI);
if (!isTriviallyVectorizable(ID)) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: Non-vectorizable call.\n");
return;
}
Function *Int = CI->getCalledFunction();
Value *A1I = nullptr;
if (hasVectorInstrinsicScalarOpd(ID, 1))
A1I = CI->getArgOperand(1);
for (unsigned i = 1, e = VL.size(); i != e; ++i) {
CallInst *CI2 = dyn_cast<CallInst>(VL[i]);
if (!CI2 || CI2->getCalledFunction() != Int ||
getIntrinsicIDForCall(CI2, TLI) != ID) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: mismatched calls:" << *CI << "!=" << *VL[i]
<< "\n");
return;
}
// ctlz,cttz and powi are special intrinsics whose second argument
// should be same in order for them to be vectorized.
if (hasVectorInstrinsicScalarOpd(ID, 1)) {
Value *A1J = CI2->getArgOperand(1);
if (A1I != A1J) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: mismatched arguments in call:" << *CI
<< " argument "<< A1I<<"!=" << A1J
<< "\n");
return;
}
}
}
newTreeEntry(VL, true);
for (unsigned i = 0, e = CI->getNumArgOperands(); i != e; ++i) {
ValueList Operands;
// Prepare the operand vector.
for (unsigned j = 0; j < VL.size(); ++j) {
CallInst *CI2 = dyn_cast<CallInst>(VL[j]);
Operands.push_back(CI2->getArgOperand(i));
}
buildTree_rec(Operands, Depth + 1);
}
return;
}
case Instruction::ShuffleVector: {
// If this is not an alternate sequence of opcode like add-sub
// then do not vectorize this instruction.
if (!isAltShuffle) {
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: ShuffleVector are not vectorized.\n");
return;
}
newTreeEntry(VL, true);
DEBUG(dbgs() << "SLP: added a ShuffleVector op.\n");
// Reorder operands if reordering would enable vectorization.
if (isa<BinaryOperator>(VL0)) {
ValueList Left, Right;
reorderAltShuffleOperands(VL, Left, Right);
buildTree_rec(Left, Depth + 1);
buildTree_rec(Right, Depth + 1);
return;
}
for (unsigned i = 0, e = VL0->getNumOperands(); i < e; ++i) {
ValueList Operands;
// Prepare the operand vector.
for (unsigned j = 0; j < VL.size(); ++j)
Operands.push_back(cast<Instruction>(VL[j])->getOperand(i));
buildTree_rec(Operands, Depth + 1);
}
return;
}
default:
BS.cancelScheduling(VL);
newTreeEntry(VL, false);
DEBUG(dbgs() << "SLP: Gathering unknown instruction.\n");
return;
}
}
int BoUpSLP::getEntryCost(TreeEntry *E) {
ArrayRef<Value*> VL = E->Scalars;
Type *ScalarTy = VL[0]->getType();
if (StoreInst *SI = dyn_cast<StoreInst>(VL[0]))
ScalarTy = SI->getValueOperand()->getType();
VectorType *VecTy = VectorType::get(ScalarTy, VL.size());
// If we have computed a smaller type for the expression, update VecTy so
// that the costs will be accurate.
if (MinBWs.count(VL[0]))
VecTy = VectorType::get(IntegerType::get(F->getContext(), MinBWs[VL[0]]),
VL.size());
if (E->NeedToGather) {
if (allConstant(VL))
return 0;
if (isSplat(VL)) {
return TTI->getShuffleCost(TargetTransformInfo::SK_Broadcast, VecTy, 0);
}
return getGatherCost(E->Scalars);
}
unsigned Opcode = getSameOpcode(VL);
assert(Opcode && getSameType(VL) && getSameBlock(VL) && "Invalid VL");
Instruction *VL0 = cast<Instruction>(VL[0]);
switch (Opcode) {
case Instruction::PHI: {
return 0;
}
case Instruction::ExtractElement: {
if (CanReuseExtract(VL)) {
int DeadCost = 0;
for (unsigned i = 0, e = VL.size(); i < e; ++i) {
ExtractElementInst *E = cast<ExtractElementInst>(VL[i]);
if (E->hasOneUse())
// Take credit for instruction that will become dead.
DeadCost +=
TTI->getVectorInstrCost(Instruction::ExtractElement, VecTy, i);
}
return -DeadCost;
}
return getGatherCost(VecTy);
}
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: {
Type *SrcTy = VL0->getOperand(0)->getType();
// Calculate the cost of this instruction.
int ScalarCost = VL.size() * TTI->getCastInstrCost(VL0->getOpcode(),
VL0->getType(), SrcTy);
VectorType *SrcVecTy = VectorType::get(SrcTy, VL.size());
int VecCost = TTI->getCastInstrCost(VL0->getOpcode(), VecTy, SrcVecTy);
return VecCost - ScalarCost;
}
case Instruction::FCmp:
case Instruction::ICmp:
case Instruction::Select:
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: {
// Calculate the cost of this instruction.
int ScalarCost = 0;
int VecCost = 0;
if (Opcode == Instruction::FCmp || Opcode == Instruction::ICmp ||
Opcode == Instruction::Select) {
VectorType *MaskTy = VectorType::get(Builder.getInt1Ty(), VL.size());
ScalarCost = VecTy->getNumElements() *
TTI->getCmpSelInstrCost(Opcode, ScalarTy, Builder.getInt1Ty());
VecCost = TTI->getCmpSelInstrCost(Opcode, VecTy, MaskTy);
} else {
// Certain instructions can be cheaper to vectorize if they have a
// constant second vector operand.
TargetTransformInfo::OperandValueKind Op1VK =
TargetTransformInfo::OK_AnyValue;
TargetTransformInfo::OperandValueKind Op2VK =
TargetTransformInfo::OK_UniformConstantValue;
TargetTransformInfo::OperandValueProperties Op1VP =
TargetTransformInfo::OP_None;
TargetTransformInfo::OperandValueProperties Op2VP =
TargetTransformInfo::OP_None;
// If all operands are exactly the same ConstantInt then set the
// operand kind to OK_UniformConstantValue.
// If instead not all operands are constants, then set the operand kind
// to OK_AnyValue. If all operands are constants but not the same,
// then set the operand kind to OK_NonUniformConstantValue.
ConstantInt *CInt = nullptr;
for (unsigned i = 0; i < VL.size(); ++i) {
const Instruction *I = cast<Instruction>(VL[i]);
if (!isa<ConstantInt>(I->getOperand(1))) {
Op2VK = TargetTransformInfo::OK_AnyValue;
break;
}
if (i == 0) {
CInt = cast<ConstantInt>(I->getOperand(1));
continue;
}
if (Op2VK == TargetTransformInfo::OK_UniformConstantValue &&
CInt != cast<ConstantInt>(I->getOperand(1)))
Op2VK = TargetTransformInfo::OK_NonUniformConstantValue;
}
// FIXME: Currently cost of model modification for division by
// power of 2 is handled only for X86. Add support for other targets.
if (Op2VK == TargetTransformInfo::OK_UniformConstantValue && CInt &&
CInt->getValue().isPowerOf2())
Op2VP = TargetTransformInfo::OP_PowerOf2;
ScalarCost = VecTy->getNumElements() *
TTI->getArithmeticInstrCost(Opcode, ScalarTy, Op1VK, Op2VK,
Op1VP, Op2VP);
VecCost = TTI->getArithmeticInstrCost(Opcode, VecTy, Op1VK, Op2VK,
Op1VP, Op2VP);
}
return VecCost - ScalarCost;
}
case Instruction::GetElementPtr: {
TargetTransformInfo::OperandValueKind Op1VK =
TargetTransformInfo::OK_AnyValue;
TargetTransformInfo::OperandValueKind Op2VK =
TargetTransformInfo::OK_UniformConstantValue;
int ScalarCost =
VecTy->getNumElements() *
TTI->getArithmeticInstrCost(Instruction::Add, ScalarTy, Op1VK, Op2VK);
int VecCost =
TTI->getArithmeticInstrCost(Instruction::Add, VecTy, Op1VK, Op2VK);
return VecCost - ScalarCost;
}
case Instruction::Load: {
// Cost of wide load - cost of scalar loads.
int ScalarLdCost = VecTy->getNumElements() *
TTI->getMemoryOpCost(Instruction::Load, ScalarTy, 1, 0);
int VecLdCost = TTI->getMemoryOpCost(Instruction::Load, VecTy, 1, 0);
return VecLdCost - ScalarLdCost;
}
case Instruction::Store: {
// We know that we can merge the stores. Calculate the cost.
int ScalarStCost = VecTy->getNumElements() *
TTI->getMemoryOpCost(Instruction::Store, ScalarTy, 1, 0);
int VecStCost = TTI->getMemoryOpCost(Instruction::Store, VecTy, 1, 0);
return VecStCost - ScalarStCost;
}
case Instruction::Call: {
CallInst *CI = cast<CallInst>(VL0);
Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI);
// Calculate the cost of the scalar and vector calls.
SmallVector<Type*, 4> ScalarTys, VecTys;
for (unsigned op = 0, opc = CI->getNumArgOperands(); op!= opc; ++op) {
ScalarTys.push_back(CI->getArgOperand(op)->getType());
VecTys.push_back(VectorType::get(CI->getArgOperand(op)->getType(),
VecTy->getNumElements()));
}
int ScalarCallCost = VecTy->getNumElements() *
TTI->getIntrinsicInstrCost(ID, ScalarTy, ScalarTys);
int VecCallCost = TTI->getIntrinsicInstrCost(ID, VecTy, VecTys);
DEBUG(dbgs() << "SLP: Call cost "<< VecCallCost - ScalarCallCost
<< " (" << VecCallCost << "-" << ScalarCallCost << ")"
<< " for " << *CI << "\n");
return VecCallCost - ScalarCallCost;
}
case Instruction::ShuffleVector: {
TargetTransformInfo::OperandValueKind Op1VK =
TargetTransformInfo::OK_AnyValue;
TargetTransformInfo::OperandValueKind Op2VK =
TargetTransformInfo::OK_AnyValue;
int ScalarCost = 0;
int VecCost = 0;
for (unsigned i = 0; i < VL.size(); ++i) {
Instruction *I = cast<Instruction>(VL[i]);
if (!I)
break;
ScalarCost +=
TTI->getArithmeticInstrCost(I->getOpcode(), ScalarTy, Op1VK, Op2VK);
}
// VecCost is equal to sum of the cost of creating 2 vectors
// and the cost of creating shuffle.
Instruction *I0 = cast<Instruction>(VL[0]);
VecCost =
TTI->getArithmeticInstrCost(I0->getOpcode(), VecTy, Op1VK, Op2VK);
Instruction *I1 = cast<Instruction>(VL[1]);
VecCost +=
TTI->getArithmeticInstrCost(I1->getOpcode(), VecTy, Op1VK, Op2VK);
VecCost +=
TTI->getShuffleCost(TargetTransformInfo::SK_Alternate, VecTy, 0);
return VecCost - ScalarCost;
}
default:
llvm_unreachable("Unknown instruction");
}
}
bool BoUpSLP::isFullyVectorizableTinyTree() {
DEBUG(dbgs() << "SLP: Check whether the tree with height " <<
VectorizableTree.size() << " is fully vectorizable .\n");
// We only handle trees of height 2.
if (VectorizableTree.size() != 2)
return false;
// Handle splat and all-constants stores.
if (!VectorizableTree[0].NeedToGather &&
(allConstant(VectorizableTree[1].Scalars) ||
isSplat(VectorizableTree[1].Scalars)))
return true;
// Gathering cost would be too much for tiny trees.
if (VectorizableTree[0].NeedToGather || VectorizableTree[1].NeedToGather)
return false;
return true;
}
int BoUpSLP::getSpillCost() {
// Walk from the bottom of the tree to the top, tracking which values are
// live. When we see a call instruction that is not part of our tree,
// query TTI to see if there is a cost to keeping values live over it
// (for example, if spills and fills are required).
unsigned BundleWidth = VectorizableTree.front().Scalars.size();
int Cost = 0;
SmallPtrSet<Instruction*, 4> LiveValues;
Instruction *PrevInst = nullptr;
for (unsigned N = 0; N < VectorizableTree.size(); ++N) {
Instruction *Inst = dyn_cast<Instruction>(VectorizableTree[N].Scalars[0]);
if (!Inst)
continue;
if (!PrevInst) {
PrevInst = Inst;
continue;
}
// Update LiveValues.
LiveValues.erase(PrevInst);
for (auto &J : PrevInst->operands()) {
if (isa<Instruction>(&*J) && ScalarToTreeEntry.count(&*J))
LiveValues.insert(cast<Instruction>(&*J));
}
DEBUG(
dbgs() << "SLP: #LV: " << LiveValues.size();
for (auto *X : LiveValues)
dbgs() << " " << X->getName();
dbgs() << ", Looking at ";
Inst->dump();
);
// Now find the sequence of instructions between PrevInst and Inst.
BasicBlock::reverse_iterator InstIt(Inst->getIterator()),
PrevInstIt(PrevInst->getIterator());
--PrevInstIt;
while (InstIt != PrevInstIt) {
if (PrevInstIt == PrevInst->getParent()->rend()) {
PrevInstIt = Inst->getParent()->rbegin();
continue;
}
if (isa<CallInst>(&*PrevInstIt) && &*PrevInstIt != PrevInst) {
SmallVector<Type*, 4> V;
for (auto *II : LiveValues)
V.push_back(VectorType::get(II->getType(), BundleWidth));
Cost += TTI->getCostOfKeepingLiveOverCall(V);
}
++PrevInstIt;
}
PrevInst = Inst;
}
return Cost;
}
int BoUpSLP::getTreeCost() {
int Cost = 0;
DEBUG(dbgs() << "SLP: Calculating cost for tree of size " <<
VectorizableTree.size() << ".\n");
// We only vectorize tiny trees if it is fully vectorizable.
if (VectorizableTree.size() < 3 && !isFullyVectorizableTinyTree()) {
if (VectorizableTree.empty()) {
assert(!ExternalUses.size() && "We should not have any external users");
}
return INT_MAX;
}
unsigned BundleWidth = VectorizableTree[0].Scalars.size();
for (TreeEntry &TE : VectorizableTree) {
int C = getEntryCost(&TE);
DEBUG(dbgs() << "SLP: Adding cost " << C << " for bundle that starts with "
<< *TE.Scalars[0] << ".\n");
Cost += C;
}
SmallSet<Value *, 16> ExtractCostCalculated;
int ExtractCost = 0;
for (ExternalUser &EU : ExternalUses) {
// We only add extract cost once for the same scalar.
if (!ExtractCostCalculated.insert(EU.Scalar).second)
continue;
// Uses by ephemeral values are free (because the ephemeral value will be
// removed prior to code generation, and so the extraction will be
// removed as well).
if (EphValues.count(EU.User))
continue;
// If we plan to rewrite the tree in a smaller type, we will need to sign
// extend the extracted value back to the original type. Here, we account
// for the extract and the added cost of the sign extend if needed.
auto *VecTy = VectorType::get(EU.Scalar->getType(), BundleWidth);
auto *ScalarRoot = VectorizableTree[0].Scalars[0];
if (MinBWs.count(ScalarRoot)) {
auto *MinTy = IntegerType::get(F->getContext(), MinBWs[ScalarRoot]);
VecTy = VectorType::get(MinTy, BundleWidth);
ExtractCost +=
TTI->getCastInstrCost(Instruction::SExt, EU.Scalar->getType(), MinTy);
}
ExtractCost +=
TTI->getVectorInstrCost(Instruction::ExtractElement, VecTy, EU.Lane);
}
int SpillCost = getSpillCost();
Cost += SpillCost + ExtractCost;
DEBUG(dbgs() << "SLP: Spill Cost = " << SpillCost << ".\n"
<< "SLP: Extract Cost = " << ExtractCost << ".\n"
<< "SLP: Total Cost = " << Cost << ".\n");
return Cost;
}
int BoUpSLP::getGatherCost(Type *Ty) {
int Cost = 0;
for (unsigned i = 0, e = cast<VectorType>(Ty)->getNumElements(); i < e; ++i)
Cost += TTI->getVectorInstrCost(Instruction::InsertElement, Ty, i);
return Cost;
}
int BoUpSLP::getGatherCost(ArrayRef<Value *> VL) {
// Find the type of the operands in VL.
Type *ScalarTy = VL[0]->getType();
if (StoreInst *SI = dyn_cast<StoreInst>(VL[0]))
ScalarTy = SI->getValueOperand()->getType();
VectorType *VecTy = VectorType::get(ScalarTy, VL.size());
// Find the cost of inserting/extracting values from the vector.
return getGatherCost(VecTy);
}
// Reorder commutative operations in alternate shuffle if the resulting vectors
// are consecutive loads. This would allow us to vectorize the tree.
// If we have something like-
// load a[0] - load b[0]
// load b[1] + load a[1]
// load a[2] - load b[2]
// load a[3] + load b[3]
// Reordering the second load b[1] load a[1] would allow us to vectorize this
// code.
void BoUpSLP::reorderAltShuffleOperands(ArrayRef<Value *> VL,
SmallVectorImpl<Value *> &Left,
SmallVectorImpl<Value *> &Right) {
const DataLayout &DL = F->getParent()->getDataLayout();
// Push left and right operands of binary operation into Left and Right
for (unsigned i = 0, e = VL.size(); i < e; ++i) {
Left.push_back(cast<Instruction>(VL[i])->getOperand(0));
Right.push_back(cast<Instruction>(VL[i])->getOperand(1));
}
// Reorder if we have a commutative operation and consecutive access
// are on either side of the alternate instructions.
for (unsigned j = 0; j < VL.size() - 1; ++j) {
if (LoadInst *L = dyn_cast<LoadInst>(Left[j])) {
if (LoadInst *L1 = dyn_cast<LoadInst>(Right[j + 1])) {
Instruction *VL1 = cast<Instruction>(VL[j]);
Instruction *VL2 = cast<Instruction>(VL[j + 1]);
if (VL1->isCommutative() && isConsecutiveAccess(L, L1, DL, *SE)) {
std::swap(Left[j], Right[j]);
continue;
} else if (VL2->isCommutative() && isConsecutiveAccess(L, L1, DL, *SE)) {
std::swap(Left[j + 1], Right[j + 1]);
continue;
}
// else unchanged
}
}
if (LoadInst *L = dyn_cast<LoadInst>(Right[j])) {
if (LoadInst *L1 = dyn_cast<LoadInst>(Left[j + 1])) {
Instruction *VL1 = cast<Instruction>(VL[j]);
Instruction *VL2 = cast<Instruction>(VL[j + 1]);
if (VL1->isCommutative() && isConsecutiveAccess(L, L1, DL, *SE)) {
std::swap(Left[j], Right[j]);
continue;
} else if (VL2->isCommutative() && isConsecutiveAccess(L, L1, DL, *SE)) {
std::swap(Left[j + 1], Right[j + 1]);
continue;
}
// else unchanged
}
}
}
}
// Return true if I should be commuted before adding it's left and right
// operands to the arrays Left and Right.
//
// The vectorizer is trying to either have all elements one side being
// instruction with the same opcode to enable further vectorization, or having
// a splat to lower the vectorizing cost.
static bool shouldReorderOperands(int i, Instruction &I,
SmallVectorImpl<Value *> &Left,
SmallVectorImpl<Value *> &Right,
bool AllSameOpcodeLeft,
bool AllSameOpcodeRight, bool SplatLeft,
bool SplatRight) {
Value *VLeft = I.getOperand(0);
Value *VRight = I.getOperand(1);
// If we have "SplatRight", try to see if commuting is needed to preserve it.
if (SplatRight) {
if (VRight == Right[i - 1])
// Preserve SplatRight
return false;
if (VLeft == Right[i - 1]) {
// Commuting would preserve SplatRight, but we don't want to break
// SplatLeft either, i.e. preserve the original order if possible.
// (FIXME: why do we care?)
if (SplatLeft && VLeft == Left[i - 1])
return false;
return true;
}
}
// Symmetrically handle Right side.
if (SplatLeft) {
if (VLeft == Left[i - 1])
// Preserve SplatLeft
return false;
if (VRight == Left[i - 1])
return true;
}
Instruction *ILeft = dyn_cast<Instruction>(VLeft);
Instruction *IRight = dyn_cast<Instruction>(VRight);
// If we have "AllSameOpcodeRight", try to see if the left operands preserves
// it and not the right, in this case we want to commute.
if (AllSameOpcodeRight) {
unsigned RightPrevOpcode = cast<Instruction>(Right[i - 1])->getOpcode();
if (IRight && RightPrevOpcode == IRight->getOpcode())
// Do not commute, a match on the right preserves AllSameOpcodeRight
return false;
if (ILeft && RightPrevOpcode == ILeft->getOpcode()) {
// We have a match and may want to commute, but first check if there is
// not also a match on the existing operands on the Left to preserve
// AllSameOpcodeLeft, i.e. preserve the original order if possible.
// (FIXME: why do we care?)
if (AllSameOpcodeLeft && ILeft &&
cast<Instruction>(Left[i - 1])->getOpcode() == ILeft->getOpcode())
return false;
return true;
}
}
// Symmetrically handle Left side.
if (AllSameOpcodeLeft) {
unsigned LeftPrevOpcode = cast<Instruction>(Left[i - 1])->getOpcode();
if (ILeft && LeftPrevOpcode == ILeft->getOpcode())
return false;
if (IRight && LeftPrevOpcode == IRight->getOpcode())
return true;
}
return false;
}
void BoUpSLP::reorderInputsAccordingToOpcode(ArrayRef<Value *> VL,
SmallVectorImpl<Value *> &Left,
SmallVectorImpl<Value *> &Right) {
if (VL.size()) {
// Peel the first iteration out of the loop since there's nothing
// interesting to do anyway and it simplifies the checks in the loop.
auto VLeft = cast<Instruction>(VL[0])->getOperand(0);
auto VRight = cast<Instruction>(VL[0])->getOperand(1);
if (!isa<Instruction>(VRight) && isa<Instruction>(VLeft))
// Favor having instruction to the right. FIXME: why?
std::swap(VLeft, VRight);
Left.push_back(VLeft);
Right.push_back(VRight);
}
// Keep track if we have instructions with all the same opcode on one side.
bool AllSameOpcodeLeft = isa<Instruction>(Left[0]);
bool AllSameOpcodeRight = isa<Instruction>(Right[0]);
// Keep track if we have one side with all the same value (broadcast).
bool SplatLeft = true;
bool SplatRight = true;
for (unsigned i = 1, e = VL.size(); i != e; ++i) {
Instruction *I = cast<Instruction>(VL[i]);
assert(I->isCommutative() && "Can only process commutative instruction");
// Commute to favor either a splat or maximizing having the same opcodes on
// one side.
if (shouldReorderOperands(i, *I, Left, Right, AllSameOpcodeLeft,
AllSameOpcodeRight, SplatLeft, SplatRight)) {
Left.push_back(I->getOperand(1));
Right.push_back(I->getOperand(0));
} else {
Left.push_back(I->getOperand(0));
Right.push_back(I->getOperand(1));
}
// Update Splat* and AllSameOpcode* after the insertion.
SplatRight = SplatRight && (Right[i - 1] == Right[i]);
SplatLeft = SplatLeft && (Left[i - 1] == Left[i]);
AllSameOpcodeLeft = AllSameOpcodeLeft && isa<Instruction>(Left[i]) &&
(cast<Instruction>(Left[i - 1])->getOpcode() ==
cast<Instruction>(Left[i])->getOpcode());
AllSameOpcodeRight = AllSameOpcodeRight && isa<Instruction>(Right[i]) &&
(cast<Instruction>(Right[i - 1])->getOpcode() ==
cast<Instruction>(Right[i])->getOpcode());
}
// If one operand end up being broadcast, return this operand order.
if (SplatRight || SplatLeft)
return;
const DataLayout &DL = F->getParent()->getDataLayout();
// Finally check if we can get longer vectorizable chain by reordering
// without breaking the good operand order detected above.
// E.g. If we have something like-
// load a[0] load b[0]
// load b[1] load a[1]
// load a[2] load b[2]
// load a[3] load b[3]
// Reordering the second load b[1] load a[1] would allow us to vectorize
// this code and we still retain AllSameOpcode property.
// FIXME: This load reordering might break AllSameOpcode in some rare cases
// such as-
// add a[0],c[0] load b[0]
// add a[1],c[2] load b[1]
// b[2] load b[2]
// add a[3],c[3] load b[3]
for (unsigned j = 0; j < VL.size() - 1; ++j) {
if (LoadInst *L = dyn_cast<LoadInst>(Left[j])) {
if (LoadInst *L1 = dyn_cast<LoadInst>(Right[j + 1])) {
if (isConsecutiveAccess(L, L1, DL, *SE)) {
std::swap(Left[j + 1], Right[j + 1]);
continue;
}
}
}
if (LoadInst *L = dyn_cast<LoadInst>(Right[j])) {
if (LoadInst *L1 = dyn_cast<LoadInst>(Left[j + 1])) {
if (isConsecutiveAccess(L, L1, DL, *SE)) {
std::swap(Left[j + 1], Right[j + 1]);
continue;
}
}
}
// else unchanged
}
}
void BoUpSLP::setInsertPointAfterBundle(ArrayRef<Value *> VL) {
Instruction *VL0 = cast<Instruction>(VL[0]);
BasicBlock::iterator NextInst(VL0);
++NextInst;
Builder.SetInsertPoint(VL0->getParent(), NextInst);
Builder.SetCurrentDebugLocation(VL0->getDebugLoc());
}
Value *BoUpSLP::Gather(ArrayRef<Value *> VL, VectorType *Ty) {
Value *Vec = UndefValue::get(Ty);
// Generate the 'InsertElement' instruction.
for (unsigned i = 0; i < Ty->getNumElements(); ++i) {
Vec = Builder.CreateInsertElement(Vec, VL[i], Builder.getInt32(i));
if (Instruction *Insrt = dyn_cast<Instruction>(Vec)) {
GatherSeq.insert(Insrt);
CSEBlocks.insert(Insrt->getParent());
// Add to our 'need-to-extract' list.
if (ScalarToTreeEntry.count(VL[i])) {
int Idx = ScalarToTreeEntry[VL[i]];
TreeEntry *E = &VectorizableTree[Idx];
// Find which lane we need to extract.
int FoundLane = -1;
for (unsigned Lane = 0, LE = VL.size(); Lane != LE; ++Lane) {
// Is this the lane of the scalar that we are looking for ?
if (E->Scalars[Lane] == VL[i]) {
FoundLane = Lane;
break;
}
}
assert(FoundLane >= 0 && "Could not find the correct lane");
ExternalUses.push_back(ExternalUser(VL[i], Insrt, FoundLane));
}
}
}
return Vec;
}
Value *BoUpSLP::alreadyVectorized(ArrayRef<Value *> VL) const {
SmallDenseMap<Value*, int>::const_iterator Entry
= ScalarToTreeEntry.find(VL[0]);
if (Entry != ScalarToTreeEntry.end()) {
int Idx = Entry->second;
const TreeEntry *En = &VectorizableTree[Idx];
if (En->isSame(VL) && En->VectorizedValue)
return En->VectorizedValue;
}
return nullptr;
}
Value *BoUpSLP::vectorizeTree(ArrayRef<Value *> VL) {
if (ScalarToTreeEntry.count(VL[0])) {
int Idx = ScalarToTreeEntry[VL[0]];
TreeEntry *E = &VectorizableTree[Idx];
if (E->isSame(VL))
return vectorizeTree(E);
}
Type *ScalarTy = VL[0]->getType();
if (StoreInst *SI = dyn_cast<StoreInst>(VL[0]))
ScalarTy = SI->getValueOperand()->getType();
VectorType *VecTy = VectorType::get(ScalarTy, VL.size());
return Gather(VL, VecTy);
}
Value *BoUpSLP::vectorizeTree(TreeEntry *E) {
IRBuilder<>::InsertPointGuard Guard(Builder);
if (E->VectorizedValue) {
DEBUG(dbgs() << "SLP: Diamond merged for " << *E->Scalars[0] << ".\n");
return E->VectorizedValue;
}
Instruction *VL0 = cast<Instruction>(E->Scalars[0]);
Type *ScalarTy = VL0->getType();
if (StoreInst *SI = dyn_cast<StoreInst>(VL0))
ScalarTy = SI->getValueOperand()->getType();
VectorType *VecTy = VectorType::get(ScalarTy, E->Scalars.size());
if (E->NeedToGather) {
setInsertPointAfterBundle(E->Scalars);
return Gather(E->Scalars, VecTy);
}
const DataLayout &DL = F->getParent()->getDataLayout();
unsigned Opcode = getSameOpcode(E->Scalars);
switch (Opcode) {
case Instruction::PHI: {
PHINode *PH = dyn_cast<PHINode>(VL0);
Builder.SetInsertPoint(PH->getParent()->getFirstNonPHI());
Builder.SetCurrentDebugLocation(PH->getDebugLoc());
PHINode *NewPhi = Builder.CreatePHI(VecTy, PH->getNumIncomingValues());
E->VectorizedValue = NewPhi;
// PHINodes may have multiple entries from the same block. We want to
// visit every block once.
SmallSet<BasicBlock*, 4> VisitedBBs;
for (unsigned i = 0, e = PH->getNumIncomingValues(); i < e; ++i) {
ValueList Operands;
BasicBlock *IBB = PH->getIncomingBlock(i);
if (!VisitedBBs.insert(IBB).second) {
NewPhi->addIncoming(NewPhi->getIncomingValueForBlock(IBB), IBB);
continue;
}
// Prepare the operand vector.
for (Value *V : E->Scalars)
Operands.push_back(cast<PHINode>(V)->getIncomingValueForBlock(IBB));
Builder.SetInsertPoint(IBB->getTerminator());
Builder.SetCurrentDebugLocation(PH->getDebugLoc());
Value *Vec = vectorizeTree(Operands);
NewPhi->addIncoming(Vec, IBB);
}
assert(NewPhi->getNumIncomingValues() == PH->getNumIncomingValues() &&
"Invalid number of incoming values");
return NewPhi;
}
case Instruction::ExtractElement: {
if (CanReuseExtract(E->Scalars)) {
Value *V = VL0->getOperand(0);
E->VectorizedValue = V;
return V;
}
return Gather(E->Scalars, VecTy);
}
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: {
ValueList INVL;
for (Value *V : E->Scalars)
INVL.push_back(cast<Instruction>(V)->getOperand(0));
setInsertPointAfterBundle(E->Scalars);
Value *InVec = vectorizeTree(INVL);
if (Value *V = alreadyVectorized(E->Scalars))
return V;
CastInst *CI = dyn_cast<CastInst>(VL0);
Value *V = Builder.CreateCast(CI->getOpcode(), InVec, VecTy);
E->VectorizedValue = V;
++NumVectorInstructions;
return V;
}
case Instruction::FCmp:
case Instruction::ICmp: {
ValueList LHSV, RHSV;
for (Value *V : E->Scalars) {
LHSV.push_back(cast<Instruction>(V)->getOperand(0));
RHSV.push_back(cast<Instruction>(V)->getOperand(1));
}
setInsertPointAfterBundle(E->Scalars);
Value *L = vectorizeTree(LHSV);
Value *R = vectorizeTree(RHSV);
if (Value *V = alreadyVectorized(E->Scalars))
return V;
CmpInst::Predicate P0 = cast<CmpInst>(VL0)->getPredicate();
Value *V;
if (Opcode == Instruction::FCmp)
V = Builder.CreateFCmp(P0, L, R);
else
V = Builder.CreateICmp(P0, L, R);
E->VectorizedValue = V;
++NumVectorInstructions;
return V;
}
case Instruction::Select: {
ValueList TrueVec, FalseVec, CondVec;
for (Value *V : E->Scalars) {
CondVec.push_back(cast<Instruction>(V)->getOperand(0));
TrueVec.push_back(cast<Instruction>(V)->getOperand(1));
FalseVec.push_back(cast<Instruction>(V)->getOperand(2));
}
setInsertPointAfterBundle(E->Scalars);
Value *Cond = vectorizeTree(CondVec);
Value *True = vectorizeTree(TrueVec);
Value *False = vectorizeTree(FalseVec);
if (Value *V = alreadyVectorized(E->Scalars))
return V;
Value *V = Builder.CreateSelect(Cond, True, False);
E->VectorizedValue = V;
++NumVectorInstructions;
return V;
}
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: {
ValueList LHSVL, RHSVL;
if (isa<BinaryOperator>(VL0) && VL0->isCommutative())
reorderInputsAccordingToOpcode(E->Scalars, LHSVL, RHSVL);
else
for (Value *V : E->Scalars) {
LHSVL.push_back(cast<Instruction>(V)->getOperand(0));
RHSVL.push_back(cast<Instruction>(V)->getOperand(1));
}
setInsertPointAfterBundle(E->Scalars);
Value *LHS = vectorizeTree(LHSVL);
Value *RHS = vectorizeTree(RHSVL);
if (LHS == RHS && isa<Instruction>(LHS)) {
assert((VL0->getOperand(0) == VL0->getOperand(1)) && "Invalid order");
}
if (Value *V = alreadyVectorized(E->Scalars))
return V;
BinaryOperator *BinOp = cast<BinaryOperator>(VL0);
Value *V = Builder.CreateBinOp(BinOp->getOpcode(), LHS, RHS);
E->VectorizedValue = V;
propagateIRFlags(E->VectorizedValue, E->Scalars);
++NumVectorInstructions;
if (Instruction *I = dyn_cast<Instruction>(V))
return propagateMetadata(I, E->Scalars);
return V;
}
case Instruction::Load: {
// Loads are inserted at the head of the tree because we don't want to
// sink them all the way down past store instructions.
setInsertPointAfterBundle(E->Scalars);
LoadInst *LI = cast<LoadInst>(VL0);
Type *ScalarLoadTy = LI->getType();
unsigned AS = LI->getPointerAddressSpace();
Value *VecPtr = Builder.CreateBitCast(LI->getPointerOperand(),
VecTy->getPointerTo(AS));
// The pointer operand uses an in-tree scalar so we add the new BitCast to
// ExternalUses list to make sure that an extract will be generated in the
// future.
if (ScalarToTreeEntry.count(LI->getPointerOperand()))
ExternalUses.push_back(
ExternalUser(LI->getPointerOperand(), cast<User>(VecPtr), 0));
unsigned Alignment = LI->getAlignment();
LI = Builder.CreateLoad(VecPtr);
if (!Alignment) {
Alignment = DL.getABITypeAlignment(ScalarLoadTy);
}
LI->setAlignment(Alignment);
E->VectorizedValue = LI;
++NumVectorInstructions;
return propagateMetadata(LI, E->Scalars);
}
case Instruction::Store: {
StoreInst *SI = cast<StoreInst>(VL0);
unsigned Alignment = SI->getAlignment();
unsigned AS = SI->getPointerAddressSpace();
ValueList ValueOp;
for (Value *V : E->Scalars)
ValueOp.push_back(cast<StoreInst>(V)->getValueOperand());
setInsertPointAfterBundle(E->Scalars);
Value *VecValue = vectorizeTree(ValueOp);
Value *VecPtr = Builder.CreateBitCast(SI->getPointerOperand(),
VecTy->getPointerTo(AS));
StoreInst *S = Builder.CreateStore(VecValue, VecPtr);
// The pointer operand uses an in-tree scalar so we add the new BitCast to
// ExternalUses list to make sure that an extract will be generated in the
// future.
if (ScalarToTreeEntry.count(SI->getPointerOperand()))
ExternalUses.push_back(
ExternalUser(SI->getPointerOperand(), cast<User>(VecPtr), 0));
if (!Alignment) {
Alignment = DL.getABITypeAlignment(SI->getValueOperand()->getType());
}
S->setAlignment(Alignment);
E->VectorizedValue = S;
++NumVectorInstructions;
return propagateMetadata(S, E->Scalars);
}
case Instruction::GetElementPtr: {
setInsertPointAfterBundle(E->Scalars);
ValueList Op0VL;
for (Value *V : E->Scalars)
Op0VL.push_back(cast<GetElementPtrInst>(V)->getOperand(0));
Value *Op0 = vectorizeTree(Op0VL);
std::vector<Value *> OpVecs;
for (int j = 1, e = cast<GetElementPtrInst>(VL0)->getNumOperands(); j < e;
++j) {
ValueList OpVL;
for (Value *V : E->Scalars)
OpVL.push_back(cast<GetElementPtrInst>(V)->getOperand(j));
Value *OpVec = vectorizeTree(OpVL);
OpVecs.push_back(OpVec);
}
Value *V = Builder.CreateGEP(
cast<GetElementPtrInst>(VL0)->getSourceElementType(), Op0, OpVecs);
E->VectorizedValue = V;
++NumVectorInstructions;
if (Instruction *I = dyn_cast<Instruction>(V))
return propagateMetadata(I, E->Scalars);
return V;
}
case Instruction::Call: {
CallInst *CI = cast<CallInst>(VL0);
setInsertPointAfterBundle(E->Scalars);
Function *FI;
Intrinsic::ID IID = Intrinsic::not_intrinsic;
Value *ScalarArg = nullptr;
if (CI && (FI = CI->getCalledFunction())) {
IID = FI->getIntrinsicID();
}
std::vector<Value *> OpVecs;
for (int j = 0, e = CI->getNumArgOperands(); j < e; ++j) {
ValueList OpVL;
// ctlz,cttz and powi are special intrinsics whose second argument is
// a scalar. This argument should not be vectorized.
if (hasVectorInstrinsicScalarOpd(IID, 1) && j == 1) {
CallInst *CEI = cast<CallInst>(E->Scalars[0]);
ScalarArg = CEI->getArgOperand(j);
OpVecs.push_back(CEI->getArgOperand(j));
continue;
}
for (Value *V : E->Scalars) {
CallInst *CEI = cast<CallInst>(V);
OpVL.push_back(CEI->getArgOperand(j));
}
Value *OpVec = vectorizeTree(OpVL);
DEBUG(dbgs() << "SLP: OpVec[" << j << "]: " << *OpVec << "\n");
OpVecs.push_back(OpVec);
}
Module *M = F->getParent();
Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI);
Type *Tys[] = { VectorType::get(CI->getType(), E->Scalars.size()) };
Function *CF = Intrinsic::getDeclaration(M, ID, Tys);
Value *V = Builder.CreateCall(CF, OpVecs);
// The scalar argument uses an in-tree scalar so we add the new vectorized
// call to ExternalUses list to make sure that an extract will be
// generated in the future.
if (ScalarArg && ScalarToTreeEntry.count(ScalarArg))
ExternalUses.push_back(ExternalUser(ScalarArg, cast<User>(V), 0));
E->VectorizedValue = V;
++NumVectorInstructions;
return V;
}
case Instruction::ShuffleVector: {
ValueList LHSVL, RHSVL;
assert(isa<BinaryOperator>(VL0) && "Invalid Shuffle Vector Operand");
reorderAltShuffleOperands(E->Scalars, LHSVL, RHSVL);
setInsertPointAfterBundle(E->Scalars);
Value *LHS = vectorizeTree(LHSVL);
Value *RHS = vectorizeTree(RHSVL);
if (Value *V = alreadyVectorized(E->Scalars))
return V;
// Create a vector of LHS op1 RHS
BinaryOperator *BinOp0 = cast<BinaryOperator>(VL0);
Value *V0 = Builder.CreateBinOp(BinOp0->getOpcode(), LHS, RHS);
// Create a vector of LHS op2 RHS
Instruction *VL1 = cast<Instruction>(E->Scalars[1]);
BinaryOperator *BinOp1 = cast<BinaryOperator>(VL1);
Value *V1 = Builder.CreateBinOp(BinOp1->getOpcode(), LHS, RHS);
// Create shuffle to take alternate operations from the vector.
// Also, gather up odd and even scalar ops to propagate IR flags to
// each vector operation.
ValueList OddScalars, EvenScalars;
unsigned e = E->Scalars.size();
SmallVector<Constant *, 8> Mask(e);
for (unsigned i = 0; i < e; ++i) {
if (i & 1) {
Mask[i] = Builder.getInt32(e + i);
OddScalars.push_back(E->Scalars[i]);
} else {
Mask[i] = Builder.getInt32(i);
EvenScalars.push_back(E->Scalars[i]);
}
}
Value *ShuffleMask = ConstantVector::get(Mask);
propagateIRFlags(V0, EvenScalars);
propagateIRFlags(V1, OddScalars);
Value *V = Builder.CreateShuffleVector(V0, V1, ShuffleMask);
E->VectorizedValue = V;
++NumVectorInstructions;
if (Instruction *I = dyn_cast<Instruction>(V))
return propagateMetadata(I, E->Scalars);
return V;
}
default:
llvm_unreachable("unknown inst");
}
return nullptr;
}
Value *BoUpSLP::vectorizeTree() {
// All blocks must be scheduled before any instructions are inserted.
for (auto &BSIter : BlocksSchedules) {
scheduleBlock(BSIter.second.get());
}
Builder.SetInsertPoint(&F->getEntryBlock().front());
auto *VectorRoot = vectorizeTree(&VectorizableTree[0]);
// If the vectorized tree can be rewritten in a smaller type, we truncate the
// vectorized root. InstCombine will then rewrite the entire expression. We
// sign extend the extracted values below.
auto *ScalarRoot = VectorizableTree[0].Scalars[0];
if (MinBWs.count(ScalarRoot)) {
if (auto *I = dyn_cast<Instruction>(VectorRoot))
Builder.SetInsertPoint(&*++BasicBlock::iterator(I));
auto BundleWidth = VectorizableTree[0].Scalars.size();
auto *MinTy = IntegerType::get(F->getContext(), MinBWs[ScalarRoot]);
auto *VecTy = VectorType::get(MinTy, BundleWidth);
auto *Trunc = Builder.CreateTrunc(VectorRoot, VecTy);
VectorizableTree[0].VectorizedValue = Trunc;
}
DEBUG(dbgs() << "SLP: Extracting " << ExternalUses.size() << " values .\n");
// Extract all of the elements with the external uses.
for (UserList::iterator it = ExternalUses.begin(), e = ExternalUses.end();
it != e; ++it) {
Value *Scalar = it->Scalar;
llvm::User *User = it->User;
// Skip users that we already RAUW. This happens when one instruction
// has multiple uses of the same value.
if (std::find(Scalar->user_begin(), Scalar->user_end(), User) ==
Scalar->user_end())
continue;
assert(ScalarToTreeEntry.count(Scalar) && "Invalid scalar");
int Idx = ScalarToTreeEntry[Scalar];
TreeEntry *E = &VectorizableTree[Idx];
assert(!E->NeedToGather && "Extracting from a gather list");
Value *Vec = E->VectorizedValue;
assert(Vec && "Can't find vectorizable value");
Value *Lane = Builder.getInt32(it->Lane);
// Generate extracts for out-of-tree users.
// Find the insertion point for the extractelement lane.
if (isa<Instruction>(Vec)){
if (PHINode *PH = dyn_cast<PHINode>(User)) {
for (int i = 0, e = PH->getNumIncomingValues(); i != e; ++i) {
if (PH->getIncomingValue(i) == Scalar) {
Builder.SetInsertPoint(PH->getIncomingBlock(i)->getTerminator());
Value *Ex = Builder.CreateExtractElement(Vec, Lane);
if (MinBWs.count(ScalarRoot))
Ex = Builder.CreateSExt(Ex, Scalar->getType());
CSEBlocks.insert(PH->getIncomingBlock(i));
PH->setOperand(i, Ex);
}
}
} else {
Builder.SetInsertPoint(cast<Instruction>(User));
Value *Ex = Builder.CreateExtractElement(Vec, Lane);
if (MinBWs.count(ScalarRoot))
Ex = Builder.CreateSExt(Ex, Scalar->getType());
CSEBlocks.insert(cast<Instruction>(User)->getParent());
User->replaceUsesOfWith(Scalar, Ex);
}
} else {
Builder.SetInsertPoint(&F->getEntryBlock().front());
Value *Ex = Builder.CreateExtractElement(Vec, Lane);
if (MinBWs.count(ScalarRoot))
Ex = Builder.CreateSExt(Ex, Scalar->getType());
CSEBlocks.insert(&F->getEntryBlock());
User->replaceUsesOfWith(Scalar, Ex);
}
DEBUG(dbgs() << "SLP: Replaced:" << *User << ".\n");
}
// For each vectorized value:
for (int EIdx = 0, EE = VectorizableTree.size(); EIdx < EE; ++EIdx) {
TreeEntry *Entry = &VectorizableTree[EIdx];
// For each lane:
for (int Lane = 0, LE = Entry->Scalars.size(); Lane != LE; ++Lane) {
Value *Scalar = Entry->Scalars[Lane];
// No need to handle users of gathered values.
if (Entry->NeedToGather)
continue;
assert(Entry->VectorizedValue && "Can't find vectorizable value");
Type *Ty = Scalar->getType();
if (!Ty->isVoidTy()) {
#ifndef NDEBUG
for (User *U : Scalar->users()) {
DEBUG(dbgs() << "SLP: \tvalidating user:" << *U << ".\n");
assert((ScalarToTreeEntry.count(U) ||
// It is legal to replace users in the ignorelist by undef.
(std::find(UserIgnoreList.begin(), UserIgnoreList.end(), U) !=
UserIgnoreList.end())) &&
"Replacing out-of-tree value with undef");
}
#endif
Value *Undef = UndefValue::get(Ty);
Scalar->replaceAllUsesWith(Undef);
}
DEBUG(dbgs() << "SLP: \tErasing scalar:" << *Scalar << ".\n");
eraseInstruction(cast<Instruction>(Scalar));
}
}
Builder.ClearInsertionPoint();
return VectorizableTree[0].VectorizedValue;
}
void BoUpSLP::optimizeGatherSequence() {
DEBUG(dbgs() << "SLP: Optimizing " << GatherSeq.size()
<< " gather sequences instructions.\n");
// LICM InsertElementInst sequences.
for (SetVector<Instruction *>::iterator it = GatherSeq.begin(),
e = GatherSeq.end(); it != e; ++it) {
InsertElementInst *Insert = dyn_cast<InsertElementInst>(*it);
if (!Insert)
continue;
// Check if this block is inside a loop.
Loop *L = LI->getLoopFor(Insert->getParent());
if (!L)
continue;
// Check if it has a preheader.
BasicBlock *PreHeader = L->getLoopPreheader();
if (!PreHeader)
continue;
// If the vector or the element that we insert into it are
// instructions that are defined in this basic block then we can't
// hoist this instruction.
Instruction *CurrVec = dyn_cast<Instruction>(Insert->getOperand(0));
Instruction *NewElem = dyn_cast<Instruction>(Insert->getOperand(1));
if (CurrVec && L->contains(CurrVec))
continue;
if (NewElem && L->contains(NewElem))
continue;
// We can hoist this instruction. Move it to the pre-header.
Insert->moveBefore(PreHeader->getTerminator());
}
// Make a list of all reachable blocks in our CSE queue.
SmallVector<const DomTreeNode *, 8> CSEWorkList;
CSEWorkList.reserve(CSEBlocks.size());
for (BasicBlock *BB : CSEBlocks)
if (DomTreeNode *N = DT->getNode(BB)) {
assert(DT->isReachableFromEntry(N));
CSEWorkList.push_back(N);
}
// Sort blocks by domination. This ensures we visit a block after all blocks
// dominating it are visited.
std::stable_sort(CSEWorkList.begin(), CSEWorkList.end(),
[this](const DomTreeNode *A, const DomTreeNode *B) {
return DT->properlyDominates(A, B);
});
// Perform O(N^2) search over the gather sequences and merge identical
// instructions. TODO: We can further optimize this scan if we split the
// instructions into different buckets based on the insert lane.
SmallVector<Instruction *, 16> Visited;
for (auto I = CSEWorkList.begin(), E = CSEWorkList.end(); I != E; ++I) {
assert((I == CSEWorkList.begin() || !DT->dominates(*I, *std::prev(I))) &&
"Worklist not sorted properly!");
BasicBlock *BB = (*I)->getBlock();
// For all instructions in blocks containing gather sequences:
for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e;) {
Instruction *In = &*it++;
if (!isa<InsertElementInst>(In) && !isa<ExtractElementInst>(In))
continue;
// Check if we can replace this instruction with any of the
// visited instructions.
for (SmallVectorImpl<Instruction *>::iterator v = Visited.begin(),
ve = Visited.end();
v != ve; ++v) {
if (In->isIdenticalTo(*v) &&
DT->dominates((*v)->getParent(), In->getParent())) {
In->replaceAllUsesWith(*v);
eraseInstruction(In);
In = nullptr;
break;
}
}
if (In) {
assert(std::find(Visited.begin(), Visited.end(), In) == Visited.end());
Visited.push_back(In);
}
}
}
CSEBlocks.clear();
GatherSeq.clear();
}
// Groups the instructions to a bundle (which is then a single scheduling entity)
// and schedules instructions until the bundle gets ready.
bool BoUpSLP::BlockScheduling::tryScheduleBundle(ArrayRef<Value *> VL,
BoUpSLP *SLP) {
if (isa<PHINode>(VL[0]))
return true;
// Initialize the instruction bundle.
Instruction *OldScheduleEnd = ScheduleEnd;
ScheduleData *PrevInBundle = nullptr;
ScheduleData *Bundle = nullptr;
bool ReSchedule = false;
DEBUG(dbgs() << "SLP: bundle: " << *VL[0] << "\n");
// Make sure that the scheduling region contains all
// instructions of the bundle.
for (Value *V : VL) {
if (!extendSchedulingRegion(V))
return false;
}
for (Value *V : VL) {
ScheduleData *BundleMember = getScheduleData(V);
assert(BundleMember &&
"no ScheduleData for bundle member (maybe not in same basic block)");
if (BundleMember->IsScheduled) {
// A bundle member was scheduled as single instruction before and now
// needs to be scheduled as part of the bundle. We just get rid of the
// existing schedule.
DEBUG(dbgs() << "SLP: reset schedule because " << *BundleMember
<< " was already scheduled\n");
ReSchedule = true;
}
assert(BundleMember->isSchedulingEntity() &&
"bundle member already part of other bundle");
if (PrevInBundle) {
PrevInBundle->NextInBundle = BundleMember;
} else {
Bundle = BundleMember;
}
BundleMember->UnscheduledDepsInBundle = 0;
Bundle->UnscheduledDepsInBundle += BundleMember->UnscheduledDeps;
// Group the instructions to a bundle.
BundleMember->FirstInBundle = Bundle;
PrevInBundle = BundleMember;
}
if (ScheduleEnd != OldScheduleEnd) {
// The scheduling region got new instructions at the lower end (or it is a
// new region for the first bundle). This makes it necessary to
// recalculate all dependencies.
// It is seldom that this needs to be done a second time after adding the
// initial bundle to the region.
for (auto *I = ScheduleStart; I != ScheduleEnd; I = I->getNextNode()) {
ScheduleData *SD = getScheduleData(I);
SD->clearDependencies();
}
ReSchedule = true;
}
if (ReSchedule) {
resetSchedule();
initialFillReadyList(ReadyInsts);
}
DEBUG(dbgs() << "SLP: try schedule bundle " << *Bundle << " in block "
<< BB->getName() << "\n");
calculateDependencies(Bundle, true, SLP);
// Now try to schedule the new bundle. As soon as the bundle is "ready" it
// means that there are no cyclic dependencies and we can schedule it.
// Note that's important that we don't "schedule" the bundle yet (see
// cancelScheduling).
while (!Bundle->isReady() && !ReadyInsts.empty()) {
ScheduleData *pickedSD = ReadyInsts.back();
ReadyInsts.pop_back();
if (pickedSD->isSchedulingEntity() && pickedSD->isReady()) {
schedule(pickedSD, ReadyInsts);
}
}
if (!Bundle->isReady()) {
cancelScheduling(VL);
return false;
}
return true;
}
void BoUpSLP::BlockScheduling::cancelScheduling(ArrayRef<Value *> VL) {
if (isa<PHINode>(VL[0]))
return;
ScheduleData *Bundle = getScheduleData(VL[0]);
DEBUG(dbgs() << "SLP: cancel scheduling of " << *Bundle << "\n");
assert(!Bundle->IsScheduled &&
"Can't cancel bundle which is already scheduled");
assert(Bundle->isSchedulingEntity() && Bundle->isPartOfBundle() &&
"tried to unbundle something which is not a bundle");
// Un-bundle: make single instructions out of the bundle.
ScheduleData *BundleMember = Bundle;
while (BundleMember) {
assert(BundleMember->FirstInBundle == Bundle && "corrupt bundle links");
BundleMember->FirstInBundle = BundleMember;
ScheduleData *Next = BundleMember->NextInBundle;
BundleMember->NextInBundle = nullptr;
BundleMember->UnscheduledDepsInBundle = BundleMember->UnscheduledDeps;
if (BundleMember->UnscheduledDepsInBundle == 0) {
ReadyInsts.insert(BundleMember);
}
BundleMember = Next;
}
}
bool BoUpSLP::BlockScheduling::extendSchedulingRegion(Value *V) {
if (getScheduleData(V))
return true;
Instruction *I = dyn_cast<Instruction>(V);
assert(I && "bundle member must be an instruction");
assert(!isa<PHINode>(I) && "phi nodes don't need to be scheduled");
if (!ScheduleStart) {
// It's the first instruction in the new region.
initScheduleData(I, I->getNextNode(), nullptr, nullptr);
ScheduleStart = I;
ScheduleEnd = I->getNextNode();
assert(ScheduleEnd && "tried to vectorize a TerminatorInst?");
DEBUG(dbgs() << "SLP: initialize schedule region to " << *I << "\n");
return true;
}
// Search up and down at the same time, because we don't know if the new
// instruction is above or below the existing scheduling region.
BasicBlock::reverse_iterator UpIter(ScheduleStart->getIterator());
BasicBlock::reverse_iterator UpperEnd = BB->rend();
BasicBlock::iterator DownIter(ScheduleEnd);
BasicBlock::iterator LowerEnd = BB->end();
for (;;) {
if (++ScheduleRegionSize > ScheduleRegionSizeLimit) {
DEBUG(dbgs() << "SLP: exceeded schedule region size limit\n");
return false;
}
if (UpIter != UpperEnd) {
if (&*UpIter == I) {
initScheduleData(I, ScheduleStart, nullptr, FirstLoadStoreInRegion);
ScheduleStart = I;
DEBUG(dbgs() << "SLP: extend schedule region start to " << *I << "\n");
return true;
}
UpIter++;
}
if (DownIter != LowerEnd) {
if (&*DownIter == I) {
initScheduleData(ScheduleEnd, I->getNextNode(), LastLoadStoreInRegion,
nullptr);
ScheduleEnd = I->getNextNode();
assert(ScheduleEnd && "tried to vectorize a TerminatorInst?");
DEBUG(dbgs() << "SLP: extend schedule region end to " << *I << "\n");
return true;
}
DownIter++;
}
assert((UpIter != UpperEnd || DownIter != LowerEnd) &&
"instruction not found in block");
}
return true;
}
void BoUpSLP::BlockScheduling::initScheduleData(Instruction *FromI,
Instruction *ToI,
ScheduleData *PrevLoadStore,
ScheduleData *NextLoadStore) {
ScheduleData *CurrentLoadStore = PrevLoadStore;
for (Instruction *I = FromI; I != ToI; I = I->getNextNode()) {
ScheduleData *SD = ScheduleDataMap[I];
if (!SD) {
// Allocate a new ScheduleData for the instruction.
if (ChunkPos >= ChunkSize) {
ScheduleDataChunks.push_back(
llvm::make_unique<ScheduleData[]>(ChunkSize));
ChunkPos = 0;
}
SD = &(ScheduleDataChunks.back()[ChunkPos++]);
ScheduleDataMap[I] = SD;
SD->Inst = I;
}
assert(!isInSchedulingRegion(SD) &&
"new ScheduleData already in scheduling region");
SD->init(SchedulingRegionID);
if (I->mayReadOrWriteMemory()) {
// Update the linked list of memory accessing instructions.
if (CurrentLoadStore) {
CurrentLoadStore->NextLoadStore = SD;
} else {
FirstLoadStoreInRegion = SD;
}
CurrentLoadStore = SD;
}
}
if (NextLoadStore) {
if (CurrentLoadStore)
CurrentLoadStore->NextLoadStore = NextLoadStore;
} else {
LastLoadStoreInRegion = CurrentLoadStore;
}
}
void BoUpSLP::BlockScheduling::calculateDependencies(ScheduleData *SD,
bool InsertInReadyList,
BoUpSLP *SLP) {
assert(SD->isSchedulingEntity());
SmallVector<ScheduleData *, 10> WorkList;
WorkList.push_back(SD);
while (!WorkList.empty()) {
ScheduleData *SD = WorkList.back();
WorkList.pop_back();
ScheduleData *BundleMember = SD;
while (BundleMember) {
assert(isInSchedulingRegion(BundleMember));
if (!BundleMember->hasValidDependencies()) {
DEBUG(dbgs() << "SLP: update deps of " << *BundleMember << "\n");
BundleMember->Dependencies = 0;
BundleMember->resetUnscheduledDeps();
// Handle def-use chain dependencies.
for (User *U : BundleMember->Inst->users()) {
if (isa<Instruction>(U)) {
ScheduleData *UseSD = getScheduleData(U);
if (UseSD && isInSchedulingRegion(UseSD->FirstInBundle)) {
BundleMember->Dependencies++;
ScheduleData *DestBundle = UseSD->FirstInBundle;
if (!DestBundle->IsScheduled) {
BundleMember->incrementUnscheduledDeps(1);
}
if (!DestBundle->hasValidDependencies()) {
WorkList.push_back(DestBundle);
}
}
} else {
// I'm not sure if this can ever happen. But we need to be safe.
// This lets the instruction/bundle never be scheduled and
// eventually disable vectorization.
BundleMember->Dependencies++;
BundleMember->incrementUnscheduledDeps(1);
}
}
// Handle the memory dependencies.
ScheduleData *DepDest = BundleMember->NextLoadStore;
if (DepDest) {
Instruction *SrcInst = BundleMember->Inst;
MemoryLocation SrcLoc = getLocation(SrcInst, SLP->AA);
bool SrcMayWrite = BundleMember->Inst->mayWriteToMemory();
unsigned numAliased = 0;
unsigned DistToSrc = 1;
while (DepDest) {
assert(isInSchedulingRegion(DepDest));
// We have two limits to reduce the complexity:
// 1) AliasedCheckLimit: It's a small limit to reduce calls to
// SLP->isAliased (which is the expensive part in this loop).
// 2) MaxMemDepDistance: It's for very large blocks and it aborts
// the whole loop (even if the loop is fast, it's quadratic).
// It's important for the loop break condition (see below) to
// check this limit even between two read-only instructions.
if (DistToSrc >= MaxMemDepDistance ||
((SrcMayWrite || DepDest->Inst->mayWriteToMemory()) &&
(numAliased >= AliasedCheckLimit ||
SLP->isAliased(SrcLoc, SrcInst, DepDest->Inst)))) {
// We increment the counter only if the locations are aliased
// (instead of counting all alias checks). This gives a better
// balance between reduced runtime and accurate dependencies.
numAliased++;
DepDest->MemoryDependencies.push_back(BundleMember);
BundleMember->Dependencies++;
ScheduleData *DestBundle = DepDest->FirstInBundle;
if (!DestBundle->IsScheduled) {
BundleMember->incrementUnscheduledDeps(1);
}
if (!DestBundle->hasValidDependencies()) {
WorkList.push_back(DestBundle);
}
}
DepDest = DepDest->NextLoadStore;
// Example, explaining the loop break condition: Let's assume our
// starting instruction is i0 and MaxMemDepDistance = 3.
//
// +--------v--v--v
// i0,i1,i2,i3,i4,i5,i6,i7,i8
// +--------^--^--^
//
// MaxMemDepDistance let us stop alias-checking at i3 and we add
// dependencies from i0 to i3,i4,.. (even if they are not aliased).
// Previously we already added dependencies from i3 to i6,i7,i8
// (because of MaxMemDepDistance). As we added a dependency from
// i0 to i3, we have transitive dependencies from i0 to i6,i7,i8
// and we can abort this loop at i6.
if (DistToSrc >= 2 * MaxMemDepDistance)
break;
DistToSrc++;
}
}
}
BundleMember = BundleMember->NextInBundle;
}
if (InsertInReadyList && SD->isReady()) {
ReadyInsts.push_back(SD);
DEBUG(dbgs() << "SLP: gets ready on update: " << *SD->Inst << "\n");
}
}
}
void BoUpSLP::BlockScheduling::resetSchedule() {
assert(ScheduleStart &&
"tried to reset schedule on block which has not been scheduled");
for (Instruction *I = ScheduleStart; I != ScheduleEnd; I = I->getNextNode()) {
ScheduleData *SD = getScheduleData(I);
assert(isInSchedulingRegion(SD));
SD->IsScheduled = false;
SD->resetUnscheduledDeps();
}
ReadyInsts.clear();
}
void BoUpSLP::scheduleBlock(BlockScheduling *BS) {
if (!BS->ScheduleStart)
return;
DEBUG(dbgs() << "SLP: schedule block " << BS->BB->getName() << "\n");
BS->resetSchedule();
// For the real scheduling we use a more sophisticated ready-list: it is
// sorted by the original instruction location. This lets the final schedule
// be as close as possible to the original instruction order.
struct ScheduleDataCompare {
bool operator()(ScheduleData *SD1, ScheduleData *SD2) {
return SD2->SchedulingPriority < SD1->SchedulingPriority;
}
};
std::set<ScheduleData *, ScheduleDataCompare> ReadyInsts;
// Ensure that all dependency data is updated and fill the ready-list with
// initial instructions.
int Idx = 0;
int NumToSchedule = 0;
for (auto *I = BS->ScheduleStart; I != BS->ScheduleEnd;
I = I->getNextNode()) {
ScheduleData *SD = BS->getScheduleData(I);
assert(
SD->isPartOfBundle() == (ScalarToTreeEntry.count(SD->Inst) != 0) &&
"scheduler and vectorizer have different opinion on what is a bundle");
SD->FirstInBundle->SchedulingPriority = Idx++;
if (SD->isSchedulingEntity()) {
BS->calculateDependencies(SD, false, this);
NumToSchedule++;
}
}
BS->initialFillReadyList(ReadyInsts);
Instruction *LastScheduledInst = BS->ScheduleEnd;
// Do the "real" scheduling.
while (!ReadyInsts.empty()) {
ScheduleData *picked = *ReadyInsts.begin();
ReadyInsts.erase(ReadyInsts.begin());
// Move the scheduled instruction(s) to their dedicated places, if not
// there yet.
ScheduleData *BundleMember = picked;
while (BundleMember) {
Instruction *pickedInst = BundleMember->Inst;
if (LastScheduledInst->getNextNode() != pickedInst) {
BS->BB->getInstList().remove(pickedInst);
BS->BB->getInstList().insert(LastScheduledInst->getIterator(),
pickedInst);
}
LastScheduledInst = pickedInst;
BundleMember = BundleMember->NextInBundle;
}
BS->schedule(picked, ReadyInsts);
NumToSchedule--;
}
assert(NumToSchedule == 0 && "could not schedule all instructions");
// Avoid duplicate scheduling of the block.
BS->ScheduleStart = nullptr;
}
unsigned BoUpSLP::getVectorElementSize(Value *V) {
auto &DL = F->getParent()->getDataLayout();
// If V is a store, just return the width of the stored value without
// traversing the expression tree. This is the common case.
if (auto *Store = dyn_cast<StoreInst>(V))
return DL.getTypeSizeInBits(Store->getValueOperand()->getType());
// If V is not a store, we can traverse the expression tree to find loads
// that feed it. The type of the loaded value may indicate a more suitable
// width than V's type. We want to base the vector element size on the width
// of memory operations where possible.
SmallVector<Instruction *, 16> Worklist;
SmallPtrSet<Instruction *, 16> Visited;
if (auto *I = dyn_cast<Instruction>(V))
Worklist.push_back(I);
// Traverse the expression tree in bottom-up order looking for loads. If we
// encounter an instruciton we don't yet handle, we give up.
auto MaxWidth = 0u;
auto FoundUnknownInst = false;
while (!Worklist.empty() && !FoundUnknownInst) {
auto *I = Worklist.pop_back_val();
Visited.insert(I);
// We should only be looking at scalar instructions here. If the current
// instruction has a vector type, give up.
auto *Ty = I->getType();
if (isa<VectorType>(Ty))
FoundUnknownInst = true;
// If the current instruction is a load, update MaxWidth to reflect the
// width of the loaded value.
else if (isa<LoadInst>(I))
MaxWidth = std::max<unsigned>(MaxWidth, DL.getTypeSizeInBits(Ty));
// Otherwise, we need to visit the operands of the instruction. We only
// handle the interesting cases from buildTree here. If an operand is an
// instruction we haven't yet visited, we add it to the worklist.
else if (isa<PHINode>(I) || isa<CastInst>(I) || isa<GetElementPtrInst>(I) ||
isa<CmpInst>(I) || isa<SelectInst>(I) || isa<BinaryOperator>(I)) {
for (Use &U : I->operands())
if (auto *J = dyn_cast<Instruction>(U.get()))
if (!Visited.count(J))
Worklist.push_back(J);
}
// If we don't yet handle the instruction, give up.
else
FoundUnknownInst = true;
}
// If we didn't encounter a memory access in the expression tree, or if we
// gave up for some reason, just return the width of V.
if (!MaxWidth || FoundUnknownInst)
return DL.getTypeSizeInBits(V->getType());
// Otherwise, return the maximum width we found.
return MaxWidth;
}
// Determine if a value V in a vectorizable expression Expr can be demoted to a
// smaller type with a truncation. We collect the values that will be demoted
// in ToDemote and additional roots that require investigating in Roots.
static bool collectValuesToDemote(Value *V, SmallPtrSetImpl<Value *> &Expr,
SmallVectorImpl<Value *> &ToDemote,
SmallVectorImpl<Value *> &Roots) {
// We can always demote constants.
if (isa<Constant>(V)) {
ToDemote.push_back(V);
return true;
}
// If the value is not an instruction in the expression with only one use, it
// cannot be demoted.
auto *I = dyn_cast<Instruction>(V);
if (!I || !I->hasOneUse() || !Expr.count(I))
return false;
switch (I->getOpcode()) {
// We can always demote truncations and extensions. Since truncations can
// seed additional demotion, we save the truncated value.
case Instruction::Trunc:
Roots.push_back(I->getOperand(0));
case Instruction::ZExt:
case Instruction::SExt:
break;
// We can demote certain binary operations if we can demote both of their
// operands.
case Instruction::Add:
case Instruction::Sub:
case Instruction::Mul:
case Instruction::And:
case Instruction::Or:
case Instruction::Xor:
if (!collectValuesToDemote(I->getOperand(0), Expr, ToDemote, Roots) ||
!collectValuesToDemote(I->getOperand(1), Expr, ToDemote, Roots))
return false;
break;
// We can demote selects if we can demote their true and false values.
case Instruction::Select: {
SelectInst *SI = cast<SelectInst>(I);
if (!collectValuesToDemote(SI->getTrueValue(), Expr, ToDemote, Roots) ||
!collectValuesToDemote(SI->getFalseValue(), Expr, ToDemote, Roots))
return false;
break;
}
// We can demote phis if we can demote all their incoming operands. Note that
// we don't need to worry about cycles since we ensure single use above.
case Instruction::PHI: {
PHINode *PN = cast<PHINode>(I);
for (Value *IncValue : PN->incoming_values())
if (!collectValuesToDemote(IncValue, Expr, ToDemote, Roots))
return false;
break;
}
// Otherwise, conservatively give up.
default:
return false;
}
// Record the value that we can demote.
ToDemote.push_back(V);
return true;
}
void BoUpSLP::computeMinimumValueSizes() {
auto &DL = F->getParent()->getDataLayout();
// If there are no external uses, the expression tree must be rooted by a
// store. We can't demote in-memory values, so there is nothing to do here.
if (ExternalUses.empty())
return;
// We only attempt to truncate integer expressions.
auto &TreeRoot = VectorizableTree[0].Scalars;
auto *TreeRootIT = dyn_cast<IntegerType>(TreeRoot[0]->getType());
if (!TreeRootIT)
return;
// If the expression is not rooted by a store, these roots should have
// external uses. We will rely on InstCombine to rewrite the expression in
// the narrower type. However, InstCombine only rewrites single-use values.
// This means that if a tree entry other than a root is used externally, it
// must have multiple uses and InstCombine will not rewrite it. The code
// below ensures that only the roots are used externally.
SmallPtrSet<Value *, 32> Expr(TreeRoot.begin(), TreeRoot.end());
for (auto &EU : ExternalUses)
if (!Expr.erase(EU.Scalar))
return;
if (!Expr.empty())
return;
// Collect the scalar values of the vectorizable expression. We will use this
// context to determine which values can be demoted. If we see a truncation,
// we mark it as seeding another demotion.
for (auto &Entry : VectorizableTree)
Expr.insert(Entry.Scalars.begin(), Entry.Scalars.end());
// Ensure the roots of the vectorizable tree don't form a cycle. They must
// have a single external user that is not in the vectorizable tree.
for (auto *Root : TreeRoot)
if (!Root->hasOneUse() || Expr.count(*Root->user_begin()))
return;
// Conservatively determine if we can actually truncate the roots of the
// expression. Collect the values that can be demoted in ToDemote and
// additional roots that require investigating in Roots.
SmallVector<Value *, 32> ToDemote;
SmallVector<Value *, 4> Roots;
for (auto *Root : TreeRoot)
if (!collectValuesToDemote(Root, Expr, ToDemote, Roots))
return;
// The maximum bit width required to represent all the values that can be
// demoted without loss of precision. It would be safe to truncate the roots
// of the expression to this width.
auto MaxBitWidth = 8u;
// We first check if all the bits of the roots are demanded. If they're not,
// we can truncate the roots to this narrower type.
for (auto *Root : TreeRoot) {
auto Mask = DB->getDemandedBits(cast<Instruction>(Root));
MaxBitWidth = std::max<unsigned>(
Mask.getBitWidth() - Mask.countLeadingZeros(), MaxBitWidth);
}
// If all the bits of the roots are demanded, we can try a little harder to
// compute a narrower type. This can happen, for example, if the roots are
// getelementptr indices. InstCombine promotes these indices to the pointer
// width. Thus, all their bits are technically demanded even though the
// address computation might be vectorized in a smaller type.
//
// We start by looking at each entry that can be demoted. We compute the
// maximum bit width required to store the scalar by using ValueTracking to
// compute the number of high-order bits we can truncate.
if (MaxBitWidth == DL.getTypeSizeInBits(TreeRoot[0]->getType())) {
MaxBitWidth = 8u;
for (auto *Scalar : ToDemote) {
auto NumSignBits = ComputeNumSignBits(Scalar, DL, 0, AC, 0, DT);
auto NumTypeBits = DL.getTypeSizeInBits(Scalar->getType());
MaxBitWidth = std::max<unsigned>(NumTypeBits - NumSignBits, MaxBitWidth);
}
}
// Round MaxBitWidth up to the next power-of-two.
if (!isPowerOf2_64(MaxBitWidth))
MaxBitWidth = NextPowerOf2(MaxBitWidth);
// If the maximum bit width we compute is less than the with of the roots'
// type, we can proceed with the narrowing. Otherwise, do nothing.
if (MaxBitWidth >= TreeRootIT->getBitWidth())
return;
// If we can truncate the root, we must collect additional values that might
// be demoted as a result. That is, those seeded by truncations we will
// modify.
while (!Roots.empty())
collectValuesToDemote(Roots.pop_back_val(), Expr, ToDemote, Roots);
// Finally, map the values we can demote to the maximum bit with we computed.
for (auto *Scalar : ToDemote)
MinBWs[Scalar] = MaxBitWidth;
}
/// The SLPVectorizer Pass.
struct SLPVectorizer : public FunctionPass {
typedef SmallVector<StoreInst *, 8> StoreList;
typedef MapVector<Value *, StoreList> StoreListMap;
typedef SmallVector<WeakVH, 8> WeakVHList;
typedef MapVector<Value *, WeakVHList> WeakVHListMap;
/// Pass identification, replacement for typeid
static char ID;
explicit SLPVectorizer() : FunctionPass(ID) {
initializeSLPVectorizerPass(*PassRegistry::getPassRegistry());
}
ScalarEvolution *SE;
TargetTransformInfo *TTI;
TargetLibraryInfo *TLI;
AliasAnalysis *AA;
LoopInfo *LI;
DominatorTree *DT;
AssumptionCache *AC;
DemandedBits *DB;
bool runOnFunction(Function &F) override {
if (skipOptnoneFunction(F))
return false;
SE = &getAnalysis<ScalarEvolutionWrapperPass>().getSE();
TTI = &getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F);
auto *TLIP = getAnalysisIfAvailable<TargetLibraryInfoWrapperPass>();
TLI = TLIP ? &TLIP->getTLI() : nullptr;
AA = &getAnalysis<AAResultsWrapperPass>().getAAResults();
LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
DT = &getAnalysis<DominatorTreeWrapperPass>().getDomTree();
AC = &getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
DB = &getAnalysis<DemandedBits>();
Stores.clear();
GEPs.clear();
bool Changed = false;
// If the target claims to have no vector registers don't attempt
// vectorization.
if (!TTI->getNumberOfRegisters(true))
return false;
// Use the vector register size specified by the target unless overridden
// by a command-line option.
// TODO: It would be better to limit the vectorization factor based on
// data type rather than just register size. For example, x86 AVX has
// 256-bit registers, but it does not support integer operations
// at that width (that requires AVX2).
if (MaxVectorRegSizeOption.getNumOccurrences())
MaxVecRegSize = MaxVectorRegSizeOption;
else
MaxVecRegSize = TTI->getRegisterBitWidth(true);
// Don't vectorize when the attribute NoImplicitFloat is used.
if (F.hasFnAttribute(Attribute::NoImplicitFloat))
return false;
DEBUG(dbgs() << "SLP: Analyzing blocks in " << F.getName() << ".\n");
// Use the bottom up slp vectorizer to construct chains that start with
// store instructions.
BoUpSLP R(&F, SE, TTI, TLI, AA, LI, DT, AC, DB);
// A general note: the vectorizer must use BoUpSLP::eraseInstruction() to
// delete instructions.
// Scan the blocks in the function in post order.
for (auto BB : post_order(&F.getEntryBlock())) {
collectSeedInstructions(BB);
// Vectorize trees that end at stores.
if (NumStores > 0) {
DEBUG(dbgs() << "SLP: Found " << NumStores << " stores.\n");
Changed |= vectorizeStoreChains(R);
}
// Vectorize trees that end at reductions.
Changed |= vectorizeChainsInBlock(BB, R);
// Vectorize the index computations of getelementptr instructions. This
// is primarily intended to catch gather-like idioms ending at
// non-consecutive loads.
if (NumGEPs > 0) {
DEBUG(dbgs() << "SLP: Found " << NumGEPs << " GEPs.\n");
Changed |= vectorizeGEPIndices(BB, R);
}
}
if (Changed) {
R.optimizeGatherSequence();
DEBUG(dbgs() << "SLP: vectorized \"" << F.getName() << "\"\n");
DEBUG(verifyFunction(F));
}
return Changed;
}
void getAnalysisUsage(AnalysisUsage &AU) const override {
FunctionPass::getAnalysisUsage(AU);
AU.addRequired<AssumptionCacheTracker>();
AU.addRequired<ScalarEvolutionWrapperPass>();
AU.addRequired<AAResultsWrapperPass>();
AU.addRequired<TargetTransformInfoWrapperPass>();
AU.addRequired<LoopInfoWrapperPass>();
AU.addRequired<DominatorTreeWrapperPass>();
AU.addRequired<DemandedBits>();
AU.addPreserved<LoopInfoWrapperPass>();
AU.addPreserved<DominatorTreeWrapperPass>();
AU.addPreserved<AAResultsWrapperPass>();
AU.addPreserved<GlobalsAAWrapperPass>();
AU.setPreservesCFG();
}
private:
/// \brief Collect store and getelementptr instructions and organize them
/// according to the underlying object of their pointer operands. We sort the
/// instructions by their underlying objects to reduce the cost of
/// consecutive access queries.
///
/// TODO: We can further reduce this cost if we flush the chain creation
/// every time we run into a memory barrier.
void collectSeedInstructions(BasicBlock *BB);
/// \brief Try to vectorize a chain that starts at two arithmetic instrs.
bool tryToVectorizePair(Value *A, Value *B, BoUpSLP &R);
/// \brief Try to vectorize a list of operands.
/// \@param BuildVector A list of users to ignore for the purpose of
/// scheduling and that don't need extracting.
/// \returns true if a value was vectorized.
bool tryToVectorizeList(ArrayRef<Value *> VL, BoUpSLP &R,
ArrayRef<Value *> BuildVector = None,
bool allowReorder = false);
/// \brief Try to vectorize a chain that may start at the operands of \V;
bool tryToVectorize(BinaryOperator *V, BoUpSLP &R);
/// \brief Vectorize the store instructions collected in Stores.
bool vectorizeStoreChains(BoUpSLP &R);
/// \brief Vectorize the index computations of the getelementptr instructions
/// collected in GEPs.
bool vectorizeGEPIndices(BasicBlock *BB, BoUpSLP &R);
/// \brief Scan the basic block and look for patterns that are likely to start
/// a vectorization chain.
bool vectorizeChainsInBlock(BasicBlock *BB, BoUpSLP &R);
bool vectorizeStoreChain(ArrayRef<Value *> Chain, int CostThreshold,
BoUpSLP &R, unsigned VecRegSize);
bool vectorizeStores(ArrayRef<StoreInst *> Stores, int costThreshold,
BoUpSLP &R);
/// The store instructions in a basic block organized by base pointer.
StoreListMap Stores;
/// The getelementptr instructions in a basic block organized by base pointer.
WeakVHListMap GEPs;
/// The number of store instructions in a basic block.
unsigned NumStores;
/// The number of getelementptr instructions in a basic block.
unsigned NumGEPs;
unsigned MaxVecRegSize; // This is set by TTI or overridden by cl::opt.
};
/// \brief Check that the Values in the slice in VL array are still existent in
/// the WeakVH array.
/// Vectorization of part of the VL array may cause later values in the VL array
/// to become invalid. We track when this has happened in the WeakVH array.
static bool hasValueBeenRAUWed(ArrayRef<Value *> VL, ArrayRef<WeakVH> VH,
unsigned SliceBegin, unsigned SliceSize) {
VL = VL.slice(SliceBegin, SliceSize);
VH = VH.slice(SliceBegin, SliceSize);
return !std::equal(VL.begin(), VL.end(), VH.begin());
}
bool SLPVectorizer::vectorizeStoreChain(ArrayRef<Value *> Chain,
int CostThreshold, BoUpSLP &R,
unsigned VecRegSize) {
unsigned ChainLen = Chain.size();
DEBUG(dbgs() << "SLP: Analyzing a store chain of length " << ChainLen
<< "\n");
unsigned Sz = R.getVectorElementSize(Chain[0]);
unsigned VF = VecRegSize / Sz;
if (!isPowerOf2_32(Sz) || VF < 2)
return false;
// Keep track of values that were deleted by vectorizing in the loop below.
SmallVector<WeakVH, 8> TrackValues(Chain.begin(), Chain.end());
bool Changed = false;
// Look for profitable vectorizable trees at all offsets, starting at zero.
for (unsigned i = 0, e = ChainLen; i < e; ++i) {
if (i + VF > e)
break;
// Check that a previous iteration of this loop did not delete the Value.
if (hasValueBeenRAUWed(Chain, TrackValues, i, VF))
continue;
DEBUG(dbgs() << "SLP: Analyzing " << VF << " stores at offset " << i
<< "\n");
ArrayRef<Value *> Operands = Chain.slice(i, VF);
R.buildTree(Operands);
R.computeMinimumValueSizes();
int Cost = R.getTreeCost();
DEBUG(dbgs() << "SLP: Found cost=" << Cost << " for VF=" << VF << "\n");
if (Cost < CostThreshold) {
DEBUG(dbgs() << "SLP: Decided to vectorize cost=" << Cost << "\n");
R.vectorizeTree();
// Move to the next bundle.
i += VF - 1;
Changed = true;
}
}
return Changed;
}
bool SLPVectorizer::vectorizeStores(ArrayRef<StoreInst *> Stores,
int costThreshold, BoUpSLP &R) {
SetVector<StoreInst *> Heads, Tails;
SmallDenseMap<StoreInst *, StoreInst *> ConsecutiveChain;
// We may run into multiple chains that merge into a single chain. We mark the
// stores that we vectorized so that we don't visit the same store twice.
BoUpSLP::ValueSet VectorizedStores;
bool Changed = false;
// Do a quadratic search on all of the given stores and find
// all of the pairs of stores that follow each other.
SmallVector<unsigned, 16> IndexQueue;
for (unsigned i = 0, e = Stores.size(); i < e; ++i) {
const DataLayout &DL = Stores[i]->getModule()->getDataLayout();
IndexQueue.clear();
// If a store has multiple consecutive store candidates, search Stores
// array according to the sequence: from i+1 to e, then from i-1 to 0.
// This is because usually pairing with immediate succeeding or preceding
// candidate create the best chance to find slp vectorization opportunity.
unsigned j = 0;
for (j = i + 1; j < e; ++j)
IndexQueue.push_back(j);
for (j = i; j > 0; --j)
IndexQueue.push_back(j - 1);
for (auto &k : IndexQueue) {
if (isConsecutiveAccess(Stores[i], Stores[k], DL, *SE)) {
Tails.insert(Stores[k]);
Heads.insert(Stores[i]);
ConsecutiveChain[Stores[i]] = Stores[k];
break;
}
}
}
// For stores that start but don't end a link in the chain:
for (SetVector<StoreInst *>::iterator it = Heads.begin(), e = Heads.end();
it != e; ++it) {
if (Tails.count(*it))
continue;
// We found a store instr that starts a chain. Now follow the chain and try
// to vectorize it.
BoUpSLP::ValueList Operands;
StoreInst *I = *it;
// Collect the chain into a list.
while (Tails.count(I) || Heads.count(I)) {
if (VectorizedStores.count(I))
break;
Operands.push_back(I);
// Move to the next value in the chain.
I = ConsecutiveChain[I];
}
// FIXME: Is division-by-2 the correct step? Should we assert that the
// register size is a power-of-2?
for (unsigned Size = MaxVecRegSize; Size >= MinVecRegSize; Size /= 2) {
if (vectorizeStoreChain(Operands, costThreshold, R, Size)) {
// Mark the vectorized stores so that we don't vectorize them again.
VectorizedStores.insert(Operands.begin(), Operands.end());
Changed = true;
break;
}
}
}
return Changed;
}
void SLPVectorizer::collectSeedInstructions(BasicBlock *BB) {
// Initialize the collections. We will make a single pass over the block.
Stores.clear();
GEPs.clear();
NumStores = NumGEPs = 0;
const DataLayout &DL = BB->getModule()->getDataLayout();
// Visit the store and getelementptr instructions in BB and organize them in
// Stores and GEPs according to the underlying objects of their pointer
// operands.
for (Instruction &I : *BB) {
// Ignore store instructions that are volatile or have a pointer operand
// that doesn't point to a scalar type.
if (auto *SI = dyn_cast<StoreInst>(&I)) {
if (!SI->isSimple())
continue;
if (!isValidElementType(SI->getValueOperand()->getType()))
continue;
Stores[GetUnderlyingObject(SI->getPointerOperand(), DL)].push_back(SI);
++NumStores;
}
// Ignore getelementptr instructions that have more than one index, a
// constant index, or a pointer operand that doesn't point to a scalar
// type.
else if (auto *GEP = dyn_cast<GetElementPtrInst>(&I)) {
auto Idx = GEP->idx_begin()->get();
if (GEP->getNumIndices() > 1 || isa<Constant>(Idx))
continue;
if (!isValidElementType(Idx->getType()))
continue;
GEPs[GetUnderlyingObject(GEP->getPointerOperand(), DL)].push_back(GEP);
++NumGEPs;
}
}
}
bool SLPVectorizer::tryToVectorizePair(Value *A, Value *B, BoUpSLP &R) {
if (!A || !B)
return false;
Value *VL[] = { A, B };
return tryToVectorizeList(VL, R, None, true);
}
bool SLPVectorizer::tryToVectorizeList(ArrayRef<Value *> VL, BoUpSLP &R,
ArrayRef<Value *> BuildVector,
bool allowReorder) {
if (VL.size() < 2)
return false;
DEBUG(dbgs() << "SLP: Vectorizing a list of length = " << VL.size() << ".\n");
// Check that all of the parts are scalar instructions of the same type.
Instruction *I0 = dyn_cast<Instruction>(VL[0]);
if (!I0)
return false;
unsigned Opcode0 = I0->getOpcode();
// FIXME: Register size should be a parameter to this function, so we can
// try different vectorization factors.
unsigned Sz = R.getVectorElementSize(I0);
unsigned VF = MinVecRegSize / Sz;
for (Value *V : VL) {
Type *Ty = V->getType();
if (!isValidElementType(Ty))
return false;
Instruction *Inst = dyn_cast<Instruction>(V);
if (!Inst || Inst->getOpcode() != Opcode0)
return false;
}
bool Changed = false;
// Keep track of values that were deleted by vectorizing in the loop below.
SmallVector<WeakVH, 8> TrackValues(VL.begin(), VL.end());
for (unsigned i = 0, e = VL.size(); i < e; ++i) {
unsigned OpsWidth = 0;
if (i + VF > e)
OpsWidth = e - i;
else
OpsWidth = VF;
if (!isPowerOf2_32(OpsWidth) || OpsWidth < 2)
break;
// Check that a previous iteration of this loop did not delete the Value.
if (hasValueBeenRAUWed(VL, TrackValues, i, OpsWidth))
continue;
DEBUG(dbgs() << "SLP: Analyzing " << OpsWidth << " operations "
<< "\n");
ArrayRef<Value *> Ops = VL.slice(i, OpsWidth);
ArrayRef<Value *> BuildVectorSlice;
if (!BuildVector.empty())
BuildVectorSlice = BuildVector.slice(i, OpsWidth);
R.buildTree(Ops, BuildVectorSlice);
// TODO: check if we can allow reordering also for other cases than
// tryToVectorizePair()
if (allowReorder && R.shouldReorder()) {
assert(Ops.size() == 2);
assert(BuildVectorSlice.empty());
Value *ReorderedOps[] = { Ops[1], Ops[0] };
R.buildTree(ReorderedOps, None);
}
R.computeMinimumValueSizes();
int Cost = R.getTreeCost();
if (Cost < -SLPCostThreshold) {
DEBUG(dbgs() << "SLP: Vectorizing list at cost:" << Cost << ".\n");
Value *VectorizedRoot = R.vectorizeTree();
// Reconstruct the build vector by extracting the vectorized root. This
// way we handle the case where some elements of the vector are undefined.
// (return (inserelt <4 xi32> (insertelt undef (opd0) 0) (opd1) 2))
if (!BuildVectorSlice.empty()) {
// The insert point is the last build vector instruction. The vectorized
// root will precede it. This guarantees that we get an instruction. The
// vectorized tree could have been constant folded.
Instruction *InsertAfter = cast<Instruction>(BuildVectorSlice.back());
unsigned VecIdx = 0;
for (auto &V : BuildVectorSlice) {
IRBuilder<true, NoFolder> Builder(
InsertAfter->getParent(), ++BasicBlock::iterator(InsertAfter));
InsertElementInst *IE = cast<InsertElementInst>(V);
Instruction *Extract = cast<Instruction>(Builder.CreateExtractElement(
VectorizedRoot, Builder.getInt32(VecIdx++)));
IE->setOperand(1, Extract);
IE->removeFromParent();
IE->insertAfter(Extract);
InsertAfter = IE;
}
}
// Move to the next bundle.
i += VF - 1;
Changed = true;
}
}
return Changed;
}
bool SLPVectorizer::tryToVectorize(BinaryOperator *V, BoUpSLP &R) {
if (!V)
return false;
// Try to vectorize V.
if (tryToVectorizePair(V->getOperand(0), V->getOperand(1), R))
return true;
BinaryOperator *A = dyn_cast<BinaryOperator>(V->getOperand(0));
BinaryOperator *B = dyn_cast<BinaryOperator>(V->getOperand(1));
// Try to skip B.
if (B && B->hasOneUse()) {
BinaryOperator *B0 = dyn_cast<BinaryOperator>(B->getOperand(0));
BinaryOperator *B1 = dyn_cast<BinaryOperator>(B->getOperand(1));
if (tryToVectorizePair(A, B0, R)) {
return true;
}
if (tryToVectorizePair(A, B1, R)) {
return true;
}
}
// Try to skip A.
if (A && A->hasOneUse()) {
BinaryOperator *A0 = dyn_cast<BinaryOperator>(A->getOperand(0));
BinaryOperator *A1 = dyn_cast<BinaryOperator>(A->getOperand(1));
if (tryToVectorizePair(A0, B, R)) {
return true;
}
if (tryToVectorizePair(A1, B, R)) {
return true;
}
}
return 0;
}
/// \brief Generate a shuffle mask to be used in a reduction tree.
///
/// \param VecLen The length of the vector to be reduced.
/// \param NumEltsToRdx The number of elements that should be reduced in the
/// vector.
/// \param IsPairwise Whether the reduction is a pairwise or splitting
/// reduction. A pairwise reduction will generate a mask of
/// <0,2,...> or <1,3,..> while a splitting reduction will generate
/// <2,3, undef,undef> for a vector of 4 and NumElts = 2.
/// \param IsLeft True will generate a mask of even elements, odd otherwise.
static Value *createRdxShuffleMask(unsigned VecLen, unsigned NumEltsToRdx,
bool IsPairwise, bool IsLeft,
IRBuilder<> &Builder) {
assert((IsPairwise || !IsLeft) && "Don't support a <0,1,undef,...> mask");
SmallVector<Constant *, 32> ShuffleMask(
VecLen, UndefValue::get(Builder.getInt32Ty()));
if (IsPairwise)
// Build a mask of 0, 2, ... (left) or 1, 3, ... (right).
for (unsigned i = 0; i != NumEltsToRdx; ++i)
ShuffleMask[i] = Builder.getInt32(2 * i + !IsLeft);
else
// Move the upper half of the vector to the lower half.
for (unsigned i = 0; i != NumEltsToRdx; ++i)
ShuffleMask[i] = Builder.getInt32(NumEltsToRdx + i);
return ConstantVector::get(ShuffleMask);
}
/// Model horizontal reductions.
///
/// A horizontal reduction is a tree of reduction operations (currently add and
/// fadd) that has operations that can be put into a vector as its leaf.
/// For example, this tree:
///
/// mul mul mul mul
/// \ / \ /
/// + +
/// \ /
/// +
/// This tree has "mul" as its reduced values and "+" as its reduction
/// operations. A reduction might be feeding into a store or a binary operation
/// feeding a phi.
/// ...
/// \ /
/// +
/// |
/// phi +=
///
/// Or:
/// ...
/// \ /
/// +
/// |
/// *p =
///
class HorizontalReduction {
SmallVector<Value *, 16> ReductionOps;
SmallVector<Value *, 32> ReducedVals;
BinaryOperator *ReductionRoot;
PHINode *ReductionPHI;
/// The opcode of the reduction.
unsigned ReductionOpcode;
/// The opcode of the values we perform a reduction on.
unsigned ReducedValueOpcode;
/// Should we model this reduction as a pairwise reduction tree or a tree that
/// splits the vector in halves and adds those halves.
bool IsPairwiseReduction;
public:
/// The width of one full horizontal reduction operation.
unsigned ReduxWidth;
HorizontalReduction()
: ReductionRoot(nullptr), ReductionPHI(nullptr), ReductionOpcode(0),
ReducedValueOpcode(0), IsPairwiseReduction(false), ReduxWidth(0) {}
/// \brief Try to find a reduction tree.
bool matchAssociativeReduction(PHINode *Phi, BinaryOperator *B) {
assert((!Phi ||
std::find(Phi->op_begin(), Phi->op_end(), B) != Phi->op_end()) &&
"Thi phi needs to use the binary operator");
// We could have a initial reductions that is not an add.
// r *= v1 + v2 + v3 + v4
// In such a case start looking for a tree rooted in the first '+'.
if (Phi) {
if (B->getOperand(0) == Phi) {
Phi = nullptr;
B = dyn_cast<BinaryOperator>(B->getOperand(1));
} else if (B->getOperand(1) == Phi) {
Phi = nullptr;
B = dyn_cast<BinaryOperator>(B->getOperand(0));
}
}
if (!B)
return false;
Type *Ty = B->getType();
if (!isValidElementType(Ty))
return false;
const DataLayout &DL = B->getModule()->getDataLayout();
ReductionOpcode = B->getOpcode();
ReducedValueOpcode = 0;
// FIXME: Register size should be a parameter to this function, so we can
// try different vectorization factors.
ReduxWidth = MinVecRegSize / DL.getTypeSizeInBits(Ty);
ReductionRoot = B;
ReductionPHI = Phi;
if (ReduxWidth < 4)
return false;
// We currently only support adds.
if (ReductionOpcode != Instruction::Add &&
ReductionOpcode != Instruction::FAdd)
return false;
// Post order traverse the reduction tree starting at B. We only handle true
// trees containing only binary operators or selects.
SmallVector<std::pair<Instruction *, unsigned>, 32> Stack;
Stack.push_back(std::make_pair(B, 0));
while (!Stack.empty()) {
Instruction *TreeN = Stack.back().first;
unsigned EdgeToVist = Stack.back().second++;
bool IsReducedValue = TreeN->getOpcode() != ReductionOpcode;
// Only handle trees in the current basic block.
if (TreeN->getParent() != B->getParent())
return false;
// Each tree node needs to have one user except for the ultimate
// reduction.
if (!TreeN->hasOneUse() && TreeN != B)
return false;
// Postorder vist.
if (EdgeToVist == 2 || IsReducedValue) {
if (IsReducedValue) {
// Make sure that the opcodes of the operations that we are going to
// reduce match.
if (!ReducedValueOpcode)
ReducedValueOpcode = TreeN->getOpcode();
else if (ReducedValueOpcode != TreeN->getOpcode())
return false;
ReducedVals.push_back(TreeN);
} else {
// We need to be able to reassociate the adds.
if (!TreeN->isAssociative())
return false;
ReductionOps.push_back(TreeN);
}
// Retract.
Stack.pop_back();
continue;
}
// Visit left or right.
Value *NextV = TreeN->getOperand(EdgeToVist);
// We currently only allow BinaryOperator's and SelectInst's as reduction
// values in our tree.
if (isa<BinaryOperator>(NextV) || isa<SelectInst>(NextV))
Stack.push_back(std::make_pair(cast<Instruction>(NextV), 0));
else if (NextV != Phi)
return false;
}
return true;
}
/// \brief Attempt to vectorize the tree found by
/// matchAssociativeReduction.
bool tryToReduce(BoUpSLP &V, TargetTransformInfo *TTI) {
if (ReducedVals.empty())
return false;
unsigned NumReducedVals = ReducedVals.size();
if (NumReducedVals < ReduxWidth)
return false;
Value *VectorizedTree = nullptr;
IRBuilder<> Builder(ReductionRoot);
FastMathFlags Unsafe;
Unsafe.setUnsafeAlgebra();
Builder.setFastMathFlags(Unsafe);
unsigned i = 0;
for (; i < NumReducedVals - ReduxWidth + 1; i += ReduxWidth) {
V.buildTree(makeArrayRef(&ReducedVals[i], ReduxWidth), ReductionOps);
V.computeMinimumValueSizes();
// Estimate cost.
int Cost = V.getTreeCost() + getReductionCost(TTI, ReducedVals[i]);
if (Cost >= -SLPCostThreshold)
break;
DEBUG(dbgs() << "SLP: Vectorizing horizontal reduction at cost:" << Cost
<< ". (HorRdx)\n");
// Vectorize a tree.
DebugLoc Loc = cast<Instruction>(ReducedVals[i])->getDebugLoc();
Value *VectorizedRoot = V.vectorizeTree();
// Emit a reduction.
Value *ReducedSubTree = emitReduction(VectorizedRoot, Builder);
if (VectorizedTree) {
Builder.SetCurrentDebugLocation(Loc);
VectorizedTree = createBinOp(Builder, ReductionOpcode, VectorizedTree,
ReducedSubTree, "bin.rdx");
} else
VectorizedTree = ReducedSubTree;
}
if (VectorizedTree) {
// Finish the reduction.
for (; i < NumReducedVals; ++i) {
Builder.SetCurrentDebugLocation(
cast<Instruction>(ReducedVals[i])->getDebugLoc());
VectorizedTree = createBinOp(Builder, ReductionOpcode, VectorizedTree,
ReducedVals[i]);
}
// Update users.
if (ReductionPHI) {
assert(ReductionRoot && "Need a reduction operation");
ReductionRoot->setOperand(0, VectorizedTree);
ReductionRoot->setOperand(1, ReductionPHI);
} else
ReductionRoot->replaceAllUsesWith(VectorizedTree);
}
return VectorizedTree != nullptr;
}
unsigned numReductionValues() const {
return ReducedVals.size();
}
private:
/// \brief Calculate the cost of a reduction.
int getReductionCost(TargetTransformInfo *TTI, Value *FirstReducedVal) {
Type *ScalarTy = FirstReducedVal->getType();
Type *VecTy = VectorType::get(ScalarTy, ReduxWidth);
int PairwiseRdxCost = TTI->getReductionCost(ReductionOpcode, VecTy, true);
int SplittingRdxCost = TTI->getReductionCost(ReductionOpcode, VecTy, false);
IsPairwiseReduction = PairwiseRdxCost < SplittingRdxCost;
int VecReduxCost = IsPairwiseReduction ? PairwiseRdxCost : SplittingRdxCost;
int ScalarReduxCost =
ReduxWidth * TTI->getArithmeticInstrCost(ReductionOpcode, VecTy);
DEBUG(dbgs() << "SLP: Adding cost " << VecReduxCost - ScalarReduxCost
<< " for reduction that starts with " << *FirstReducedVal
<< " (It is a "
<< (IsPairwiseReduction ? "pairwise" : "splitting")
<< " reduction)\n");
return VecReduxCost - ScalarReduxCost;
}
static Value *createBinOp(IRBuilder<> &Builder, unsigned Opcode, Value *L,
Value *R, const Twine &Name = "") {
if (Opcode == Instruction::FAdd)
return Builder.CreateFAdd(L, R, Name);
return Builder.CreateBinOp((Instruction::BinaryOps)Opcode, L, R, Name);
}
/// \brief Emit a horizontal reduction of the vectorized value.
Value *emitReduction(Value *VectorizedValue, IRBuilder<> &Builder) {
assert(VectorizedValue && "Need to have a vectorized tree node");
assert(isPowerOf2_32(ReduxWidth) &&
"We only handle power-of-two reductions for now");
Value *TmpVec = VectorizedValue;
for (unsigned i = ReduxWidth / 2; i != 0; i >>= 1) {
if (IsPairwiseReduction) {
Value *LeftMask =
createRdxShuffleMask(ReduxWidth, i, true, true, Builder);
Value *RightMask =
createRdxShuffleMask(ReduxWidth, i, true, false, Builder);
Value *LeftShuf = Builder.CreateShuffleVector(
TmpVec, UndefValue::get(TmpVec->getType()), LeftMask, "rdx.shuf.l");
Value *RightShuf = Builder.CreateShuffleVector(
TmpVec, UndefValue::get(TmpVec->getType()), (RightMask),
"rdx.shuf.r");
TmpVec = createBinOp(Builder, ReductionOpcode, LeftShuf, RightShuf,
"bin.rdx");
} else {
Value *UpperHalf =
createRdxShuffleMask(ReduxWidth, i, false, false, Builder);
Value *Shuf = Builder.CreateShuffleVector(
TmpVec, UndefValue::get(TmpVec->getType()), UpperHalf, "rdx.shuf");
TmpVec = createBinOp(Builder, ReductionOpcode, TmpVec, Shuf, "bin.rdx");
}
}
// The result is in the first element of the vector.
return Builder.CreateExtractElement(TmpVec, Builder.getInt32(0));
}
};
/// \brief Recognize construction of vectors like
/// %ra = insertelement <4 x float> undef, float %s0, i32 0
/// %rb = insertelement <4 x float> %ra, float %s1, i32 1
/// %rc = insertelement <4 x float> %rb, float %s2, i32 2
/// %rd = insertelement <4 x float> %rc, float %s3, i32 3
///
/// Returns true if it matches
///
static bool findBuildVector(InsertElementInst *FirstInsertElem,
SmallVectorImpl<Value *> &BuildVector,
SmallVectorImpl<Value *> &BuildVectorOpds) {
if (!isa<UndefValue>(FirstInsertElem->getOperand(0)))
return false;
InsertElementInst *IE = FirstInsertElem;
while (true) {
BuildVector.push_back(IE);
BuildVectorOpds.push_back(IE->getOperand(1));
if (IE->use_empty())
return false;
InsertElementInst *NextUse = dyn_cast<InsertElementInst>(IE->user_back());
if (!NextUse)
return true;
// If this isn't the final use, make sure the next insertelement is the only
// use. It's OK if the final constructed vector is used multiple times
if (!IE->hasOneUse())
return false;
IE = NextUse;
}
return false;
}
static bool PhiTypeSorterFunc(Value *V, Value *V2) {
return V->getType() < V2->getType();
}
/// \brief Try and get a reduction value from a phi node.
///
/// Given a phi node \p P in a block \p ParentBB, consider possible reductions
/// if they come from either \p ParentBB or a containing loop latch.
///
/// \returns A candidate reduction value if possible, or \code nullptr \endcode
/// if not possible.
static Value *getReductionValue(const DominatorTree *DT, PHINode *P,
BasicBlock *ParentBB, LoopInfo *LI) {
// There are situations where the reduction value is not dominated by the
// reduction phi. Vectorizing such cases has been reported to cause
// miscompiles. See PR25787.
auto DominatedReduxValue = [&](Value *R) {
return (
dyn_cast<Instruction>(R) &&
DT->dominates(P->getParent(), dyn_cast<Instruction>(R)->getParent()));
};
Value *Rdx = nullptr;
// Return the incoming value if it comes from the same BB as the phi node.
if (P->getIncomingBlock(0) == ParentBB) {
Rdx = P->getIncomingValue(0);
} else if (P->getIncomingBlock(1) == ParentBB) {
Rdx = P->getIncomingValue(1);
}
if (Rdx && DominatedReduxValue(Rdx))
return Rdx;
// Otherwise, check whether we have a loop latch to look at.
Loop *BBL = LI->getLoopFor(ParentBB);
if (!BBL)
return nullptr;
BasicBlock *BBLatch = BBL->getLoopLatch();
if (!BBLatch)
return nullptr;
// There is a loop latch, return the incoming value if it comes from
// that. This reduction pattern occassionaly turns up.
if (P->getIncomingBlock(0) == BBLatch) {
Rdx = P->getIncomingValue(0);
} else if (P->getIncomingBlock(1) == BBLatch) {
Rdx = P->getIncomingValue(1);
}
if (Rdx && DominatedReduxValue(Rdx))
return Rdx;
return nullptr;
}
/// \brief Attempt to reduce a horizontal reduction.
/// If it is legal to match a horizontal reduction feeding
/// the phi node P with reduction operators BI, then check if it
/// can be done.
/// \returns true if a horizontal reduction was matched and reduced.
/// \returns false if a horizontal reduction was not matched.
static bool canMatchHorizontalReduction(PHINode *P, BinaryOperator *BI,
BoUpSLP &R, TargetTransformInfo *TTI) {
if (!ShouldVectorizeHor)
return false;
HorizontalReduction HorRdx;
if (!HorRdx.matchAssociativeReduction(P, BI))
return false;
// If there is a sufficient number of reduction values, reduce
// to a nearby power-of-2. Can safely generate oversized
// vectors and rely on the backend to split them to legal sizes.
HorRdx.ReduxWidth =
std::max((uint64_t)4, PowerOf2Floor(HorRdx.numReductionValues()));
return HorRdx.tryToReduce(R, TTI);
}
bool SLPVectorizer::vectorizeChainsInBlock(BasicBlock *BB, BoUpSLP &R) {
bool Changed = false;
SmallVector<Value *, 4> Incoming;
SmallSet<Value *, 16> VisitedInstrs;
bool HaveVectorizedPhiNodes = true;
while (HaveVectorizedPhiNodes) {
HaveVectorizedPhiNodes = false;
// Collect the incoming values from the PHIs.
Incoming.clear();
for (Instruction &I : *BB) {
PHINode *P = dyn_cast<PHINode>(&I);
if (!P)
break;
if (!VisitedInstrs.count(P))
Incoming.push_back(P);
}
// Sort by type.
std::stable_sort(Incoming.begin(), Incoming.end(), PhiTypeSorterFunc);
// Try to vectorize elements base on their type.
for (SmallVector<Value *, 4>::iterator IncIt = Incoming.begin(),
E = Incoming.end();
IncIt != E;) {
// Look for the next elements with the same type.
SmallVector<Value *, 4>::iterator SameTypeIt = IncIt;
while (SameTypeIt != E &&
(*SameTypeIt)->getType() == (*IncIt)->getType()) {
VisitedInstrs.insert(*SameTypeIt);
++SameTypeIt;
}
// Try to vectorize them.
unsigned NumElts = (SameTypeIt - IncIt);
DEBUG(errs() << "SLP: Trying to vectorize starting at PHIs (" << NumElts << ")\n");
if (NumElts > 1 && tryToVectorizeList(makeArrayRef(IncIt, NumElts), R)) {
// Success start over because instructions might have been changed.
HaveVectorizedPhiNodes = true;
Changed = true;
break;
}
// Start over at the next instruction of a different type (or the end).
IncIt = SameTypeIt;
}
}
VisitedInstrs.clear();
for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; it++) {
// We may go through BB multiple times so skip the one we have checked.
if (!VisitedInstrs.insert(&*it).second)
continue;
if (isa<DbgInfoIntrinsic>(it))
continue;
// Try to vectorize reductions that use PHINodes.
if (PHINode *P = dyn_cast<PHINode>(it)) {
// Check that the PHI is a reduction PHI.
if (P->getNumIncomingValues() != 2)
return Changed;
Value *Rdx = getReductionValue(DT, P, BB, LI);
// Check if this is a Binary Operator.
BinaryOperator *BI = dyn_cast_or_null<BinaryOperator>(Rdx);
if (!BI)
continue;
// Try to match and vectorize a horizontal reduction.
if (canMatchHorizontalReduction(P, BI, R, TTI)) {
Changed = true;
it = BB->begin();
e = BB->end();
continue;
}
Value *Inst = BI->getOperand(0);
if (Inst == P)
Inst = BI->getOperand(1);
if (tryToVectorize(dyn_cast<BinaryOperator>(Inst), R)) {
// We would like to start over since some instructions are deleted
// and the iterator may become invalid value.
Changed = true;
it = BB->begin();
e = BB->end();
continue;
}
continue;
}
if (ShouldStartVectorizeHorAtStore)
if (StoreInst *SI = dyn_cast<StoreInst>(it))
if (BinaryOperator *BinOp =
dyn_cast<BinaryOperator>(SI->getValueOperand())) {
if (canMatchHorizontalReduction(nullptr, BinOp, R, TTI) ||
tryToVectorize(BinOp, R)) {
Changed = true;
it = BB->begin();
e = BB->end();
continue;
}
}
// Try to vectorize horizontal reductions feeding into a return.
if (ReturnInst *RI = dyn_cast<ReturnInst>(it))
if (RI->getNumOperands() != 0)
if (BinaryOperator *BinOp =
dyn_cast<BinaryOperator>(RI->getOperand(0))) {
DEBUG(dbgs() << "SLP: Found a return to vectorize.\n");
if (tryToVectorizePair(BinOp->getOperand(0),
BinOp->getOperand(1), R)) {
Changed = true;
it = BB->begin();
e = BB->end();
continue;
}
}
// Try to vectorize trees that start at compare instructions.
if (CmpInst *CI = dyn_cast<CmpInst>(it)) {
if (tryToVectorizePair(CI->getOperand(0), CI->getOperand(1), R)) {
Changed = true;
// We would like to start over since some instructions are deleted
// and the iterator may become invalid value.
it = BB->begin();
e = BB->end();
continue;
}
for (int i = 0; i < 2; ++i) {
if (BinaryOperator *BI = dyn_cast<BinaryOperator>(CI->getOperand(i))) {
if (tryToVectorizePair(BI->getOperand(0), BI->getOperand(1), R)) {
Changed = true;
// We would like to start over since some instructions are deleted
// and the iterator may become invalid value.
it = BB->begin();
e = BB->end();
break;
}
}
}
continue;
}
// Try to vectorize trees that start at insertelement instructions.
if (InsertElementInst *FirstInsertElem = dyn_cast<InsertElementInst>(it)) {
SmallVector<Value *, 16> BuildVector;
SmallVector<Value *, 16> BuildVectorOpds;
if (!findBuildVector(FirstInsertElem, BuildVector, BuildVectorOpds))
continue;
// Vectorize starting with the build vector operands ignoring the
// BuildVector instructions for the purpose of scheduling and user
// extraction.
if (tryToVectorizeList(BuildVectorOpds, R, BuildVector)) {
Changed = true;
it = BB->begin();
e = BB->end();
}
continue;
}
}
return Changed;
}
bool SLPVectorizer::vectorizeGEPIndices(BasicBlock *BB, BoUpSLP &R) {
auto Changed = false;
for (auto &Entry : GEPs) {
// If the getelementptr list has fewer than two elements, there's nothing
// to do.
if (Entry.second.size() < 2)
continue;
DEBUG(dbgs() << "SLP: Analyzing a getelementptr list of length "
<< Entry.second.size() << ".\n");
// We process the getelementptr list in chunks of 16 (like we do for
// stores) to minimize compile-time.
for (unsigned BI = 0, BE = Entry.second.size(); BI < BE; BI += 16) {
auto Len = std::min<unsigned>(BE - BI, 16);
auto GEPList = makeArrayRef(&Entry.second[BI], Len);
// Initialize a set a candidate getelementptrs. Note that we use a
// SetVector here to preserve program order. If the index computations
// are vectorizable and begin with loads, we want to minimize the chance
// of having to reorder them later.
SetVector<Value *> Candidates(GEPList.begin(), GEPList.end());
// Some of the candidates may have already been vectorized after we
// initially collected them. If so, the WeakVHs will have nullified the
// values, so remove them from the set of candidates.
Candidates.remove(nullptr);
// Remove from the set of candidates all pairs of getelementptrs with
// constant differences. Such getelementptrs are likely not good
// candidates for vectorization in a bottom-up phase since one can be
// computed from the other. We also ensure all candidate getelementptr
// indices are unique.
for (int I = 0, E = GEPList.size(); I < E && Candidates.size() > 1; ++I) {
auto *GEPI = cast<GetElementPtrInst>(GEPList[I]);
if (!Candidates.count(GEPI))
continue;
auto *SCEVI = SE->getSCEV(GEPList[I]);
for (int J = I + 1; J < E && Candidates.size() > 1; ++J) {
auto *GEPJ = cast<GetElementPtrInst>(GEPList[J]);
auto *SCEVJ = SE->getSCEV(GEPList[J]);
if (isa<SCEVConstant>(SE->getMinusSCEV(SCEVI, SCEVJ))) {
Candidates.remove(GEPList[I]);
Candidates.remove(GEPList[J]);
} else if (GEPI->idx_begin()->get() == GEPJ->idx_begin()->get()) {
Candidates.remove(GEPList[J]);
}
}
}
// We break out of the above computation as soon as we know there are
// fewer than two candidates remaining.
if (Candidates.size() < 2)
continue;
// Add the single, non-constant index of each candidate to the bundle. We
// ensured the indices met these constraints when we originally collected
// the getelementptrs.
SmallVector<Value *, 16> Bundle(Candidates.size());
auto BundleIndex = 0u;
for (auto *V : Candidates) {
auto *GEP = cast<GetElementPtrInst>(V);
auto *GEPIdx = GEP->idx_begin()->get();
assert(GEP->getNumIndices() == 1 || !isa<Constant>(GEPIdx));
Bundle[BundleIndex++] = GEPIdx;
}
// Try and vectorize the indices. We are currently only interested in
// gather-like cases of the form:
//
// ... = g[a[0] - b[0]] + g[a[1] - b[1]] + ...
//
// where the loads of "a", the loads of "b", and the subtractions can be
// performed in parallel. It's likely that detecting this pattern in a
// bottom-up phase will be simpler and less costly than building a
// full-blown top-down phase beginning at the consecutive loads.
Changed |= tryToVectorizeList(Bundle, R);
}
}
return Changed;
}
bool SLPVectorizer::vectorizeStoreChains(BoUpSLP &R) {
bool Changed = false;
// Attempt to sort and vectorize each of the store-groups.
for (StoreListMap::iterator it = Stores.begin(), e = Stores.end(); it != e;
++it) {
if (it->second.size() < 2)
continue;
DEBUG(dbgs() << "SLP: Analyzing a store chain of length "
<< it->second.size() << ".\n");
// Process the stores in chunks of 16.
// TODO: The limit of 16 inhibits greater vectorization factors.
// For example, AVX2 supports v32i8. Increasing this limit, however,
// may cause a significant compile-time increase.
for (unsigned CI = 0, CE = it->second.size(); CI < CE; CI+=16) {
unsigned Len = std::min<unsigned>(CE - CI, 16);
Changed |= vectorizeStores(makeArrayRef(&it->second[CI], Len),
-SLPCostThreshold, R);
}
}
return Changed;
}
} // end anonymous namespace
char SLPVectorizer::ID = 0;
static const char lv_name[] = "SLP Vectorizer";
INITIALIZE_PASS_BEGIN(SLPVectorizer, SV_NAME, lv_name, false, false)
INITIALIZE_PASS_DEPENDENCY(AAResultsWrapperPass)
INITIALIZE_PASS_DEPENDENCY(TargetTransformInfoWrapperPass)
INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
INITIALIZE_PASS_DEPENDENCY(ScalarEvolutionWrapperPass)
INITIALIZE_PASS_DEPENDENCY(LoopSimplify)
INITIALIZE_PASS_END(SLPVectorizer, SV_NAME, lv_name, false, false)
namespace llvm {
Pass *createSLPVectorizerPass() { return new SLPVectorizer(); }
}
Reference in New Issue View Git Blame Copy Permalink