RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-31 12:41:49 +01:00
Code Issues Projects Releases Wiki Activity

[LoopUnroll] Pass SCEV to getUnrollingPreferences hook. NFCI.

Browse Source 
Reviewers: sanjoy, anna, reames, apilipenko, igor-laevsky, mkuper

Subscribers: jholewinski, arsenm, mzolotukhin, nemanjai, nhaehnle, javed.absar, mcrosier, llvm-commits

Differential Revision: https://reviews.llvm.org/D34531

llvm-svn: 306554
This commit is contained in:
Geoff Berry 2017-06-28 15:53:17 +00:00
parent f7b4320d1e
commit 9c1c467f23
17 changed files with 48 additions and 37 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
include/llvm/Analysis/TargetTransformInfo.h
View File

@ -366,7 +366,8 @@ public:
/// \brief Get target-customized preferences for the generic loop unrolling
/// transformation. The caller will initialize UP with the current
/// target-independent defaults.
void getUnrollingPreferences(Loop *L, UnrollingPreferences &UP) const;
void getUnrollingPreferences(Loop *L, ScalarEvolution &,
UnrollingPreferences &UP) const;
/// @}
@ -829,7 +830,8 @@ public:
virtual bool isAlwaysUniform(const Value *V) = 0;
virtual unsigned getFlatAddressSpace() = 0;
virtual bool isLoweredToCall(const Function *F) = 0;
virtual void getUnrollingPreferences(Loop *L, UnrollingPreferences &UP) = 0;
virtual void getUnrollingPreferences(Loop *L, ScalarEvolution &,
UnrollingPreferences &UP) = 0;
virtual bool isLegalAddImmediate(int64_t Imm) = 0;
virtual bool isLegalICmpImmediate(int64_t Imm) = 0;
virtual bool isLegalAddressingMode(Type *Ty, GlobalValue *BaseGV,
@ -1015,8 +1017,9 @@ public:
bool isLoweredToCall(const Function *F) override {
return Impl.isLoweredToCall(F);
}
void getUnrollingPreferences(Loop *L, UnrollingPreferences &UP) override {
return Impl.getUnrollingPreferences(L, UP);
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
UnrollingPreferences &UP) override {
return Impl.getUnrollingPreferences(L, SE, UP);
}
bool isLegalAddImmediate(int64_t Imm) override {
return Impl.isLegalAddImmediate(Imm);

3
include/llvm/Analysis/TargetTransformInfoImpl.h
View File

@ -217,7 +217,8 @@ public:
return true;
}
void getUnrollingPreferences(Loop *, TTI::UnrollingPreferences &) {}
void getUnrollingPreferences(Loop *, ScalarEvolution &,
TTI::UnrollingPreferences &) {}
bool isLegalAddImmediate(int64_t Imm) { return false; }

3
include/llvm/CodeGen/BasicTTIImpl.h
View File

@ -277,7 +277,8 @@ public:
unsigned getInliningThresholdMultiplier() { return 1; }
void getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP) {
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
// This unrolling functionality is target independent, but to provide some
// motivation for its intended use, for x86:

4
lib/Analysis/TargetTransformInfo.cpp
View File

@ -116,8 +116,8 @@ bool TargetTransformInfo::isLoweredToCall(const Function *F) const {
}
void TargetTransformInfo::getUnrollingPreferences(
Loop *L, UnrollingPreferences &UP) const {
return TTIImpl->getUnrollingPreferences(L, UP);
Loop *L, ScalarEvolution &SE, UnrollingPreferences &UP) const {
return TTIImpl->getUnrollingPreferences(L, SE, UP);
}
bool TargetTransformInfo::isLegalAddImmediate(int64_t Imm) const {

4
lib/Target/AArch64/AArch64TargetTransformInfo.cpp
View File

@ -645,10 +645,10 @@ unsigned AArch64TTIImpl::getMaxInterleaveFactor(unsigned VF) {
return ST->getMaxInterleaveFactor();
}
void AArch64TTIImpl::getUnrollingPreferences(Loop *L,
void AArch64TTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
// Enable partial unrolling and runtime unrolling.
BaseT::getUnrollingPreferences(L, UP);
BaseT::getUnrollingPreferences(L, SE, UP);
// For inner loop, it is more likely to be a hot one, and the runtime check
// can be promoted out from LICM pass, so the overhead is less, let's try

3
lib/Target/AArch64/AArch64TargetTransformInfo.h
View File

@ -122,7 +122,8 @@ public:
int getCostOfKeepingLiveOverCall(ArrayRef<Type *> Tys);
void getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP);
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
Value *getOrCreateResultFromMemIntrinsic(IntrinsicInst *Inst,
Type *ExpectedType);

2
lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
View File

@ -63,7 +63,7 @@ static bool dependsOnLocalPhi(const Loop *L, const Value *Cond,
return false;
}
void AMDGPUTTIImpl::getUnrollingPreferences(Loop *L,
void AMDGPUTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
UP.Threshold = 300; // Twice the default.
UP.MaxCount = UINT_MAX;

3
lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
View File

@ -68,7 +68,8 @@ public:
bool hasBranchDivergence() { return true; }
void getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP);
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
TTI::PopcntSupportKind getPopcntSupport(unsigned TyWidth) {
assert(isPowerOf2_32(TyWidth) && "Ty width must be power of 2");

2
lib/Target/Hexagon/HexagonTargetTransformInfo.cpp
View File

@ -29,7 +29,7 @@ HexagonTTIImpl::getPopcntSupport(unsigned IntTyWidthInBit) const {
}
// The Hexagon target can unroll loops with run-time trip counts.
void HexagonTTIImpl::getUnrollingPreferences(Loop *L,
void HexagonTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
UP.Runtime = UP.Partial = true;
}

3
lib/Target/Hexagon/HexagonTargetTransformInfo.h
View File

@ -46,7 +46,8 @@ public:
TTI::PopcntSupportKind getPopcntSupport(unsigned IntTyWidthInBit) const;
// The Hexagon target can unroll loops with run-time trip counts.
void getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP);
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
// L1 cache prefetch.
unsigned getPrefetchDistance() const;

4
lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
View File

@ -141,9 +141,9 @@ int NVPTXTTIImpl::getArithmeticInstrCost(
}
}
void NVPTXTTIImpl::getUnrollingPreferences(Loop *L,
void NVPTXTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
BaseT::getUnrollingPreferences(L, UP);
BaseT::getUnrollingPreferences(L, SE, UP);
// Enable partial unrolling and runtime unrolling, but reduce the
// threshold. This partially unrolls small loops which are often

3
lib/Target/NVPTX/NVPTXTargetTransformInfo.h
View File

@ -61,7 +61,8 @@ public:
TTI::OperandValueProperties Opd2PropInfo = TTI::OP_None,
ArrayRef<const Value *> Args = ArrayRef<const Value *>());
void getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP);
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
};
} // end namespace llvm

4
lib/Target/PowerPC/PPCTargetTransformInfo.cpp
View File

@ -189,7 +189,7 @@ int PPCTTIImpl::getIntImmCost(unsigned Opcode, unsigned Idx, const APInt &Imm,
return PPCTTIImpl::getIntImmCost(Imm, Ty);
}
void PPCTTIImpl::getUnrollingPreferences(Loop *L,
void PPCTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
if (ST->getDarwinDirective() == PPC::DIR_A2) {
// The A2 is in-order with a deep pipeline, and concatenation unrolling
@ -201,7 +201,7 @@ void PPCTTIImpl::getUnrollingPreferences(Loop *L,
UP.AllowExpensiveTripCount = true;
}
BaseT::getUnrollingPreferences(L, UP);
BaseT::getUnrollingPreferences(L, SE, UP);
}
bool PPCTTIImpl::enableAggressiveInterleaving(bool LoopHasReductions) {

3
lib/Target/PowerPC/PPCTargetTransformInfo.h
View File

@ -52,7 +52,8 @@ public:
Type *Ty);
TTI::PopcntSupportKind getPopcntSupport(unsigned TyWidth);
void getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP);
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
/// @}

2
lib/Target/SystemZ/SystemZTargetTransformInfo.cpp
View File

@ -238,7 +238,7 @@ SystemZTTIImpl::getPopcntSupport(unsigned TyWidth) {
return TTI::PSK_Software;
}
void SystemZTTIImpl::getUnrollingPreferences(Loop *L,
void SystemZTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP) {
// Find out if L contains a call, what the machine instruction count
// estimate is, and how many stores there are.

3
lib/Target/SystemZ/SystemZTargetTransformInfo.h
View File

@ -45,7 +45,8 @@ public:
TTI::PopcntSupportKind getPopcntSupport(unsigned TyWidth);
void getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP);
void getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
TTI::UnrollingPreferences &UP);
/// @}

28
lib/Transforms/Scalar/LoopUnrollPass.cpp
View File

@ -131,7 +131,7 @@ static const unsigned NoThreshold = UINT_MAX;
/// Gather the various unrolling parameters based on the defaults, compiler
/// flags, TTI overrides and user specified parameters.
static TargetTransformInfo::UnrollingPreferences gatherUnrollingPreferences(
Loop *L, const TargetTransformInfo &TTI, int OptLevel,
Loop *L, ScalarEvolution &SE, const TargetTransformInfo &TTI, int OptLevel,
Optional<unsigned> UserThreshold, Optional<unsigned> UserCount,
Optional<bool> UserAllowPartial, Optional<bool> UserRuntime,
Optional<bool> UserUpperBound) {
@ -158,7 +158,7 @@ static TargetTransformInfo::UnrollingPreferences gatherUnrollingPreferences(
UP.AllowPeeling = true;
// Override with any target specific settings
TTI.getUnrollingPreferences(L, UP);
TTI.getUnrollingPreferences(L, SE, UP);
// Apply size attributes
if (L->getHeader()->getParent()->optForSize()) {
@ -699,7 +699,7 @@ static uint64_t getUnrolledLoopSize(
// Calculates unroll count and writes it to UP.Count.
static bool computeUnrollCount(
Loop *L, const TargetTransformInfo &TTI, DominatorTree &DT, LoopInfo *LI,
ScalarEvolution *SE, OptimizationRemarkEmitter *ORE, unsigned &TripCount,
ScalarEvolution &SE, OptimizationRemarkEmitter *ORE, unsigned &TripCount,
unsigned MaxTripCount, unsigned &TripMultiple, unsigned LoopSize,
TargetTransformInfo::UnrollingPreferences &UP, bool &UseUpperBound) {
// Check for explicit Count.
@ -770,7 +770,7 @@ static bool computeUnrollCount(
// helps to remove a significant number of instructions.
// To check that, run additional analysis on the loop.
if (Optional<EstimatedUnrollCost> Cost = analyzeLoopUnrollCost(
L, FullUnrollTripCount, DT, *SE, TTI,
L, FullUnrollTripCount, DT, SE, TTI,
UP.Threshold * UP.MaxPercentThresholdBoost / 100)) {
unsigned Boost =
getFullUnrollBoostingFactor(*Cost, UP.MaxPercentThresholdBoost);
@ -926,7 +926,7 @@ static bool computeUnrollCount(
}
static bool tryToUnrollLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
ScalarEvolution *SE, const TargetTransformInfo &TTI,
ScalarEvolution &SE, const TargetTransformInfo &TTI,
AssumptionCache &AC, OptimizationRemarkEmitter &ORE,
bool PreserveLCSSA, int OptLevel,
Optional<unsigned> ProvidedCount,
@ -948,8 +948,8 @@ static bool tryToUnrollLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
bool NotDuplicatable;
bool Convergent;
TargetTransformInfo::UnrollingPreferences UP = gatherUnrollingPreferences(
L, TTI, OptLevel, ProvidedThreshold, ProvidedCount, ProvidedAllowPartial,
ProvidedRuntime, ProvidedUpperBound);
L, SE, TTI, OptLevel, ProvidedThreshold, ProvidedCount,
ProvidedAllowPartial, ProvidedRuntime, ProvidedUpperBound);
// Exit early if unrolling is disabled.
if (UP.Threshold == 0 && (!UP.Partial || UP.PartialThreshold == 0))
return false;
@ -977,8 +977,8 @@ static bool tryToUnrollLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
if (!ExitingBlock || !L->isLoopExiting(ExitingBlock))
ExitingBlock = L->getExitingBlock();
if (ExitingBlock) {
TripCount = SE->getSmallConstantTripCount(L, ExitingBlock);
TripMultiple = SE->getSmallConstantTripMultiple(L, ExitingBlock);
TripCount = SE.getSmallConstantTripCount(L, ExitingBlock);
TripMultiple = SE.getSmallConstantTripMultiple(L, ExitingBlock);
}
// If the loop contains a convergent operation, the prelude we'd add
@ -1000,8 +1000,8 @@ static bool tryToUnrollLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
// count.
bool MaxOrZero = false;
if (!TripCount) {
MaxTripCount = SE->getSmallConstantMaxTripCount(L);
MaxOrZero = SE->isBackedgeTakenCountMaxOrZero(L);
MaxTripCount = SE.getSmallConstantMaxTripCount(L);
MaxOrZero = SE.isBackedgeTakenCountMaxOrZero(L);
// We can unroll by the upper bound amount if it's generally allowed or if
// we know that the loop is executed either the upper bound or zero times.
// (MaxOrZero unrolling keeps only the first loop test, so the number of
@ -1030,7 +1030,7 @@ static bool tryToUnrollLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
// Unroll the loop.
if (!UnrollLoop(L, UP.Count, TripCount, UP.Force, UP.Runtime,
UP.AllowExpensiveTripCount, UseUpperBound, MaxOrZero,
TripMultiple, UP.PeelCount, LI, SE, &DT, &AC, &ORE,
TripMultiple, UP.PeelCount, LI, &SE, &DT, &AC, &ORE,
PreserveLCSSA))
return false;
@ -1073,7 +1073,7 @@ public:
auto &DT = getAnalysis<DominatorTreeWrapperPass>().getDomTree();
LoopInfo *LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
ScalarEvolution *SE = &getAnalysis<ScalarEvolutionWrapperPass>().getSE();
ScalarEvolution &SE = getAnalysis<ScalarEvolutionWrapperPass>().getSE();
const TargetTransformInfo &TTI =
getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F);
auto &AC = getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
@ -1157,7 +1157,7 @@ PreservedAnalyses LoopUnrollPass::run(Loop &L, LoopAnalysisManager &AM,
if (!AllowPartialUnrolling)
AllowPartialParam = RuntimeParam = UpperBoundParam = false;
bool Changed = tryToUnrollLoop(
&L, AR.DT, &AR.LI, &AR.SE, AR.TTI, AR.AC, *ORE,
&L, AR.DT, &AR.LI, AR.SE, AR.TTI, AR.AC, *ORE,
/*PreserveLCSSA*/ true, OptLevel, /*Count*/ None,
/*Threshold*/ None, AllowPartialParam, RuntimeParam, UpperBoundParam);
if (!Changed)