summaryrefslogtreecommitdiffstats
path: root/lib/Target/ARM/ARMTargetTransformInfo.cpp
diff options
context:
space:
mode:
Diffstat (limited to 'lib/Target/ARM/ARMTargetTransformInfo.cpp')
-rw-r--r--lib/Target/ARM/ARMTargetTransformInfo.cpp147
1 files changed, 88 insertions, 59 deletions
diff --git a/lib/Target/ARM/ARMTargetTransformInfo.cpp b/lib/Target/ARM/ARMTargetTransformInfo.cpp
index 6bbb38f..a2ace62 100644
--- a/lib/Target/ARM/ARMTargetTransformInfo.cpp
+++ b/lib/Target/ARM/ARMTargetTransformInfo.cpp
@@ -14,17 +14,18 @@
///
//===----------------------------------------------------------------------===//
-#define DEBUG_TYPE "armtti"
#include "ARM.h"
#include "ARMTargetMachine.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/Support/Debug.h"
-#include "llvm/Target/TargetLowering.h"
#include "llvm/Target/CostTable.h"
+#include "llvm/Target/TargetLowering.h"
using namespace llvm;
+#define DEBUG_TYPE "armtti"
+
// Declare the pass initialization routine locally as target-specific passes
-// don't havve a target-wide initialization entry point, and so we rely on the
+// don't have a target-wide initialization entry point, and so we rely on the
// pass constructor initialization.
namespace llvm {
void initializeARMTTIPass(PassRegistry &);
@@ -32,7 +33,7 @@ void initializeARMTTIPass(PassRegistry &);
namespace {
-class ARMTTI : public ImmutablePass, public TargetTransformInfo {
+class ARMTTI final : public ImmutablePass, public TargetTransformInfo {
const ARMBaseTargetMachine *TM;
const ARMSubtarget *ST;
const ARMTargetLowering *TLI;
@@ -42,7 +43,7 @@ class ARMTTI : public ImmutablePass, public TargetTransformInfo {
unsigned getScalarizationOverhead(Type *Ty, bool Insert, bool Extract) const;
public:
- ARMTTI() : ImmutablePass(ID), TM(0), ST(0), TLI(0) {
+ ARMTTI() : ImmutablePass(ID), TM(nullptr), ST(nullptr), TLI(nullptr) {
llvm_unreachable("This pass cannot be directly constructed");
}
@@ -52,15 +53,11 @@ public:
initializeARMTTIPass(*PassRegistry::getPassRegistry());
}
- virtual void initializePass() {
+ void initializePass() override {
pushTTIStack(this);
}
- virtual void finalizePass() {
- popTTIStack();
- }
-
- virtual void getAnalysisUsage(AnalysisUsage &AU) const {
+ void getAnalysisUsage(AnalysisUsage &AU) const override {
TargetTransformInfo::getAnalysisUsage(AU);
}
@@ -68,7 +65,7 @@ public:
static char ID;
/// Provide necessary pointer adjustments for the two base classes.
- virtual void *getAdjustedAnalysisPointer(const void *ID) {
+ void *getAdjustedAnalysisPointer(const void *ID) override {
if (ID == &TargetTransformInfo::ID)
return (TargetTransformInfo*)this;
return this;
@@ -76,8 +73,8 @@ public:
/// \name Scalar TTI Implementations
/// @{
-
- virtual unsigned getIntImmCost(const APInt &Imm, Type *Ty) const;
+ using TargetTransformInfo::getIntImmCost;
+ unsigned getIntImmCost(const APInt &Imm, Type *Ty) const override;
/// @}
@@ -85,7 +82,7 @@ public:
/// \name Vector TTI Implementations
/// @{
- unsigned getNumberOfRegisters(bool Vector) const {
+ unsigned getNumberOfRegisters(bool Vector) const override {
if (Vector) {
if (ST->hasNEON())
return 16;
@@ -94,10 +91,10 @@ public:
if (ST->isThumb1Only())
return 8;
- return 16;
+ return 13;
}
- unsigned getRegisterBitWidth(bool Vector) const {
+ unsigned getRegisterBitWidth(bool Vector) const override {
if (Vector) {
if (ST->hasNEON())
return 128;
@@ -107,7 +104,7 @@ public:
return 32;
}
- unsigned getMaximumUnrollFactor() const {
+ unsigned getMaximumUnrollFactor() const override {
// These are out of order CPUs:
if (ST->isCortexA15() || ST->isSwift())
return 2;
@@ -115,23 +112,27 @@ public:
}
unsigned getShuffleCost(ShuffleKind Kind, Type *Tp,
- int Index, Type *SubTp) const;
+ int Index, Type *SubTp) const override;
unsigned getCastInstrCost(unsigned Opcode, Type *Dst,
- Type *Src) const;
+ Type *Src) const override;
- unsigned getCmpSelInstrCost(unsigned Opcode, Type *ValTy, Type *CondTy) const;
+ unsigned getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
+ Type *CondTy) const override;
- unsigned getVectorInstrCost(unsigned Opcode, Type *Val, unsigned Index) const;
+ unsigned getVectorInstrCost(unsigned Opcode, Type *Val,
+ unsigned Index) const override;
- unsigned getAddressComputationCost(Type *Val, bool IsComplex) const;
+ unsigned getAddressComputationCost(Type *Val,
+ bool IsComplex) const override;
- unsigned getArithmeticInstrCost(unsigned Opcode, Type *Ty,
- OperandValueKind Op1Info = OK_AnyValue,
- OperandValueKind Op2Info = OK_AnyValue) const;
+ unsigned
+ getArithmeticInstrCost(unsigned Opcode, Type *Ty,
+ OperandValueKind Op1Info = OK_AnyValue,
+ OperandValueKind Op2Info = OK_AnyValue) const override;
unsigned getMemoryOpCost(unsigned Opcode, Type *Src, unsigned Alignment,
- unsigned AddressSpace) const;
+ unsigned AddressSpace) const override;
/// @}
};
@@ -162,25 +163,25 @@ unsigned ARMTTI::getIntImmCost(const APInt &Imm, Type *Ty) const {
(ARM_AM::getSOImmVal(~ZImmVal) != -1))
return 1;
return ST->hasV6T2Ops() ? 2 : 3;
- } else if (ST->isThumb2()) {
+ }
+ if (ST->isThumb2()) {
if ((SImmVal >= 0 && SImmVal < 65536) ||
(ARM_AM::getT2SOImmVal(ZImmVal) != -1) ||
(ARM_AM::getT2SOImmVal(~ZImmVal) != -1))
return 1;
return ST->hasV6T2Ops() ? 2 : 3;
- } else /*Thumb1*/ {
- if (SImmVal >= 0 && SImmVal < 256)
- return 1;
- if ((~ZImmVal < 256) || ARM_AM::isThumbImmShiftedVal(ZImmVal))
- return 2;
- // Load from constantpool.
- return 3;
}
- return 2;
+ // Thumb1.
+ if (SImmVal >= 0 && SImmVal < 256)
+ return 1;
+ if ((~ZImmVal < 256) || ARM_AM::isThumbImmShiftedVal(ZImmVal))
+ return 2;
+ // Load from constantpool.
+ return 3;
}
unsigned ARMTTI::getCastInstrCost(unsigned Opcode, Type *Dst,
- Type *Src) const {
+ Type *Src) const {
int ISD = TLI->InstructionOpcodeToISD(Opcode);
assert(ISD && "Invalid opcode");
@@ -442,34 +443,62 @@ unsigned ARMTTI::getAddressComputationCost(Type *Ty, bool IsComplex) const {
unsigned ARMTTI::getShuffleCost(ShuffleKind Kind, Type *Tp, int Index,
Type *SubTp) const {
- // We only handle costs of reverse shuffles for now.
- if (Kind != SK_Reverse)
+ // We only handle costs of reverse and alternate shuffles for now.
+ if (Kind != SK_Reverse && Kind != SK_Alternate)
return TargetTransformInfo::getShuffleCost(Kind, Tp, Index, SubTp);
- static const CostTblEntry<MVT::SimpleValueType> NEONShuffleTbl[] = {
- // Reverse shuffle cost one instruction if we are shuffling within a double
- // word (vrev) or two if we shuffle a quad word (vrev, vext).
- { ISD::VECTOR_SHUFFLE, MVT::v2i32, 1 },
- { ISD::VECTOR_SHUFFLE, MVT::v2f32, 1 },
- { ISD::VECTOR_SHUFFLE, MVT::v2i64, 1 },
- { ISD::VECTOR_SHUFFLE, MVT::v2f64, 1 },
-
- { ISD::VECTOR_SHUFFLE, MVT::v4i32, 2 },
- { ISD::VECTOR_SHUFFLE, MVT::v4f32, 2 },
- { ISD::VECTOR_SHUFFLE, MVT::v8i16, 2 },
- { ISD::VECTOR_SHUFFLE, MVT::v16i8, 2 }
- };
+ if (Kind == SK_Reverse) {
+ static const CostTblEntry<MVT::SimpleValueType> NEONShuffleTbl[] = {
+ // Reverse shuffle cost one instruction if we are shuffling within a
+ // double word (vrev) or two if we shuffle a quad word (vrev, vext).
+ {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1},
+ {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1},
+ {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1},
+ {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1},
- std::pair<unsigned, MVT> LT = TLI->getTypeLegalizationCost(Tp);
+ {ISD::VECTOR_SHUFFLE, MVT::v4i32, 2},
+ {ISD::VECTOR_SHUFFLE, MVT::v4f32, 2},
+ {ISD::VECTOR_SHUFFLE, MVT::v8i16, 2},
+ {ISD::VECTOR_SHUFFLE, MVT::v16i8, 2}};
- int Idx = CostTableLookup(NEONShuffleTbl, ISD::VECTOR_SHUFFLE, LT.second);
- if (Idx == -1)
- return TargetTransformInfo::getShuffleCost(Kind, Tp, Index, SubTp);
+ std::pair<unsigned, MVT> LT = TLI->getTypeLegalizationCost(Tp);
+
+ int Idx = CostTableLookup(NEONShuffleTbl, ISD::VECTOR_SHUFFLE, LT.second);
+ if (Idx == -1)
+ return TargetTransformInfo::getShuffleCost(Kind, Tp, Index, SubTp);
- return LT.first * NEONShuffleTbl[Idx].Cost;
+ return LT.first * NEONShuffleTbl[Idx].Cost;
+ }
+ if (Kind == SK_Alternate) {
+ static const CostTblEntry<MVT::SimpleValueType> NEONAltShuffleTbl[] = {
+ // Alt shuffle cost table for ARM. Cost is the number of instructions
+ // required to create the shuffled vector.
+
+ {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1},
+ {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1},
+ {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1},
+ {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1},
+
+ {ISD::VECTOR_SHUFFLE, MVT::v4i32, 2},
+ {ISD::VECTOR_SHUFFLE, MVT::v4f32, 2},
+ {ISD::VECTOR_SHUFFLE, MVT::v4i16, 2},
+
+ {ISD::VECTOR_SHUFFLE, MVT::v8i16, 16},
+
+ {ISD::VECTOR_SHUFFLE, MVT::v16i8, 32}};
+
+ std::pair<unsigned, MVT> LT = TLI->getTypeLegalizationCost(Tp);
+ int Idx =
+ CostTableLookup(NEONAltShuffleTbl, ISD::VECTOR_SHUFFLE, LT.second);
+ if (Idx == -1)
+ return TargetTransformInfo::getShuffleCost(Kind, Tp, Index, SubTp);
+ return LT.first * NEONAltShuffleTbl[Idx].Cost;
+ }
+ return TargetTransformInfo::getShuffleCost(Kind, Tp, Index, SubTp);
}
-unsigned ARMTTI::getArithmeticInstrCost(unsigned Opcode, Type *Ty, OperandValueKind Op1Info,
+unsigned ARMTTI::getArithmeticInstrCost(unsigned Opcode, Type *Ty,
+ OperandValueKind Op1Info,
OperandValueKind Op2Info) const {
int ISDOpcode = TLI->InstructionOpcodeToISD(Opcode);
@@ -533,7 +562,7 @@ unsigned ARMTTI::getArithmeticInstrCost(unsigned Opcode, Type *Ty, OperandValueK
// creates a sequence of shift, and, or instructions to construct values.
// These sequences are recognized by the ISel and have zero-cost. Not so for
// the vectorized code. Because we have support for v2i64 but not i64 those
- // sequences look particularily beneficial to vectorize.
+ // sequences look particularly beneficial to vectorize.
// To work around this we increase the cost of v2i64 operations to make them
// seem less beneficial.
if (LT.second == MVT::v2i64 &&
OpenPOWER on IntegriCloud