virtual unsigned getNumberOfRegisters(bool Vector) const;
virtual unsigned getRegisterBitWidth(bool Vector) const;
virtual unsigned getMaximumUnrollFactor() const;
- virtual unsigned getArithmeticInstrCost(unsigned Opcode, Type *Ty) const;
+ virtual unsigned getArithmeticInstrCost(unsigned Opcode, Type *Ty,
+ OperandValueKind,
+ OperandValueKind) const;
virtual unsigned getShuffleCost(ShuffleKind Kind, Type *Tp,
int Index, Type *SubTp) const;
virtual unsigned getCastInstrCost(unsigned Opcode, Type *Dst,
PPCTTI::PopcntSupportKind PPCTTI::getPopcntSupport(unsigned TyWidth) const {
assert(isPowerOf2_32(TyWidth) && "Ty width must be power of 2");
- // FIXME: PPC currently does not have custom popcnt lowering even though
- // there is hardware support. Once this is fixed, update this function
- // to reflect the real capabilities of the hardware.
+ if (ST->hasPOPCNTD() && TyWidth <= 64)
+ return PSK_FastHardware;
return PSK_Software;
}
return 2;
}
-unsigned PPCTTI::getArithmeticInstrCost(unsigned Opcode, Type *Ty) const {
+unsigned PPCTTI::getArithmeticInstrCost(unsigned Opcode, Type *Ty,
+ OperandValueKind Op1Info,
+ OperandValueKind Op2Info) const {
assert(TLI->InstructionOpcodeToISD(Opcode) && "Invalid opcode");
// Fallback to the default implementation.
- return TargetTransformInfo::getArithmeticInstrCost(Opcode, Ty);
+ return TargetTransformInfo::getArithmeticInstrCost(Opcode, Ty, Op1Info,
+ Op2Info);
}
unsigned PPCTTI::getShuffleCost(ShuffleKind Kind, Type *Tp, int Index,
unsigned Index) const {
assert(Val->isVectorTy() && "This must be a vector type");
- const unsigned Awful = 1000;
-
- // Vector element insert/extract with Altivec is very expensive.
- // Until VSX is available, avoid vectorizing loops that require
- // these operations.
- if (Opcode == ISD::EXTRACT_VECTOR_ELT ||
- Opcode == ISD::INSERT_VECTOR_ELT)
- return Awful;
-
- // We don't vectorize SREM/UREM so well. Constrain the vectorizer
- // for those as well.
- if (Opcode == ISD::SREM || Opcode == ISD::UREM)
- return Awful;
-
- // VSELECT is not yet implemented, leading to use of insert/extract
- // and ISEL, hence not a good idea.
- if (Opcode == ISD::VSELECT)
- return Awful;
+ int ISD = TLI->InstructionOpcodeToISD(Opcode);
+ assert(ISD && "Invalid opcode");
+
+ // Estimated cost of a load-hit-store delay. This was obtained
+ // experimentally as a minimum needed to prevent unprofitable
+ // vectorization for the paq8p benchmark. It may need to be
+ // raised further if other unprofitable cases remain.
+ unsigned LHSPenalty = 12;
+
+ // Vector element insert/extract with Altivec is very expensive,
+ // because they require store and reload with the attendant
+ // processor stall for load-hit-store. Until VSX is available,
+ // these need to be estimated as very costly.
+ if (ISD == ISD::EXTRACT_VECTOR_ELT ||
+ ISD == ISD::INSERT_VECTOR_ELT)
+ return LHSPenalty +
+ TargetTransformInfo::getVectorInstrCost(Opcode, Val, Index);
return TargetTransformInfo::getVectorInstrCost(Opcode, Val, Index);
}