Rename the operation of turning a float type into an

integer of the same type.  Before it was "promotion",
but this is confusing because it is quite different
to promotion of integers.  Call it "softening" instead,
inspired by "soft float".

llvm-svn: 52546
This commit is contained in:
Duncan Sands 2008-06-20 17:49:55 +00:00
parent 3792c470d5
commit 49295b48eb
5 changed files with 76 additions and 77 deletions

View File

@ -7,13 +7,12 @@
//
//===----------------------------------------------------------------------===//
//
// This file implements float type expansion and conversion of float types to
// integer types on behalf of LegalizeTypes.
// Converting to integer is the act of turning a computation in an illegal
// floating point type into a computation in an integer type of the same size.
// For example, turning f32 arithmetic into operations using i32. Also known as
// "soft float". The result is equivalent to bitcasting the float value to the
// integer type.
// This file implements float type expansion and softening for LegalizeTypes.
// Softening is the act of turning a computation in an illegal floating point
// type into a computation in an integer type of the same size; also known as
// "soft float". For example, turning f32 arithmetic into operations using i32.
// The resulting integer value is the same as what you would get by performing
// the floating point operation and bitcasting the result to the integer type.
// Expansion is the act of changing a computation in an illegal type to be a
// computation in multiple registers of a smaller type. For example,
// implementing ppcf128 arithmetic in two f64 registers.
@ -44,8 +43,8 @@ static RTLIB::Libcall GetFPLibCall(MVT VT,
// Result Float to Integer Conversion.
//===----------------------------------------------------------------------===//
void DAGTypeLegalizer::PromoteFloatResult(SDNode *N, unsigned ResNo) {
DEBUG(cerr << "Promote float result " << ResNo << ": "; N->dump(&DAG);
void DAGTypeLegalizer::SoftenFloatResult(SDNode *N, unsigned ResNo) {
DEBUG(cerr << "Soften float result " << ResNo << ": "; N->dump(&DAG);
cerr << "\n");
SDOperand R = SDOperand();
@ -67,37 +66,37 @@ void DAGTypeLegalizer::PromoteFloatResult(SDNode *N, unsigned ResNo) {
switch (N->getOpcode()) {
default:
#ifndef NDEBUG
cerr << "PromoteFloatResult #" << ResNo << ": ";
cerr << "SoftenFloatResult #" << ResNo << ": ";
N->dump(&DAG); cerr << "\n";
#endif
assert(0 && "Do not know how to convert the result of this operator!");
abort();
case ISD::BIT_CONVERT: R = PromoteFloatRes_BIT_CONVERT(N); break;
case ISD::BUILD_PAIR: R = PromoteFloatRes_BUILD_PAIR(N); break;
case ISD::BIT_CONVERT: R = SoftenFloatRes_BIT_CONVERT(N); break;
case ISD::BUILD_PAIR: R = SoftenFloatRes_BUILD_PAIR(N); break;
case ISD::ConstantFP:
R = PromoteFloatRes_ConstantFP(cast<ConstantFPSDNode>(N));
R = SoftenFloatRes_ConstantFP(cast<ConstantFPSDNode>(N));
break;
case ISD::FCOPYSIGN: R = PromoteFloatRes_FCOPYSIGN(N); break;
case ISD::LOAD: R = PromoteFloatRes_LOAD(N); break;
case ISD::FCOPYSIGN: R = SoftenFloatRes_FCOPYSIGN(N); break;
case ISD::LOAD: R = SoftenFloatRes_LOAD(N); break;
case ISD::SINT_TO_FP:
case ISD::UINT_TO_FP: R = PromoteFloatRes_XINT_TO_FP(N); break;
case ISD::UINT_TO_FP: R = SoftenFloatRes_XINT_TO_FP(N); break;
case ISD::FADD: R = PromoteFloatRes_FADD(N); break;
case ISD::FMUL: R = PromoteFloatRes_FMUL(N); break;
case ISD::FSUB: R = PromoteFloatRes_FSUB(N); break;
case ISD::FADD: R = SoftenFloatRes_FADD(N); break;
case ISD::FMUL: R = SoftenFloatRes_FMUL(N); break;
case ISD::FSUB: R = SoftenFloatRes_FSUB(N); break;
}
// If R is null, the sub-method took care of registering the result.
if (R.Val)
SetPromotedFloat(SDOperand(N, ResNo), R);
SetSoftenedFloat(SDOperand(N, ResNo), R);
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_BIT_CONVERT(SDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatRes_BIT_CONVERT(SDNode *N) {
return BitConvertToInteger(N->getOperand(0));
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_BUILD_PAIR(SDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatRes_BUILD_PAIR(SDNode *N) {
// Convert the inputs to integers, and build a new pair out of them.
return DAG.getNode(ISD::BUILD_PAIR,
TLI.getTypeToTransformTo(N->getValueType(0)),
@ -105,15 +104,15 @@ SDOperand DAGTypeLegalizer::PromoteFloatRes_BUILD_PAIR(SDNode *N) {
BitConvertToInteger(N->getOperand(1)));
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_ConstantFP(ConstantFPSDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatRes_ConstantFP(ConstantFPSDNode *N) {
return DAG.getConstant(N->getValueAPF().convertToAPInt(),
TLI.getTypeToTransformTo(N->getValueType(0)));
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_FADD(SDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatRes_FADD(SDNode *N) {
MVT NVT = TLI.getTypeToTransformTo(N->getValueType(0));
SDOperand Ops[2] = { GetPromotedFloat(N->getOperand(0)),
GetPromotedFloat(N->getOperand(1)) };
SDOperand Ops[2] = { GetSoftenedFloat(N->getOperand(0)),
GetSoftenedFloat(N->getOperand(1)) };
return MakeLibCall(GetFPLibCall(N->getValueType(0),
RTLIB::ADD_F32,
RTLIB::ADD_F64,
@ -122,8 +121,8 @@ SDOperand DAGTypeLegalizer::PromoteFloatRes_FADD(SDNode *N) {
NVT, Ops, 2, false/*sign irrelevant*/);
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_FCOPYSIGN(SDNode *N) {
SDOperand LHS = GetPromotedFloat(N->getOperand(0));
SDOperand DAGTypeLegalizer::SoftenFloatRes_FCOPYSIGN(SDNode *N) {
SDOperand LHS = GetSoftenedFloat(N->getOperand(0));
SDOperand RHS = BitConvertToInteger(N->getOperand(1));
MVT LVT = LHS.getValueType();
@ -161,10 +160,10 @@ SDOperand DAGTypeLegalizer::PromoteFloatRes_FCOPYSIGN(SDNode *N) {
return DAG.getNode(ISD::OR, LVT, LHS, SignBit);
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_FMUL(SDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatRes_FMUL(SDNode *N) {
MVT NVT = TLI.getTypeToTransformTo(N->getValueType(0));
SDOperand Ops[2] = { GetPromotedFloat(N->getOperand(0)),
GetPromotedFloat(N->getOperand(1)) };
SDOperand Ops[2] = { GetSoftenedFloat(N->getOperand(0)),
GetSoftenedFloat(N->getOperand(1)) };
return MakeLibCall(GetFPLibCall(N->getValueType(0),
RTLIB::MUL_F32,
RTLIB::MUL_F64,
@ -173,10 +172,10 @@ SDOperand DAGTypeLegalizer::PromoteFloatRes_FMUL(SDNode *N) {
NVT, Ops, 2, false/*sign irrelevant*/);
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_FSUB(SDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatRes_FSUB(SDNode *N) {
MVT NVT = TLI.getTypeToTransformTo(N->getValueType(0));
SDOperand Ops[2] = { GetPromotedFloat(N->getOperand(0)),
GetPromotedFloat(N->getOperand(1)) };
SDOperand Ops[2] = { GetSoftenedFloat(N->getOperand(0)),
GetSoftenedFloat(N->getOperand(1)) };
return MakeLibCall(GetFPLibCall(N->getValueType(0),
RTLIB::SUB_F32,
RTLIB::SUB_F64,
@ -185,7 +184,7 @@ SDOperand DAGTypeLegalizer::PromoteFloatRes_FSUB(SDNode *N) {
NVT, Ops, 2, false/*sign irrelevant*/);
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_LOAD(SDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatRes_LOAD(SDNode *N) {
LoadSDNode *L = cast<LoadSDNode>(N);
MVT VT = N->getValueType(0);
MVT NVT = TLI.getTypeToTransformTo(VT);
@ -206,7 +205,7 @@ SDOperand DAGTypeLegalizer::PromoteFloatRes_LOAD(SDNode *N) {
return BitConvertToInteger(DAG.getNode(ISD::FP_EXTEND, VT, NL));
}
SDOperand DAGTypeLegalizer::PromoteFloatRes_XINT_TO_FP(SDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatRes_XINT_TO_FP(SDNode *N) {
bool isSigned = N->getOpcode() == ISD::SINT_TO_FP;
MVT DestVT = N->getValueType(0);
SDOperand Op = N->getOperand(0);
@ -311,8 +310,8 @@ SDOperand DAGTypeLegalizer::PromoteFloatRes_XINT_TO_FP(SDNode *N) {
// Operand Float to Integer Conversion..
//===----------------------------------------------------------------------===//
bool DAGTypeLegalizer::PromoteFloatOperand(SDNode *N, unsigned OpNo) {
DEBUG(cerr << "Promote float operand " << OpNo << ": "; N->dump(&DAG);
bool DAGTypeLegalizer::SoftenFloatOperand(SDNode *N, unsigned OpNo) {
DEBUG(cerr << "Soften float operand " << OpNo << ": "; N->dump(&DAG);
cerr << "\n");
SDOperand Res(0, 0);
@ -327,13 +326,13 @@ bool DAGTypeLegalizer::PromoteFloatOperand(SDNode *N, unsigned OpNo) {
switch (N->getOpcode()) {
default:
#ifndef NDEBUG
cerr << "PromoteFloatOperand Op #" << OpNo << ": ";
cerr << "SoftenFloatOperand Op #" << OpNo << ": ";
N->dump(&DAG); cerr << "\n";
#endif
assert(0 && "Do not know how to convert this operator's operand!");
abort();
case ISD::BIT_CONVERT: Res = PromoteFloatOp_BIT_CONVERT(N); break;
case ISD::BIT_CONVERT: Res = SoftenFloatOp_BIT_CONVERT(N); break;
}
}
@ -357,9 +356,9 @@ bool DAGTypeLegalizer::PromoteFloatOperand(SDNode *N, unsigned OpNo) {
return false;
}
SDOperand DAGTypeLegalizer::PromoteFloatOp_BIT_CONVERT(SDNode *N) {
SDOperand DAGTypeLegalizer::SoftenFloatOp_BIT_CONVERT(SDNode *N) {
return DAG.getNode(ISD::BIT_CONVERT, N->getValueType(0),
GetPromotedFloat(N->getOperand(0)));
GetSoftenedFloat(N->getOperand(0)));
}

View File

@ -232,9 +232,9 @@ SDOperand DAGTypeLegalizer::PromoteIntRes_BIT_CONVERT(SDNode *N) {
// The input promotes to the same size. Convert the promoted value.
return DAG.getNode(ISD::BIT_CONVERT, OutVT, GetPromotedInteger(InOp));
break;
case PromoteFloat:
case SoftenFloat:
// Promote the integer operand by hand.
return DAG.getNode(ISD::ANY_EXTEND, OutVT, GetPromotedFloat(InOp));
return DAG.getNode(ISD::ANY_EXTEND, OutVT, GetSoftenedFloat(InOp));
case ExpandInteger:
case ExpandFloat:
break;
@ -963,9 +963,9 @@ void DAGTypeLegalizer::ExpandIntRes_BIT_CONVERT(SDNode *N,
case Legal:
case PromoteInteger:
break;
case PromoteFloat:
case SoftenFloat:
// Convert the integer operand instead.
SplitInteger(GetPromotedFloat(InOp), Lo, Hi);
SplitInteger(GetSoftenedFloat(InOp), Lo, Hi);
Lo = DAG.getNode(ISD::BIT_CONVERT, NVT, Lo);
Hi = DAG.getNode(ISD::BIT_CONVERT, NVT, Hi);
return;

View File

@ -80,8 +80,8 @@ void DAGTypeLegalizer::run() {
case ExpandInteger:
ExpandIntegerResult(N, i);
goto NodeDone;
case PromoteFloat:
PromoteFloatResult(N, i);
case SoftenFloat:
SoftenFloatResult(N, i);
goto NodeDone;
case ExpandFloat:
ExpandFloatResult(N, i);
@ -113,8 +113,8 @@ void DAGTypeLegalizer::run() {
case ExpandInteger:
NeedsRevisit = ExpandIntegerOperand(N, i);
break;
case PromoteFloat:
NeedsRevisit = PromoteFloatOperand(N, i);
case SoftenFloat:
NeedsRevisit = SoftenFloatOperand(N, i);
break;
case ExpandFloat:
NeedsRevisit = ExpandFloatOperand(N, i);
@ -393,8 +393,8 @@ void DAGTypeLegalizer::ExpungeNode(SDOperand N) {
I->second = Replacement;
}
for (DenseMap<SDOperand, SDOperand>::iterator I = PromotedFloats.begin(),
E = PromotedFloats.end(); I != E; ++I) {
for (DenseMap<SDOperand, SDOperand>::iterator I = SoftenedFloats.begin(),
E = SoftenedFloats.end(); I != E; ++I) {
assert(I->first != N);
if (I->second == N)
I->second = Replacement;
@ -446,11 +446,11 @@ void DAGTypeLegalizer::SetPromotedInteger(SDOperand Op, SDOperand Result) {
OpEntry = Result;
}
void DAGTypeLegalizer::SetPromotedFloat(SDOperand Op, SDOperand Result) {
void DAGTypeLegalizer::SetSoftenedFloat(SDOperand Op, SDOperand Result) {
ExpungeNode(Result);
AnalyzeNewNode(Result.Val);
SDOperand &OpEntry = PromotedFloats[Op];
SDOperand &OpEntry = SoftenedFloats[Op];
assert(OpEntry.Val == 0 && "Node is already converted to integer!");
OpEntry = Result;
}

View File

@ -62,7 +62,7 @@ private:
Legal, // The target natively supports this type.
PromoteInteger, // Replace this integer type with a larger one.
ExpandInteger, // Split this integer type into two of half the size.
PromoteFloat, // Convert this float type to a same size integer type.
SoftenFloat, // Convert this float type to a same size integer type.
ExpandFloat, // Split this float type into two of half the size.
Scalarize, // Replace this one-element vector type with its element type.
Split // This vector type should be split into smaller vectors.
@ -95,7 +95,7 @@ private:
return ExpandInteger;
else if (VT.getSizeInBits() ==
TLI.getTypeToTransformTo(VT).getSizeInBits())
return PromoteFloat;
return SoftenFloat;
else
return ExpandFloat;
} else if (VT.getVectorNumElements() == 1) {
@ -119,9 +119,9 @@ private:
/// indicates which operands are the expanded version of the input.
DenseMap<SDOperand, std::pair<SDOperand, SDOperand> > ExpandedIntegers;
/// PromotedFloats - For floating point nodes converted to integers of
/// SoftenedFloats - For floating point nodes converted to integers of
/// the same size, this map indicates the converted value to use.
DenseMap<SDOperand, SDOperand> PromotedFloats;
DenseMap<SDOperand, SDOperand> SoftenedFloats;
/// ExpandedFloats - For float nodes that need to be expanded this map
/// indicates which operands are the expanded version of the input.
@ -321,29 +321,29 @@ private:
// Float to Integer Conversion Support: LegalizeFloatTypes.cpp
//===--------------------------------------------------------------------===//
SDOperand GetPromotedFloat(SDOperand Op) {
SDOperand &PromotedOp = PromotedFloats[Op];
RemapNode(PromotedOp);
assert(PromotedOp.Val && "Operand wasn't converted to integer?");
return PromotedOp;
SDOperand GetSoftenedFloat(SDOperand Op) {
SDOperand &SoftenedOp = SoftenedFloats[Op];
RemapNode(SoftenedOp);
assert(SoftenedOp.Val && "Operand wasn't converted to integer?");
return SoftenedOp;
}
void SetPromotedFloat(SDOperand Op, SDOperand Result);
void SetSoftenedFloat(SDOperand Op, SDOperand Result);
// Result Float to Integer Conversion.
void PromoteFloatResult(SDNode *N, unsigned OpNo);
SDOperand PromoteFloatRes_BIT_CONVERT(SDNode *N);
SDOperand PromoteFloatRes_BUILD_PAIR(SDNode *N);
SDOperand PromoteFloatRes_ConstantFP(ConstantFPSDNode *N);
SDOperand PromoteFloatRes_FADD(SDNode *N);
SDOperand PromoteFloatRes_FCOPYSIGN(SDNode *N);
SDOperand PromoteFloatRes_FMUL(SDNode *N);
SDOperand PromoteFloatRes_FSUB(SDNode *N);
SDOperand PromoteFloatRes_LOAD(SDNode *N);
SDOperand PromoteFloatRes_XINT_TO_FP(SDNode *N);
void SoftenFloatResult(SDNode *N, unsigned OpNo);
SDOperand SoftenFloatRes_BIT_CONVERT(SDNode *N);
SDOperand SoftenFloatRes_BUILD_PAIR(SDNode *N);
SDOperand SoftenFloatRes_ConstantFP(ConstantFPSDNode *N);
SDOperand SoftenFloatRes_FADD(SDNode *N);
SDOperand SoftenFloatRes_FCOPYSIGN(SDNode *N);
SDOperand SoftenFloatRes_FMUL(SDNode *N);
SDOperand SoftenFloatRes_FSUB(SDNode *N);
SDOperand SoftenFloatRes_LOAD(SDNode *N);
SDOperand SoftenFloatRes_XINT_TO_FP(SDNode *N);
// Operand Float to Integer Conversion.
bool PromoteFloatOperand(SDNode *N, unsigned OpNo);
SDOperand PromoteFloatOp_BIT_CONVERT(SDNode *N);
bool SoftenFloatOperand(SDNode *N, unsigned OpNo);
SDOperand SoftenFloatOp_BIT_CONVERT(SDNode *N);
//===--------------------------------------------------------------------===//
// Float Expansion Support: LegalizeFloatTypes.cpp

View File

@ -503,7 +503,7 @@ void DAGTypeLegalizer::SplitRes_BIT_CONVERT(SDNode *N,
assert(false && "Unknown type action!");
case Legal:
case PromoteInteger:
case PromoteFloat:
case SoftenFloat:
case Scalarize:
break;
case ExpandInteger: