Skip to content

[RISCV][TTI] Support fdiv/udiv/sdiv/srem/urem in getArithmeticInstrCost #89170

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
May 13, 2024
Merged

[RISCV][TTI] Support fdiv/udiv/sdiv/srem/urem in getArithmeticInstrCost #89170

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
7e943ef
[RISCV][TTI] Scale the cost of ArithmeticInstr with LMUL
arcbbb Apr 18, 2024
33c75d2
Refine comments
arcbbb Apr 29, 2024
f0cc926
Refine comments
arcbbb Apr 30, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
47 changes: 38 additions & 9 deletions llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1612,29 +1612,58 @@ InstructionCost RISCVTTIImpl::getArithmeticInstrCost(
if (Op2Info.isConstant())
ConstantMatCost += getConstantMatCost(1, Op2Info);

unsigned Op;
switch (TLI->InstructionOpcodeToISD(Opcode)) {
case ISD::ADD:
case ISD::SUB:
case ISD::AND:
case ISD::OR:
case ISD::XOR:
Op = RISCV::VADD_VV;
break;
case ISD::SHL:
case ISD::SRL:
case ISD::SRA:
Op = RISCV::VSLL_VV;
break;
case ISD::AND:
case ISD::OR:
case ISD::XOR:
Op = (Ty->getScalarSizeInBits() == 1) ? RISCV::VMAND_MM : RISCV::VAND_VV;
break;
case ISD::MUL:
case ISD::MULHS:
case ISD::MULHU:
Op = RISCV::VMUL_VV;
break;
case ISD::SDIV:
case ISD::UDIV:
Op = RISCV::VDIV_VV;
break;
case ISD::SREM:
case ISD::UREM:
Op = RISCV::VREM_VV;
break;
case ISD::FADD:
case ISD::FSUB:
// TODO: Address FP16 with VFHMIN
Op = RISCV::VFADD_VV;
case ISD::FMUL:
case ISD::FNEG: {
return ConstantMatCost + TLI->getLMULCost(LT.second) * LT.first * 1;
}
// TODO: Address FP16 with VFHMIN
Op = RISCV::VFMUL_VV;
break;
case ISD::FDIV:
Op = RISCV::VFDIV_VV;
break;
case ISD::FNEG:
Op = RISCV::VFSGNJN_VV;
break;
default:
return ConstantMatCost +
BaseT::getArithmeticInstrCost(Opcode, Ty, CostKind, Op1Info, Op2Info,
Args, CxtI);
// Assuming all other instructions have the same cost until a need arises to
// differentiate them.
return ConstantMatCost + BaseT::getArithmeticInstrCost(Opcode, Ty, CostKind,
Op1Info, Op2Info,
Args, CxtI);
}
return ConstantMatCost +
LT.first * getRISCVInstructionCost(Op, LT.second, CostKind);
}

// TODO: Deduplicate from TargetTransformInfoImplCRTPBase.
Expand Down
48 changes: 24 additions & 24 deletions llvm/test/Analysis/CostModel/RISCV/arith-fp.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -248,36 +248,36 @@ define i32 @fdiv() {
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %F16 = fdiv half undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %F32 = fdiv float undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %F64 = fdiv double undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V1F16 = fdiv <1 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F16 = fdiv <2 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4F16 = fdiv <4 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F16 = fdiv <8 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16F16 = fdiv <16 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V1F16 = fdiv <1 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F16 = fdiv <2 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F16 = fdiv <4 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F16 = fdiv <8 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16F16 = fdiv <16 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V32F16 = fdiv <32 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV1F16 = fdiv <vscale x 1 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV2F16 = fdiv <vscale x 2 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV4F16 = fdiv <vscale x 4 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %NXV1F16 = fdiv <vscale x 1 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %NXV2F16 = fdiv <vscale x 2 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %NXV4F16 = fdiv <vscale x 4 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV8F16 = fdiv <vscale x 8 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV16F16 = fdiv <vscale x 16 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV32F16 = fdiv <vscale x 32 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V1F32 = fdiv <1 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fdiv <2 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fdiv <4 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fdiv <8 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %NXV16F16 = fdiv <vscale x 16 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %NXV32F16 = fdiv <vscale x 32 x half> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V1F32 = fdiv <1 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fdiv <2 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fdiv <4 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fdiv <8 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fdiv <16 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV1F32 = fdiv <vscale x 1 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV2F32 = fdiv <vscale x 2 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %NXV1F32 = fdiv <vscale x 1 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %NXV2F32 = fdiv <vscale x 2 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV4F32 = fdiv <vscale x 4 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV8F32 = fdiv <vscale x 8 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV16F32 = fdiv <vscale x 16 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V1F64 = fdiv <1 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fdiv <2 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fdiv <4 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %NXV8F32 = fdiv <vscale x 8 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %NXV16F32 = fdiv <vscale x 16 x float> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V1F64 = fdiv <1 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fdiv <2 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fdiv <4 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fdiv <8 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV1F64 = fdiv <vscale x 1 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %NXV1F64 = fdiv <vscale x 1 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV2F64 = fdiv <vscale x 2 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV4F64 = fdiv <vscale x 4 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %NXV8F64 = fdiv <vscale x 8 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %NXV4F64 = fdiv <vscale x 4 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %NXV8F64 = fdiv <vscale x 8 x double> undef, undef
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
%F16 = fdiv half undef, undef
Expand Down
Loading
Loading