Skip to content

[RISCV] Fold vector shift of sext/zext to widening multiply #121563

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
pfusik merged 12 commits into from
Feb 4, 2025
Merged

[RISCV] Fold vector shift of sext/zext to widening multiply #121563

Show file tree
Hide file tree
Changes from 5 commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
15412e7
[RISCV] Fold vector shift of sext/zext to widening multiply
pfusik Jan 3, 2025
5aa9da4
[RISCV] Fold vector shift to widening multiply on rv32
pfusik Jan 8, 2025
1868507
[RISCV] Use `getSimpleValueType`
pfusik Jan 8, 2025
3a59255
[RISCV] Add a TODO comment
pfusik Jan 8, 2025
174e59e
[RISCV] Handle fixed-length vectors
pfusik Jan 9, 2025
fa18fca
[RISCV] Handle VSEXT_VL/VZEXT_VL
pfusik Jan 10, 2025
0da5f8c
[RISCV] Delay the transform so that vwsll has priority
pfusik Jan 10, 2025
d84396c
[RISCV] Remove dead code handling fixed-length vectors after DAG lega…
pfusik Jan 13, 2025
6df5335
[RISCV] Fix formatting
pfusik Jan 13, 2025
63f902d
[RISCV] Disable this transform by default, except for SpacemiT X60
pfusik Jan 28, 2025
b243fbc
[RISCV] Rename function for consistency
pfusik Jan 30, 2025
bddd7fd
Revert "[RISCV] Disable this transform by default, except for Spacemi…
pfusik Feb 4, 2025
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
94 changes: 92 additions & 2 deletions llvm/lib/Target/RISCV/RISCVISelLowering.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17341,6 +17341,96 @@ static SDValue combineScalarCTPOPToVCPOP(SDNode *N, SelectionDAG &DAG,
return DAG.getZExtOrTrunc(Pop, DL, VT);
}

static SDValue combineSHL(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
Copy link
Collaborator

@preames preames Jan 29, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The naming convention in this file would have this routine named performSHLCombine

Copy link
Contributor Author

@pfusik pfusik Jan 30, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Renamed, thanks!

const RISCVSubtarget &Subtarget) {
if (DCI.isBeforeLegalize())
return SDValue();

// (shl (zext x), y) -> (vwsll x, y)
if (SDValue V = combineOp_VLToVWOp_VL(N, DCI, Subtarget))
return V;

// (shl (sext x), C) -> (vwmulsu x, 1u << C)
// (shl (zext x), C) -> (vwmulu x, 1u << C)

SDValue LHS = N->getOperand(0);
if (!LHS.hasOneUse())
return SDValue();
Copy link
Contributor Author

@pfusik pfusik Jan 3, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For now, this only handles single-use of sext/zext.
I can rewrite it to be part of combineOp_VLToVWOp_VL so that it handles multi-use too.

unsigned Opcode;
switch (LHS.getOpcode()) {
case ISD::SIGN_EXTEND:
Opcode = RISCVISD::VWMULSU_VL;
break;
case ISD::ZERO_EXTEND:
Opcode = RISCVISD::VWMULU_VL;
break;
// TODO:
// case RISCVISD::VSEXT_VL:
// case RISCVISD::VZEXT_VL:
default:
return SDValue();
}

SDValue RHS = N->getOperand(1);
APInt ShAmt;
uint64_t ShAmtInt;
if (ISD::isConstantSplatVector(RHS.getNode(), ShAmt))
ShAmtInt = ShAmt.getZExtValue();
else if (RHS.getOpcode() == RISCVISD::VMV_V_X_VL &&
RHS.getOperand(1).getOpcode() == ISD::Constant)
ShAmtInt = RHS.getConstantOperandVal(1);
else
return SDValue();

// Better foldings:
// (shl (sext x), 1) -> (vwadd x, x)
// (shl (zext x), 1) -> (vwaddu x, x)
if (ShAmtInt <= 1)
return SDValue();

SDValue NarrowOp = LHS.getOperand(0);
MVT NarrowVT = NarrowOp.getSimpleValueType();
uint64_t NarrowBits = NarrowVT.getScalarSizeInBits();
if (ShAmtInt >= NarrowBits)
return SDValue();
MVT VT = N->getSimpleValueType(0);
if (NarrowBits * 2 != VT.getScalarSizeInBits())
return SDValue();

SelectionDAG &DAG = DCI.DAG;
MVT NarrowContainerVT = NarrowVT;
MVT ContainerVT = VT;
SDLoc DL(N);
SDValue Passthru, Mask, VL;
switch (N->getOpcode()) {
case ISD::SHL:
if (VT.isFixedLengthVector()) {
NarrowContainerVT =
getContainerForFixedLengthVector(DAG, NarrowVT, Subtarget);
NarrowOp =
convertToScalableVector(NarrowContainerVT, NarrowOp, DAG, Subtarget);
ContainerVT = getContainerForFixedLengthVector(DAG, VT, Subtarget);
}
Passthru = DAG.getUNDEF(VT);
std::tie(Mask, VL) = getDefaultVLOps(VT, ContainerVT, DL, DAG, Subtarget);
break;
case RISCVISD::SHL_VL:
Passthru = N->getOperand(2);
Mask = N->getOperand(3);
VL = N->getOperand(4);
break;
default:
llvm_unreachable("Expected SHL");
}
SDValue Mul =
DAG.getNode(Opcode, DL, ContainerVT, NarrowOp,
DAG.getConstant(1ULL << ShAmtInt, SDLoc(RHS), ContainerVT),
Passthru, Mask, VL);
if (VT.isFixedLengthVector())
return convertFromScalableVector(VT, Mul, DAG, Subtarget);
return Mul;
}

SDValue RISCVTargetLowering::PerformDAGCombine(SDNode *N,
DAGCombinerInfo &DCI) const {
SelectionDAG &DAG = DCI.DAG;
Expand Down Expand Up @@ -17970,7 +18060,7 @@ SDValue RISCVTargetLowering::PerformDAGCombine(SDNode *N,
break;
}
case RISCVISD::SHL_VL:
if (SDValue V = combineOp_VLToVWOp_VL(N, DCI, Subtarget))
if (SDValue V = combineSHL(N, DCI, Subtarget))
return V;
[[fallthrough]];
case RISCVISD::SRA_VL:
Expand All @@ -17995,7 +18085,7 @@ SDValue RISCVTargetLowering::PerformDAGCombine(SDNode *N,
case ISD::SRL:
case ISD::SHL: {
if (N->getOpcode() == ISD::SHL) {
if (SDValue V = combineOp_VLToVWOp_VL(N, DCI, Subtarget))
if (SDValue V = combineSHL(N, DCI, Subtarget))
return V;
}
SDValue ShAmt = N->getOperand(1);
Expand Down
Loading
Loading