Skip to content

Refactor diff rules to prepare for more rules #89

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.

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 2 commits into from
Jun 1, 2024
Merged
Changes from all commits
Commits
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
240 changes: 112 additions & 128 deletions src/enzyme_ad/jax/Implementations/HLODerivatives.td
Original file line number Diff line number Diff line change
Expand Up @@ -104,6 +104,9 @@ def UniformQuantize : HLOInst<"UniformQuantizeOp">;
def While : HLOInst<"WhileOp">;
def Xor : HLOInst<"XorOp">;

/// StableHLO - WIP operations
def ReturnOp : HLOInst<"ReturnOp">;

// CHLO
/// CHLO - binary elementwise operations
def BroadcastAdd : HLOInst<"BroadcastAddOp">;
Expand Down Expand Up @@ -180,111 +183,7 @@ def UnaryEinsum : HLOInst<"UnaryEinsumOp">;
def CheckedMul : HLOInst<"MulOp">;
def CheckedDiv : HLOInst<"DivOp">;

def : HLODerivative<"AddOp", (Op $x, $y),
[
(DiffeRet),
(DiffeRet),
]
>;

def : HLODerivative<"SubtractOp", (Op $x, $y),
[
(DiffeRet),
(Neg (DiffeRet)),
]
>;

def : HLODerivative<"NegOp", (Op $x),
[
(Neg (DiffeRet)),
]
>;

def : HLODerivative<"MulOp", (Op $x, $y),
[
(CheckedMul (DiffeRet), $y),
(CheckedMul (DiffeRet), $x)
]
>;

def : HLODerivative<"DivOp", (Op $x, $y),
[
(CheckedDiv (DiffeRet), $y),
(Neg (Mul (CheckedDiv (DiffeRet), $y), (Div $x, $y)))
]
// (CheckedDiv (FSub (SelectIfActive $x, (FMul (Shadow $x), $y), (Zero $x)), (SelectIfActive $y, (FMul (Shadow $y), $x), (Zero $y))), (FMul $y, $y))
>;

def : HLODerivative<"PowOp", (Op $x, $y),
[
(CheckedMul (DiffeRet), (Mul $y, (Pow $x, (Sub $y, (HLOConstantFP<"1"> $y))))),
(CheckedMul (DiffeRet), (Mul (Pow $x, $y), (Log $x)
))
]
>;

def : HLODerivative<"CosineOp", (Op $x),
[
(CheckedMul (DiffeRet), (Neg (Sin $x)))
]
>;

def : HLODerivative<"ExpOp", (Op $x),
[
(CheckedMul (DiffeRet), (Exp $x))
]
>;

def : HLODerivative<"SineOp", (Op $x),
[
(CheckedMul (DiffeRet), (Cos $x))
]
>;

def : HLODerivative<"SqrtOp", (Op $x),
[
// (Select (FCmpUEQ $x, (ConstantFP<"0"> $x)), (ConstantFP<"0"> $x), (FDiv (DiffeRet), (FMul (ConstantFP<"2"> $x), (Call<(SameFunc), [ReadNone,NoUnwind]> $x))))
(Div (DiffeRet), (Mul (HLOConstantFP<"2"> $x), (Sqrt $x)))
]
>;

def : HLODerivative<"RsqrtOp", (Op $x),
[
// (Select (FCmpUEQ $x, (ConstantFP<"0"> $x)), (ConstantFP<"0"> $x), (FDiv (DiffeRet), (FMul (ConstantFP<"2"> $x), (Call<(SameFunc), [ReadNone,NoUnwind]> $x))))
(Div (DiffeRet), (Mul (HLOConstantFP<"2"> $x), (Mul $x, (Sqrt $x))))
]
>;

def : HLODerivative<"TanhOp", (Op $x),
[
(CheckedMul (DiffeRet), (Sub (HLOConstantFP<"1">), (Mul (Tanh $x), (Tanh $x))))
]
>;

def : HLODerivative<"LogOp", (Op $x),
[
(CheckedDiv (DiffeRet), $x)
]
>;

def LT : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, "ComparisonDirection::LT">;

def : HLODerivative<"MaxOp", (Op $x, $y),
[
(Select (Compare $x, $y, (LT)), (HLOConstantFP<"0"> $x), (DiffeRet)),
(Select (Compare $x, $y, (LT)), (DiffeRet), (HLOConstantFP<"0"> $x))
],
(Select (Compare $x, $y, (LT)), (SelectIfActive $y, (Shadow $y), (HLOConstantFP<"0"> $y)), (SelectIfActive $x, (Shadow $x), (HLOConstantFP<"0"> $x)))
>;

def : HLOReadOnlyIdentityOp<"ReshapeOp", [0], (Op $x), [(Reshape (TypeOf $x), (DiffeRet))]>;

def : HLOReadOnlyIdentityOp<"SliceOp">;

def : HLOReadOnlyIdentityOp<"BroadcastInDimOp">;

def : HLOMemoryIdentityOp<"ConcatenateOp", [], [-1]>;

// Helpers
def PadToSliceStart : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, [{
op.getEdgePaddingLow();
}]>;
Expand All @@ -303,11 +202,6 @@ def PadToSliceStride : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, [{
getI64Attr(builder, strides);
}]>;

def : HLOMemoryIdentityOp<"PadOp", [], [-1], (Op $op, $padval), [
(Slice (TypeOf $op), (DiffeRet), (PadToSliceStart), (PadToSliceLimit), (PadToSliceStride)),
(AssertingInactiveArg)
]>;

// convert

def ResultDotDim : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, "op.getDotDimensionNumbersAttr()">;
Expand Down Expand Up @@ -508,14 +402,6 @@ def ShadowRHSTranspose : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, [{
builder.getNamedAttr(TransposeOp::getAttributeNames()[0], builder.getDenseI64ArrayAttr(transposes));
}]>;

def : HLODerivative<"DotGeneralOp", (Op $lhs, $rhs),
[
(Transpose (TypeOf $lhs), (DotGeneral (ShadowLHSDotRes), (DiffeRet), $rhs, (ShadowLHSDotDim), (ResultDotPrec)), (ShadowLHSTranspose)),
(Transpose (TypeOf $rhs), (DotGeneral (ShadowRHSDotRes), $lhs, (DiffeRet), (ShadowRHSDotDim), (ResultDotPrec)), (ShadowRHSTranspose))
],
(Add (SelectIfActive $lhs, (DotGeneral (ResultTypes), (Shadow $lhs), $rhs, (ResultDotDim), (ResultDotPrec)), (HLOConstantFP<"0">)), (SelectIfActive $rhs, (DotGeneral (ResultTypes), $lhs, (Shadow $rhs), (ResultDotDim), (ResultDotPrec)), (HLOConstantFP<"0">)))
>;

def Permutation : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, [{
op.getPermutation();
}]>;
Expand All @@ -528,13 +414,6 @@ def InversePermutation : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, [{
getI64Attr(builder, res);
}]>;

def : HLODerivative<"TransposeOp", (Op $x),
[
(Transpose (TypeOf $x), (DiffeRet), (InversePermutation)),
],
(SelectIfActive $x, (Transpose (ResultTypes), (Shadow $x), (Permutation)), (HLOConstantFP<"0">))
>;

def ConvWindowStrides : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, [{
op.getWindowStridesAttr();
}]>;
Expand Down Expand Up @@ -635,18 +514,94 @@ def GradDataConvBatchGroupCount : GlobalExpr</*needsprimal*/0, /*needsshadow*/0,
op.getBatchGroupCountAttr();
}]>;

def LT : GlobalExpr</*needsprimal*/0, /*needsshadow*/0, "ComparisonDirection::LT">;

// Derivative rules
def : HLODerivative<"AddOp", (Op $x, $y),
[
(DiffeRet),
(DiffeRet),
]
>;

def : HLOReadOnlyIdentityOp<"BroadcastInDimOp">;

def : HLOMemoryIdentityOp<"ConcatenateOp", [], [-1]>;

def : HLODerivative<"ConvolutionOp", (Op $lhs, $rhs),
[
(Convolution (TypeOf $lhs), (DiffeRet), $lhs, (GradFilterConvWindowStrides), (GradFilterConvPadding), (GradFilterConvLhsDialation), (GradFilterConvRhsDialation), (GradFilterConvWindowReversal), (GradFilterConvDimensionNumbers), (GradFilterConvFeatureGroupCount), (GradFilterConvBatchGroupCount), (ResultDotPrec)),
(Convolution (TypeOf $rhs), (DiffeRet), $rhs, (GradDataConvWindowStrides), (GradDataConvPadding), (GradDataConvLhsDialation), (GradDataConvRhsDialation), (GradDataConvWindowReversal), (GradDataConvDimensionNumbers), (GradDataConvFeatureGroupCount), (GradDataConvBatchGroupCount), (ResultDotPrec))
],
(Add (SelectIfActive $lhs,
],
(Add (SelectIfActive $lhs,
(Convolution (ResultTypes), (Shadow $lhs), $rhs, (ConvWindowStrides), (ConvPadding), (ConvLhsDialation), (ConvRhsDialation), (ConvWindowReversal), (ConvDimensionNumbers), (ConvFeatureGroupCount), (ConvBatchGroupCount), (ResultDotPrec)),
(HLOConstantFP<"0">)), (SelectIfActive $rhs,
(Convolution (ResultTypes), $lhs, (Shadow $rhs), (ConvWindowStrides), (ConvPadding), (ConvLhsDialation), (ConvRhsDialation), (ConvWindowReversal), (ConvDimensionNumbers), (ConvFeatureGroupCount), (ConvBatchGroupCount), (ResultDotPrec)),
(HLOConstantFP<"0">)))
>;

def : HLODerivative<"CosineOp", (Op $x), [(CheckedMul (DiffeRet), (Neg (Sin $x)))]>;

def : HLODerivative<"DivOp", (Op $x, $y),
[
(CheckedDiv (DiffeRet), $y),
(Neg (Mul (CheckedDiv (DiffeRet), $y), (Div $x, $y)))
]
// (CheckedDiv (FSub (SelectIfActive $x, (FMul (Shadow $x), $y), (Zero $x)), (SelectIfActive $y, (FMul (Shadow $y), $x), (Zero $y))), (FMul $y, $y))
>;

def : HLODerivative<"DotGeneralOp", (Op $lhs, $rhs),
[
(Transpose (TypeOf $lhs), (DotGeneral (ShadowLHSDotRes), (DiffeRet), $rhs, (ShadowLHSDotDim), (ResultDotPrec)), (ShadowLHSTranspose)),
(Transpose (TypeOf $rhs), (DotGeneral (ShadowRHSDotRes), $lhs, (DiffeRet), (ShadowRHSDotDim), (ResultDotPrec)), (ShadowRHSTranspose))
],
(Add (SelectIfActive $lhs, (DotGeneral (ResultTypes), (Shadow $lhs), $rhs, (ResultDotDim), (ResultDotPrec)), (HLOConstantFP<"0">)), (SelectIfActive $rhs, (DotGeneral (ResultTypes), $lhs, (Shadow $rhs), (ResultDotDim), (ResultDotPrec)), (HLOConstantFP<"0">)))
>;

def : HLODerivative<"ExpOp", (Op $x), [(CheckedMul (DiffeRet), (Exp $x))]>;

def : HLODerivative<"LogOp", (Op $x), [(CheckedDiv (DiffeRet), $x)]>;

def : HLODerivative<"MaxOp", (Op $x, $y),
[
(Select (Compare $x, $y, (LT)), (HLOConstantFP<"0"> $x), (DiffeRet)),
(Select (Compare $x, $y, (LT)), (DiffeRet), (HLOConstantFP<"0"> $x))
],
(Select (Compare $x, $y, (LT)), (SelectIfActive $y, (Shadow $y), (HLOConstantFP<"0"> $y)), (SelectIfActive $x, (Shadow $x), (HLOConstantFP<"0"> $x)))
>;

def : HLODerivative<"MulOp", (Op $x, $y),
[
(CheckedMul (DiffeRet), $y),
(CheckedMul (DiffeRet), $x)
]
>;

def : HLODerivative<"NegOp", (Op $x), [(Neg (DiffeRet))]>;

def : HLOMemoryIdentityOp<"PadOp", [], [-1], (Op $op, $padval), [
(Slice (TypeOf $op), (DiffeRet), (PadToSliceStart), (PadToSliceLimit), (PadToSliceStride)),
(AssertingInactiveArg)
]>;

def : HLODerivative<"PowOp", (Op $x, $y),
[
(CheckedMul (DiffeRet), (Mul $y, (Pow $x, (Sub $y, (HLOConstantFP<"1"> $y))))),
(CheckedMul (DiffeRet), (Mul (Pow $x, $y), (Log $x)))
]
>;

def : HLOReadOnlyIdentityOp<"ReshapeOp", [0], (Op $x), [(Reshape (TypeOf $x), (DiffeRet))]>;

def : HLORegionTerminatorOp<"ReturnOp">;

def : HLODerivative<"RsqrtOp", (Op $x),
[
// (Select (FCmpUEQ $x, (ConstantFP<"0"> $x)), (ConstantFP<"0"> $x), (FDiv (DiffeRet), (FMul (ConstantFP<"2"> $x), (Call<(SameFunc), [ReadNone,NoUnwind]> $x))))
(Div (DiffeRet), (Mul (HLOConstantFP<"2"> $x), (Mul $x, (Sqrt $x))))
]
>;

def : HLODerivative<"SelectOp", (Op $cond, $lhs, $rhs),
[
(AssertingInactiveArg),
Expand All @@ -656,4 +611,33 @@ def : HLODerivative<"SelectOp", (Op $cond, $lhs, $rhs),
(Select $cond, (SelectIfActive $lhs, (Shadow $lhs), (HLOConstantFP<"0">)), (SelectIfActive $rhs, (Shadow $rhs), (HLOConstantFP<"0">)))
>;

def : HLORegionTerminatorOp<"ReturnOp">;
def : HLODerivative<"SineOp", (Op $x), [(CheckedMul (DiffeRet), (Cos $x))]>;

def : HLOReadOnlyIdentityOp<"SliceOp">;

def : HLODerivative<"SubtractOp", (Op $x, $y),
[
(DiffeRet),
(Neg (DiffeRet)),
]
>;

def : HLODerivative<"SqrtOp", (Op $x),
[
// (Select (FCmpUEQ $x, (ConstantFP<"0"> $x)), (ConstantFP<"0"> $x), (FDiv (DiffeRet), (FMul (ConstantFP<"2"> $x), (Call<(SameFunc), [ReadNone,NoUnwind]> $x))))
(Div (DiffeRet), (Mul (HLOConstantFP<"2"> $x), (Sqrt $x)))
]
>;

def : HLODerivative<"TanhOp", (Op $x),
[
(CheckedMul (DiffeRet), (Sub (HLOConstantFP<"1">), (Mul (Tanh $x), (Tanh $x))))
]
>;

def : HLODerivative<"TransposeOp", (Op $x),
[
(Transpose (TypeOf $x), (DiffeRet), (InversePermutation)),
],
(SelectIfActive $x, (Transpose (ResultTypes), (Shadow $x), (Permutation)), (HLOConstantFP<"0">))
>;