'chlo' Dialect

Stay organized with collections Save and categorize content based on your preferences.

Client HLO Ops

This dialect contains ops that align closely with the API surface area of the XlaBuilder C++ API, where such ops have semantics that go beyond what exists in the lower level dialects (such as stablehlo). Essentially, whenever the client library uses syntactic sugar or composition of multiple ops for an API call, this dialect tries to model the API call and provide conversion patterns to fully materialize into lower level dialects.

Operations

chlo._asin_acos_kernel (chlo::AsinAcosKernelOp)

AsinAcosKernel operator

Syntax:

operation ::= `chlo._asin_acos_kernel` $operand attr-dict `:` type($operand) `->` type($result)

Returns AsinAcosKernel(operand) element-wise.

If
  w = _asin_acos_kernel(z)
  w' = _asin_acos_kernel(I * z)
Then
  asin(z) = complex(atan2(z.real, w.real), sign(z.imag) * w.imag)
  acos(z) = complex(atan2(w.real, z.real), -sign(z.imag) * w.imag)
  asinh(z) = complex(sign(z.real) * w'.imag, atan2(z.imag, w'.real))
  acosh(z) = complex(w.imag, sign(z.imag) * atan2(w.real, z.real))

This op is used as an intermediate value in decompositions and should never be constructed directly by frameworks or consumed by backends.

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.acos (chlo::AcosOp)

Acos operator

Syntax:

operation ::= `chlo.acos` $operand attr-dict `:` type($operand) `->` type($result)

Returns Acos(operand) element-wise.

\[ \acos(x) = 2 * \atan(\sqrt(1 - x^2) / (1 + x)) if x != -1 = pi if x == -1 \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.acosh (chlo::AcoshOp)

Acosh operation

Syntax:

operation ::= `chlo.acosh` $operand attr-dict `:` type($operand) `->` type($result)

Returns Acosh(operand) element-wise.

\[ \acosh(x) = log(x + sqrt(x^2 - 1)) if x >= -1 \acosh(x) = nan if x < -1 \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.asin (chlo::AsinOp)

Asin operator

Syntax:

operation ::= `chlo.asin` $operand attr-dict `:` type($operand) `->` type($result)

Returns Asin(operand) element-wise.

\[ \asin(x) = 2 * atan(x / (1 + sqrt(1 - x^2))) \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.asinh (chlo::AsinhOp)

Asinh operation

Syntax:

operation ::= `chlo.asinh` $operand attr-dict `:` type($operand) `->` type($result)

Returns Asinh(operand) element-wise.

\[ \asinh(x) = log(x + sqrt(x^2 + 1)) \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.atan (chlo::AtanOp)

Atan operator

Syntax:

operation ::= `chlo.atan` $operand attr-dict `:` type($operand) `->` type($result)

Returns Atan(operand) element-wise.

\[ \atan(x) = \atan2(x, 1) \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.atanh (chlo::AtanhOp)

Atanh operator

Syntax:

operation ::= `chlo.atanh` $operand attr-dict `:` type($operand) `->` type($result)

Returns Atanh(operand) element-wise.

\[ \atanh(x) = 0.5 * log((1 + x) / (1 - x)) if abs(x) <= 1 = nan otherwise \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.bessel_i1e (chlo::BesselI1eOp)

Bessel function of order 1

Syntax:

operation ::= `chlo.bessel_i1e` $operand attr-dict `:` type($operand) `->` type($result)

Returns bessel_i1e(operand) element-wise.

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.broadcast_add (chlo::BroadcastAddOp)

Addition operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_add` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs + rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, Commutative, HLO_BroadcastingElementwise, HLO_CompatibleOperandsAndResultElementType, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_and (chlo::BroadcastAndOp)

Logical and operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_and` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns logical_and(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, Commutative, HLO_BroadcastingElementwise, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_atan2 (chlo::BroadcastAtan2Op)

Atan2 operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_atan2` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns atan2(lhs/rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_compare (chlo::BroadcastCompareOp)

Compare operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_compare` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Compares lhs and rhs elementwise according to comparison_direction and compare_type. If unspecified, compare_type is FLOAT for float element types, SIGNED for signed element types and UNSIGNED for unsigned element types.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_comparison_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_complex (chlo::BroadcastComplexOp)

Complex operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_complex` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Performs element-wise conversion of a pair of real and imaginary values to a complex value.

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_divide (chlo::BroadcastDivOp)

Division operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_divide` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs / rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_maximum (chlo::BroadcastMaxOp)

Maximum operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_maximum` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns max(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, Commutative, HLO_BroadcastingElementwise, HLO_CompatibleOperandsAndResultElementType, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_minimum (chlo::BroadcastMinOp)

Minimum operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_minimum` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns min(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, Commutative, HLO_BroadcastingElementwise, HLO_CompatibleOperandsAndResultElementType, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_multiply (chlo::BroadcastMulOp)

Multiplication operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_multiply` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs * rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, Commutative, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_next_after (chlo::BroadcastNextAfterOp)

Std::nextafter operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_next_after` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns the next representable value of lhs in the direction of rhs, element-wise. It can also return a subnormal number.

Equivalent to the C++ std::nextafter function.

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_or (chlo::BroadcastOrOp)

Logical or operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_or` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns logical_or(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, Commutative, HLO_BroadcastingElementwise, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_polygamma (chlo::BroadcastPolygammaOp)

Polygamma function (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_polygamma` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns Polygamma(operand, operand) element-wise.

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_power (chlo::BroadcastPowOp)

Power operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_power` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs ^ rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_remainder (chlo::BroadcastRemOp)

Remainder operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_remainder` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs % rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_select (chlo::BroadcastSelectOp)

Select operator (with optional numpy-style broadcasting)

Syntax:

operation ::= `chlo.broadcast_select` $pred `,` $on_true `,` $on_false attr-dict `:`
              `(` type($pred) `,` type($on_true) `,` type($on_false) `)` `->` type(results)

Constructs an output array from elements of two input arrays, based on the values of a predicate array.

See https://www.tensorflow.org/xla/operation_semantics#select

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.broadcast_shift_left (chlo::BroadcastShiftLeftOp)

Shift left operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_shift_left` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs << rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_shift_right_arithmetic (chlo::BroadcastShiftRightArithmeticOp)

Shift right arithmetic operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_shift_right_arithmetic` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs >> rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_shift_right_logical (chlo::BroadcastShiftRightLogicalOp)

Shift right logical operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_shift_right_logical` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs >> rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_subtract (chlo::BroadcastSubOp)

Subtraction operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_subtract` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns lhs - rhs element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_xor (chlo::BroadcastXorOp)

Logical xor operator (with optional broadcasting)

Syntax:

operation ::= `chlo.broadcast_xor` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns logical_xor(lhs, rhs) element-wise.

See https://www.tensorflow.org/xla/operation_semantics#element-wise_binary_arithmetic_operations

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, Commutative, HLO_BroadcastingElementwise, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.broadcast_zeta (chlo::BroadcastZetaOp)

Hurwitz zeta function

Syntax:

operation ::= `chlo.broadcast_zeta` $lhs `,` $rhs attr-dict `:`
              `(` type($lhs) `,` type($rhs) `)` `->` type(results)

Returns Zeta(operand, operand) element-wise.

\[ \(\zeta(x, q) = \sum_{n=0}^{\infty} (q + n)^{-x}\) \]

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.conj (chlo::ConjOp)

Conj operator

Syntax:

operation ::= `chlo.conj` $operand attr-dict `:` type($operand) `->` type($result)

Returns Conj(operand) element-wise.

\[ \conj(x) = (\real(x), \neg(\imag(x))) \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.constant (chlo::ConstantOp)

Constant operator

Syntax:

operation ::= `chlo.constant` attr-dict $value

Represents a constant value.

Traits: AlwaysSpeculatableImplTrait, ConstantLike

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Results:

chlo.constant_like (chlo::ConstantLikeOp)

Constant like operator

Returns a splat constant of the same shape as the operand.

Traits: AlwaysSpeculatableImplTrait, CHLO_Broadcasting, HLO_BroadcastingElementwise, InferTensorType, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.cosh (chlo::CoshOp)

Cosh operator

Syntax:

operation ::= `chlo.cosh` $operand attr-dict `:` type($operand) `->` type($result)

Returns Cosh(operand) element-wise.

\[ \cosh(x) = (e^x + e^-x) / 2 \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.digamma (chlo::DigammaOp)

Digamma function

Syntax:

operation ::= `chlo.digamma` $operand attr-dict `:` type($operand) `->` type($result)

Returns Digamma(operand) element-wise.

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.erf (chlo::ErfOp)

Erfc operator

Syntax:

operation ::= `chlo.erf` $operand attr-dict `:` type($operand) `->` type($result)

Computes the Gauss error function of x element-wise.

erf(x) = erf_impl(x) if |x| < 1 = 1 - erfc_impl(x) otherwise

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.erf_inv (chlo::ErfInvOp)

Inverse Erf

Syntax:

operation ::= `chlo.erf_inv` $operand attr-dict `:` type($operand) `->` type($result)

Returns ErfInv(operand) element-wise.

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.erfc (chlo::ErfcOp)

Erfc operator

Syntax:

operation ::= `chlo.erfc` $operand attr-dict `:` type($operand) `->` type($result)

Computes an approximation of the error function complement (1 - erf(x)).

erfc(x) = erfc_impl(x) if |x| > 1 = 1 - erf_impl(x) otherwise

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.is_inf (chlo::IsInfOp)

IsInf predicate

Syntax:

operation ::= `chlo.is_inf` $operand attr-dict `:` type($operand) `->` type($result)

Returns if a value is +/-inf element-wise.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.is_neg_inf (chlo::IsNegInfOp)

IsNegInf predicate

Syntax:

operation ::= `chlo.is_neg_inf` $operand attr-dict `:` type($operand) `->` type($result)

Returns if a value is -inf element-wise.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.is_pos_inf (chlo::IsPosInfOp)

IsPosInf predicate

Syntax:

operation ::= `chlo.is_pos_inf` $operand attr-dict `:` type($operand) `->` type($result)

Returns if a value is +inf element-wise.

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.lgamma (chlo::LgammaOp)

Lgamma function

Syntax:

operation ::= `chlo.lgamma` $operand attr-dict `:` type($operand) `->` type($result)

Returns Lgamma(operand) element-wise.

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.next_after (chlo::NextAfterOp)

Std::nextafter operator

Syntax:

operation ::= `chlo.next_after` $x `,` $y attr-dict `:` type($x) `,` type($y) `->` type(results)

Returns the next representable value of x in the direction of y, element-wise. It can also return a subnormal number.

Equivalent to the C++ std::nextafter function.

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.polygamma (chlo::PolygammaOp)

Polygamma function

Syntax:

operation ::= `chlo.polygamma` $n `,` $x attr-dict `:` type($n) `,` type($x) `->` type(results)

Returns Polygamma(operand, operand) element-wise.

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.ragged_dot (chlo::RaggedDotOp)

Computes a matmul over a single ragged dimension

This operation takes three tensor args---lhs, rhs, and group_sizes---and a "ragged_dot_dimension_numbers" attribute. Like dot_general, the lhs and rhs are allowed arbitrary batch and contracting dimensions. Additionally, the lhs is required to have one ragged dimension, and the rhs may have at most one group dimension. The op has three modes, depending on the kind of the lhs ragged dimension.

In mode 1, the shape-signature is [b,m,k], [g,b,k,n], [b,g] -> [b,m,n]. Here the ragged dimension is an lhs non-contracting dimension (m). The dimensions b and k represent batch and contracting dimensions respectively. The rhs is required to have a group dimension (g).

In mode 2, the shape-signature is [b,m,k], [b,k,n], [b,g] -> [g,b,m,n]. Here the ragged dimension is an lhs/rhs contracting dimension (k).

In mode 3, the shape-signature is [b,m,k], [b,k,n], [g] -> [b,m,n]. Here the ragged dimension is an lhs/rhs batch dimension (b).

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.sinh (chlo::SinhOp)

Sinh operation

Syntax:

operation ::= `chlo.sinh` $operand attr-dict `:` type($operand) `->` type($result)

Returns Sinh(operand) element-wise.

\[ \sinh(x) = (e^x - e^-x) / 2 if |x| < 1 = e^(x + log(1/2)) - e^(-x + log(1/2)) otherwise. \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.square (chlo::SquareOp)

Square operation

Syntax:

operation ::= `chlo.square` $operand attr-dict `:` type($operand) `->` type($result)

Returns Square(operand) element-wise.

\[ \square(x) = complex((x.real - x.imag) * (x.real + x.imag), x.real * x.imag * 2) if x is a complex number = x * x otherwise \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.tan (chlo::TanOp)

Tan operation

Syntax:

operation ::= `chlo.tan` $operand attr-dict `:` type($operand) `->` type($result)

Returns Tan(operand) element-wise.

\[ \tan(x) = \sin(x) / \cos(x) \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

chlo.top_k (chlo::TopKOp)

Finds values and indices of the k largest elements for the last dimension

Syntax:

operation ::= `chlo.top_k` `(`$operand `,` `k` `=` $k`)` attr-dict `:`
              type($operand) `->` `(`type($values)`,` type($indices)`)`

If the input is a vector (rank-1), finds the k largest entries in the vector and outputs their values and indices as vectors. Thus values[j] is the j-th largest entry in input, and its index is indices[j].

For matrices (resp. higher rank input), computes the top k entries in each row (resp. vector along the last dimension). Thus,

values.shape = indices.shape = input.shape[:-1] + [k]

If two elements are equal, the lower-index element appears first.

Traits: AlwaysSpeculatableImplTrait, InferTensorType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Attributes:

Operands:

Results:

chlo.zeta (chlo::ZetaOp)

Hurwitz zeta function

Syntax:

operation ::= `chlo.zeta` $x `,` $q attr-dict `:` type($x) `,` type($q) `->` type(results)

Returns Zeta(operand, operand) element-wise.

\[ \(\zeta(x, q) = \sum_{n=0}^{\infty} (q + n)^{-x}\) \]

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Results:

Attributes

ComparisonDirectionAttr

Which comparison operation to perform.

Syntax:

#chlo.comparison_direction<
  ::mlir::chlo::ComparisonDirection   # value
>

Enum cases:

• EQ (EQ)
• NE (NE)
• GE (GE)
• GT (GT)
• LE (LE)
• LT (LT)

Parameters:

ComparisonTypeAttr

Which comparison type to use.

Syntax:

#chlo.comparison_type<
  ::mlir::chlo::ComparisonType   # value
>

Enum cases:

• NOTYPE (NOTYPE)
• FLOAT (FLOAT)
• TOTALORDER (TOTALORDER)
• SIGNED (SIGNED)
• UNSIGNED (UNSIGNED)

Parameters:

PrecisionAttr

XLA precision for an operand. Has backend specific meaning.

Syntax:

#chlo.precision<
  ::mlir::chlo::Precision   # value
>

Enum cases:

• DEFAULT (DEFAULT)
• HIGH (HIGH)
• HIGHEST (HIGHEST)

Parameters:

RaggedDotDimensionNumbersAttr

Attribute that models the dimension information for ragged dot.

Parameters:

Enums

ComparisonDirection

Which comparison operation to perform.

Cases:

ComparisonType

Which comparison type to use.

Cases:

Precision

XLA precision for an operand. Has backend specific meaning.

Cases: