'mhlo' Dialecto

Definición de operación

mhlo.abs (mhlo::AbsOp)

operación de abdominales

Sintaxis:

operation ::= `mhlo.abs` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Realiza una operación abs por elementos en el tensor operand y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#abs

Ejemplo:

%result = mhlo.abs %operand : tensor<3xi32>

Rasgos: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Operandos:

Resultados:

mhlo.add (mhlo::AddOp)

Agregar operación

Sintaxis:

operation ::= `mhlo.add` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Realiza la suma por elementos de dos tensores lhs y rhs y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#add

Ejemplo:

%result = mhlo.add %lhs, %rhs : tensor<2x2xi32>

Rasgos: AlwaysSpeculatableImplTrait, Conmutativo, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Operandos:

Resultados:

mhlo.add_dependency (mhlo::AddDependencyOp)

Operación Agregar Dependencia

Sintaxis:

operation ::= `mhlo.add_dependency` operands attr-dict `:` functional-type(operands, results)

Esta operación es privada del compilador XLA, por lo que aún no tiene una especificación.

De manera informal, esta operación tiene dos operandos: un operando de datos y un token. La salida de la operación es el operando de datos. Cuando se usa con AfterAll, esta operación permite ordenar operaciones sin efectos secundarios (aquellas que no producen valores simbólicos).

Ejemplo:

%1 = mhlo.add_dependency %arg0, %0 : (tensor<3x4xf32>, !mhlo.token) -> tensor<3x4xf32>

Rasgos: siempre especulableImplRasgo

Interfaces: condicionalmente especulable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Operandos:

Resultados:

mhlo.after_all (mhlo::AfterAllOp)

Operación después de todo

Sintaxis:

operation ::= `mhlo.after_all` $inputs attr-dict
              `:` custom<VariadicSameOperandsAndResultType>(ref($inputs), type($inputs), type($result))

Garantiza que las operaciones que producen las inputs se ejecuten antes que cualquier operación que dependa del result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#after_all

Ejemplo:

%result = mhlo.after_all %input0, %input1 : !mhlo.token

Rasgos: siempre especulableImplRasgo

Interfaces: condicionalmente especulable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Operandos:

Resultados:

mhlo.all_gather (mhlo::AllGatherOp)

Operación AllGather

Dentro de cada grupo de procesos en la cuadrícula de procesos, concatena los valores del tensor operand de cada proceso a lo largo de all_gather_dim y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_gather

Ejemplo:

%result = "mhlo.all_gather"(%operand) {
  all_gather_dim = 1 : i64,
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
  // channel_id = 0
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>,
  // use_global_device_ids = false
} : (tensor<2x2xf32>) -> tensor<2x4xf32>

Rasgos: SameOperandsAndResultElementType

Atributos:

Operandos:

Resultados:

mhlo.all_reduce (mhlo::AllReduceOp)

Operación AllReduce

Dentro de cada grupo de procesos en la cuadrícula de procesos, aplica un computation función de reducción a los valores de un tensor de operando de cada proceso y produce un tensor de resultado. El computation se aplica por separado para cada operando en operands , produciendo un resultado por operando.

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_reduce

Ejemplo:

%result = "mhlo.all_reduce"(%operand) ({
  ^bb0(%arg0: tensor<f32>, %arg1: tensor<f32>):
    %0 = mhlo.add %arg1, %arg2 : tensor<f32>
    mhlo.return %0 : tensor<f32>
}) {
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
  // channel_id = 0
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
  // use_global_device_ids = false
} : (tensor<4xf32>) -> tensor<4xf32>

Rasgos: InferTensorType, SameOperandsAndResultElementType, SingleBlock, SingleBlockImplicitTerminator

Interfaces: InferShapedTypeOpInterface, InferTypeOpInterface

Atributos:

Operandos:

Resultados:

mhlo.all_to_all (mhlo::AllToAllOp)

Operación todo a todo

Dentro de cada grupo de procesos en la cuadrícula de procesos, divide los valores del tensor operand a lo largo split_dimension en partes, dispersa las partes divididas entre los procesos, concatena las partes dispersas a lo largo de concat_dimension y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_to_all

Ejemplo:

%result = "mhlo.all_to_all"(%operand) {
  split_dimension = 1 : i64,
  concat_dimension = 0 : i64,
  split_count = 2 : i64,
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
} : (tensor<2x4xf32>) -> tensor<4x2xf32>

Rasgos: AlwaysSpeculatableImplTrait, InferTensorType, SameOperandsElementType, SameOperandsShape, SameVariadicOperandSize

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Atributos:

Operandos:

Resultados:

mhlo.and (mhlo::AndOp)

y operación

Sintaxis:

operation ::= `mhlo.and` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Realiza AND por elementos de dos tensores lhs y rhs y produce un tensor result

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#and

Ejemplo:

%result = mhlo.and %lhs, %rhs : tensor<2x2xi32>

Rasgos: AlwaysSpeculatableImplTrait, Conmutativo, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Operandos:

Resultados:

mhlo.async_done (mhlo::AsyncDoneOp)

Operación asíncrona

Esta operación es privada del compilador XLA, por lo que aún no tiene una especificación.

De manera informal, esta operación se bloquea hasta el final de un cálculo asincrónico. Devuelve el resultado final del cálculo asincrónico.

Consulte la documentación de AsyncStart para obtener más información.

Interfaces: InferTypeOpInterface

Atributos:

Operandos:

Resultados:

mhlo.async_start (mhlo::AsyncStartOp)

Operación de inicio asíncrono

Esta operación es privada del compilador XLA, por lo que aún no tiene una especificación.

De manera informal, esta operación inicia un cálculo asincrónico.

Esto se utiliza cuando hay funciones que contienen esperas asincrónicas (como DMA) y cálculos en subprocesos. Por ejemplo, una función podría consistir en un cálculo, un DMA, otro cálculo, un segundo DMA y un cálculo final. Esto se representaría como async_start seguido de async_update y async_done. async_start haría el primer cálculo en el subproceso y luego iniciaría el DMA. async_update esperaría a que se complete el DMA si aún no se ha hecho, luego ejecutará el segundo cálculo en la función e iniciará el segundo DMA. Finalmente, async_done esperaría en este último DMA y luego ejecutaría el último cálculo que debe ejecutarse en el subproceso y devolvería el resultado de ese cálculo final.

operands se pasan al cálculo directamente called_computation es la función que se ejecutará de forma asincrónica execution_thread es el nombre del hilo en el que se ejecutará. El hilo principal se llama "principal". Todos los hilos tienen nombres. group_id etiqueta un conjunto de operaciones de inicio asíncrono, de finalización asíncrona y cero o más operaciones de actualización asíncrona correspondientes al mismo cálculo. Representamos un group_id faltante con un valor negativo o Ninguno.

Esto devuelve todo el estado necesario entre operaciones asíncronas. Después de la asignación del búfer, los valores devueltos representan el espacio necesario para contener la entrada, los resultados y cualquier bloc de notas necesario o editado por la operación asíncrona.

Atributos:

Operandos:

Resultados:

mhlo.async_update (mhlo::AsyncUpdateOp)

Operación de actualización asíncrona

Esta operación es privada del compilador XLA, por lo que aún no tiene una especificación.

De manera informal, esta operación se bloquea en un cálculo asincrónico hasta una barrera de sincronización. Este bundle devuelve después de operarlo.

Consulte la documentación de AsyncStart para obtener más información.

Interfaces: InferTypeOpInterface

Atributos:

Operandos:

Resultados:

mhlo.atan2 (mhlo::Atan2Op)

operación atan2

Sintaxis:

operation ::= `mhlo.atan2` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Realiza la operación atan2 por elementos en los tensores lhs y rhs y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#atan2

Ejemplo:

%result = mhlo.atan2 %lhs, %rhs : tensor<3xf32>

Rasgos: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Operandos:

Resultados:

mhlo.batch_norm_grad (mhlo::BatchNormGradOp)

Operación BatchNormGrad

Calcula los gradientes de varias entradas de BatchNormTrainingOp que se propagan hacia atrás desde grad_output y produce tensores grad_operand , grad_scale y grad_offset .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_grad

Ejemplo:

%grad_operand, %grad_scale, %grad_offset =
"mhlo.batch_norm_grad"(%operand, %scale, %mean, %variance, %grad_output) {
  epsilon = 0.0 : f32,
  feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>,
    tensor<2x2x2xf32>) -> (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>)

Rasgos: AlwaysSpeculatableImplTrait, InferTensorType

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Atributos:

Operandos:

Resultados:

mhlo.batch_norm_inference (mhlo::BatchNormInferenceOp)

Operación BatchNormInference

Normaliza el tensor operand en todas las dimensiones excepto en la dimensión feature_index y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_inference

Ejemplo:

%result = "mhlo.batch_norm_inference"(%operand, %scale, %offset, %mean, %variance) {
  epsilon = 0.0 : f32,
  feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>) -> tensor<2x2x2xf32>

Rasgos: AlwaysSpeculatableImplTrait, InferTensorType

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Atributos:

Operandos:

Resultados:

mhlo.batch_norm_training (mhlo::BatchNormTrainingOp)

Operación de entrenamiento BatchNorm

Calcula la media y la varianza entre dimensiones espaciales y por lotes y normaliza el tensor operand para cada característica en la dimensión feature_index y produce tensores output , batch_mean y batch_var .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_training

Ejemplo:

%output, %batch_mean, %batch_var = "mhlo.batch_norm_training"(%operand, %scale, %offset) {
  epsilon = 0.0 : f32,
  feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>) -> (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>)

Rasgos: AlwaysSpeculatableImplTrait, InferTensorType

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Atributos:

Operandos:

Resultados:

mhlo.bitcast (mhlo::BitcastOp)

Operación de transmisión de bits

Sintaxis:

operation ::= `mhlo.bitcast` operands attr-dict `:` functional-type(operands, results)

Esta operación es privada del compilador XLA, por lo que aún no tiene una especificación.

De manera informal, esta operación cambia la forma de la entrada de la misma manera que la disposición física de los elementos no cambia.

Esta operación necesita información de diseño para entender la "disposición física de los elementos", y el soporte de diseño en MHLO es actualmente un trabajo en progreso.

Ejemplo:

%0 = mhlo.bitcast %arg0 : (tensor<3x4xf32>) -> tensor<3x4x1xf32>

Rasgos: siempre especulableImplRasgo

Interfaces: condicionalmente especulable, sin memoria (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Operandos:

Resultados:

mhlo.bitcast_convert (mhlo::BitcastConvertOp)

Operación BitcastConvert

Sintaxis:

operation ::= `mhlo.bitcast_convert` operands attr-dict `:` functional-type(operands, results)

Realiza una operación de difusión de bits en el tensor operand y produce un tensor result donde los bits de todo el tensor operand se reinterpretan utilizando el tipo de tensor de result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#bitcast_convert

Ejemplo:

%result = mhlo.bitcast_convert %operand : (tensor<2xf32>) -> tensor<2x4xi8>

Rasgos: siempre especulableImplRasgo

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Operandos:

Resultados:

mhlo.broadcast (mhlo::BroadcastOp)

Operación de transmisión

Esta operación está saliendo de StableHLO, por lo que no está incluida en la especificación: https://github.com/openxla/stablehlo/issues/3

De manera informal, esta operación hace lo mismo que la transmisión de XLA: https://www.tensorflow.org/xla/operation_semantics#broadcast

Ejemplo:

%result = mhlo.broadcast %operand, sizes = [1, 2] : (tensor<3xi32>) -> tensor<1x2x3xi32>

Rasgos: AlwaysSpeculatableImplTrait, InferTensorType, SameOperandsAndResultElementType

Interfaces: condicionalmente especulable, InferShapedTypeOpInterface, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Atributos:

Operandos:

Resultados:

mhlo.broadcast_in_dim (mhlo::BroadcastInDimOp)

Operación BroadcastInDim

Expande las dimensiones y/o el rango de un tensor de entrada duplicando los datos en el tensor operand y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#broadcast_in_dim

Ejemplo:

%result = mhlo.broadcast_in_dim %operand, dims = [2, 1] : (tensor<1x3xi32>) -> tensor<2x3x2xi32>

Rasgos: AlwaysSpeculatableImplTrait, HLO_CompatibleOperandsAndResultElementType

Interfaces: condicionalmente especulable, sin memoria (MemoryEffectOpInterface)

Efectos: MemoryEffects::Efecto{}

Atributos:

Operandos:

Resultados:

mhlo.case (mhlo::CaseOp)

Operación de caso

Produce el resultado de ejecutar exactamente una function de branches dependiendo del valor de index .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#case

Ejemplo:

%result0, %result1 = "mhlo.case"(%index) ({
  mhlo.return %result_branch0, %result_branch0 : tensor<2xi64>, tensor<2xi64>
}, {
  mhlo.return %result_branch1, %result_branch1 : tensor<2xi64>, tensor<2xi64>
}) : (tensor<i32>) -> (tensor<2xi64>, tensor<2xi64>)

Rasgos: recursivemoryEffects, un solo bloqueo, un solo bloqueo de mínimo

Interfaces: infertypeopeinterface

Operands:

Resultados:

mhlo.cbrt (mhlo :: cbrtop)

Operación CBRT

Sintaxis:

operation ::= `mhlo.cbrt` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Realiza el funcionamiento de la raíz cúbica en forma de elemento en el tensor operand y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#cbrt

Ejemplo:

%result = mhlo.cbrt %operand : tensor<4xf32>

Rasgos: AlwayspeculatableImtrait, compatibleperandsandresulttype, elementwise, lo mismo ocurre y resulta

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.ceil (Mhlo :: Ceilop)

Operación de techo

Sintaxis:

operation ::= `mhlo.ceil` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Realiza un techo de elementos de tensor de operand y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#ceil

Ejemplo:

%result = mhlo.ceil %operand : tensor<5xf32>

Rasgos: AlwayspeculatableImtrait, compatibleperandsandresulttype, elementwise, lo mismo ocurre y resulta

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.cholesky (mhlo :: choleskyop)

Operación de Cholesky

Calcula la descomposición de Cholesky de un lote de matrices.

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#cholesky

Ejemplo:

%result = mhlo.cholesky %a, lower = true : tensor<3x3xf32>

Rasgos: AlwaysSpeculatableImtrait, infertenSortype, lo mismo ocurre y compleja

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Operands:

Resultados:

mhlo.clamp (Mhlo :: Clampop)

Operación de sujeción

Sintaxis:

operation ::= `mhlo.clamp` $min `,` $operand `,` $max attr-dict
              `:` custom<SameOperandsAndResultType>(type($min), type($operand), type($max), type($result))

Agua cada elemento del tensor operand entre un valor mínimo y máximo y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#clamp

Ejemplo:

%result = mhlo.clamp %min, %operand, %max : tensor<3xi32>

Rasgos: AlwaysSpeculatableImpltrait, hlo_broadcastingElementwise, infertenSortype, lo mismo ocurre y compleja

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.collective_permute (mhlo :: colectivepermuteop)

Operación colectiva por periódico

Dentro de cada grupo de proceso en la cuadrícula de proceso, envía el valor del tensor operand desde el proceso de origen al proceso de destino y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#collective_permute

Ejemplo:

%result = "mhlo.collective_permute"(%operand) {
  source_target_pairs = dense<[[0, 1], [1, 2]]> : tensor<2x2xi64>,
  // channel_id = 0
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
} : (tensor<4x2xf32>) -> tensor<4x2xf32>

Rasgos: AlwaysSpeculatableImtrait, compatibleperandsandresultType

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Operands:

Resultados:

mhlo.compare (mhlo :: compareop)

Comparar operación

Sintaxis:

operation ::= `mhlo.compare` $comparison_direction `,` $lhs `,` $rhs (`,` $compare_type^)?
              attr-dict `:` functional-type(operands, results)

Realiza la comparación de elementos de los tensores lhs y rhs de acuerdo con comparison_direction y compare_type , y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#compare

Ejemplo:

%result = mhlo.compare LT, %lhs, %rhs, FLOAT : (tensor<2xf32>, tensor<2xf32>) -> tensor<2xi1>

Rasgos: AlwayspeculatableImplitTait, Elementwise, infertenSortype, lo mismo ocurre y se resulta, lo mismo ocurre y sellaype

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Operands:

Resultados:

mhlo.complex (mhlo :: compleP)

Operación compleja

Sintaxis:

operation ::= `mhlo.complex` operands attr-dict
              `:` custom<ComplexOpType>(type($lhs), type($rhs), type($result))

Realiza una conversión de elementos a un valor complejo de un par de valores reales e imaginarios, lhs y rhs , y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#complex

Ejemplo:

%result = mhlo.complex %lhs, %rhs : tensor<2xcomplex<f32>>

Rasgos: AlwayspeculatableImtrait, elementwise, lo mismo ocurre y resulta, lo mismo ocurre y selectionType

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.compute_reshape_shape (mhlo :: computereshapeShapeop)

Operación de compensación

Sintaxis:

operation ::= `mhlo.compute_reshape_shape` operands attr-dict `:` functional-type(operands, results)

Esta operación es un trabajo en progreso, por lo que aún no se incluye en la especificación: https://github.com/openxla/stablehlo/issues/8

Informalmente, esta operación calcula una salida_shape para DynamicReshapeop a partir del número de elementos num_elements en un operando de DynamicReshapeop y la forma dynamic_shape proporcionada a la reashape de TF: https://www.tensorflow.org/api_docs/pythontt/reshape

Por ejemplo, para num_elements = 12 y dynamic_shape = [2, -1] , el result será [2, 6] . Si los operandos no son válidos (por ejemplo, si las dimensiones no dividen uniformemente el número de elementos, o si hay múltiples valores de -1 en las dimensiones), esto conduce a un comportamiento indefinido.

Ejemplo:

%result = mhlo.compute_reshape_shape %num_elements, %dynamic_shape
       : (index, tensor<2xi32>) -> tensor<2xi32>

Rasgos: AlwaysSpeculatableImtrait

Interfaces: ConditionlySpeculatable, NomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.concatenate (mhlo :: concatenateop)

Concatenato de operación

Concatena un número variádico de tensores en inputs a lo largo de la dimensión dimension en el mismo orden que los argumentos dados y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#concatenate

Ejemplo:

%result = mhlo.concatenate %input0, %input1, dim = 0 : (tensor<3x2xi64>, tensor<1x2xi64>) -> tensor<4x2xi64>

Rasgos: AlwaysSpeculatableImtrait, Sameperands yResultElentType

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Operands:

Resultados:

mhlo.constant (mhlo :: constantop)

Operación constante

Produce un tensor output a partir de un value constante.

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#constant

Ejemplo:

%output = mhlo.constant dense<[[0.0, 1.0], [2.0, 3.0]]> : tensor<2x2xf32>

Rasgos: AlwaysSpeculatableImtrait, constante

Interfaces: condicionalmente especulable, infertypeopinterface, nomemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Resultados:

mhlo.convert (mhlo :: convertop)

Convertir operación

Sintaxis:

operation ::= `mhlo.convert` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Realiza una conversión de elemento de un tipo de elemento a otro en el tensor operand y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#convert

Ejemplo:

%result = mhlo.convert %operand : (tensor<3xi32>) -> tensor<3xcomplex<f32>>

Rasgos: AlwayspeculatableImtrait, elementwise, lo mismo ocurre y se produce

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, NomemoryEffect (MemoryEffectopinterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.convolution (mhlo :: convolución)

Operación de convolución

Sintaxis:

operation ::= `mhlo.convolution` `(`operands`)`
              `dim_numbers` `=` custom<ConvolutionDimensions>($dimension_numbers) `,`
              `window` `=` `{` custom<WindowAttributes>($window_strides, $padding,
              $lhs_dilation, $rhs_dilation,
              $window_reversal) `}`
              attr-dict `:` functional-type(operands, results)

Calcula productos DOT entre ventanas de lhs y rodajas de rhs y produce result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#convolution

Ejemplo:

%result = "mhlo.convolution"(%lhs, %rhs) {
  window_strides = dense<4> : tensor<2xi64>,
  padding = dense<0> : tensor<2x2xi64>,
  lhs_dilation = dense<2> : tensor<2xi64>,
  rhs_dilation = dense<1> : tensor<2xi64>,
  window_reversal = dense<false> : tensor<2xi1>,
  dimension_numbers = #mhlo.conv<[b, 0, 1, f]x[0, 1, i, o]->[b, 0, 1, f]>,
  feature_group_count = 1 : i64,
  batch_group_count = 1 : i64,
  precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<1x4x4x1xi32>, tensor<3x3x1x1xi32>) -> tensor<1x2x2x1xi32>

Rasgos: AlwaysSpeculatableImtrait

Interfaces: ConditionlySpeculatable, NomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Operands:

Resultados:

mhlo.copy (mhlo :: copyop)

Operación de copia

Sintaxis:

operation ::= `mhlo.copy` operands attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Esta operación es privada para el compilador XLA, por lo que aún no tiene una especificación.

Informalmente, esta operación es una copia de operand . Dependiendo de los metadatos adjuntos a la operación, puede comportarse de manera muy diferente a una OP de OP.

Ejemplo:

%0 = mhlo.copy %arg0 : tensor<f32>

Rasgos: AlwaysSpeculatableImtrait, compatibleperandsandresultType

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Operands:

Resultados:

mhlo.cosine (Mhlo :: Cosineop)

Operación coseno

Sintaxis:

operation ::= `mhlo.cosine` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Realiza la operación de coseno en forma de elemento en el tensor operand y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#cosine

Ejemplo:

%result = mhlo.cosine %operand : tensor<2xf32>

Rasgos: AlwayspeculatableImtrait, compatibleperandsandresulttype, elementwise, lo mismo ocurre y resulta

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.count_leading_zeros (mhlo :: clzop)

Operación CLZ

Sintaxis:

operation ::= `mhlo.count_leading_zeros` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Realiza un recuento de elementos del número de bits cero principales en el tensor operand y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#count_leading_zeross

Ejemplo:

%result = mhlo.count_leading_zeros %operand : tensor<2x2xi8>

Rasgos: AlwayspeculatableImtrait, compatibleperandsandresulttype, elementwise, lo mismo ocurre y resulta

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.create_token (mhlo :: createTokenop)

Operación de createToken

Sintaxis:

operation ::= `mhlo.create_token` attr-dict `:` type(results)

Esta operación está saliendo de Stablehlo, por lo que no está incluida en la especificación: https://github.com/openxla/stablehlo/issues/3

Informalmente, esta operación hace lo mismo que AfterAllop con 0 entradas: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#after_all

Ejemplo:

%output = mhlo.create_token : !mhlo.token

Rasgos: AlwaysSpeculatableImtrait

Interfaces: condicionalmente especulable, infertypeopinterface, nomemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects :: Effect {}

Resultados:

mhlo.cross-replica-sum (Mhlo :: CrossReplicAsumop)

Operación CrossReplicAsum

Esta operación está saliendo de Stablehlo, por lo que no está incluida en la especificación: https://github.com/openxla/stablehlo/issues/3

Informalmente, esta operación hace lo mismo que Allreduceop con channel_id = 0 , use_global_device_ids = false e implementación de computation Adición: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_reduce

Ejemplo:

%result = "mhlo.cross-replica-sum"(%operand) {
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
} : (tensor<4xf32>) -> tensor<4xf32>

Rasgos: AlwaysSpeculatableImtrait, compatibleperandsandresultType

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Operands:

Resultados:

mhlo.cstr_reshapable (mhlo :: cstrreshapableop)

Operación Cstrreshapable

Sintaxis:

operation ::= `mhlo.cstr_reshapable` operands attr-dict `:` functional-type(operands, results)

Esta operación es un trabajo en progreso, por lo que aún no se incluye en la especificación: https://github.com/openxla/stablehlo/issues/8

Informalmente, esta operación crea un testigo sobre la restricción de que ComputerShapeshape tendrá éxito con los operandos proporcionados.

Ejemplo:

%result = mhlo.cstr_reshapable %num_elements, %dynamic_shape
       : (index, tensor<3xi32>) -> !shape.witness

Rasgos: AlwaysSpeculatableImtrait

Interfaces: condicionalmente especulable, infertypeopinterface, nomemoryEffect (MemoryEffectOpInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.custom_call (mhlo :: customcallop)

Operación personalizada

Sintaxis:

operation ::= `mhlo.custom_call` custom<CustomCallTarget>($call_target_name) `(` $inputs `)`
              attr-dict `:` functional-type(operands, results)

Encapsula una operación definida por implementación call_target_name que toma inputs y called_computations y produce results .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#custom_call

Ejemplo:

%results = "mhlo.custom_call"(%input0) {
  call_target_name = "foo",
  has_side_effect = false,
  backend_config = "bar",
  api_version = 1 : i32,
  called_computations = [@foo]
} : (tensor<f32>) -> tensor<f32>

A custom call invokes code external to XLA. The `inputs` are passed to the
external code, and the external code is expected to produce a result of the
given type. The exact mechanism is backend-specific. For example, in the CPU
backend, a call instruction is emitted which targets a symbol with the name
`call_target_name`.

If XLA runtime is enabled for a backend, then custom calls use the runtime
custom call calling convention to call into the external functions. This
calling convention defines an ABI for encoding arguments, attributes and
results.

Depending on the API version there are two ways to pass extra bits of static
information to the external function:

1. For `API_VERSION_TYPED_FFI` custom calls `backend_config` must be a
   dictionary attribute, that will be encoded according to the custom call
   calling convention and passed to the external function as the attributes
   argument. External code is expected to use declarative bindings (see
   `xla/runtime/custom_call.h`) to decode them at run time. These custom
   calls are only supported if XLA uses XLA runtime.

2. For previous API versions it is the user responsibility to encode extra
   bits of static information as a string `backend_config` attribute, and
   decode it at run time.

Interfaces: MemoryEffectopinterface

Atributos:

Operands:

Resultados:

mhlo.divide (mhlo :: divop)

Operación Div

Sintaxis:

operation ::= `mhlo.divide` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Realiza la división de elementos de dividendos lhs y los tensores del divisor rhs y produce un tensor result .

Ver: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#divide

Ejemplo:

%result = mhlo.divide %lhs, %rhs : tensor<4xf32>

Rasgos: AlwayspeculatableImtrait, compatibleperandsandresulttype, elementwise, lo mismo ocurre y resulta

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Operands:

Resultados:

mhlo.domain (mhlo :: domainop)

Operación de dominio

Esta operación es privada para el compilador XLA, por lo que aún no tiene una especificación.

Informalmente, estas operaciones se utilizan para agrupar las instrucciones con la misma propiedad DomainMetadata. ShardingMetadata es el caso de uso principal hoy para agrupar las instrucciones en el mismo dispositivo. Las instrucciones de dominio proporcionan dos beneficios principales:

• Evite la optimización involuntaria de las instrucciones entre los dominios.
• Asigne automáticamente los metadatos de las instrucciones creadas en el dominio. Sin instrucciones de dominio, cada pase de optimización HLO tendría que verificar y propagar los metadatos, lo que sería fácil de perder y también agrega complejidad al compilador. Dado que las instrucciones de dominio conectan dos dominios diferentes, cada instrucción de dominio está asociada con dos metadatos de dominio, uno en el lado del operando y otro en el lado del usuario del dominio.

Rasgos: AlwaysSpeculatableImtrait, compatibleperandsandresultType

Interfaces: condicionalmente especulable, infershapedTypeopeRinterface, infertypeopeRinterface, nomemoryEffect (MemoryEffectopInterface)

Efectos: MemoryEffects :: Effect {}

Atributos:

Operands:

Resultados:

mhlo.dot (mhlo :: dotop)

Operación de puntos

Esta operación está saliendo de Stablehlo, por lo que no está incluida en la especificación: https://github.com/openxla/stablehlo/issues/3

Informally, this operation does the same thing as XLA's Dot: https://www.tensorflow.org/xla/operation_semantics#dot

Ejemplo:

%0 = mhlo.dot %arg0, %arg1 : (tensor<1x2xi32>, tensor<2x1xi32>) -> tensor<1x1xi32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.dot_general (mhlo::DotGeneralOp)

DotGeneral operation

Computes dot products between slices of lhs and slices of rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dot_general

Ejemplo:

%result = "mhlo.dot_general"(%lhs, %rhs) {
  dot_dimension_numbers = #mhlo.dot<
    lhs_batching_dimensions = [0],
    rhs_batching_dimensions = [0],
    lhs_contracting_dimensions = [2],
    rhs_contracting_dimensions = [1]
  >,
  precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<2x2x2xi32>, tensor<2x2x2xi32>) -> tensor<2x2x2xi32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.dynamic_broadcast_in_dim (mhlo::DynamicBroadcastInDimOp)

DynamicBroadcastInDim operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as BroadcastInDimOp except that the result shape is specified dynamically via output_dimensions : https://github.com/openxla/stablehlo/blob/main/docs/spec.md#broadcast_in_dim

It also accepts optional attributes to express static knowledge about the expanding behavior of dimensions. If not specified, all dimensions are assumed to be possibly expanding. The sets of dimensions that are known to be expanding and the set of dimensions that are known to be non-expanding must be disjoint and they must be a subset of the operand's dimensions.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.dynamic_conv (mhlo::DynamicConvOp)

DynamicConv operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as ConvolutionOp except that padding is specified dynamically via d_padding : https://github.com/openxla/stablehlo/blob/main/docs/spec.md#convolution

Ejemplo:

%result = "mhlo.dynamic_conv"(%lhs, %rhs, %d_padding) {
  window_strides = dense<4> : tensor<2xi64>,
  lhs_dilation = dense<2> : tensor<2xi64>,
  rhs_dilation = dense<1> : tensor<2xi64>,
  window_reversal = dense<false> : tensor<2xi1>,
  dimension_numbers = #mhlo.conv<[b, 0, 1, f]x[0, 1, i, o]->[b, 0, 1, f]>,
  feature_group_count = 1 : i64,
  batch_group_count = 1 : i64,
  precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<1x4x4x1xi32>, tensor<3x3x1x1xi32>, tensor<2x2xi64>) -> tensor<1x2x2x1xi32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.dynamic_gather (mhlo::DynamicGatherOp)

DynamicGather operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as GatherOp except that slice_sizes are specified dynamically: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#gather

Ejemplo:

%result = "mhlo.dynamic_gather"(%operand, %start_indices, %slice_sizes) {
  dimension_numbers = #mhlo.gather<
    offset_dims = [2, 3],
    collapsed_slice_dims = [0],
    start_index_map = [0, 2],
    index_vector_dim = 2>,
  indices_are_sorted = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>, tensor<3xi64>) -> tensor<2x3x2x2xi32>

Traits: AlwaysSpeculatableImplTrait, InferTensorType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.dynamic_iota (mhlo::DynamicIotaOp)

DynamicIota operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as IotaOp except that the result shape is specified dynamically via output_shape : https://github.com/openxla/stablehlo/blob/main/docs/spec.md#iota

Ejemplo:

%0 = mhlo.dynamic_iota %arg0, dim = 0 : (tensor<1xindex>) -> tensor<4xi32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.dynamic_pad (mhlo::DynamicPadOp)

DynamicPad operation

Sintaxis:

operation ::= `mhlo.dynamic_pad` operands attr-dict `:` functional-type(operands, results)

Dynamically Pads the operand , with amount of padding added at low-end/high-end/interior is passed through input tensors.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.dynamic_reshape (mhlo::DynamicReshapeOp)

DynamicReshape operation

Sintaxis:

operation ::= `mhlo.dynamic_reshape` operands attr-dict `:` functional-type(operands, results)

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as ReshapeOp except that the result shape is specified dynamically via output_shape : https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reshape

Ejemplo:

%0 = mhlo.dynamic_reshape %arg0, %shape : (tensor<?xf32>, tensor<2xindex>) -> tensor<?x?xf32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.dynamic_slice (mhlo::DynamicSliceOp)

DynamicSlice operation

Extracts a slice from the operand using dynamically-computed starting indices and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dynamic_slice

Ejemplo:

%result = mhlo.dynamic_slice %operand, %start_indices0, %start_indices1, sizes = [2, 2]
  : (tensor<4x4xi32>, tensor<i64>, tensor<i64>) -> tensor<2x2xi32>

Traits: AlwaysSpeculatableImplTrait, InferTensorType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.dynamic_update_slice (mhlo::DynamicUpdateSliceOp)

DynamicUpdateSlice operation

Sintaxis:

operation ::= `mhlo.dynamic_update_slice` operands attr-dict `:` functional-type(operands, results)

Produces a result tensor which is equal to the operand tensor except that the slice starting at start_indices is updated with the values in update .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dynamic_update_slice

Ejemplo:

%result = mhlo.dynamic_update_slice %operand, %update, %start_indices0, %start_indices1
  : (tensor<4x4xi32>, tensor<2x2xi32>, tensor<i64>, tensor<i64>) -> tensor<4x4xi32>

Traits: AlwaysSpeculatableImplTrait, InferTensorType

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.einsum (mhlo::EinsumOp)

Einsum operation

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3

Informally, this operation does the same thing as TF's einsum: https://www.tensorflow.org/api_docs/python/tf/einsum

Ejemplo:

%result = "mhlo.einsum"(%lhs, %rhs) {
  einsum_config = "ab,bc->ac"
} : (tensor<4x16xf32>, tensor<16x4xf32>) -> tensor<4x4xf32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.exponential (mhlo::ExpOp)

Exp operation

Sintaxis:

operation ::= `mhlo.exponential` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise exponential operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#exponential

Ejemplo:

%result = mhlo.exponential %operand : tensor<2x2xf64>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.exponential_minus_one (mhlo::Expm1Op)

Expm1 operation

Sintaxis:

operation ::= `mhlo.exponential_minus_one` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise exponential minus one operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#exponential_minus_one

Ejemplo:

%result = mhlo.exponential_minus_one %operand : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.fft (mhlo::FftOp)

Fft operation

Performs the forward and inverse Fourier transforms for real and complex inputs/outputs.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#fft

Ejemplo:

%result = mhlo.fft %operand, type = FFT, length = [4] : (tensor<4xcomplex<f32>>) -> tensor<4xcomplex<f32>>

Traits: AlwaysSpeculatableImplTrait, InferTensorType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.floor (mhlo::FloorOp)

Floor operation

Sintaxis:

operation ::= `mhlo.floor` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise floor of operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#floor

Ejemplo:

%result = mhlo.floor %operand : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.fusion (mhlo::FusionOp)

Fusion operation

This operation is private to the XLA compiler, so it is does not yet have a specification.

Informally, this operation consists of a group of basic ops (represented as a region attached to it). It serves as a hint to the backend that it is beneficial to emit the contained ops into a single loop nest or kernel.

Atributos:

Operands:

Resultados:

mhlo.gather (mhlo::GatherOp)

Gather operation

Gathers slices from operand tensor from offsets specified in start_indices and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#gather

Ejemplo:

%result = "mhlo.gather"(%operand, %start_indices) {
  dimension_numbers = #mhlo.gather<
    offset_dims = [2, 3],
    collapsed_slice_dims = [0],
    start_index_map = [0, 2],
    index_vector_dim = 2>,
  slice_sizes = dense<[0, 2, 2]> : tensor<3xi64>,
  indices_are_sorted = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>) -> tensor<2x3x2x2xi32>

Traits: AlwaysSpeculatableImplTrait, InferTensorType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.get_dimension_size (mhlo::GetDimensionSizeOp)

GetDimensionSize operation

Produces the size of the given dimension of the operand .

See https://github.com/openxla/stablehlo/blob/main/docs/spec.md#get_dimension_size

Ejemplo:

%result = mhlo.get_dimension_size %operand, dim = 1 : (tensor<2x3xf32>) -> tensor<i32>

Traits: AlwaysSpeculatableImplTrait, InferTensorType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.get_tuple_element (mhlo::GetTupleElementOp)

GetTupleElement operation

Sintaxis:

operation ::= `mhlo.get_tuple_element` $operand `[` $index `]` attr-dict `:` functional-type(operands, results)

Extracts element at index position of the operand tuple and produces a result .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#get_tuple_element

Ejemplo:

%result = mhlo.get_tuple_element %operand[0] : (tuple<tensor<2xf32>, tuple<tensor<i32>>>) -> tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.if (mhlo::IfOp)

If operation

Produces the output from executing exactly one branch from true_branch or false_branch depending on the value of pred .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#if

Example: %result = "mhlo.if"(%pred) ({ "mhlo.return"(%result_true_branch) : (tensor ) -> () }, { "mhlo.return"(%result_false_branch) : (tensor ) -> () }) : (tensor ) -> tensor

Traits: RecursiveMemoryEffects, SingleBlock, SingleBlockImplicitTerminator

Interfaces: InferTypeOpInterface

Operands:

Resultados:

mhlo.imag (mhlo::ImagOp)

Imag operation

Sintaxis:

operation ::= `mhlo.imag` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Extracts the imaginary part, element-wise, from the operand and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#imag

Ejemplo:

%result = mhlo.imag %operand : (tensor<2xcomplex<f32>>) -> tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.infeed (mhlo::InfeedOp)

Infeed operation

Reads data from the infeed and produces results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#infeed

Ejemplo:

%results:2 = "mhlo.infeed"(%token) {
  infeed_config = ""
} : (!mhlo.token) -> (tensor<3x3x3xi32>, !mhlo.token)

Atributos:

Operands:

Resultados:

mhlo.iota (mhlo::IotaOp)

Iota operation

Fills an output tensor with values in increasing order starting from zero along the iota_dimension dimension.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#iota

Ejemplo:

%output = mhlo.iota dim = 0 : tensor<4x5xi32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Resultados:

mhlo.is_finite (mhlo::IsFiniteOp)

IsFinite operation

Sintaxis:

operation ::= `mhlo.is_finite` $x attr-dict `:` functional-type(operands, results)

Performs element-wise check whether the value in x is finite (ie is neither +Inf, -Inf, nor NaN) and produces a y tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#is_finite

Ejemplo:

%y = mhlo.is_finite %x : (tensor<7xf32>) -> tensor<7xi1>

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.log (mhlo::LogOp)

Log operation

Sintaxis:

operation ::= `mhlo.log` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise logarithm operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#log

Ejemplo:

%result = mhlo.log %operand : tensor<2x2xf64>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.log_plus_one (mhlo::Log1pOp)

Log1p operation

Sintaxis:

operation ::= `mhlo.log_plus_one` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise logarithm plus one operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#log_plus_one

Ejemplo:

%result = mhlo.log_plus_one %operand : tensor<6xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.logistic (mhlo::LogisticOp)

Logistic operation

Sintaxis:

operation ::= `mhlo.logistic` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise logistic operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#logistic

Ejemplo:

%result = mhlo.logistic %operand : tensor<2x2xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.map (mhlo::MapOp)

Map operation

Applies a map function computation to inputs along the dimensions and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#map

Ejemplo:

%result = "mhlo.map"(%input0, %input1) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = mhlo.multiply %arg0, %arg1 : tensor<i32>
    mhlo.return %0 : tensor<i32>
}) {
  dimensions = dense<[0, 1]> : tensor<2xi64>
} : (tensor<2x2xi32>, tensor<2x2xi32>) -> tensor<2x2xi32>

Traits: InferTensorType, RecursiveMemoryEffects, SameOperandsAndResultShape, SingleBlock, SingleBlockImplicitTerminator

Interfaces: InferShapedTypeOpInterface, InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.maximum (mhlo::MaxOp)

Max operation

Sintaxis:

operation ::= `mhlo.maximum` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise max operation on tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#maximum

Ejemplo:

%result = mhlo.maximum %lhs, %rhs : tensor<4xf32>

Traits: AlwaysSpeculatableImplTrait, Commutative, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.minimum (mhlo::MinOp)

Min operation

Sintaxis:

operation ::= `mhlo.minimum` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise min operation on tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#minimum

Ejemplo:

%result = mhlo.minimum %lhs, %rhs : tensor<4xf32>

Traits: AlwaysSpeculatableImplTrait, Commutative, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.multiply (mhlo::MulOp)

Mul operation

Sintaxis:

operation ::= `mhlo.multiply` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise product of two tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#multiply

Ejemplo:

%result = mhlo.multiply %lhs, %rhs : tensor<2xi32>

Traits: AlwaysSpeculatableImplTrait, Commutative, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.negate (mhlo::NegOp)

Neg operation

Sintaxis:

operation ::= `mhlo.negate` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise negation of operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#negate

Ejemplo:

%result = mhlo.negate %operand : tensor<2x3xi32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.not (mhlo::NotOp)

Not operation

Sintaxis:

operation ::= `mhlo.not` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise NOT of tensor operand of type integer and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#not

Ejemplo:

%result = mhlo.not %operand : tensor<5x3x1xi1>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.optimization_barrier (mhlo::OptimizationBarrierOp)

OptimizationBarrier operation

Sintaxis:

operation ::= `mhlo.optimization_barrier` attr-dict ($operand^ `:` custom<PairwiseOpType>(type($operand), type($result))):(`(` `)`)?

Ensures that the operations that produce the operand are executed before any operations that depend on the result and prevents compiler transformations from moving operations across the barrier. Other than that, the operation is an identity, ie result = operand .

See https://github.com/openxla/stablehlo/blob/main/docs/spec.md#optimization_barrier

Ejemplo:

%result0, %result1 = mhlo.optimization_barrier %operand0, %operand1 : tensor<f32>, tensor<f32>

Traits: AlwaysSpeculatableImplTrait, HLO_PairwiseSameOperandAndResultType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.or (mhlo::OrOp)

Or operation

Sintaxis:

operation ::= `mhlo.or` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise OR of two tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#or

Ejemplo:

%result = mhlo.or %lhs, %rhs : tensor<2xi1>

Traits: AlwaysSpeculatableImplTrait, Commutative, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.outfeed (mhlo::OutfeedOp)

Outfeed operation

Writes inputs to the outfeed and produces a result token.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#outfeed

Ejemplo:

%result = "mhlo.outfeed"(%input0, %token) {
  outfeed_config = ""
} : (tensor<3x3x3xi32>, !mhlo.token) -> !mhlo.token

Interfaces: InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.pad (mhlo::PadOp)

Pad operation

Expands operand by padding around the tensor as well as between the elements of the tensor with the given padding_value .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#pad

Ejemplo:

%0 = mhlo.pad %arg0, %arg1, low = [0, 1], high = [2, 1], interior = [1, 2]
  : (tensor<2x3xi32>, tensor<i32>) -> tensor<5x9xi32>

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultElementType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.partition_id (mhlo::PartitionIdOp)

PartitionId operation

Sintaxis:

operation ::= `mhlo.partition_id` attr-dict `:` type(results)

Produces partition_id of the current process.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#partition_id

Ejemplo:

%result = mhlo.partition_id : tensor<ui32>

Interfaces: InferTypeOpInterface

Resultados:

mhlo.popcnt (mhlo::PopulationCountOp)

PopulationCount operation

Sintaxis:

operation ::= `mhlo.popcnt` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise count of the number of bits set in the operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#popcnt

Ejemplo:

%result = mhlo.popcnt %operand : tensor<4xi8>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.power (mhlo::PowOp)

Pow operation

Sintaxis:

operation ::= `mhlo.power` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise exponentiation of lhs tensor by rhs tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#power

Ejemplo:

%result = mhlo.power %lhs, %rhs : tensor<6xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.real (mhlo::RealOp)

Real operation

Sintaxis:

operation ::= `mhlo.real` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Extracts the real part, element-wise, from the operand and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#real

Ejemplo:

%result = mhlo.real %operand : (tensor<2xcomplex<f32>>) -> tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.real_dynamic_slice (mhlo::RealDynamicSliceOp)

RealDynamicSlice operation

Sintaxis:

operation ::= `mhlo.real_dynamic_slice` operands attr-dict `:` functional-type(operands, results)

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as SliceOp except that start_indices , limit_indices and strides are specified dynamically: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#slice

Ejemplo:

%result = mhlo.real_dynamic_slice %operand,
            %start_indices, %limit_indices, %strides
       : (tensor<256x?xf32>, tensor<2xindex>, tensor<2xindex>, tensor<2xindex>) -> tensor<256x?xf32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.recv (mhlo::RecvOp)

Recv operation

Receives data from a channel with channel_id and produces results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#recv

Ejemplo:

%results:2 = "mhlo.recv"(%token) {
  // channel_id = 5 : i64,
  // channel_type = #stablehlo<channel_type HOST_TO_DEVICE>,
  channel_handle = #mhlo.channel_handle<handle = 5, type = 3>,
  is_host_transfer = true
} : (!mhlo.token) -> (tensor<3x4xi32>, !mhlo.token)

Atributos:

Operands:

Resultados:

mhlo.reduce (mhlo::ReduceOp)

Reduce operation

Applies a reduction function body to inputs and init_values along the dimensions and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce

Ejemplo:

%result = "mhlo.reduce"(%input, %init_value) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
    "mhlo.return"(%0) : (tensor<i32>) -> ()
}) {
  dimensions = dense<1> : tensor<1xi64>
} : (tensor<1x6xi32>, tensor<i32>) -> tensor<1xi32>

Traits: InferTensorType, RecursiveMemoryEffects, SameVariadicOperandSize, SingleBlock, SingleBlockImplicitTerminator

Interfaces: InferShapedTypeOpInterface, InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.reduce_precision (mhlo::ReducePrecisionOp)

ReducePrecision operation

Sintaxis:

operation ::= `mhlo.reduce_precision` $operand `,` `format` `=` custom<ExponentMantissa>($exponent_bits, $mantissa_bits)
              attr-dict `:` custom<SameOperandsAndResultType>(type($operand), type($output))

Performs element-wise conversion of operand to another floating-point type that uses exponent_bits and mantissa_bits and back to the original floating-point type and produces an output tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_precision

Ejemplo:

%output = mhlo.reduce_precision %operand, format = e5m2 : tensor<6xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.reduce_scatter (mhlo::ReduceScatterOp)

ReduceScatter operation

Within each process group in the process grid, performs reduction, using computations , over the values of the operand tensor from each process, splits the reduction result along scatter_dimension into parts, and scatters the split parts between the processes to produce the result .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_scatter

Ejemplo:

%result = "mhlo.reduce_scatter"(%operand) ({
  ^bb0(%arg0: tensor<f32>, %arg1: tensor<f32>):
  %0 = mhlo.add %arg0, %arg1 : tensor<f32>
  mhlo.return %0 : tensor<f32>
}) {
  scatter_dimension = 1 : i64,
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>,
  // channel_id = 0
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
  // use_global_device_ids = false
} : (tensor<2x4xf32>) -> tensor<2x2xf32>

Traits: SameOperandsAndResultElementType

Atributos:

Operands:

Resultados:

mhlo.reduce_window (mhlo::ReduceWindowOp)

ReduceWindow operation

Applies a reduction function body to windows of inputs and init_values and produces results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_window

Ejemplo:

%result = "mhlo.reduce_window"(%input, %init_value) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = mhlo.add %arg0, %arg1 : tensor<i32>
    mhlo.return %0 : tensor<i32>
}) {
  window_dimensions = dense<[2, 1]> : tensor<2xi64>,
  window_strides = dense<[4, 1]> : tensor<2xi64>,
  base_dilations = dense<[2, 1]> : tensor<2xi64>,
  window_dilations = dense<[3, 1]> : tensor<2xi64>,
  padding = dense<[[2, 1], [0, 0]]> : tensor<2x2xi64>
} : (tensor<3x2xi32>, tensor<i32>) -> tensor<2x2xi32>

Traits: InferTensorType, RecursiveMemoryEffects, SameVariadicOperandSize, SingleBlock, SingleBlockImplicitTerminator

Interfaces: InferShapedTypeOpInterface, InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.remainder (mhlo::RemOp)

Rem operation

Sintaxis:

operation ::= `mhlo.remainder` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise remainder of dividend lhs and divisor rhs tensors and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#remainder

Ejemplo:

%result = mhlo.remainder %lhs, %rhs : tensor<4xi64>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.replica_id (mhlo::ReplicaIdOp)

ReplicaId operation

Sintaxis:

operation ::= `mhlo.replica_id` attr-dict `:` type(results)

Produces replica_id of the current process.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#replica_id

Ejemplo:

%result = mhlo.replica_id : tensor<ui32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Resultados:

mhlo.reshape (mhlo::ReshapeOp)

Reshape operation

Sintaxis:

operation ::= `mhlo.reshape` operands attr-dict `:` functional-type(operands, results)

Performs reshape of operand tensor to a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reshape

Ejemplo:

%result = mhlo.reshape %operand : (tensor<2xf32>) -> tensor<1x2xf32>

Traits: AlwaysSpeculatableImplTrait, HLO_CompatibleOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.return (mhlo::ReturnOp)

_This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/425

Informally, this operation serves as a terminator for regions defined by
the StableHLO ops. Non-StableHLO ops, e.g. `func.func`, have their own
terminators, e.g. `func.return`.

Example:

    ```mlir
    %result = "mhlo.reduce"(%input, %init_value) ({
      ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
        %0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
        "mhlo.return"(%0) : (tensor<i32>) -> ()
    }) {
      dimensions = dense<1> : tensor<1xi64>
    } : (tensor<1x6xi32>, tensor<i32>) -> tensor<1xi32>
    ```_


Syntax:

```

operation ::= mhlo.return $results attr-dict ( : type($results)^)?

Traits: AlwaysSpeculatableImplTrait, Terminator

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

#### Operands:

| Operand | Description |
| :-----: | ----------- |
| `results` | variadic of tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or token or nested tuple with any combination of tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or token values


### `mhlo.reverse` (mhlo::ReverseOp)

_Reverse operation_

Reverses the order of elements in the `operand` along the specified
`dimensions` and produces a `result` tensor.

See:
<a href="https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reverse">https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reverse</a>

Example:
```mlir
%result = mhlo.reverse %operand, dims = [1] : tensor<3x2xi32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.rng (mhlo::RngOp)

Rng operation

Generates random numbers using the rng_distribution algorithm and produces a result tensor of a given shape shape .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rng

Ejemplo:

%result = mhlo.rng %a, %b, %shape, distribution = NORMAL : (tensor<i32>, tensor<i32>, tensor<2xi64>) -> tensor<3x3xi32>

Traits: InferTensorType

Interfaces: InferShapedTypeOpInterface, InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.rng_bit_generator (mhlo::RngBitGeneratorOp)

RngBitGenerator operation

Returns an output filled with uniform random data and an updated output state output_state given an initial state initial_state using the pseudorandom number generator algorithm rng_algorithm .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rng_bit_generator

Ejemplo:

%output_state, %output = mhlo.rng_bit_generator %initial_state, algorithm = THREE_FRY : (tensor<2xui64>) -> (tensor<2xui64>, tensor<2x2xui64>)

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.round_nearest_afz (mhlo::RoundOp)

Round operation

Sintaxis:

operation ::= `mhlo.round_nearest_afz` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise rounding towards the nearest integer, breaking ties away from zero, on the operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#round_nearest_afz

Ejemplo:

%result = mhlo.round_nearest_afz %operand : tensor<5xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.round_nearest_even (mhlo::RoundNearestEvenOp)

RoundNearestEven operation

Sintaxis:

operation ::= `mhlo.round_nearest_even` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise rounding towards the nearest integer, breaking ties towards the even integer, on the operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#round_nearest_even

Ejemplo:

%result = mhlo.round_nearest_even %operand : tensor<5xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.rsqrt (mhlo::RsqrtOp)

Rsqrt operation

Sintaxis:

operation ::= `mhlo.rsqrt` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise reciprocal square root operation on operand tensor and produces a result tensor, implementing the rSqrt operation from the IEEE-754 specification.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rsqrt

Ejemplo:

%result = mhlo.rsqrt %operand : tensor<2x2xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.scatter (mhlo::ScatterOp)

Scatter operation

Produces results tensors which are equal to inputs tensors except that several slices specified by scatter_indices are updated with the values updates using update_computation .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#scatter

Ejemplo:

%result = "mhlo.scatter"(%input, %scatter_indices, %update) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = mhlo.add %arg0, %arg1 : tensor<i32>
    mhlo.return %0 : tensor<i32>
}) {
  scatter_dimension_numbers = #mhlo.scatter<
    update_window_dims = [2,3],
    inserted_window_dims = [0],
    scatter_dims_to_operand_dims = [1, 0],
    index_vector_dim = 2>,
  indices_are_sorted = false,
  unique_indices = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>, tensor<2x3x2x2xi32>) -> tensor<3x4x2xi32>

Traits: RecursiveMemoryEffects, SameVariadicOperandSize

Interfaces: InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.select (mhlo::SelectOp)

Select operation

Sintaxis:

operation ::= `mhlo.select` operands attr-dict `:`
              custom<SelectOpType>(type($pred), type($on_true), type($on_false), type($result))

Produces a result tensor where each element is selected from on_true or on_false tensor based on the value of the corresponding element of pred .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#select

Ejemplo:

%result = mhlo.select %pred, %on_true, %on_false : tensor<2x2xi1>, tensor<2x2xi32>

Traits: AlwaysSpeculatableImplTrait, HLO_BroadcastingElementwise, InferTensorType

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.select_and_scatter (mhlo::SelectAndScatterOp)

SelectAndScatter operation

Scatters the values from the source tensor using scatter based on the outcome of reduce_window of the input tensor using select and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#select_and_scatter

Ejemplo:

%result = "mhlo.select_and_scatter"(%operand, %source, %init_value) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.compare"(%arg0, %arg1) {
      comparison_direction = #stablehlo<comparison_direction GE>
    } : (tensor<i32>, tensor<i32>) -> tensor<i1>
    "mhlo.return"(%0) : (tensor<i1>) -> ()
}, {
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
    "mhlo.return"(%0) : (tensor<i32>) -> ()
}) {
  window_dimensions = dense<[3, 1]> : tensor<2xi64>,
  window_strides = dense<[2, 1]> : tensor<2xi64>,
  padding = dense<[[0, 1], [0, 0]]> : tensor<2x2xi64>
} : (tensor<4x2xi32>, tensor<2x2xi32>, tensor<i32>) -> tensor<4x2xi32>

Traits: RecursiveMemoryEffects

Interfaces: InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.send (mhlo::SendOp)

Send operation

Sends inputs to a channel channel_id and produces a result token.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#send

Ejemplo:

%result = "mhlo.send"(%operand, %token) {
  // channel_id = 5 : i64,
  // channel_type = #stablehlo<channel_type DEVICE_TO_HOST>,
  channel_handle = #mhlo.channel_handle<handle = 5, type = 2>,
  is_host_transfer = true
} : (tensor<3x4xi32>, !mhlo.token) -> !mhlo.token

Interfaces: InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.set_dimension_size (mhlo::SetDimensionSizeOp)

SetDimensionSize operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as XLA's SetDimensionSize: https://www.tensorflow.org/xla/operation_semantics#setdimensionsize

Ejemplo:

%0 = mhlo.set_dimension_size %arg0, %arg1, dim = 1 : (tensor<4x2xf32>, tensor<i32>) -> tensor<4x2xf32>

Traits: AlwaysSpeculatableImplTrait, InferTensorType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.shift_left (mhlo::ShiftLeftOp)

ShiftLeft operation

Sintaxis:

operation ::= `mhlo.shift_left` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise left-shift operation on the lhs tensor by rhs number of bits and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_left

Ejemplo:

%result = mhlo.shift_left %lhs, %rhs : tensor<6xi8>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.shift_right_arithmetic (mhlo::ShiftRightArithmeticOp)

ShiftRightArithmetic operation

Sintaxis:

operation ::= `mhlo.shift_right_arithmetic` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise arithmetic right-shift operation on the lhs tensor by rhs number of bits and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_right_arithmetic

Ejemplo:

%result = mhlo.shift_right_arithmetic %lhs, %rhs : tensor<6xi8>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.shift_right_logical (mhlo::ShiftRightLogicalOp)

ShiftRightLogical operation

Sintaxis:

operation ::= `mhlo.shift_right_logical` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise logical right-shift operation on the lhs tensor by rhs number of bits and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_right_logical

Ejemplo:

%result = mhlo.shift_right_logical %lhs, %rhs : tensor<6xi8>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.sign (mhlo::SignOp)

Sign operation

Sintaxis:

operation ::= `mhlo.sign` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Returns the sign of the operand element-wise and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sign

Ejemplo:

%result = mhlo.sign %operand : tensor<7xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.sine (mhlo::SineOp)

Sine operation

Sintaxis:

operation ::= `mhlo.sine` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise sine operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sine

Ejemplo:

%result = mhlo.sine %operand : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.slice (mhlo::SliceOp)

Slice operation

Extracts a slice from the operand using statically-computed starting indices and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#slice

Ejemplo:

%result = "mhlo.slice" (%operand) {
  start_indices = dense<[1, 2]> : tensor<2xi64>,
  limit_indices = dense<[3, 4]> : tensor<2xi64>,
  strides = dense<1> : tensor<2xi64>
} : (tensor<3x4xi64>) -> tensor<2x2xi64>

Traits: AlwaysSpeculatableImplTrait, SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.sort (mhlo::SortOp)

Sort operation

Sorts a variadic number of tensors in inputs together, according to a custom comparator , along the given dimension and produces a variadic number of tensors as results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sort

Ejemplo:

%result0, %result1 = "mhlo.sort"(%input0, %input1) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>, %arg2: tensor<i32>, %arg3: tensor<i32>):
    %predicate = "mhlo.compare"(%arg0, %arg1) {
      comparison_direction = #stablehlo<comparison_direction GT>
      } : (tensor<i32>, tensor<i32>) -> tensor<i1>
    "mhlo.return"(%predicate) : (tensor<i1>) -> ()
}) {
  dimension = 0 : i64,
  is_stable = true
} : (tensor<2x3xi32>, tensor<2x3xi32>) -> (tensor<2x3xi32>, tensor<2x3xi32>)

Traits: InferTensorType, RecursiveMemoryEffects, SameOperandsAndResultShape

Interfaces: InferShapedTypeOpInterface, InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.sqrt (mhlo::SqrtOp)

Sqrt operation

Sintaxis:

operation ::= `mhlo.sqrt` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise square root operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sqrt

Ejemplo:

%result = mhlo.sqrt %operand : tensor<2x2xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.stochastic_convert (mhlo::StochasticConvertOp)

StochasticConvert operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/295

Informally, this operation performs element-wise conversion of values from a bigger type to a smaller one with stochastic rounding using the random number passed in.

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.subtract (mhlo::SubtractOp)

Subtract operation

Sintaxis:

operation ::= `mhlo.subtract` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise subtraction of two tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#subtract

Ejemplo:

%result = mhlo.subtract %lhs, %rhs : tensor<2xi32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.tan (mhlo::TanOp)

Tan operation

Sintaxis:

operation ::= `mhlo.tan` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/954

Informally, this operation returns Tan(operand) element-wise.

Ejemplo:

%0 = mhlo.tan %arg0 : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.tanh (mhlo::TanhOp)

Tanh operation

Sintaxis:

operation ::= `mhlo.tanh` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise hyperbolic tangent operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#tanh

Ejemplo:

%result = mhlo.tanh %operand : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.topk (mhlo::TopKOp)

TopK operation

Sintaxis:

operation ::= `mhlo.topk` `(`$operand `,` `k` `=` $k (`,` `largest` `=` $largest^)? `)` attr-dict `:`
              type($operand) `->` `(`type($values)`,` type($indices)`)`

Returns top k values and their indices, along the last dimension of the operand if largest=true or the bottom k values if largest=false .

See: https://www.tensorflow.org/xla/operation_semantics#top-k

Ejemplo:

%values, %indices = mhlo.topk(%operand, k=5, largest=true)
  : tensor<100xf32> -> (tensor<5xf32>, tensor<5xi32>)

Traits: InferTensorType, RecursiveMemoryEffects

Interfaces: InferShapedTypeOpInterface, InferTypeOpInterface

Atributos:

Operands:

Resultados:

mhlo.torch_index_select (mhlo::TorchIndexSelectOp)

TorchIndexSelect operation

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3

Informally, this operation does the same thing as PyTorch's index_select, augmented with support for batch dimensions: https://pytorch.org/docs/stable/generated/torch.index_select.html

The batch_dims attribute specifies the number of major batch dimensions (0 or more) that act like a multidimensional loop over both the operand and the index.

Ejemplo:

%result = "mhlo.torch_index_select"(%operand, %index) {
  dim = 2 : i64,
  batch_dims = 1 : i64
} : (tensor<8x128x3072x64xf32>, tensor<8x16x1024xi32>) -> tensor<8x128x16x1024x64xf32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.trace (mhlo::TraceOp)

Trace operation

Sintaxis:

operation ::= `mhlo.trace` $operand `,` $tag attr-dict `:` type($operand)

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/604

It is not used by JAX, PyTorch or TensorFlow, so it looks like we should've classified it as "Private to XLA" and not included it in StableHLO in the first place. With that in mind, its semantics will not be documented here.

Ejemplo:

mhlo.trace %arg0, "In test code." : tensor<5x1x5xi32>

Atributos:

Operands:

mhlo.transpose (mhlo::TransposeOp)

Transpose operation

Permutes the dimensions of operand tensor using permutation and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#transpose

Ejemplo:

%0 = mhlo.transpose %arg0, dims = [2, 1, 0] : (tensor<1x2x3xi32>) -> tensor<3x2x1xi32>

Traits: AlwaysSpeculatableImplTrait, HLO_CompatibleOperandsAndResultElementType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.triangular_solve (mhlo::TriangularSolveOp)

TriangularSolve operation

Solves batches of systems of linear equations with lower or upper triangular coefficient matrices.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#triangular_solve

Ejemplo:

%result = "mhlo.triangular_solve"(%a, %b) {
  left_side = true,
  lower = true,
  unit_diagonal = false,
  transpose_a = #stablehlo<transpose NO_TRANSPOSE>
} : (tensor<3x3xf32>, tensor<3x3xf32>) -> tensor<3x3xf32>

Traits: AlwaysSpeculatableImplTrait, InferTensorType, SameOperandsAndResultElementType

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

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.tuple (mhlo::TupleOp)

Tuple operation

Sintaxis:

operation ::= `mhlo.tuple` $val attr-dict `:` custom<TupleOpType>(type($val), type($result))

Produces a result tuple from values val .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#tuple

Ejemplo:

%result = mhlo.tuple %val0, %val1 : tuple<tensor<2xf32>, tuple<tensor<i32>>>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, InferTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.unary_einsum (mhlo::UnaryEinsumOp)

UnaryEinsum operation

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3

Informally, this operation does the same thing as TF's einsum: https://www.tensorflow.org/api_docs/python/tf/einsum

Ejemplo:

%result = "mhlo.unary_einsum"(%operand) {
  einsum_config = "ab->a"
} : (tensor<4x16xf32>) -> tensor<4xf32>

Traits: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Atributos:

Operands:

Resultados:

mhlo.uniform_dequantize (mhlo::UniformDequantizeOp)

UniformDequantize operation

Sintaxis:

operation ::= `mhlo.uniform_dequantize` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise conversion of quantized tensor operand to a floating-point tensor result according to the quantization parameters defined by the operand type.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#uniform_dequantize

Ejemplo:

%result = mhlo.uniform_dequantize %operand : (tensor<16x16x!quant.uniform<i8:f32, 34.0:16>>) -> tensor<16x16xf32>

Traits: AlwaysSpeculatableImplTrait, Elementwise, InferTensorType, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.uniform_quantize (mhlo::UniformQuantizeOp)

UniformQuantize operation

Sintaxis:

operation ::= `mhlo.uniform_quantize` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise conversion of floating-point tensor or quantized tensor operand to a quantized tensor result according to the quantization parameters defined by the result type.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#uniform_quantize

Ejemplo:

%result = mhlo.uniform_quantize %operand : (tensor<16x16xf32>) -> tensor<16x16x!quant.uniform<ui8:f32, 34.0:16>>

Traits: AlwaysSpeculatableImplTrait, Elementwise, SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable, InferShapedTypeOpInterface, NoMemoryEffect (MemoryEffectOpInterface)

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

mhlo.while (mhlo::WhileOp)

While operation

Produces the output from executing body function 0 or more times while the cond function outputs true .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#while

Ejemplo:

%results0, %results1 = "mhlo.while"(%operand0, %operand1) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.compare"(%arg0, %arg1) {
      comparison_direction = #stablehlo<comparison_direction LT>
    } : (tensor<i32>, tensor<i32>) -> tensor<i1>
    "mhlo.return"(%0) : (tensor<i1>) -> ()
}, {
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.add"(%arg0, %constant0) : (tensor<i32>, tensor<i32>) -> tensor<i32>
    "mhlo.return"(%0, %arg1) : (tensor<i32>, tensor<i32>) -> ()
}) : (tensor<i32>, tensor<i32>) -> (tensor<i32>, tensor<i32>)

Traits: RecursiveMemoryEffects, SingleBlock, SingleBlockImplicitTerminator

Interfaces: InferTypeOpInterface, OpAsmOpInterface

Operands:

Resultados:

mhlo.xla.rng_get_and_update_state (mhlo::XlaRngGetAndUpdateStateOp)

XlaRngGetAndUpdateState operation

Sintaxis:

operation ::= `mhlo.xla.rng_get_and_update_state` attr-dict

This operation is private to the XLA compiler, so it is does not yet have a specification.

Informally, this operation represents the change of the global random number generator state for rng instructions. The global state is incremented by delta and the old state is returned.

The output is currently defined for a single output type. If this changes in the future to support multiple types, lowering to use of a global memref must ensure that a single memref is still used and updated appropriately.

Interfaces: InferTypeOpInterface

Atributos:

Resultados:

mhlo.xor (mhlo::XorOp)

Xor operation

Sintaxis:

operation ::= `mhlo.xor` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise XOR of two tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#xor

Ejemplo:

%result = mhlo.xor %lhs, %rhs : tensor<2xi32>

Traits: AlwaysSpeculatableImplTrait, Commutative, CompatibleOperandsAndResultType, Elementwise, SameOperandsAndResultShape

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

Effects: MemoryEffects::Effect{}

Operands:

Resultados:

Attribute definition

ArgResultAliasAttr

Attribute that models the alias relationship of entry function argument

This attribute captures the alias relationship of an MHLO main function argument to one of the results, denoted by resultIndex . The argTupleIndices and resultTupleIndices are used to index into nested tuples in operand and result respectively. If isMustAlias is true then the operand-result pair must alias.

This is meant to be used as an attribute on a function argument in MHLO. For example, in the following code it expresses that %arg1 may alias 0-th result.

func @main(%arg0: tensor<2xf32>, %arg1: tensor<3xf32> {mhlo.result_alias =
    mhlo.result_alias<result_index = [2], ...>}
  ) -> tensor<2xf32>, tensor<3xf32> {
  // function body ...
}

Parámetros:

ChannelHandleAttr

two 64-bit integers 'handle' and 'type'

Sintaxis:

#mhlo.channel_handle<
  int64_t,   # handle
  int64_t   # type
>

Parámetros:

ComparisonDirectionAttr

Which comparison operation to perform.

Sintaxis:

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

Enum cases:

• EQ ( EQ )
• NE ( NE )
• GE ( GE )
• GT ( GT )
• LE ( LE )
• LT ( LT ) #### Parameters:

ComparisonTypeAttr

Which comparison type to use.

Sintaxis:

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

Enum cases:

• NOTYPE ( NOTYPE )
• FLOAT ( FLOAT )
• TOTALORDER ( TOTALORDER )
• SIGNED ( SIGNED )
• UNSIGNED ( UNSIGNED ) #### Parameters:

ConvDimensionNumbersAttr

Structure of dimension information for conv op

Parámetros:

CrossProgramPrefetchAttr

Argument that is prefetched from another program

Sintaxis:

#mhlo.cross_program_prefetch<
  int64_t,   # parameter
  ::llvm::ArrayRef<int64_t>,   # indices
  std::optional<int64_t>   # offset
>

This attribute captures an argument that is prefetched from another program. For a given CrossProgramPrefetchAttr , parameter tells us which argument of the main function of the module is prefetched, and indices is a shape index telling us what subshape of that argument is prefetched.

A shape has a subshape iff it is a tuple. In that case, the subshape of the tuple by indices is the shape achieved after indexing by each element of indices in turn. For example, the [1,0] subshape of tuple<tuple<token, token>, tuple<tensor<i32>, token>> is tensor<i32> .

An empty value for indices means the whole shape is prefetched.

Por ejemplo,

module attributes { mhlo.cross_program_prefetch = [ #mhlo.cross_program_prefetch< parameter = 0, indices = [0]> ]} {
  func.func @copy(%arg0 : tuple<tensor<2x3xi32>, tensor<i32>>) -> tuple<tensor<2x3xi32>, tensor<i32>> {
    %0 = "mhlo.copy"(%arg0) {is_cross_program_prefetch}
    return %0 : tuple<tensor<2x3xi32>, tensor<i32>>
  }
  func.func @main(%arg0 : tuple<tensor<2x3xi32>, tensor<i32>>) -> tuple<tensor<2x3xi32>, tensor<i32>> {
    %1 = "mhlo.async_start"(%arg0) {called_computation=@copy}
    %2 = "mhlo.async_done"(%1) {called_computation=@copy}
    return %2 : tuple<tensor<2x3xi32>, tensor<i32>>
  }
}

The parameter = 0 tells us that the async copy of the 0 th parameter is a cross_program_prefetch , while the index of [0] tells us that the 0 th element of the tuple is prefetched while the other element of the tuple is not.

Parámetros:

CustomCallScheduleAttr

Specifies the desired schedule for the custom-call.

Sintaxis:

#mhlo.custom_call_schedule<
  ::mlir::mhlo::CustomCallSchedule   # value
>

Enum cases:

• NINGUNO NONE )
• LATEST ( LATEST )
• EARLIEST ( EARLIEST ) #### Parameters:

DequantizeModeAttr

Dequantization mode. Only MIN_COMBINED is supported.

Sintaxis:

#mhlo.dequantize_mode<
  ::mlir::mhlo::DequantizeMode   # value
>

Enum cases:

• MIN_COMBINED ( MIN_COMBINED ) #### Parameters:

DomainKindAttr

Kind of domain metatdata attached to an HLO domain.

Sintaxis:

#mhlo.kind<
  ::mlir::mhlo::DomainKind   # value
>

Enum cases:

• sharding ( sharding ) #### Parameters:

DotDimensionNumbersAttr

Attribute that models the dimension information for dot.

Parámetros:

FftTypeAttr

XLA fast fourier transform type.

Sintaxis:

#mhlo.fft_type<
  ::mlir::mhlo::FftType   # value
>

Enum cases:

• FFT ( FFT )
• IFFT ( IFFT )
• RFFT ( RFFT )
• IRFFT ( IRFFT ) #### Parameters:

FusionKindAttr

fusion kind

Sintaxis:

#mhlo.fusion_kind<
  ::mlir::mhlo::FusionKind   # value
>

Enum cases:

• kLoop ( kLoop )
• kInput ( kInput )
• kOutput ( kOutput )
• kCustom ( kCustom ) #### Parameters:

GatherDimensionNumbersAttr

Attribute that models the dimension information for gather

Parámetros:

OutputOperandAliasAttr

Attribute that models the alias relationship of output and operand of a CustomCall op

Sintaxis:

#mhlo.output_operand_alias<
  ::llvm::ArrayRef<int64_t>,   # outputTupleIndices
  int64_t,   # operandIndex
  ::llvm::ArrayRef<int64_t>   # operandTupleIndices
>

This attribute captures the alias relationship of the output to one of the operands for a CustomCall op, denoted by operand_index . The output_tuple_indices and operand_tuple_indices are used to index into output and operand types. These indices lists are empty if the corresponding types are not tuple types, and can be arbitrarily long in case of arbitrarily nested tuple types.

See https://www.tensorflow.org/xla/aliasing

Example when used as array with in mhlo.custom-call:

%0 = "mhlo.custom_call"(%arg0, %arg1) {
  // other attributes
  output_operand_alias = [
    #mhlo.output_operand_alias<output_tuple_indices = [0],
                               operand_index = 0,
                               operand_tuple_indices = [1]>
  ]
} : (tuple<tensor<1x1xf32>, tensor<2x3xf32>>, tensor<5x5xf32>) -> tuple<tensor<2x3xf32>>

The output and the 0th operand are both tuples. The aliasing shows the
relationship between the 0th element in output tuple with the 1st element in
the 0th operand. And both of them are of the same type: tensor<2x3xf32>.

Parámetros:

PrecisionAttr

XLA precision for an operand. Has backend specific meaning.

Sintaxis:

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

Enum cases:

• DEFAULT ( DEFAULT )
• ALTA HIGH )
• HIGHEST ( HIGHEST )
• PACKED_NIBBLE ( PACKED_NIBBLE ) #### Parameters:

RngAlgorithmAttr

XLA PRNG algorithm to be used.

Sintaxis:

#mhlo.rng_algorithm<
  ::mlir::mhlo::RngAlgorithm   # value
>

Enum cases:

• DEFAULT ( DEFAULT )
• THREE_FRY ( THREE_FRY )
• PHILOX ( PHILOX ) #### Parameters:

RngDistributionAttr

XLA PRNG distribution to be used.

Sintaxis:

#mhlo.rng_distribution<
  ::mlir::mhlo::RngDistribution   # value
>

Enum cases:

• UNIFORM ( UNIFORM )
• NORMAL ( NORMAL ) #### Parameters:

ScatterDimensionNumbersAttr

Attribute that models the dimension information for scatter

Parámetros:

TransposeAttr

Transpose options

Sintaxis:

#mhlo.transpose<
  ::mlir::mhlo::Transpose   # value
>

Enum cases:

• TRANSPOSE_INVALID ( TRANSPOSE_INVALID )
• NO_TRANSPOSE ( NO_TRANSPOSE )
• TRANSPOSE ( TRANSPOSE )
• ADJOINT ( ADJOINT ) #### Parameters:

TypeExtensionsAttr

Attribute that extends tensor type with MHLO type properties.

Sintaxis:

#mhlo.type_extensions<
  ::llvm::ArrayRef<int64_t>   # bounds
>

This attribute is used to extend MLIR tensor type with MHLO tensor specific properties. These properties aren't modeled in the MLIR type. This attribute is set in the encoding field of the tensor type.

See HLO_BoundedAttrInterface for documentation for bounds .

Parámetros:

Type definition

AsyncBundleType

Opaque collection of other types

Sintaxis:

!mhlo.async_bundle<
  ::llvm::ArrayRef<Type>   # types
>

Parámetros: