MLIR 22.0.0git
mlir::TensorType Class Reference

Tensor types represent multi-dimensional arrays, and have two variants: RankedTensorType and UnrankedTensorType. More...

#include "mlir/IR/BuiltinTypes.h"

Inheritance diagram for mlir::TensorType:

Public Member Functions

Type getElementType () const
 Returns the element type of this tensor type.
bool hasRank () const
 Returns if this type is ranked, i.e. it has a known number of dimensions.
ArrayRef< int64_tgetShape () const
 Returns the shape of this tensor type.
TensorType cloneWith (std::optional< ArrayRef< int64_t > > shape, Type elementType) const
 Clone this type with the given shape and element type.
RankedTensorType clone (ArrayRef< int64_t > shape, Type elementType) const
 Return a clone of this type with the given new shape and element type.
RankedTensorType clone (ArrayRef< int64_t > shape) const
 Return a clone of this type with the given new shape.
 operator ShapedType () const
 Allow implicit conversion to ShapedType.
constexpr Type ()=default
 Type (const ImplType *impl)
 Type (const Type &other)=default
Public Member Functions inherited from mlir::Type
constexpr Type ()=default
 Type (const ImplType *impl)
 Type (const Type &other)=default
Typeoperator= (const Type &other)=default
bool operator== (Type other) const
bool operator!= (Type other) const
 operator bool () const
bool operator! () const
TypeID getTypeID ()
 Return a unique identifier for the concrete type.
MLIRContextgetContext () const
 Return the MLIRContext in which this type was uniqued.
DialectgetDialect () const
 Get the dialect this type is registered to.
bool isIndex () const
bool isBF16 () const
bool isF16 () const
bool isTF32 () const
bool isF32 () const
bool isF64 () const
bool isF80 () const
bool isF128 () const
bool isFloat () const
 Return true if this is an float type (with the specified width).
bool isFloat (unsigned width) const
 Return true if this is a float type with the specified width.
bool isInteger () const
 Return true if this is an integer type (with the specified width).
bool isInteger (unsigned width) const
bool isSignlessInteger () const
 Return true if this is a signless integer type (with the specified width).
bool isSignlessInteger (unsigned width) const
bool isSignedInteger () const
 Return true if this is a signed integer type (with the specified width).
bool isSignedInteger (unsigned width) const
bool isUnsignedInteger () const
 Return true if this is an unsigned integer type (with the specified width).
bool isUnsignedInteger (unsigned width) const
unsigned getIntOrFloatBitWidth () const
 Return the bit width of an integer or a float type, assert failure on other types.
bool isSignlessIntOrIndex () const
 Return true if this is a signless integer or index type.
bool isSignlessIntOrIndexOrFloat () const
 Return true if this is a signless integer, index, or float type.
bool isSignlessIntOrFloat () const
 Return true of this is a signless integer or a float type.
bool isIntOrIndex () const
 Return true if this is an integer (of any signedness) or an index type.
bool isIntOrFloat () const
 Return true if this is an integer (of any signedness) or a float type.
bool isIntOrIndexOrFloat () const
 Return true if this is an integer (of any signedness), index, or float type.
void print (raw_ostream &os) const
 Print the current type.
void print (raw_ostream &os, AsmState &state) const
void dump () const
friend::llvm::hash_code hash_value (Type arg)
const voidgetAsOpaquePointer () const
 Methods for supporting PointerLikeTypeTraits.
template<typename InterfaceT>
bool hasPromiseOrImplementsInterface ()
 Returns true if InterfaceT has been promised by the dialect or implemented.
template<template< typename T > class Trait>
bool hasTrait ()
 Returns true if the type was registered with a particular trait.
const AbstractTygetAbstractType () const
 Return the abstract type descriptor for this type.
ImplTypegetImpl () const
 Return the Type implementation.
void walkImmediateSubElements (function_ref< void(Attribute)> walkAttrsFn, function_ref< void(Type)> walkTypesFn) const
 Walk all of the immediately nested sub-attributes and sub-types.
auto replaceImmediateSubElements (ArrayRef< Attribute > replAttrs, ArrayRef< Type > replTypes) const
 Replace the immediately nested sub-attributes and sub-types with those provided.
template<WalkOrder Order = WalkOrder::PostOrder, typename... WalkFns>
auto walk (WalkFns &&...walkFns)
 Walk this type and all attibutes/types nested within using the provided walk functions.
template<typename... ReplacementFns>
auto replace (ReplacementFns &&...replacementFns)
 Recursively replace all of the nested sub-attributes and sub-types using the provided map functions.

Static Public Member Functions

static bool isValidElementType (Type type)
 Return true if the specified element type is ok in a tensor.
static bool classof (Type type)
 Methods for support type inquiry through isa, cast, and dyn_cast.
Static Public Member Functions inherited from mlir::Type
static Type getFromOpaquePointer (const void *pointer)

Additional Inherited Members

Public Types inherited from mlir::Type
template<typename ConcreteType, typename BaseType, typename StorageType, template< typename T > class... Traits>
using TypeBase
 Utility class for implementing types.
using ImplType = TypeStorage
using AbstractTy = AbstractType
Protected Attributes inherited from mlir::Type
ImplTypeimpl {nullptr}

Detailed Description

Tensor types represent multi-dimensional arrays, and have two variants: RankedTensorType and UnrankedTensorType.

Note: This class attaches the ShapedType trait to act as a mixin to provide many useful utility functions. This inheritance has no effect on derived tensor types.

Definition at line 55 of file BuiltinTypes.h.

Member Function Documentation

◆ classof()

bool mlir::TensorType::classof ( Type type)
inlinestatic

Methods for support type inquiry through isa, cast, and dyn_cast.

Definition at line 420 of file BuiltinTypes.h.

References Type().

◆ clone() [1/2]

RankedTensorType mlir::TensorType::clone ( ArrayRef< int64_t > shape) const

Return a clone of this type with the given new shape.

The returned type is ranked, even if this type is unranked.

References Type().

◆ clone() [2/2]

RankedTensorType mlir::TensorType::clone ( ArrayRef< int64_t > shape,
Type elementType ) const

Return a clone of this type with the given new shape and element type.

The returned type is ranked, even if this type is unranked.

References Type().

◆ cloneWith()

TensorType TensorType::cloneWith ( std::optional< ArrayRef< int64_t > > shape,
Type elementType ) const

Clone this type with the given shape and element type.

If the provided shape is std::nullopt, the current shape of the type is used.

Definition at line 310 of file BuiltinTypes.cpp.

References Type().

◆ getElementType()

Type TensorType::getElementType ( ) const

Returns the element type of this tensor type.

Definition at line 296 of file BuiltinTypes.cpp.

References Type().

Referenced by joinShapes().

◆ getShape()

ArrayRef< int64_t > TensorType::getShape ( ) const

Returns the shape of this tensor type.

Definition at line 306 of file BuiltinTypes.cpp.

Referenced by verifyReduceOp().

◆ hasRank()

bool TensorType::hasRank ( ) const

Returns if this type is ranked, i.e. it has a known number of dimensions.

Definition at line 302 of file BuiltinTypes.cpp.

Referenced by joinShapes(), and verifyReduceOp().

◆ isValidElementType()

bool TensorType::isValidElementType ( Type type)
static

Return true if the specified element type is ok in a tensor.

Definition at line 345 of file BuiltinTypes.cpp.

References mlir::Type::getDialect(), and Type().

Referenced by checkTensorElementType(), and mlir::detail::Parser::parseTensorType().

◆ operator ShapedType()

mlir::TensorType::operator ShapedType ( ) const
inline

Allow implicit conversion to ShapedType.

Definition at line 91 of file BuiltinTypes.h.

◆ Type() [1/3]

mlir::Type::Type ( )
constexprdefault

◆ Type() [2/3]

mlir::Type::Type ( const ImplType * impl)
inline

Definition at line 87 of file Types.h.

◆ Type() [3/3]

mlir::Type::Type ( const Type & other)
default

The documentation for this class was generated from the following files: