Mila::Dnn::CpuTensorDataTypeTraits Class Reference

CPU-specific traits for abstract tensor data types. More...

Public Types
template<TensorDataType TDataType>
using	native_type
	Maps abstract tensor data types to concrete CPU native types.

Static Public Member Functions
template<TensorDataType TDataType>
static consteval bool	supports ()
	Validates CPU support for abstract tensor data types.

Detailed Description

CPU-specific traits for abstract tensor data types.

Provides device-specific characteristics and native type mappings for CPU-supported tensor data types, enabling type-safe value initialization and host-specific optimizations.

Member Typedef Documentation

native_type

template<TensorDataType TDataType>

using Mila::Dnn::CpuTensorDataTypeTraits::native_type

Initial value:

 
            std::conditional_t<TDataType == TensorDataType::FP32, float,
            std::conditional_t<TDataType == TensorDataType::INT8, int8_t,
            std::conditional_t<TDataType == TensorDataType::INT16, int16_t,
            std::conditional_t<TDataType == TensorDataType::INT32, int32_t,
            std::conditional_t<TDataType == TensorDataType::UINT8, uint8_t,
            std::conditional_t<TDataType == TensorDataType::UINT16, uint16_t,
            std::conditional_t<TDataType == TensorDataType::UINT32, uint32_t,
            void>>>>>>>

Maps abstract tensor data types to concrete CPU native types.

Provides compile-time mapping from abstract TensorDataType enumeration to concrete standard C++ types supported on CPU, enabling type-safe value initialization and host operations.

Template Parameters

TDataType

Abstract tensor data type to map

Returns

Concrete C++ type corresponding to the abstract type

Note

Only supports types where supports<TDataType>() returns true

All mapped types are standard C++ types compatible with host code

Member Function Documentation

supports()

template<TensorDataType TDataType>

consteval bool Mila::Dnn::CpuTensorDataTypeTraits::supports ( )

inlinestaticconsteval

Validates CPU support for abstract tensor data types.

Determines whether CPU devices and memory resources support the specified abstract tensor data type. CPU typically supports standard C++ types but not device-specific precision formats.

Template Parameters

TDataType

Abstract tensor data type to validate

Returns

true if CPU supports the data type, false otherwise

---

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

• /__w/Mila/Mila/Mila/Src/Dnn/Compute/Devices/Cpu/Tensors/CpuTensorDataTypeTraits.ixx