array.hpp File Reference

Contains a proxy class for CUDA arrays - GPU memory with 2-D or 3-D locality and hardware support for interpolated value retrieval); see also texture_view.hpp . More...

#include "context.hpp"
#include "error.hpp"
#include <cuda_fp16.h>

Include dependency graph for array.hpp:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
class	cuda::array_t< T, NumDimensions >
	Owning wrapper for CUDA 2D and 3D arrays. More...
class	cuda::array_t< T, NumDimensions >
	Owning wrapper for CUDA 2D and 3D arrays. More...

Namespaces
	cuda
	Definitions and functionality wrapping CUDA APIs.
	cuda::array
	CUDA facilities for interpolating access to multidimensional array objects, in particular via the array_t class.

Typedefs
using	cuda::array::handle_t = CUarray
	Raw CUDA driver handle for arrays (of any dimension)
template<dimensionality_t NumDimensions>
using	cuda::array::descriptor_t = typename ::std::conditional< NumDimensions==2, CUDA_ARRAY_DESCRIPTOR, CUDA_ARRAY3D_DESCRIPTOR >::type
	Raw CUDA driver descriptor structure for an array of dimension. More...

Functions
template<typename T , dimensionality_t NumDimensions>
array_t< T, NumDimensions >	cuda::array::wrap (device::id_t device_id, context::handle_t context_handle, handle_t handle, dimensions_t< NumDimensions > dimensions) noexcept
	Wrap an existing CUDA array in an array_t instance.
template<typename T , dimensionality_t NumDimensions>
array_t< T, NumDimensions >	cuda::array::create (const context_t &context, dimensions_t< NumDimensions > dimensions)
	Create a new (typed) CUDA array of the specified dimensions. More...
template<typename T , dimensionality_t NumDimensions>
array_t< T, NumDimensions >	cuda::array::create (const device_t &device, dimensions_t< NumDimensions > dimensions)

Detailed Description

Contains a proxy class for CUDA arrays - GPU memory with 2-D or 3-D locality and hardware support for interpolated value retrieval); see also texture_view.hpp .

Note

Not all kinds of arrays are supported: Only non-layered, non-cubemap arrays of 2 or 3 dimensions.