pointer.hpp File Reference

A wrapper class for host and/or device pointers, allowing easy access to CUDA's pointer attributes. More...

#include "constants.hpp"
#include "error.hpp"
#include "types.hpp"
#include <cassert>

Include dependency graph for pointer.hpp:

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

Go to the source code of this file.

Classes
class	cuda::memory::pointer_t< T >
	A convenience wrapper around a raw pointer "known" to the CUDA runtime and which thus has various kinds of associated information which this wrapper allows access to. More...

Namespaces
	cuda
	Definitions and functionality wrapping CUDA APIs.
	cuda::memory
	Representation, allocation and manipulation of CUDA-related memory, of different.

Enumerations
enum	cuda::memory::type_t : ::std::underlying_type< CUmemorytype >::type {   host_ = CU_MEMORYTYPE_HOST,   device_ = CU_MEMORYTYPE_DEVICE,   array = CU_MEMORYTYPE_ARRAY,   unified_ = CU_MEMORYTYPE_UNIFIED,   managed_ = CU_MEMORYTYPE_UNIFIED,   non_cuda = ~(::std::underlying_type<CUmemorytype>::type{0}) }
	The CUDA execution ecosystem involves different memory spaces in their relation to a GPU device or their treatment by the CUDA driver; this type distinguishes among them.

Functions
memory::type_t	cuda::memory::type_of (const void *ptr)
	Determine the type of memory at a given address vis-a-vis the CUDA ecosystem: Was it allocated by the CUDA driver? Does it reside solely on a GPU device'? Solely on the host? Movable between locations? etc.
context_t	cuda::memory::context_of (void const *ptr)
	Obtain (a non-owning wrapper for) the CUDA context with which a memory address is associated (e.g. More...
template<typename T >
pointer_t< T >	cuda::memory::pointer::wrap (T *ptr) noexcept
	Wraps an existing pointer in a pointer_t wrapper. More...

Detailed Description

A wrapper class for host and/or device pointers, allowing easy access to CUDA's pointer attributes.

Note

at the moment, this class is not used by other sections of the API wrappers; specifically, freestanding functions and methods returning pointers return raw T*'s rather than pointer_t<T>'s. This may change in the future.

Todo:

Consider allowing for storing attributes within the class, lazily (e.g. with an ::std::optional).

Function Documentation

wrap()

template<typename T >

pointer_t<T> cuda::memory::pointer::wrap ( T *  ptr )

inlinenoexcept

Wraps an existing pointer in a pointer_t wrapper.

Parameters

ptr	a pointer - into either device or host memory - to be wrapped.