cuda-api-wrappers
Thin C++-flavored wrappers for the CUDA Runtime API
cuda::memory::mapped::region_pair_t Struct Reference

A pair of memory regions, one in system (=host) memory and one on a CUDA device's memory - mapped to each other. More...

#include <memory.hpp>

Public Member Functions

template<typename T >
constexpr span_pair_t< T > as_spans () const
 

Public Attributes

memory::region_t host_side
 The two regions mapped to each other by the CUDA driver; they must be identical in size. More...
 
memory::region_t device_side
 

Detailed Description

A pair of memory regions, one in system (=host) memory and one on a CUDA device's memory - mapped to each other.

Note
this is the mapped-pair equivalent of a void *; it is not a proper memory region abstraction, i.e. it has no size information

Member Function Documentation

◆ as_spans()

template<typename T >
constexpr span_pair_t<T> cuda::memory::mapped::region_pair_t::as_spans ( ) const
inline
Returns
two spans, one for the each of the host-side and device-side regions

Member Data Documentation

◆ host_side

memory::region_t cuda::memory::mapped::region_pair_t::host_side

The two regions mapped to each other by the CUDA driver; they must be identical in size.


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