cuda::memory::mapped Namespace Reference

Memory regions appearing in both on the host-side and device-side address spaces with the regions in both spaces mapped to each other (i.e. More...

Classes
struct	region_pair_t
	A pair of memory regions, one in system (=host) memory and one on a CUDA device's memory - mapped to each other. More...
struct	span_pair_t
	A pair of memory spans, one in device-global memory and one in host/system memory, mapped to it. More...

Functions
template<typename T >
T *	device_side_pointer_for (T *host_memory_ptr)
	Obtain a pointer in the device-side memory space (= address range) given given a host-side pointer mapped to it. More...
region_t	device_side_region_for (region_t region)
	Get the memory region mapped to a given host-side region. More...
const_region_t	device_side_region_for (const_region_t region)
	Get the memory region mapped to a given host-side region. More...
region_pair_t	allocate (cuda::context_t &context, size_t size_in_bytes, allocation_options options)
	Allocate a memory region on the host, which is also mapped to a memory region in a context of some CUDA device - so that changes to one will be reflected in the other. More...
region_pair_t	allocate (cuda::device_t &device, size_t size_in_bytes, allocation_options options=allocation_options{})
	Allocate a memory region on the host, which is also mapped to a memory region in the global memory of a CUDA device - so that changes to one will be reflected in the other. More...
void	free (region_pair_t pair)
	Free a pair of mapped memory regions. More...
void	free_region_pair_of (void *ptr)
	Free a pair of mapped memory regions using just one of them. More...
bool	is_part_of_a_region_pair (const void *ptr)
	Determine whether a given stretch of memory was allocated as part of a mapped pair of host and device memory regions. More...

Detailed Description

Memory regions appearing in both on the host-side and device-side address spaces with the regions in both spaces mapped to each other (i.e.

guaranteed to have the same contents on access up to synchronization details). Consult the CUDA C programming guide section on mapped memory for more details.

Function Documentation

allocate() [1/2]

region_pair_t cuda::memory::mapped::allocate ( cuda::context_t &  context, size_t  size_in_bytes, allocation_options  options  )

inline

Allocate a memory region on the host, which is also mapped to a memory region in a context of some CUDA device - so that changes to one will be reflected in the other.

Parameters

context	The device context in which the device-side region in the pair will be allocated.
size_in_bytes	amount of memory to allocate (in each of the regions)
options	see memory::allocation_options

allocate() [2/2]

region_pair_t cuda::memory::mapped::allocate ( cuda::device_t &  device, size_t  size_in_bytes, allocation_options  options = allocation_options{}  )

inline

Allocate a memory region on the host, which is also mapped to a memory region in the global memory of a CUDA device - so that changes to one will be reflected in the other.

Parameters

device	The device on which the device-side region in the pair will be allocated
size_in_bytes	amount of memory to allocate (in each of the regions)
options	see memory::allocation_options

device_side_pointer_for()

template<typename T >

T* cuda::memory::mapped::device_side_pointer_for ( T *  host_memory_ptr )

inline

Obtain a pointer in the device-side memory space (= address range) given given a host-side pointer mapped to it.

Parameters

[in]

host_memory_ptr

a pointer to host-side memory which has been allocated as one side of a CUDA mapped memory region pair

device_side_region_for() [1/2]

region_t cuda::memory::mapped::device_side_region_for ( region_t  region )

inline

Get the memory region mapped to a given host-side region.

Note

if the input is already a device-side region, this function is idempotent

device_side_region_for() [2/2]

const_region_t cuda::memory::mapped::device_side_region_for ( const_region_t  region )

inline

Get the memory region mapped to a given host-side region.

Note

if the input is already a device-side region, this function is idempotent

free()

void cuda::memory::mapped::free ( region_pair_t  pair )

inline

Free a pair of mapped memory regions.

Parameters

pair	a pair of mapped host- and device-side regions

free_region_pair_of()

void cuda::memory::mapped::free_region_pair_of ( void *  ptr )

inline

Free a pair of mapped memory regions using just one of them.

Parameters

ptr	a pointer to one of the mapped regions (can be either the device-side or the host-side)

is_part_of_a_region_pair()

bool cuda::memory::mapped::is_part_of_a_region_pair ( const void *  ptr )

inline

Determine whether a given stretch of memory was allocated as part of a mapped pair of host and device memory regions.

Todo:

What if it's a managed pointer?

Parameters

ptr	the beginning of a memory region - in either host or device memory - to check

Returns

true iff the region was allocated as one side of a mapped memory region pair