rmm::mr::pinned_memory_resource Class Reference

A host_memory_resource that uses cudaMallocHost to allocate pinned/page-locked host memory. More...

#include <pinned_memory_resource.hpp>

Inheritance diagram for rmm::mr::pinned_memory_resource:

Collaboration diagram for rmm::mr::pinned_memory_resource:

Public Member Functions
	pinned_memory_resource (pinned_memory_resource const &)=default
	Default copy constructor.
	pinned_memory_resource (pinned_memory_resource &&)=default
	Default move constructor.
pinned_memory_resource &	operator= (pinned_memory_resource const &)=default
	Default copy assignment operator. More...
pinned_memory_resource &	operator= (pinned_memory_resource &&)=default
	Default move assignment operator. More...
void *	allocate_async (std::size_t bytes, std::size_t alignment, cuda_stream_view)
	Pretend to support the allocate_async interface, falling back to stream 0. More...
void *	allocate_async (std::size_t bytes, cuda_stream_view)
	Pretend to support the allocate_async interface, falling back to stream 0. More...
void	deallocate_async (void *ptr, std::size_t bytes, std::size_t alignment, cuda_stream_view)
	Pretend to support the deallocate_async interface, falling back to stream 0. More...
Public Member Functions inherited from rmm::mr::host_memory_resource
	host_memory_resource (host_memory_resource const &)=default
	Default copy constructor.
	host_memory_resource (host_memory_resource &&) noexcept=default
	Default move constructor.
host_memory_resource &	operator= (host_memory_resource const &)=default
	Default copy assignment operator. More...
host_memory_resource &	operator= (host_memory_resource &&) noexcept=default
	Default move assignment operator. More...
void *	allocate (std::size_t bytes, std::size_t alignment=alignof(std::max_align_t))
	Allocates memory on the host of size at least bytes bytes. More...
void	deallocate (void *ptr, std::size_t bytes, std::size_t alignment=alignof(std::max_align_t))
	Deallocate memory pointed to by ptr. More...
bool	is_equal (host_memory_resource const &other) const noexcept
	Compare this resource to another. More...
bool	operator== (host_memory_resource const &other) const noexcept
	Comparison operator with another device_memory_resource. More...
bool	operator!= (host_memory_resource const &other) const noexcept
	Comparison operator with another device_memory_resource. More...

Friends
void	get_property (pinned_memory_resource const &, cuda::mr::device_accessible) noexcept
	Enables the cuda::mr::device_accessible property. More...

Detailed Description

A host_memory_resource that uses cudaMallocHost to allocate pinned/page-locked host memory.

See https://devblogs.nvidia.com/how-optimize-data-transfers-cuda-cc/

Member Function Documentation

allocate_async() [1/2]

void* rmm::mr::pinned_memory_resource::allocate_async ( std::size_t  bytes, cuda_stream_view    )

inline

Pretend to support the allocate_async interface, falling back to stream 0.

Exceptions

rmm::bad_alloc

When the requested bytes cannot be allocated on the specified stream.

Parameters

bytes

The size of the allocation

Returns

void* Pointer to the newly allocated memory

allocate_async() [2/2]

void* rmm::mr::pinned_memory_resource::allocate_async ( std::size_t  bytes, std::size_t  alignment, cuda_stream_view    )

inline

Pretend to support the allocate_async interface, falling back to stream 0.

Exceptions

rmm::bad_alloc

When the requested bytes cannot be allocated on the specified stream.

Parameters

bytes	The size of the allocation
alignment	The expected alignment of the allocation

Returns

void* Pointer to the newly allocated memory

deallocate_async()

void rmm::mr::pinned_memory_resource::deallocate_async ( void *  ptr, std::size_t  bytes, std::size_t  alignment, cuda_stream_view    )

inline

Pretend to support the deallocate_async interface, falling back to stream 0.

Parameters

ptr	Pointer to be deallocated
bytes	The size in bytes of the allocation. This must be equal to the value of bytes that was passed to the allocate call that returned p.
alignment	The alignment that was passed to the allocate call that returned p

operator=() [1/2]

pinned_memory_resource& rmm::mr::pinned_memory_resource::operator= ( pinned_memory_resource &&  )

default

Default move assignment operator.

Returns

pinned_memory_resource& Reference to the assigned object

operator=() [2/2]

pinned_memory_resource& rmm::mr::pinned_memory_resource::operator= ( pinned_memory_resource const &  )

default

Default copy assignment operator.

Returns

pinned_memory_resource& Reference to the assigned object

Friends And Related Function Documentation

get_property

void get_property ( pinned_memory_resource const &  , cuda::mr::device_accessible    )

friend

Enables the cuda::mr::device_accessible property.

This property declares that a pinned_memory_resource provides device accessible memory

---

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

• pinned_memory_resource.hpp