rmm::mr::cuda_async_view_memory_resource Class Reference

Memory resource that uses cudaMallocAsync/cudaFreeAsync for allocation/deallocation. More...

#include <cuda_async_view_memory_resource.hpp>

Public Member Functions
	cuda_async_view_memory_resource (cudaMemPool_t pool_handle)
	Constructs a cuda_async_view_memory_resource which uses an existing CUDA memory pool. The provided pool is not owned by cuda_async_view_memory_resource and must remain valid during the lifetime of the memory resource.
cudaMemPool_t	pool_handle () const noexcept
	Returns the underlying native handle to the CUDA pool.
	cuda_async_view_memory_resource (cuda_async_view_memory_resource const &)=default
	Default copy constructor.
	cuda_async_view_memory_resource (cuda_async_view_memory_resource &&)=default
	Default move constructor.
cuda_async_view_memory_resource &	operator= (cuda_async_view_memory_resource const &)=default
	Default copy assignment operator.
cuda_async_view_memory_resource &	operator= (cuda_async_view_memory_resource &&)=default
	Default move assignment operator.
void *	allocate (cuda::stream_ref stream, std::size_t bytes, std::size_t alignment=rmm::CUDA_ALLOCATION_ALIGNMENT)
	Allocates memory of size at least bytes.
void	deallocate (cuda::stream_ref stream, void *ptr, std::size_t bytes, std::size_t alignment=rmm::CUDA_ALLOCATION_ALIGNMENT) noexcept
	Deallocate memory pointed to by ptr.
void *	allocate_sync (std::size_t bytes, std::size_t alignment=rmm::CUDA_ALLOCATION_ALIGNMENT)
	Allocates memory of size at least bytes synchronously.
void	deallocate_sync (void *ptr, std::size_t bytes, std::size_t alignment=rmm::CUDA_ALLOCATION_ALIGNMENT) noexcept
	Deallocate memory pointed to by ptr synchronously.
bool	operator== (cuda_async_view_memory_resource const &other) const noexcept
	Compare this resource to another.
bool	operator!= (cuda_async_view_memory_resource const &other) const noexcept
	Compare this resource to another.

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

Detailed Description

Memory resource that uses cudaMallocAsync/cudaFreeAsync for allocation/deallocation.

Constructor & Destructor Documentation

cuda_async_view_memory_resource()

rmm::mr::cuda_async_view_memory_resource::cuda_async_view_memory_resource

(

cudaMemPool_t

pool_handle

)

Constructs a cuda_async_view_memory_resource which uses an existing CUDA memory pool. The provided pool is not owned by cuda_async_view_memory_resource and must remain valid during the lifetime of the memory resource.

Exceptions

rmm::logic_error

if the CUDA version does not support cudaMallocAsync

Parameters

pool_handle

Handle to a CUDA memory pool which will be used to serve allocation requests.

Member Function Documentation

allocate()

void * rmm::mr::cuda_async_view_memory_resource::allocate ( cuda::stream_ref stream, std::size_t bytes, std::size_t alignment = rmm::CUDA_ALLOCATION_ALIGNMENT )

nodiscard

Allocates memory of size at least bytes.

The returned pointer will have at minimum 256 byte alignment.

Parameters

stream	Stream on which to perform allocation
bytes	The size of the allocation
alignment	The alignment of the allocation

Returns

void* Pointer to the newly allocated memory

allocate_sync()

void * rmm::mr::cuda_async_view_memory_resource::allocate_sync ( std::size_t bytes, std::size_t alignment = rmm::CUDA_ALLOCATION_ALIGNMENT )

nodiscard

Allocates memory of size at least bytes synchronously.

Parameters

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

Returns

void* Pointer to the newly allocated memory

deallocate()

void rmm::mr::cuda_async_view_memory_resource::deallocate ( cuda::stream_ref stream, void * ptr, std::size_t bytes, std::size_t alignment = rmm::CUDA_ALLOCATION_ALIGNMENT )

noexcept

Deallocate memory pointed to by ptr.

Parameters

stream	Stream on which to perform deallocation
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 ptr.
alignment	The alignment that was passed to the allocate call that returned ptr

deallocate_sync()

void rmm::mr::cuda_async_view_memory_resource::deallocate_sync ( void * ptr, std::size_t bytes, std::size_t alignment = rmm::CUDA_ALLOCATION_ALIGNMENT )

noexcept

Deallocate memory pointed to by ptr synchronously.

Parameters

ptr	Pointer to be deallocated
bytes	The size in bytes of the allocation
alignment	The alignment that was passed to the allocate call that returned ptr

operator!=()

bool rmm::mr::cuda_async_view_memory_resource::operator!= ( cuda_async_view_memory_resource const & other ) const

nodiscardnoexcept

Compare this resource to another.

Parameters

other

The other resource to compare to

Returns

true If the two resources are equivalent

false If the two resources are not equal

operator=() [1/2]

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

default

Default move assignment operator.

Returns

cuda_async_view_memory_resource& Reference to the assigned object

operator=() [2/2]

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

default

Default copy assignment operator.

Returns

cuda_async_view_memory_resource& Reference to the assigned object

operator==()

bool rmm::mr::cuda_async_view_memory_resource::operator== ( cuda_async_view_memory_resource const & other ) const

nodiscardnoexcept

Compare this resource to another.

Parameters

other

The other resource to compare to

Returns

true If the two resources are equivalent

false If the two resources are not equal

pool_handle()

cudaMemPool_t rmm::mr::cuda_async_view_memory_resource::pool_handle ( ) const

nodiscardnoexcept

Returns the underlying native handle to the CUDA pool.

Returns

cudaMemPool_t Handle to the underlying CUDA pool

Friends And Related Symbol Documentation

get_property

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

friend

Enables the cuda::mr::device_accessible property.

This property declares that a cuda_async_view_memory_resource provides device accessible memory

---

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

• cuda_async_view_memory_resource.hpp