rmm::mr::system_memory_resource Class Reference

Memory resource that uses malloc/free for allocation/deallocation. More...

#include <system_memory_resource.hpp>

Public Member Functions
	system_memory_resource (system_memory_resource const &)=default
	Default copy constructor.
	system_memory_resource (system_memory_resource &&)=default
	Default copy constructor.
system_memory_resource &	operator= (system_memory_resource const &)=default
	Default copy assignment operator. More...
system_memory_resource &	operator= (system_memory_resource &&)=default
	Default move assignment operator. More...
void *	allocate ([[maybe_unused]] cuda::stream_ref stream, std::size_t bytes, [[maybe_unused]] std::size_t alignment=rmm::CUDA_ALLOCATION_ALIGNMENT)
	Allocates memory of size at least bytes. More...
void	deallocate (cuda::stream_ref stream, void *ptr, std::size_t bytes, [[maybe_unused]] std::size_t alignment=rmm::CUDA_ALLOCATION_ALIGNMENT) noexcept
	Deallocate memory pointed to by ptr. More...
void *	allocate_sync (std::size_t bytes, std::size_t alignment=rmm::CUDA_ALLOCATION_ALIGNMENT)
	Allocates memory of size at least bytes synchronously. More...
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. More...
bool	operator== (system_memory_resource const &) const noexcept
	Compare this resource to another. More...
bool	operator!= (system_memory_resource const &) const noexcept
	Compare this resource to another. More...

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

Detailed Description

Memory resource that uses malloc/free for allocation/deallocation.

There are two flavors of hardware/software environments that support accessing system allocated memory (SAM) from the GPU: HMM and ATS.

Heterogeneous Memory Management (HMM) is a software-based solution for PCIe-connected GPUs on x86 systems. Requirements:

• NVIDIA CUDA 12.2 with the open-source r535_00 driver or newer.
• A sufficiently recent Linux kernel: 6.1.24+, 6.2.11+, or 6.3+.
• A GPU with one of the following supported architectures: NVIDIA Turing, NVIDIA Ampere, NVIDIA Ada Lovelace, NVIDIA Hopper, or newer.
• A 64-bit x86 CPU.

For more information, see https://developer.nvidia.com/blog/simplifying-gpu-application-development-with-heterogeneous-memory-management/.

Address Translation Services (ATS) is a hardware/software solution for the Grace Hopper Superchip that uses the NVLink Chip-2-Chip (C2C) interconnect to provide coherent memory. For more information, see https://developer.nvidia.com/blog/nvidia-grace-hopper-superchip-architecture-in-depth/.

Member Function Documentation

allocate()

void* rmm::mr::system_memory_resource::allocate ( [[maybe_unused] ] cuda::stream_ref  stream, std::size_t  bytes, [[maybe_unused] ] std::size_t  alignment = rmm::CUDA_ALLOCATION_ALIGNMENT  )

inline

Allocates memory of size at least bytes.

The returned pointer will have at minimum 256 byte alignment.

The stream argument is ignored.

Parameters

stream	This argument is ignored
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::system_memory_resource::allocate_sync ( std::size_t  bytes, std::size_t  alignment = rmm::CUDA_ALLOCATION_ALIGNMENT  )

inline

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::system_memory_resource::deallocate ( cuda::stream_ref  stream, void *  ptr, std::size_t  bytes, [[maybe_unused] ] std::size_t  alignment = rmm::CUDA_ALLOCATION_ALIGNMENT  )

inlinenoexcept

Deallocate memory pointed to by ptr.

This function synchronizes the stream before deallocating the memory.

Parameters

stream	The stream in which to order this 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::system_memory_resource::deallocate_sync ( void *  ptr, std::size_t  bytes, std::size_t  alignment = rmm::CUDA_ALLOCATION_ALIGNMENT  )

inlinenoexcept

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::system_memory_resource::operator!= ( system_memory_resource const &  ) const

inlinenoexcept

Compare this resource to another.

All instances of system_memory_resource are equivalent.

Returns

true Always

operator=() [1/2]

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

default

Default move assignment operator.

Returns

system_memory_resource& Reference to the assigned object

operator=() [2/2]

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

default

Default copy assignment operator.

Returns

system_memory_resource& Reference to the assigned object

operator==()

bool rmm::mr::system_memory_resource::operator== ( system_memory_resource const &  ) const

inlinenoexcept

Compare this resource to another.

All instances of system_memory_resource are equivalent.

Returns

true Always

Friends And Related Function Documentation

get_property [1/2]

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

friend

Enables the cuda::mr::device_accessible property.

This property declares that a system_memory_resource provides device-accessible memory

get_property [2/2]

void get_property ( system_memory_resource const &  , cuda::mr::host_accessible    )

friend

Enables the cuda::mr::host_accessible property.

This property declares that a system_memory_resource provides host-accessible memory

---

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

• system_memory_resource.hpp