base::core::GPUAllocator Class Reference

An allocator that allocates memory on a GPU device. More...

#include <gpu_allocator.h>

Inheritance diagram for base::core::GPUAllocator:

Collaboration diagram for base::core::GPUAllocator:

Public Member Functions
void *	allocate (size_t size) override
	Allocate a block of memory with the given size and default alignment on GPU.
void *	allocate (size_t size, size_t alignment) override
	Allocate a block of memory with the given size and alignment on GPU.
void	free (void *ptr) override
	Free a block of GPU memory that was previously allocated by this allocator.
container::DeviceType	GetDeviceType () override
	Get the type of memory used by the TensorBuffer.
Public Member Functions inherited from base::core::Allocator
virtual size_t	AllocatedSize (void *ptr)
	Get the allocated size of a given pointer.
virtual	~Allocator ()=default

Additional Inherited Members
Protected Attributes inherited from base::core::Allocator
size_t	allocated_size_ = 0
	The total number of bytes allocated by this allocator.

Detailed Description

An allocator that allocates memory on a GPU device.

This class provides an implementation of the Allocator interface that allocates memory on a GPU device using CUDA APIs.

Member Function Documentation

allocate() [1/2]

void * base::core::GPUAllocator::allocate ( size_t  size )

overridevirtual

Allocate a block of memory with the given size and default alignment on GPU.

Parameters

size	The size of the memory block to allocate.

Returns

A pointer to the allocated memory block, or nullptr if the allocation fails.

Implements base::core::Allocator.

allocate() [2/2]

void * base::core::GPUAllocator::allocate ( size_t  size, size_t  alignment  )

overridevirtual

Allocate a block of memory with the given size and alignment on GPU.

Parameters

size	The size of the memory block to allocate.
alignment	The alignment of the memory block to allocate.

Returns

A pointer to the allocated memory block, or nullptr if the allocation fails.

Implements base::core::Allocator.

free()

void base::core::GPUAllocator::free ( void *  ptr )

overridevirtual

Free a block of GPU memory that was previously allocated by this allocator.

Parameters

ptr	A pointer to the memory block to free.

Implements base::core::Allocator.

GetDeviceType()

container::DeviceType base::core::GPUAllocator::GetDeviceType ( )

overridevirtual

Get the type of memory used by the TensorBuffer.

Returns

MemoryType The type of memory used by the TensorBuffer.

Implements base::core::Allocator.

---

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

• /home/runner/work/abacus-develop/abacus-develop/source/source_base/module_container/base/core/gpu_allocator.h
• /home/runner/work/abacus-develop/abacus-develop/source/source_base/module_container/base/core/gpu_allocator.cpp