BaseCUDAMemoryManager#

class nvmath.BaseCUDAMemoryManager(device_id: int, logger: Logger)[source]#

Protocol for memory manager plugins.

See also

numba.cuda.BaseCUDAMemoryManager

Methods

abstract __init__(
device_id: int,
logger: Logger,
)[source]#
abstract memalloc(
size: int,
) → MemoryPointer[source]#

Allocate device memory synchronously or on the current stream.

Parameters:

size – The size of the memory buffer in bytes.

Returns:

An object that owns the allocated memory and is responsible for releasing it (to the OS or a pool). The object must have an attribute named device_ptr, device_pointer, or ptr specifying the pointer to the allocated memory buffer. See MemoryPointer for an example interface.

Note

Objects of type numba.cuda.MemoryPointer as well as cupy.cuda.MemoryPointer meet the requirements listed above for the device memory pointer object.