cuda.core.LegacyPinnedMemoryResource#

class cuda.core.LegacyPinnedMemoryResource#

Create a pinned memory resource that uses legacy cuMemAllocHost/cudaMallocHost APIs.

Methods

__init__()#
allocate(
size,
*,
stream: Stream | GraphBuilder | None = None,
) → Buffer#

Allocate a buffer of the requested size.

cuMemAllocHost is synchronous, so this resource ignores any supplied stream. The argument is accepted (and validated when non-None) for interface conformance with stream-ordered memory resources.

Parameters:

  • size (int) – The size of the buffer to allocate, in bytes.

  • stream (Stream, optional) – Keyword-only. Validated when provided but otherwise unused.

Returns:

The allocated buffer object, which is accessible on both host and device.

Return type:

Buffer

deallocate(
ptr: DevicePointerType,
size,
*,
stream: Stream | GraphBuilder | None = None,
)#

Deallocate a buffer previously allocated by this resource.

Parameters:

  • ptr (DevicePointerType) – The pointer or handle to the buffer to deallocate.

  • size (int) – The size of the buffer to deallocate, in bytes.

  • stream (Stream, optional) – Keyword-only. If provided, stream.sync() is called before the host allocation is freed. None skips the sync.

Attributes

device_id#

This memory resource is not bound to any GPU.

is_device_accessible#

this memory resource provides device-accessible buffers.

Type:

bool

is_host_accessible#

this memory resource provides host-accessible buffers.

Type:

bool

is_managed#

bool

Whether buffers allocated by this resource are CUDA managed (unified) memory.

Type:

MemoryResource.is_managed