cuda.core.experimental.VirtualMemoryResourceOptions#

class cuda.core.experimental.VirtualMemoryResourceOptions( allocation_type: ~typing.Literal['pinned', 'managed'] = 'pinned', location_type: ~typing.Literal['device', 'host', 'host_numa', 'host_numa_current'] = 'device', handle_type: ~typing.Literal['posix_fd', 'generic', 'win32', 'win32_kmt', 'fabric'] | None = 'posix_fd', granularity: ~typing.Literal['minimum', 'recommended'] = 'recommended', gpu_direct_rdma: bool = False, addr_hint: int | None = 0, addr_align: int | None = None, peers: ~typing.Iterable[int] = <factory>, self_access: ~typing.Literal['rw', 'r'] | None = 'rw', peer_access: ~typing.Literal['rw', 'r'] | None = 'rw', )#

A configuration object for the VirtualMemoryResource: Stores configuration information which tells the resource how to use the CUDA VMM APIs

allocation_type#

Controls the type of allocation.

Type:: VirtualMemoryAllocationTypeT

location_type#

Controls the location of the allocation.

Type:: VirtualMemoryLocationTypeT

handle_type#

Export handle type for the physical allocation. Use "posix_fd" on Linux if you plan to import/export the allocation (required for cuMemRetainAllocationHandle). Use None if you don’t need an exportable handle.

Type:: VirtualMemoryHandleTypeT

gpu_direct_rdma#

Hint that the allocation should be GDR-capable (if supported).

Type:: bool

granularity#

Controls granularity query and size rounding.

Type:: VirtualMemoryGranularityT

addr_hint#

A (optional) virtual address hint to try to reserve at. Setting it to 0 lets the CUDA driver decide.

Type:: int

addr_align#

Alignment for the VA reservation. If None, use the queried granularity.

Type:: int

peers#

Extra device IDs that should be granted access in addition to device.

Type:: Iterable[int]

self_access#

Access flags for the owning device.

Type:: VirtualMemoryAccessTypeT

peer_access#

Access flags for peers.

Type:: VirtualMemoryAccessTypeT