cuda.bindings.nvml.PlatformInfo_v2

class cuda.bindings.nvml.PlatformInfo_v2

Empty-initialize an instance of nvmlPlatformInfo_v2_t.

See also

nvmlPlatformInfo_v2_t

__init__(*args, **kwargs)

Methods

__init__(*args, **kwargs)
from_buffer(buffer)	Create an PlatformInfo_v2 instance with the memory from the given buffer.
from_data(data)	Create an PlatformInfo_v2 instance wrapping the given NumPy array.
from_ptr(intptr_t ptr, bool readonly=False)	Create an PlatformInfo_v2 instance wrapping the given pointer.

Attributes

chassis_serial_number	(array of length 16).Serial number of the chassis containing this GPU (for Blackwell it is 13 bytes so indices 13-15 are zero)
host_id	Index of the node within the slot containing this GPU.
ib_guid	(array of length 16).Infiniband GUID reported by platform (for Blackwell, ibGuid is 8 bytes so indices 8-15 are zero)
module_id	ID of this GPU within the node.
peer_type	Platform indicated NVLink-peer type (e.g. switch present or not).
ptr	Get the pointer address to the data as Python int.
slot_number	The slot number in the chassis containing this GPU (includes switches)
tray_ind_ex	The tray index within the compute slots in the chassis containing this GPU (does not include switches)
version	the API version number