cuda.bindings.nvml.device_get_topology_nearest_gpus

cuda.bindings.nvml.device_get_topology_nearest_gpus(intptr_t device, int level)

Retrieve the set of GPUs that are nearest to a given device at a specific interconnectivity level For all products. Supported on Linux only.

Parameters:

• device (intptr_t) – The identifier of the first device.

• level (GpuTopologyLevel) – The nvmlGpuTopologyLevel_t level to search for other GPUs.

Returns:

An array of device handles for GPUs found at

level.

Return type:

intptr_t

See also

nvmlDeviceGetTopologyNearestGpus