cuda.bindings.nvml.gpu_instance_create_compute_instance_with_placement#

cuda.bindings.nvml.gpu_instance_create_compute_instance_with_placement(
intptr_t gpu_instance,
unsigned int profile_id,
intptr_t placement,
) → intptr_t#

Create compute instance with the specified placement.

Parameters:

  • gpu_instance (intptr_t) – The identifier of the target GPU instance.

  • profile_id (unsigned int) – The compute instance profile ID. See nvmlGpuInstanceGetComputeInstanceProfileInfo.

  • placement (intptr_t) – The requested placement. See nvmlGpuInstanceGetComputeInstancePossiblePlacements.

Returns:

Returns the compute instance handle.

Return type:

intptr_t

See also

nvmlGpuInstanceCreateComputeInstanceWithPlacement