cuda.bindings.nvml.device_create_gpu_instance_with_placement#

cuda.bindings.nvml.device_create_gpu_instance_with_placement( intptr_t device, unsigned int profile_id, placement, ) → intptr_t#

Create GPU instance with the specified placement.

For Ampere ™️ or newer fully supported devices. Supported on Linux only. Requires privileged user. If the parent device is unbound, reset or the GPU instance is destroyed explicitly, the GPU instance handle would become invalid. The GPU instance must be recreated to acquire a valid handle.

Parameters:

device (intptr_t) – The identifier of the target device.
profile_id (unsigned int) – The GPU instance profile ID. See nvmlDeviceGetGpuInstanceProfileInfo.
placement (intptr_t) – The requested placement. See nvmlDeviceGetGpuInstancePossiblePlacements_v2.

Returns:

Returns the GPU instance handle.

Return type:

intptr_t