cuda.bindings.nvml.device_get_utilization_rates#

cuda.bindings.nvml.device_get_utilization_rates(intptr_t device)#

Retrieves the current utilization rates for the device’s major subsystems.

For Fermi ™️ or newer fully supported devices. See nvmlUtilization_t for details on available utilization rates.

Parameters:

device (intptr_t) – The identifier of the target device.

Returns:

Reference in which to return the utilization: information.

Return type:

nvmlUtilization_t

Note

During driver initialization when ECC is enabled one can see high GPU and Memory Utilization readings. This is caused by ECC Memory Scrubbing mechanism that is performed during driver initialization.

Note

On MIG-enabled GPUs, querying device utilization rates is not currently supported.