Kubernetes Engine support for deploying the AMD GPU Operator cluster add-on

You can now use the AMD GPU Operator add-on to automate the deployment and management of AMD GPU software components. These components include the device plugin, node labeller, metrics exporter, test runner, and device configuration manager, which support AMD GPU workloads, monitoring, and GPU sharing in Kubernetes clusters.

For more information, see Overview of Cluster Add-ons and AMD GPU Operator add-on configuration arguments.