This is to create a blog or a documentation guide consisting of all the steps required to bring up the OCP cluster with a GPU, all the required prerequisites installed and leverage this feature using the featuregate to request a GPU via extended resources mechanism something very much similar to today's device plugin way

This doc can also contain the steps to migrate/upgrade the setup from using the device plugin to the DRA

Upstream Doc Ref: https://kubernetes.io/docs/concepts/scheduling-eviction/dynamic-resource-allocation/#extended-resource