Anthos feedback

It starts with a very detailed instruction on how to deploy Anthos, that is good but unnecessary.

    I would move the Anthos material to a separate instruction page and put link to the new page as reference for those not used to deploying Anthos.

The next section on adding an NVIDIA GPU is certainly where the interest for guidance starts. However even that has its issues. It starts with all the manual work for deploying drivers, tool-kits, device plug-ins, node-discovery-feature, etc.

    Again I would move this detail to a different reference page. I had previously assumed those instructions were the first steps, when in actual fact the GPU-Operator will handle all of that automatically.

The section on GPU-Operator installation is where the magic happens, and if not for the above two issues, we would have realized this much sooner.

    Just an update to that section, to make it clear that the basic command will work in most cases, I would move the default instruction above the chart customization options, then a user can see the default but also know how to modify the command to change some of the default options if required.

helm install --wait --generate-name \

 -n gpu-operator --create-namespace \

 nvidia/gpu-operator