gpu operator docs update

docs/NVIDIA-GPU-OPERATOR.md

@@ -1,8 +1,41 @@
 # NVIDIA GPU Operator
 
-`sudo rm -rf /run/nvidia`
+The NVIDIA GPU Operator provides a single-pane management resources that follows the Kuernetes operator pattern. When configured properly, capabilities like time-slicing and multi-instance GPUs can be provisioned on existing GPU resources on nodes within the cluster. Additionally, some optional components within the NVIDIA GPU Operator allow engineers to maintain NVIDIA dependencies in a Kubernetes-native way.
+
+## Node Feature Discovery
+
+The Kubernetes Node Feature Discovery component allows other Kubrenetes resources define and consume hardware and software resources available on a node. The NVIDIA GPU Operator requires this to be installed on the cluster beforehand so that NVIDIA GPUs can be characterized properly
+
+## Optional Components
+
+### NVIDIA Container Toolkit
+
+The NVIDIA Container Toolkit allows containerized applications and services to consume NVIDIA GPUs as resources. This is usually pre-installed on the host node prior to air-gapping or via an internally mirrored package repository or by bringing in the dependencies into the air-gap. The NVIDIA GPU Operator includes a DaemonSet that can be enabled to install the NVIDIA Container Toolkit on the host as a Kubernetes resource, allowing engineers the flexibility of deploying and updating the toolkit in a Kubernetes-native way.
+
+### NVIDIA GPU Drivers
+
+NVIDIA's Driver GPUs are usually pre-installed on the host node, similar to the NVIDIA Container Toolkit. Upgrading GPU drivers and ensuring they remain up-to-date can be done using the NVIDIA GPU Operator as well. By providing the correct pre-compiled drivers within the [`nvidia-gpu-operator-values.yaml`](../packages/nvidia-gpu-operator/values/nvidia-gpu-operator-values.yaml), and ensuring the host meets minimum requirements for installing these drivers via a Kubernetes pod, engineers can maintain and deploy drivers to host node sin a Kubernetes-native way.
+
+## Usage
+
+### Configuration
+
+Create and deploy-time configuration of the Zarf package is done mainly on the following components, and is dependent on the engineer's final desired configuration and the cluster's available node resources:
+
+1. (Create-time) [NVIDIA GPU Drivers](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/precompiled-drivers.html)
+2. (Deploy-time) [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html)
+3. (Deploy-time) [Time-Slicing](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/gpu-sharing.html)
+4. (Deploy-time) [Multi-Instance GPUs](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/gpu-operator-mig.html)
+
+### Cleanup
+
+In order to perform a fresh install after a previous deployment of the NVIDIA GPU operator, the engineer must remove the following directory from the host:
+
+```bash
+sudo rm -rf /run/nvidia
+```
 
 ## Additional Info
 
-- [LeapfrogAI Repository](https://github.com/NVIDIA/gpu-operator)
-- [LeapfrogAI Documentation Website](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/overview.html)
+- [NVIDIA GPU Operator Repository](https://github.com/NVIDIA/gpu-operator)
+- [NVIDIA GPU Operator Documentation Website](https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/overview.html)