Add MIG setup to the GPU page

[nwneisen]

This change is still waiting for tickets and configuration to be finalized

Document MIG (Multi-Instance GPU) functionality
https://mirantis.jira.com/browse/BOP-1310
https://github.com/MirantisContainers/mke/pull/437
https://github.com/MirantisContainers/mke-operator/pull/194

MIG is a GPU setting that allows a GPU to be partitioned so that multiple workloads can be ran in parallel without interfering with one another. Nvidia has two strategies for doing this

• single - The GPU partitions are all the same size
• mixed - The GPU partitions can be of various sizes

The initial implementation only supports the single strategy to match MKE3 functionality.

Configuration
MIG is disabled default in MKE. It can be enabled by setting the MIG section's enabled field to true in the mkeconfig.

Official documentation: https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/gpu-operator-mig.html#about-multi-instance-gpu

Note
Enabling MIG on a GPU that was previously running will cause the pods to reboot even if it has workloads running.

devicePlugins:
  nvidiaGPU:
    enabled: true
    mig:
      enabled: true

A profile can then be applied to a node using the nvidia.com/mig.config: <MIG profile>. Supported profiles vary by GPU and can be found at: https://docs.nvidia.com/datacenter/tesla/mig-user-guide/#getting-started-with-mig