Resolve Whether to Install NVidia GPU operator or Nvidia Device Plugin

[aedenj]

This remark would indicate that the gpu operator doesn't need to be installed and only the device plugin, but does that apply to the older AL2_x86_64_GPU AMI. Additionally the AWS Labs repo only installs the device plugin so that's most likely correct.

This announcement would indicate perhaps the operator is the way to go now with the newer AL2023_x86_64_NVIDIA AMI

[aedenj]

This error

      Message:      failed to create containerd task: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #0: error running hook: exit status 1, stdout: , stderr: /usr/local/nvidia/toolkit/nvidia-container-cli.real: /lib64/libc.so.6: version `GLIBC_2.27' not found (required by /usr/local/nvidia/toolkit/libnvidia-container.so.1): unknown

is resolved with the use of the gpu-operator by using latest AL2023_x86_64_NVIDIA AMI.