Test new GPU image installed with all prod packages.

After installing all the packages in the packer img to match with Cheaha compute node, install latest cuda drivers to use it as a VGPU node. We need to spin up a VM with the newly built image and then add this to slurm.conf for it to be a part of the cluster. We need to do some basic tests on this VM node by treating it as a GPU node. Check if we can load CUDA modules, Submit a job as we would for a GPU node. If all acceptance tests pass this issue can be closed.