These notes assume you followed the steps outlined here e.g ParallelCluster creation and then the GPU instance. If you do the other way around you may need to change the order of these steps (and those in the clean up setup in Section 3).
If you wish to simply stop the instance for when you’re not using it, then you can just head to the EC2 console and your instances and then click the instance you wish to stop (i.e g4d.4xlarge) and click Instance State, Stop Instance as shown below:
If you wish to terminate the instance, then you follow the same steps but select Terminate Instance. However you first need to deattach the ParallelCluster security group from the security group assigned to the g4dn.4xlarge instance. This will ensure that when you delete the parallelcluster it will also delete it’s own security group. To do this click Actions, Security, Change Security Groups
Then simply click remove next to the ParallelCluster MasterGroup and then Save.
You can then terminate the instance.
Please follow the steps in Section 3 to delete your HPC cluster.