How to manage and monitor local cluster resource allocation ?
We are setting up two GPU server rigs ( https://in.pcpartpicker.com/list/GJstm8 ). These GPU servers should be accessible from anywhere over SSH. The GPU servers will be used for Deep Learning. We want to allocate 25% of GPU resource (out of the total 24GB) at once to each user, on request, we’ll assign more if free, maximum of 50%. How can we manage the User authentication and roles ?