Conduct Research on Ai localized data center

AI Discussions
1

I want to do PhD on ai localized data center using LSTM and RL for evaluating performance metrics like cpu, gpu performance and memory operations.

  1. How should I setup the environment and tools required for it.
  2. How should I train my ml model for this and when to use LSTM and RL in my ml model.

I have already setup the environment using Proxmox and Ubuntu which uses nested virtualization.

I would be grateful for your guidance

Regards

2

Heads up: I don’t have a PhD.

Afaik, publishing the metrics you’ve listed would probably count towards a survey paper or as part of a paper where existing state is described. Getting a PhD with just a survey paper isn’t something I’ve heard of. To publish a paper at a PhD level, you’ll have to either contribute an improvement over an existing state or come up with a breakthrough idea (with metrics to back it up of course).

It would help readers get a better grip of your ask if the problem statement is clearly stated.

Based on your questions, this should help:

I don’t think proxmox is free. Try docker / kind (k8s).

3

Well, Proxmox VE itself is free. But with their subscription plans, the main difference is that the enterprise repositories are tested, unlike the normal ones you get with the free tier.