Getting Started
Welcome to the Caltech Central HPC Cluster. If you’re new, work through the steps below in order; if you’re already set up, the quick links jump to what you need.
Quick Links
Request an account and set up the login credentials you’ll connect with.
See the compute nodes, GPUs, and memory the cluster offers.
See how compute and storage are priced, and how allocations work.
How to reach us when you get stuck.
First Steps
Get an account. Everyone starts here — request access through Account Information.
Set up multi-factor authentication. The cluster uses Duo MFA, so you’ll enroll a device before your first login.
Connect over SSH. From your own terminal, log in to the head node at
login.hpc.caltech.edu.Get comfortable with SLURM. Work runs through the SLURM scheduler rather than directly on the login node — SLURM Commands covers the handful of commands you’ll use most.
Move your data over. Copy your code and datasets onto the cluster using the methods in Transferring Files.
Logging In
Once your account is active, you can connect from any terminal — macOS, Linux, or Windows:
ssh username@login.hpc.caltech.edu
The first time through, you’ll enter your cluster password and then approve a Duo prompt on your phone. Once you’re in, you’ll land on a login node — a place to prepare and submit work, not to run heavy computation directly.
Tip
If you log in many times a day, SSH multiplexing lets you reuse a single authenticated connection so you’re not approving Duo over and over. The SSH Password FAQ walks through the setup.
Next Steps
Browse the Available Software to see what’s already installed and ready to load as a module.
Get to know your Storage Options — where to keep code, working data, and large scratch files.
Skim the FAQ for the questions new users run into most.