Connecting and Using the HPC

This page describes how to connect to the High Performance Computing Cluster, and how to use the features available on the login nodes.

Connecting

To connect to the HPC, you must either be on-campus or be connected to the FSU VPN. Use an SSH client connect:

ssh USERNAME@hpc-login.rcc.fsu.edu

For more information using SSH in general, refer to our guide.

HPC Overview

Once you authenticate via SSH, you will be connected to one of our three HPC Login nodes. When you first login, you are placed in your home directory, which resides on our Panasas filesystem. Your disk quota is 100gb. If you temporarily need more storage, you may be able to use our scratch space.

If you are a member of a research group with extra storage space, you may have access to other directories than your home directory. Check with your lab manager, or send us a note to see what storage you have access to.

Tools Available on the HPC

The HPC Login nodes are loaded with many open source and commercial software packages and libraries. In addition, all of the standard Linux utilities are available (text editors such as VI and Nano, CLI tools, etc). See all available software.

Some software packages require you to load modules before you use them. To better understand how modules work, refer to our brief guide.

Copying Data to and From the HPC

You may wish to copy data from external systems (such as your workstation or another server) onto the HPC, or vice versa. For details about how to move data to and from the HPC Panasas filesystem, refer to our guide.

Compiling Code

There are three compiler packages installed on the HPC Login Nodes, the GNU GCC Compiler, Intel Compilers, and Portland Group (PGI Compilers). In addition, there are a number of libraries availalble, notably MPI and OpenMP.

Refer to our guide for compiling code if you wish to learn how to do so on the HPC Login Nodes.

Submitting HPC Jobs

The primary purpose of the HPC Login nodes is to enable job submission and management on our powerful HPC cluster. Compute jobs do not run directly on the login nodes; once you compile your code and configure your job submission parameters, they run non-interactively on the HPC.

Refer to our guide for submitting and managing jobs to learn how to submit HPC jobs.