What is the SSCC?

What is the SSCC and how does it work?

The Social Sciences Computing Cluster (SSCC) can help you run mathematical and statistical applications for computationally intensive research. There is a large collection of Statistical Software available. You can use the cluster to run large computational jobs that would take too long on you computer at home or in your office.

The figure below offers specific information on how the SSCC works.

Definitions

Your Computer is any computer that lets you access the NU Internet and that has the program SSH Secure Shell installed on it. SSH Secure Shell is installed on most computers in labs on the university campus and is available free of charge to install on your own computer. Off-campus connections require additional software to establish a secure extension to the NU campus network. See Authenticating to the NU Network with VPN.

SSH Secure Shell is the program which enables you to connect to the interactive machines. See How to Connect to the SSCC to install SSH and create an SSH profile.

Two Interactive Machines are the computers you log on to. They are called interactive machines because you are able to run jobs that require interaction with the program that you are using — you type commands to the program and immediately see the results on your screen.

Note that "hardin" replaced "hardin2" in October 2006. If you find SSCC graphical documentation still using "hardin2" please substitute "hardin" instead.

The Batch Cluster consists of 16 units called "nodes". Each node has two CPUs (processors), 2 GB RAM (memory) and 200 GB disk space used for temporary files.You can submit computing jobs to the Batch Cluster that do not require interaction (e.g. you have to run a long regression that would take 15 hours on an interactive computer. So just submit it to the Batch Cluster, let it run for 15 hours, and after the job is done you can login again and retrieve the results.)

The software that controls the Batch Cluster is configured to run just one job on a CPU. So, on the whole, 32 jobs can run simultaneously. If every processor is busy, you can still submit a job, which will wait in a queue until a processor is free, and then it will start automatically. You can be notified via email when your jobs start and finish. See Submit a Single Batch Job to learn how to submit a job to the Batch Cluster.

The NFS (Network File System) is where your home directory files are stored. No matter which computer you login to, your files will always be in the same central place (i.e. they are not stored on hardin2or seldon or any node in the Batch Cluster individually). The files in your home directory are backed up to tape five nights a week, from Tuesday morning to Saturday morning. Backup tapes are re-used after four weeks. No long-term, archival tape backups are made. As a result, you are responsible to archive important files yourself on your own backup media.