Kellogg Linux Cluster

Access

The Kellogg Linux Cluster (KLC) is available to all Kellogg researchers (faculty, PhD students, post-docs) and their collaborators. Current faculty and PhD students should already have access. If you have any trouble connecting or would like to grant access to a collaborator please contact Kellogg Research Support.

Installed Software

  • RStudio
  • Stata
  • Sas
  • Matlab
  • Python 2 and 3
  • GNU and Intel C, C++ and Fortan compilers
  • ... and much more

Technical Specifications

KLC consists of five machines, each of which has:

  • 28 Xeon CPU cores
  • 1.5TB of RAM

Storage on KLC is shared among the five machines, and it includes:

  • an 80GB home directory for each user
  • a shared 80 terabyte volume for shared data sets and for project data that does not fit in users' home directories

Connect to KLC with FastX

It is easy for Windows and Mac users to run statistical analysis programs on KLC.  Users may simply launch remote KLC programs using FastX.

Alternatively, those familiar with Linux may connect using their netId and one of the SSH hostnames:
  • klc01.ci.northwestern.edu
  • klc02.ci.northwestern.edu
  • klc03.ci.northwestern.edu
  • klc04.ci.northwestern.edu
  • klc05.ci.northwestern.edu

Access files on KLC

You can use the free programs WinSCP (for Windows) or Cyberduck (for Windows and Mac) to access the filesystem on KLC. Simply choose the SFTP protocol and connect to the hostname klc.northwestern.edu.

Software Modules

Unlike the old Skew5 server, you must use the module command to load software packages. This extra step allows the system to support multiple versions of the same software packages. The essential commands are:

  • module avail
  • module add [module_name]

For example, you can run Python 3 with the following two commands

  • module add python/anaconda3.6
  • python

Quest interoperability

KLC is actually a special part of NUIT's Quest High Performance Computing facility. Your files on KLC are also available on Quest, and this connectivity allows you to easily scale your computation up to thousands of cores on Quest, if necessary.

Load

You can use this chart to determine which of the machines has capacity available to run your code.