Connecting | Cluster Layout | Running Jobs | Compilers and Libraries | Further Reading
Getting Started
In this document, names beginning with $ refer to shell variables. These refer to values that depend on your login environment. For example the $USER variable contains your username.
Connecting to the Clusters
The clusters are enabled for Secure Shell (SSH) access only. If you are a Windows user you can download the SSH Client for Windows. If you prefer, you could also use eXceed, cygwin or PuTTY to connect to the clusters. Linux users should find that SSH is installed by default on their system.
If you are at St. Andrews University you can make use of a site licence for eXceed and wonload it from the St. Andrews IT Services website.
There is a How To Connect to the Clusters document that provides step-by-step instructions on connecting to and transferring files to/from the cluster using the SSH Client for Windows.
Cluster Layout
For the physical layout of the clusters see the Facilities Pages.
Home directories for users are found in the /users/$USER directory. Please ensure that you keep your directory space clean and tidy. Do not keep large useless files lying around as this slows down the system for everyone.
Please note that the home directories are not backed-up in any shape or form. It is your responsibility to keep copies of your important data somewhere safe.
Everybody has temporary disk space for use by various programs and to store temporary files. This can be found in /scratch/$GROUP/$USER and accessed by the $scratch shell variable.
All computational software can be found in the /usr/local directory. You must be a member of the relevant group in order to access a particular piece of code. Please apply to your RCO for this service if you are not a member of a group for the software you want to use.
Running Compuational Jobs
All computational jobs must be submitted through the queuing system.
The standard software packages all have their own submission scripts that you can find out about in the Software Information topic.
There are also scripts that allow you to submit jobs using your own code to the queuing system - either parallel or serial. For more information, see the Job Submission topic.
Compilers and Libraries
Different (parallel) Fortran and C compilers are available on each of the two clusters. The Portland Group (PGI) compilers are available on hare and the Intel Compilers on burke. The documentation on these compilers is available in our User Guides section.
In addition, the clusters have Java compilers installed as well as access to Perl and Python. More information on these languages is available on the web. Just Google for them.
Both clusters have the BLAS, LAPACK and FFTW3 libraries installed. The documentation for these libraries is available in our User Guides section.
Your RCO will also be happy to answer any questions you may have on compilation, libraries and programming issues.
Other things to read
All users should probably read the Job Submission topic which introduces the use of the SGE queuing system for submitting computational jobs to the clusters.
People who are new to using the Linux command line should definitely read the Linux and Text Editors topic which outlines the basic Linux commands and points to other resources on the web. This also includes an introduction to the Vi text editor that you will spend a lot of time using on the clusters.
For information on running specific software on the clusters please see the Software Information topic. There you will find specific help on using the computational chemistry software on the system as well as links to the online versions of the manuals.