Distributed computing with Condor - How-to do it
What is this about?
The Biocomputation and Adaptive Systems groups of the University of Hertfordshire have recently joined their spare computers to form a big cluster. Everybody can now run jobs and quickly get results, making work more efficient.
On this page you will soon find an installation guide for Windows and Linux as well as tutorials on submitting jobs to the cluster. The information is tailored for people located at the University of Hertfordshire, however other people might find the information provided useful, too.
The software used here is Condor; installing it is easy and your computer will only run external jobs when it is idle and you can submit jobs to the cluster yourself. If you want to know more details about the underlying principles, read "What is Condor?".
Acknowledgment: Apart from the author, Alexander Klyubin, Mark Robinson and Moritz Buck are all very important parts of this collaborative effort against the waste of computing power.
There are the following sections:If you are not at the University of Hertfordshire read up about installing your own cluster; it probably pays for most universities and companies!
- Install Condor on a Windows machine
- Install Condor on a Linux machine (to come)
- Basic commands (work in progress, more to come)
- Create and submit jobs (work in progress, more to come)
- Make Condor return Java job results from a directory structure
official Condor manual
Infos from Cambridge University
Special info on flocking between two clusters (e.g. between departments...)