The cluster is dedicated to computing in support of research. It is not to be used in direct support of course-work or teaching.
For now we have not implemented any kind of quota or queuing systems on the cluster. Rather, we are relying upon the users to take a friendly and collegial approach to sharing the cluster.
That is to say that for now we are relying on individual users to nice their processes. This can be done in various ways, such as with one of the commands nice, renice, or top. Nice values, as displayed by top, range from 0 to 19, with 0 being given highest priority and 19 the lowest. The operating system uses these nice values to determine how much computing time each process receives.
Note, however, that if a node has excessively many processes running on it, then all processes will suffer as the node devotes more time to the task of juggling the processes. If this should become a problem that we face, then a more formal system to regulate the scheduling of CPU time may be implemented.