Policies for Use of the Turing Cluster

Please observe the following policies in using the Turing Cluster.  Failure to do so may result in account termination.
  1. Sharing of accounts is forbidden.  Each user must have his or her own individual account and password, and these are not to be shared under any circumstances.
  2. Each user must be personally responsible for his or her own jobs. If you cannot be around to monitor your jobs, then don't submit them.
  3. The only machines you should ever log into directly are the front end array machines, turing-1, turing-2, turing-3, and turing-4, which can be reached via hostname turing.cse.uiuc.edu. With the exception of compiling, no computationally intensive activity is allowed on the front-end array.
  4. Do not use ssh keys or any other method to set up passwordless login to the cluster.
  5. The cluster is to be used for scientific and engineering computational jobs only; no other work is allowed.  Usage that requires modification of the operating system, networking software, or system libraries is not permitted without explicit permission; other clusters are available and configured for such experiments.
  6. Always use the job queue to submit your computational jobs.  Do not attempt to subvert the queue by using mpirun directly.  Violators will be promptly executed.
  7. For MPI jobs for which optimum performance is not an issue (most class jobs fit here), please use both processors on each node when possible. This will help keep as many cluster nodes as possible available for performance sensitive jobs.
  8. Remember that you are sharing this resource with other users.  Monopolistic or other antisocial behavior will not be tolerated.

There are no automatic mechanisms in place to prevent you from violating the above guidelines.  However, if you are found to be in violation of them, or otherwise using the cluster in an inappropriate manner, your account may be terminated. If you have special need to run jobs which exceed these limits, please contact turing-help@cse.uiuc.edu as it is possible we can accomidate your need during or immediately following Turing weekly preventive maintenance (PM) sessions.

To get started running MPI jobs on Turing, see the tutorial.

If you have any problems using the cluster, please send a description of the problem, along with any error messages, to turing-help@cse.uiuc.edu.