|
|
Statistics and Computation ServiceThe ITCS Statistics and Computation Service (SCS) offers servers to regular students, faculty, and staff of the University of Michigan to run computationally intensive or long-running programs. These vary from using mathematical and statistical programs to compiling and running custom programs. SCS on LinuxThe SCS was migrated from Solaris to Linux in January 2006. The current production pool of SCS servers runs a customized Linux distribution created at the University. The testing pool runs on Red Hat Linux. In the coming months the whole SCS service will be migrated to Red Hat. For information about changes in hardware and software, please see our SCS on Linux page. Connecting to the ServerIf you have the University's Basic Computing Package that is provided to all registered students, regular faculty, and regular staff, you can use the SCS. Access to the SCS is included in the package. You must use SSH (secure shell) to connect to the SCS, and you must use SFTP (secure ftp) to copy files to and from the server. For more information about SSH, and why we implemented this requirement, please see our SSH web page. The hostname for connecting to the SCS is scs.itd.umich.eduConnecting to this hostname will log you into one of the machines in the SCS pool. Depending on the program you're running and whether you're writing to the local disk, you may need to log into the same machine multiple times. To determine which machine you've logged into, type hostnameat the command prompt. To connect to a specific machine, point your SSH client at the hostname for that machine. For example, gauntlet.gpcc.itd.umich.edu Enabling X11 ForwardingFor instructions on using graphical programs on the Statistics and Computation machines while connected from your local computer, click here.SCS SoftwareMost of the programs that run on the Login Service machines are also available on the SCS server, along with a variety of mathematics and statistics software. Long-Running Jobs and Local File SpaceFor information about local file access for long-running jobs, and other ways to make running long jobs easier, please read the Long-Running Jobs page. Long Term Storage for Large Data SetsIf your analyses require a lot of long term file space, please send a request via e-mail to scs-admins@umich.edu and we will create space in IFS for your use. There will be no additional fees for this space, provided that it is used for research purposes. Resource Utilization and Memory LimitBecause there are a limited number of SCS machines with a limited amount of memory, it is important to be conscious of your use of resources on these machines. Please do not run many instances of the same program on one machine. If you need to run four instances of a program, you should run one instance on each machine in the pool rather than running all of them on one machine, and using up all of the CPU and memory available on that machine. Also, when you are finished with a job, please make sure that you don't leave it running. An unattended process can run for weeks or even months, using up resources. There is currently a memory limit of 8GB per login per user. This limit prevents a runaway process from accidentally consuming all of the memory on one of the machines. If you need to run a job that requires more than 8 GB of RAM, you can email scs-admins@umich.edu to arrange a temporary lifting of the limit for your account on one of the machines. Software SupportIf you have questions about, or encounter errors in, the software or operating system on the SCS machines, please contact scs-admins@umich.edu. Likewise, requests for the installation of new software and suggestions on improving the SCS should be sent to scs-admins@umich.edu. If you have any questions regarding how to use the statistics or computational software available on these machines, please contact the Center for Statistical Consultation and Research (CSCAR).
Email: cscar@umich.edu
Email: online.stat@umich.edu
Web: http://www.umich.edu/~cscar/
Phone: (734) 764-STAT(7828)
The hours of operation for CSCAR are
|
|
Comments, questions, or suggestions? Please e-mail
them to
ITCS Statistics and Computation Service
scs-admins@umich.edu
![]()
![]()