You are here: Home Experiment Information US ATLAS Getting Started Getting Started

Getting Started

by John S. De Stefano Jr. last modified Oct 02, 2007 09:46 AM
Using the US ATLAS Computing Facility at BNL.
  • If you don't have one yet, you need to apply for an account to gain access to the facility.
  • Disk space: New accounts are automatically given an NFS home directory [/usatlas/u/<username>] with an initial quota of 2 GB and an AFS [/afs/usatlas/users/<username>] directory with an initial quota of 400 MB. If you need more space, please submit a ticket to request an increase.
  • The ACF uses Kerberos 5 for authentication. You will need to read that for information on how to change your password or if you would like to authenticate from your desktop.
  • To access the computers in the ACF, first you need to SSH into one of the gateways at
  • From there you can either log into the general purpose interactive workstation at, or the Linux analysis farm nodes [where XXXX is a four digit number, currently 0001-0584, but only a subset of these allow interactive logins]. A complete list of all analysis nodes is also available.
  • To log into the farm nodes from the gateways, you can either use the rterm command, which will log you into the least loaded interactive node (run rterm -h on the gateway for more info), or you can log directly into the interactive node of your choice. Follow these steps for easy access to the interactive nodes:
    rterm acas+       # for an xterm window, or:
    rterm -i acas+          # no xterm
  • Access to the analysis farm is also available via the LSF and Condor batch systems, which is the only way to run jobs on the non-interactive nodes. As Condor is the preferred batch system at the ACF, users are encouraged to use (or migrate to) Condor, since LSF jobs get suspended when Condor jobs begin to execute on the same CPU resource. For information on using Condor, please look at the Condor Quick Start Guide. If you are unfamiliar with LSF please look at the Quick Start Guide and the list of the queues that are currently configured. You always have to specify the queue you want to submit to, if you are not sure which one then use either at_cas for regular jobs or at_cas_long for long running jobs.
  • If you need to archive your files then you can use the HPSS system.
  • Various grid services are provided by the RACF.
  • If you have any problems with the facility, please submit a ticket.
  • You should also look at the information on how to use the ATLAS software available from the PAS group.
Document Actions