Farms used by Birmingham Particle Physics system

Farms used by Birmingham Particle Physics system

Filename locfarms. Last update 20110707. Part of Guide to the Local System.

Farm namePrefixImplementation dateNumber of nodes, coresspec SI2000base per core Memory GB per node, per core Local disk GB per node  
Etwin (non-grid) epef 2009/12 16, 128 3006+ 24, 3 500
Dtwin (grid) epgd 2008/08 24, 192 3006 16, 2 160
BlueBEAR (non-grid) u?n 2007/09 336, 1344 1606 8, 2 100
BlueBEAR (grid) u4n 2009/01 48, 192 1606 8, 2 100

Using Torque in general

For information on using the PBS/Torque batch system, take a look at our PBS/Torque pages.

Submitting jobs to the Etwin farm (non-GRID)

Jobs can be submitted from eprexb or eprexa or most PCs to the "E" twin farm (for example):
          qsub4 -j oe -l walltime=24:00:00
          qsub5 -j oe -l walltime=24:00:00
depending on whether you want your job to run under SL4 or SL5. Notice the use of the -l (lower case L) option to specify limits.

The job output file will be stored in the directory that was the current directory when you submitted the job, as usual. This has to be a writeable directory in /home/. (And not the local /tmp directory, for example). The default walltime is small so do not rely on the default.

This farm can see all files in the /home/ file systems.

Submitting jobs to BlueBEAR (non-GRID)

See this local web page.