The y-jobs are run through the use of a csh script file that is submitted onto the machine. The jobs are run on Wednesdays in the very early morning (between 12 a.m. and 12 p.m. - see http://computing.oac.ucla.edu/ for more information on all this). The csh script below contains the commands that are submitted, as a whole, onto the job queue. By convention, you should name it with a .q extension. The job files are of this form:
--- next line is first line of queue file --- #!/bin/csh # @ initialdir = /u/data01/iegbtcp/trials_4_14_98/sim/02hosts # @ error = sweep_sim_loc64.us.2hosts.1.2.1.1.50cubed.error # @ output = sweep_sim_loc64.us.2hosts.1.2.1.1.50cubed.out # @ group = y # @ class = y # @ min_processors = 2 # @ max_processors = 2 # @ job_type = parallel # @ requirements = (Adapter=="hps_user") # @ notification = never # @ notify_user = phantom@cs.ucla.edu # @ queue unsetenv MP_HOSTFILE cd /u/data01/iegbtcp/trials_4_14_98/sim/02hosts rm -f hostfile poe /u/data01/iegbtcp/trials_4_14_98/sim/02hosts/sweep_sim_loc64 -euilib us 2 -ttp 1 2 1 1 50 -ttp > sweep_sim_loc64.us.2hosts.1.2.1.1.50cubed.data ---Important issues to note are the following: