If you are not very familiar with OpenPBS and want to schedule
some task to be run on cluster without writing your own PBS script,
then you can try to use ravanize command with supplied fullname
of the command to run across the cluster. For the options of
ravanize run ravanize --help.
Ravanize can help you to run simple batch jobs as well as to run
interactive X applications available across the cluster (if X application is a
wrapper around computationally intensive software like FSL, you should not
 run it on the MainNoderavanize). For every ravanize invocation on
command COMMAND, there will be COMMAND.PID.pbs file created, so, in
future, you can rerun the job simply by running qsub COMMAND.PID.pbs
Whenever commands commands succeeds its standard output will be
provided in the file JOBID.OU  and rather use

$> ravanize -q verylong $PWD/cpi_cut
Name: 'yoh.21974'  Queue: verylong   Nodes: 1   PPN: 1
Command: /home/yoh/admin/testm/cpi_cut
348766.itanix.ravana.rutgers.edu

That tells you the job's number is 348766. Now you can check the status
of your job by using qstat 348766 (or qstat -f 348766. for more information)

You must be working in a VNC session. If you are working
through pure ssh with X forwarding, it would not work. Simply
from the main node (it would fail if ran on another node)
run ravanize -X APPLICATION_NAME

$> ravanize -q long -X fsl
Name: 'yoh.26441'  Queue: long   Nodes: 1   PPN: 1
Command: fsl
382784.itanix.ravana.rutgers.edu

And if we look at qstat -f we can see that actually FSL is running on node13

*$> qstat -f 382784 | grep exec_host
    exec_host = node13/0

A bit more evolved command to reserve 2GB of RAM and run FSL4.0 is
ravanize -l mem=2GB -q long -X "export FSLDIR=/usr/apps/fsl.4.0/; fsl"