How to Use the Molecular Modeling Facility Cluster
If you want to get an account on the cluster, please contact the Director at 4-4508 or e-mail to
From a workstation X11 terminal, you can open a secure shell connection:
ssh -Y firstname.lastname@example.org.
Change your temporary password immediately.
Do not share your account with anyone else.
Do not run any calculations on the head node -- use the queuing system.
Do not run any calculations that read and write big files in your home directory -- use your assigned scratch space (/work/username for single node jobs and /work/cluster/username for multinode jobs).
Login Nodes (gplogin1.ps.uci.edu and gplogin2)
Non-compute-intensive programs can be run on the login nodes. These include result-visualization and molecule-building programs like:
The following packages are set up to run from the queue on the compute nodes:
Torque/Maui Queuing System
can be found on the Adaptive Computing website.
but here is a short list:
qsub submission_script - submits a job
qdel job_number - kills a job
qstat -u $USER - shows your current jobs
qstat -n - gives detailed info on your jobs
qstat - shows everyone's jobs
You will also need to copy the example run script to your calculation setup directory:
cp /modfac/etc/run_pbsjob ~/path/to/your/calculation/input/directory/
Go to your calculation directory and edit the run script per its internal documentation (use gedit, gvim, emacs, etc.)
The job submission queues are listed below:
|Name ||Time limit ||Notes|
|mf_short ||3h ||can run in any free space|
|mf_prefinity ||none ||can be put to sleep by high-priority jobs|
Once your job is ready to go, type "qsub run_pbsjob" and patiently wait for your results to come back.