How to submit jobs on Combo

From Computational Biophysics and Materials Science Group
Jump to: navigation, search

Queuing System

For the efficient use of the cluster, two Monitoring/Job Management software (PBS/Torque and Maui) have been installed.

After logging into to the cluster, the user is on the master node. When a program is run, it is also immediately run on the master. This is the "interactive mode", which is convenient for running simple commands like ls, vi, etc. or for editing/compiling a program. But, long computing jobs should be submitted through the queuing system. The submitted job will be in a queue waiting for its turn, then will be sent to one or more compute node(s), which the job will have dedicated access to until it finishes. Therefore, the job will run faster and the cluster will be more efficiently utilized.

Sample PBS job scripts

PBS job script for Parallel OPENMPI

You'll need to use PBS (Portable Batch System) scripts to set up and launch jobs on any cluster. While it is possible to submit batch requests using an elaborate command line invocation, it is much easier to use PBS scripts, which are more transparent and can be reapplied for sets of slightly different jobs. A PBS script performs these two key jobs:

1. It tells the scheduler about your job, such as:

  • The name of the program executable
  • How many CPUs you need and length of time to run the job
  • What to do if something goes wrong

2. The scheduler will 'run' your script when it comes time to launch your job.

A typical PBS script looks like this:


#PBS -l nodes=1:ppn=1,walltime=5:00:00
#PBS -N jobname
#PBS -M [email protected]
#PBS -m abe

cd /your/job/directory/

./Command &> output

exit 0;

The first "#PBS -l" line tells the scheduler to use one node with one processor per node (1 CPU in total), and this job will abort if not completed in 12 hours. You should put your job's name after "#PBS -N". If you would like to receive emails regarding to this job, you may leave your email address after "#PBS -M". The "#PBS -m abe" asks the system to email you when the job Aborts, Begins, and Ends.

Estimating Resources Requested

Estimating walltime as accurately as possible will help MOAB/Torque to schedule your job more efficiently. If your job requires 10-20 hours to finish do not ask for a much longer walltime. Please review available queues and queue parameters offered by Combo.

Estimating the number of nodes and the number of CPU cores is equally important. Requesting more nodes or more CPU cores than the job needs will remove these resources from the available pool.

Specify the nodes to be used

You may specify which node to be used by options in the submission script.


qsub -l nodes=compute-0-2:ppn=4+compute-0-3:ppn=3

This requests 4 processors on compute-0-2 and 3 processors on compute-0-3.

Submit Your Jobs

Submitting Serial Jobs

A serial job on Combo is defined as a job that does not require more than one node, which do not involve any inter-compute node data communications either.

Although OpenMP (NOT OpenMPI) jobs can use more than one CPU cores, all such cores are within a node. The OpenMP jobs, as a result, are serial jobs.

A serial job usually takes 1 to 16 CPU core in a node. We specify this in the "#PBS -l" line. The PBS script should be like this,


#PBS -l nodes=1:ppn=1,walltime=5:00:00
#PBS -N jobname
#PBS -M [email protected]
#PBS -m abe

cd /your/job/directory/

./Command &> output

exit 0;

Submit your batch job from the frontend with the command

qsub [job_script]

You get the job_name and job_id assigned, which can be used with various command.

Monitor Your Jobs

To see the progress information of running jobs, the command showq(Maui) and qstat(Torque) can be used. Both commands give you a summary of the status of submitted jobs and queues They give slightly different types of information. qstat shows a list of all running and waiting jobs in the queue, sorted by job identifier.

Please note that sometimes it takes a minute for submitted job to showq up under showq.

Another difference is that qstat shows time used for running jobs, while showq displays time left until the job will be killed by the queue system. When a job has finished it will no longer appear in the qstat or showq output.

Besides, the web based cluster monitor Ganglia is a very helpful tools to monitor the compute-node loading/status. Go to and from there you could view the status of Combo

To delete a running job, use

qdel [jobid]

Frequently Used PBS Command

PBS supplies a command line interface. This is used to submit, monitor, modify, and delete jobs. The following are some frequent used PBS user commands and their functions:

Command Description
qsub Submit a job
qstat List all information of queues and jobs
qdel jobid Delete a job

Frequently Used qsub option

Command Description
qsub -l list Set job resource list
qsub -N <jobname> Set job name to <jobname>
qsub -q <queue_name> Submit to queue <queue_name>

The resource requested on command line has a high preference than the directive line in the script file. For an example, submit job by command

qsub -l nodes=2:ppn=4 [jobscript]

this job will run on 2 compute nodes with 4 processors each instead of what stated in the script file.

qsub -l nodes=compute-0-0:ppn=16 [jobscript]

this job will run on the specified node (compute-0-0 in this case) with 16 processors.

Frequently Used qstat option

Command Description
qstat -a List all jobs with details
qstat -q List all queues on the system
qstat -n List all jobs with node information
qstat -u userid List all jobs owned by user userid
qstat -r List all running jobs
qstat -f jobid List all information known about specified job(jobid)

Must Read


With reference to: