IBM LoadLeveler is the workload manager on URSA. Users access it directly by loging to ursa.gsu.edu via SSH or they can submit their jobs via the EnginFrame portal.
The portal lets users submit, monitor, and control jobs and data files through a simple web interface. There are services for most frequently used applications: Gaussian, Amber, and SAS. Additionally, there is a service that lets you build your own submission script using a template and upload any data files for your job.

Use

Basic user commands:

llsubmit command_script - submits a job described in command_script

llcancel job_id - cancels a job

llclass - displays classes (queues) existing on the system and their descriptions

llq - lists queued jobs

llhold job_id - holds or releases a job

Find an online book with a wealth of information regarding LoadLeveler LoadLeveler">here.

A generic command script for a parallel job:

#!/bin/ksh

#####################################################

# This is a generic submission script fora parallel

# job. 

#####################################################

# @ error = default.$(jobid).err

# @ output = default.$(jobid).out

# @ class = verylong 

# @ job_type = parallel

# @ env_copy = all

# @ network.MPI = sn_single,shared,us

######################################################

# @ initialdir =   

######################################################

# @ restart = no

# @ node = 1

# @ total_tasks = 2

# @ queue

######################################################

######################################################

your_executable 

-- JaroKlc - 19 Jun 2008

Queues

Below are the current queues (known as 'classes' in Loadleveler) currently configured on the cluster:

*Name*	MaxJobCPU d+hh:mm:ss	*Max Slots*	*Description*
small	00:15:00	504	15 minute runtime limit queue
medium	02:00:00	504	2 hour runtime limit queue
large	1+00:00:00	496	24 hour runtime limit queue
large5	5+00:00:00	480	5 day runtime limit queue
large10	10+00:00:00	480	10 day runtime limit queue
large15	15+00:00:00	480	15 day runtime limit queue
large20	20+00:00:00	480	20 day runtime limit queue
large25	25+00:00:00	480	25 day runtime limit queue
verylong	30+00:00:00	480	30 day runtime limit queue
largememory64	30+04:00:00	16	30 day queue for the expanded memory node, 64GB
largememory256	30+04:00:00	16	30 day queue for the expanded memory node, 256GB
preemptME	unlimited	568	Unlimited runtime, however this class runs at lowest priority and will ALWAYS be preempted by any other queue
dev_small	00:15:00	16	15 minute runtime limit development queue
dev_medium	02:00:00	16	2 hour runtime limit development queue
sura_small	00:15:00	32	15 minute runtime limit SURAgrid queue
sura_large	5+00:00:00	32	5 day runtime limit SURAgrid queue
sura_priority	5+00:00:00	32	5 day runtime limit SURAgrid queue - preempts other SURA classes
georgia_southern	unlimited	16	Georgia Southern University restricted access queue 16 cpus, Node 36

-- VictorBolet - 12 Aug 2009