Running jobs
Job Launch Overview
Parallel applications can not run on the login nodes. They must be launched onto the compute nodes via the aprun command. Read More »
Interactive Jobs
Interactive jobs may be run on Franklin by requesting resources from the batch system. "qsub -I -V -q interactive -lmppwidth=[num_cores]" is the basic command to request interactive resources. Read More »
Submitting Batch Jobs
Basic batch scripts, torque keywords. Sample scripts for advanced work flows: running multiple jobs, MPMD jobs, job dependencies. Read More »
Queues and Policies
Queue configuration and limits, policies and tips for getting your job through the queue faster. Read More »
Monitoring Jobs
Once a job is submitted it can be monitored, held, deleted and in some cases altered. Read More »
Memory Considerations
Each Franklin compute node has 8 GB (8192 MB) of physical memory, but, not all that memory is available to user programs. A user can change MPI buffer sizes by setting certain MPICH environment variables. Read More »
Runtime Tuning Options
Optimizations and tuning parameters which can be set at run time to improve the performance of your code. Read More »
Dedicated Time Request
NERSC is testing a new service which will allow users who need access to a large number of nodes to make a reservation for a certain date, time and duration. Read More »
How Usage Is Charged
When a job runs on a NERSC MPP system charges accrue against one of the user's repository allocations. Read More »
