CAMM‎ > ‎Resources‎ > ‎Hardware‎ > ‎

NERSC

The National Energy Research Scientific Computing Center (NERSC) is the primary scientific computing facility for the Office of Science in the U.S. Department of Energy.

Extensive documentation to get you started and every detail is available at https://www.nersc.gov/users/

CAMM users have access to the following clusters under project m1503:

Barebone profile files & Setting you default shell

What follows are instructions to get you started with the BASH shell. NERSC IT provides readable-only .bash_profile and .bashrc files. Users are required to create their own customization files .bash_profile.ext and .bashrc.ext. Here are files to get you started:

If you rather use your own files, please add the following two lines to your .bashrc.ext file in order to be able to load software compiled for CAMM users.

export MODULEPATH=$MODULEPATH:/project/projectdirs/m1503/camm/${NERSC_HOST}/modulefiles
module load camm

In addition, you must set you default login shell as BASH. Follow these instructions:
  1. Log in to https://nim.nersc.gov/loginform.phtml with your NIM usename/password.
  2. Select "change shell" under "Actions" pull-down menu
  3. For each machine that you plan on login in, select "bash" as the Login Shell, then click on the "Save" button.

Using compiled software in PBS scripts

To invoque a particular piece of compiled software in a PBS script, insert a call to the appropriate module file:
source $MODULESHOME/init/bash
module load desiredApplication # loads executables and environment variables for 'desiredApplication'

Here are the first lines of a typical PBS script:
#!/bin/bash
#PBS -N test                                                                                                 
#PBS -j oe                                                                                                   
#PBS -l walltime=3:30:00                                                                                    
#PBS -l mppwidth=48                                                                                          
#PBS -l mppnppn=24                                                                                           
#PBS -l mppdepth=1                                                                                           

module load epw

OMP_NUM_THREADS=1
NPROC=48
NPOOL=16
set APRUN  = "aprun -n $NPROC"


Consult the manual for the meaning of the different PBS directives and how to submit jobs to each machine:

Submit a job dependent on another

ID=`qsub my_first_job.pbs`
qsub -W depend=afterok:$ID my_dependent_job.pbs

Command to show unallocated nodes

This is Useful if you can run your job accommodating with available resources
edison09:epw$ showbf
Partition     Tasks  Nodes      Duration   StartOffset       StartDate
---------     -----  -----  ------------  ------------  --------------
ALL            1471     23      INFINITY      00:00:00  08:56:36_02/16
login          1471     23      INFINITY      00:00:00  08:56:36_02/16

Subpages (1): Edison