running older version of CESM (1.3) on Cheyenne @NCAR

06 Feb 2019

Here I kept some notes on how to run an older version of CESM on the new HPC at NCAR (i.e., Cheyenne). I mainly followed the guideline on Google doc, which is organized by Jim Edwards.

I used mainly 4 steps:

1. copy two files to source code folder/scripts/ccsm_utils/Machines

maybe need to use this to download files: guestuser friendly

2. add the following machine description to config_machines.xml

<machine MACH="cheyenne">  
<DESC>NCAR SGI, os is Linux, 36 pes/node, batch system is PBS</DESC>  

3. check & revise a few script that may have a bad syntax

look for the following code (or similar structure) foreach $var qw(a b c),
change it to foreach $var (qw(a b c))
i.e. add () outside qw

The following files are worth checking, and I only found one case in models/drv/bld/build-namelist

4. after create a new case, add #PBS -A XXXXXAccountNumberXXXXX to

Note that, after submit the job, there will be only one active job, which is different from CESM2.0 that has an additional archive job.