Difference between revisions of "CESM"
m |
|||
Line 9: | Line 9: | ||
=== Getting Started === | === Getting Started === | ||
==== SLURM | ==== CESM example using a SLURM interactive session ==== | ||
<br> | |||
Running CESM requires a more complicated set up process than the typical software package on the cluster. CESM needs to be rebuilt separately for each job, which will require an interactive session to be started. In this example, we use 8 CPU cores with 8 GB of memory on 1 node for 1 hour. Your session parameters will vary. | |||
<pre style="white-space:pre-wrap; width:45%; border:1px solid lightgrey; background:#E0E0E0; color:black;"> | |||
<pre style="white-space:pre-wrap; width:45%; border:1px solid lightgrey; background:# | |||
srun -p circe --nodes=1 --ntasks-per-node=8 --mem-per-cpu=1024 --time=01:00:00 --pty /bin/bash | srun -p circe --nodes=1 --ntasks-per-node=8 --mem-per-cpu=1024 --time=01:00:00 --pty /bin/bash | ||
</pre> | </pre> | ||
Line 47: | Line 47: | ||
./case.submit | ./case.submit | ||
</pre> | </pre> | ||
==== Troubleshooting ==== | |||
There are multiple log files written during a CESM job, and you will probably need to look through several of them in order to find the root cause. These include the log directory in your $CASE directory, and the output/$CASE directory within your $WORK/cesm or $WORK_BGFS/cesm directory structture. | |||
If it looks like a model error, we strongly recommend checking the CESM discussion forums: [https://bb.cgd.ucar.edu/cesm/ CESM Forums] | |||
Revision as of 20:20, 15 April 2021
Description
From the CESM homepage: CESM is a fully-coupled, community, global climate model that provides state-of-the-art computer simulations of the Earth's past, present, and future climate states.
Version
- 2.2.0
Authorized Users
CIRCE
account holdersRRA
account holdersSC
account holders
Platforms
CIRCE
clusterRRA
clusterSC
cluster
Modules
CESM requires the following module file to run:
apps/cesm/2.2.0
- See Modules for more information.
Running CESM on CIRCE/SC
The CESM user guide is essential to understanding the application and making the most of it. The guide and this page should help you to get started with your simulations. Please refer to the Documentation section for a link to the guide.
- Note on CIRCE: Make sure to run your jobs from your $WORK directory!
- Note: Scripts are provided as examples only. Your SLURM executables, tools, and options may vary from the example below. For help on submitting jobs to the queue, see our SLURM User’s Guide.
Getting Started
CESM example using a SLURM interactive session
Running CESM requires a more complicated set up process than the typical software package on the cluster. CESM needs to be rebuilt separately for each job, which will require an interactive session to be started. In this example, we use 8 CPU cores with 8 GB of memory on 1 node for 1 hour. Your session parameters will vary.
srun -p circe --nodes=1 --ntasks-per-node=8 --mem-per-cpu=1024 --time=01:00:00 --pty /bin/bash
You will then need to load the CESM module to configure your environment.
module purge module load apps/cesm/2.2.0
Move into your intended job directory location in either $WORK or $WORK_BGFS, depending on which partition you started your interactive session on. For this example, we will use $WORK and mach=circe. For BGFS partitions, use $WORK_BGFS and mach=bgfs instead (https://wiki.rc.usf.edu/index.php/SLURM_Partitions). Create a cesm/cases directory path, if not already existing. Once there, you can run the create_newcase script to begin the setup process for a new simulation. We also set $CASE to the name of the job directory.
cd $WORK mkdir -p cesm/cases cd cesm/cases export CASE=B1850.f19_g17.test1 create_newcase -case $CASE --compset B1850 --res f19_g17 --mach=circe cd $CASE
At this point, make any changes to the namelist files, if needed. Then run the case setup and build scripts.
./case.setup ./case.build
Once the build process has completed, use the CESM submit script to run the case.
./case.submit
Troubleshooting
There are multiple log files written during a CESM job, and you will probably need to look through several of them in order to find the root cause. These include the log directory in your $CASE directory, and the output/$CASE directory within your $WORK/cesm or $WORK_BGFS/cesm directory structture.
If it looks like a model error, we strongly recommend checking the CESM discussion forums: CESM Forums
Documentation
Home Page, User Guides, and Manuals
- CESM Home Page:
Benchmarks, Known Tests, Examples, Tutorials, and Other Resources
- CESM Guide and Examples:
More Job Information
See the following for more detailed job submission information:
Reporting Bugs
Report bugs with CESM to the IT Help Desk: rc-help@usf.edu