wiki:General

General information on running the model


1. Arborescence

2. First runs

2.1. Run a simple simulation

Note that the installation of iloveclim is detailed here: http://forge.ipsl.jussieu.fr/ludus/wiki/basicInstall. This should be done before running a simulation.

  • To get the help and details on the different options:
    ./bin/run-iloveclim -h
    
  • Interactive run: this will create the proper directories, compile and run.

Exmple for a default pre-industrial:

./bin/run-iloveclim -l PI-test -n 10 -r 10 -k -v 1

-n number of years of the simulation (here 10 years)
-r frequency of the restarts (here written every 10 years)
-k to remove the old files
-v to write info

2.2. For developers

You can create the directory in your work directory and copy all the files without compling nor running with the -z option:
Exemple:

./bin/run-iloveclim -l PI-test -n 10 -r 10 -z -k -v 1

To compile and run you then have to go to the work directory.
First to compile:

cd wkdir/PI-test_ecbilt_clio/compil/
make

Then go to the run directory, create a soft link and run:

cd ../run/ 
ln -s ../compil/emic.x .
./emic.x 

2.3. Run with a job (for example for long runs at LSCE)

Look at the examples of jobs given in the iloveclim directory such as:
example_ppn08-scratchspace-migration.job: to run in parallel with 8 processors and with synchronised files
Submit your job, for example at LSCE using the batch system:

qsub example_ppn08-scratchspace-migration.job

To check your job

qstat -u login

To delete your job

qdel number_of_simulation

3. Useful directories and files

  • wkdir/

This is your working directory with two directories:

wkdir/name_of_your_simulation/compil/sources contains all the source files of your simulation
wkdir/name_of_your_simulation/run/outputdata contains all the data outputs

  • data/

At the end of your simulation, all the outputs can be found in data/name_of_your_simulation/outputTIME
While the simulation is still running, some outputs can be found in data/name_of_your_simulation/synced_files if you used a job allowing a synchronisation

  • scenario/

You can use scenario directories to run with specific inputs, parameters or sources (rather than the default ones). Specify in your command line :

./bin/run-iloveclim -l PI-test -n 10 -r 10 -I scenario/scenario_name/inputdata -P scenario/scenario_name/parameters -F scenario/scenario_name/sources -z -k -v 1
  • icTIME/ in data/ or wkdir/name_of_your_simulation/run/restartdata/resTIME/

To run your simulation with a specific restart (rather than the default one), specify in your command line :

./bin/loveclim-new-run -l PI-test -n 10 -r 10 -s TIME -S /home/path/to/iloveclim/data/name_of_restart_simulation/icTIME -z -k -v 1

Make sure than the TIME number in -s TIME does not start with 0.

4. Outputs

The outputs can be found in: workdir/NameOfYourSimulation_ecbilt_clio/run/outputdata or data/NameOfYourSimulation_ecbilt_clio/outputXXX.

Each composante of the model has its own output directory:

  • global for global variables and info:

bookXXX is a text file with: year day global mean temperature (degrees)

  • ocean for the ocean:

CLIO3_test0_ann_avrg_013910.nc is a netcdf file with 3D temperature, salinity...

  • carbon for the carbon cycle:

C_reservoirs.txt is a text file with carbon contents

  • each clio file has 100 years by default
  • postprocessing: see help on previous page (Tools)
  • run param can be changed
  • netcdf out param to have more variables
Last modified 7 months ago Last modified on 03/15/21 11:34:22

Attachments (1)

Download all attachments as: .zip