Setting up the global_settings.py
Contents
Setting up the global_settings.py¶
The file global_settings.py is the central configuration file of the IOW ESM.
This file has to keep its name and its location, which is the folder input in the root directory
on your target machine.
It consists of the following sections.
Specify the machine you are workin on¶
Here you have to specifiy on which machine you are working. This section might look like:
####################################################
# Global settings for the IOW-ESM model run #
####################################################
##################################
# STEP 0: specify the machine #
##################################
machine = "hlrn" # this will ensure that the correct MPI variant is used and the correct queueing system if present
This ensures that MPI (Intel-MPI or OpenMPI) and the queueing system (if present) are correctly used.
Currently available machine keywords are
hlrnfor the two HLRN clusters in Göttingen and Berlinhaumeafor the Uni Rostock’s cluster
According to your setting one of the scripts/run/machine_settings_*.py is loaded.
If you want to overwrite some of the predefined settings you can set the variable explicitely in the global settings.
Modeller’s information¶
The modeller’s section contains your personal information. Some of these information can be found later on in some output files, such that your work can be related to you.
This section might look like:
###################################
# STEP 1: Info about the modeller #
###################################
modeller = "Your Name" # string, name of the modeller who is responsible
email = "your.name@io-warnemuende.de" # string, contact of the responsible modeller
institution = "Leibniz Institute for Baltic Sea Research Warnemunde (IOW)" # string, name of the institute
Run information¶
The next sections determines the run you want to perform.
Especially important is the run_name variable.
This determines the folder, where you will find your results in the output folder in the root directory
on your target machine.
The hotstart files of your run are also stored folder named run_name in the hotstart folder in the root directory
on your target machine.
Another important configuration are the init_date and final_date variables.
They determine the running period and they must be in the “YYYYMMDD” format, e.g. “20220324” for the 24th of March in 2022.
The model will be cold-started at the init_date if there are no hotstart files present.
(If there are hotstart files, the model will start from the last common hotstart date for all models.)
The model will then run until it reaches the final_date unless it crashes in the meantime.
The debug_mode determines which executables are used.
For each model there is a debug and a production (or release) version available.
Usually you use here False which refers to the production executables.
This section in the global_settings.py might look like:
###############################
# STEP 2: Info about the run #
###############################
run_name = "TEST_RUN" # string without space, name of the current run
run_description = "run for a first test" # string, description: what is this run good for?
init_date = "19800101" # string, date when model is/was cold-started (YYYYMMDD)
final_date = "20100101" # string, when will the model run finally end? (YYYYMMDD)
debug_mode = False # boolean, whether or not to use executables compiled with debugging options (slow)
Time step information¶
TODO: Add further description here…
#################################################
# STEP 3: Time stepping info #
#################################################
coupling_time_step = 600 # integer, time step when fluxes are calculated and exchanged (s)
run_duration = "1 month" # string, duration of one model run (day/days, month/months, year/years)
runs_per_job = 1 # how many runs will be done in one job script