Changes in ECLIS with V5.8
Article mis en ligne le 31 octobre 2012
dernière modification le 19 février 2013

par senesi

V5.8 introduces Surfex V7 and Offline Surfex. It allows to also run old experiments the old way. Reference ECLIS directory is renamed ’eclis’. Yuki is supported. Forcing files size is checked. A user command can be executed at experiment’s run end. And more ....

V5.8 introduces

  • the size of forcing files is checked according to the resolution (tl63 and tl127 are managed, yet) ; this apply only to the forcing files usually taken on the local filesystem ; for some cases (BCOND) , only the filename is cheked w.r.t. resolution ; for Ozone, the variety of cases is not yet handled ; for GHG, size check is irrelevant ; aerosols, and volcanic aerosols, are duly checked
  • support for version 7 of Surfex, which introduces a PGD file. In that case, you have to set LSFXPGD to 1 (if your PGD is in the same directory as your PREP file, and is compressed by gzip) or to tell where your PGD is by setting SFXPGD. The PREP can be gzipped or not, as can also be an explicitly named PGD. A DATA_SFX parameters allows to tell which ECOCLIMAP files should be linked in run directory (default value is fine). See the surfex section of parameters page. See also examples of parameter files in ECLIS/V5.8/testing/tori : param_arp6_sfx7 and param_cpl_sfxv7
  • support for Surfex Offline mode ; the offline binary can be a parallel one ; you can make spinups runs by using SCYC_YEAR1_YEAR2 as value of YEAR_ATM ; see the relevant section ; see also the example of parameter file : ECLIS/V5.8/testing/tori/param_sfx_offline
  • renaming of ECLIS home directory from ’cnrm-cm’ to ’eclis’ ; see ref
  • introducing a symbolic link ’current’, in ECLIS directory ; this link designates the last fully validated version of ECLIS scripts ; it allows you to stay at the forefront of ECLIS development without changing your parameters files ; we try to ensure upward compatibility !
  • the ’cm_setup’ (see ref) content has been re-designed in order to avoid interactions with the environment you may already use for running CNRM-CM, e.g. by sourcing a .cnrm-cmrc file from earlier ECLIS versions ; please take care to source cm_setup AFTER having sourced the older environment ; you will then be able to run both your old and new experiments
  • the installer script now accepts new arguments (useful if you param file does pass their own arguments to the install script)
    • -noask : for avoiding to be asked wether an existing experiment should be re-installed
    • -go : for avoiding to be asked wether to launch the experiment at the end of the (hopefully successful) installation
  • new parameter EXP_END_CMD can hold the value of a command that will be run at the end of the experiment run ; this can be useful to launch the install and run of a another experiment, which installation phase may depend on the last restarts generated by your experiment
  • yuki is now fully supported
  • script cm_files debugged re. migration of files ; it also allows to migrate the files of another user ; see relevant page
  • the ocean model (forced or coupled) can now run without any sea-ice model
  • part of code of plugin cpl_medit was put in plugin oce_medit, for running NemoMed also in forced mode
  • For Nemo runs, IO Server namelist no more mandatory
  • introduces parameters YEAR_PATTERN and MONTH_PATTERN, in order to have full control on the forcing file names (previously, they had to cope with using YYYY and MM, which prevented to have files named like ’’) ; default values of course are YYYY and MM
  • Nemo forced mode can use a non-MPI binary