Changes between Version 3 and Version 4 of schwarz


Ignore:
Timestamp:
07/12/19 14:37:58 (4 years ago)
Author:
snguyen
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • schwarz

    v3 v4  
    99Get the NEMO Schwarz sources from the repository: 
    1010{{{ 
    11 svn co svn+ssh://forge/ipsl/forge/projets/cocoa/svn/ 
     11svn co svn+ssh://forge/ipsl/forge/projets/cocoa/svn/SCHWARZ_IPSLCM6VLRNOP/SOURCES/NEMO MY_NEMO 
    1212}}} 
     13 
     14All the needed files to modify NEMO for the Schwarz algorithm are in this directory: `MY_NEMO`. 
     15 
     16The rest of the work consist in checking the subroutines in `MY_NEMO` vs those in `SOURCES/NEMO`, `NEMOGCM/CONFIG/ORCA2_LIM3_PISCES/MY_SRC` or `NEMOGCM/NEMO/OPA_SRC`. 
     17 
     18Review each `.f90` in `MY_NEMO`:  
     191. If it is present in `SOURCES/NEMO` do a vimdiff between the two and make the needed changes to the file in `SOURCES/NEMO` to include Schwarz modifications. 
     202. If it is present in `NEMOGCM/CONFIG/ORCA2_LIM3_PISCES/MY_SRC` but not in `SOURCES/NEMO` copy the file to `SOURCES/NEMO` do a vimdiff with the `MY_NEMO` file and include Schwarz modifications. 
     213. If it is neither in `ORCA2_LIM3_PISCES/MY_SRC` nor in `SOURCES/NEMO` copy the file from `NEMOGCM/NEMO/OPA_SRC` review the changes in `MY_NEMO` with vimdiff and include them in the `SOURCES/NEMO` file. 
     22 
     23When you have an updated version of every `.f90` files from `MY_NEMO` in `SOURCES/NEMO` you can recompile NEMO. 
     24 
     25You will still have to modify the NEMO namelist: `namelist_ORCA2_cfg`. 
     26 
     27In your submission directory (where you have `config.card` and `Job_MYEXPERIMENT` files), go to `PARAM` and modify `namelist_ORCA2_cfg`. 
     28 
     29Add `mswr`, `ksout`, `kall` and `cplfrq` to the `&namrun` section: 
     30{{{ 
     31   mswr        =  010   !  number of scwharz iterations 
     32   ksout       =  010   !  Schwarz iteration for output 
     33   kall        =  001   !  first iteration with all physics 
     34   ncplfrq     =  86400 !  coupling frequency in seconds 
     35}}} 
     36 
     37You should try to run a simulation with `mswr=1` `ksout=1`, `kall=1` and `ncplfrq=` the coupling frequency of your reference simulation as a test. 
     38 
     39It should give you the same result as without Schwarz. 
     40 
     41== updating LMDZ == 
     42 
     43Go to