To come back to the upper stage

Developments for the year 2011/2012

OASIS4
Personal documentation of all the problems encountered.
Personal documentation of OASIS4 on the web (http://www.cerfacs.fr/oa4web/oasis4_dev/oasis4_index.html)
Put all the CICLE interpolations in the test_interpolations environment to test I and IJ partitioning for BT42 and Orca2T
(see results on the web : http://www.cerfacs.fr/oa4web/projet_cicle_devr3219/RESULTS_BASSE_RESOLUTION_PARALLELE3PROCS/projet_cicle.html).
Merge of the trunk and the development branch development_4 (branch containing the multiblock developments and the gaussian reduced grids developments) retesting all the low resolution interpolations on the trunk (https://oasistrac.cerfacs.fr/wiki/cicledev15) and all the examples on the trunk (using  Buildbot).
Creation of the of the branch development_5 (for modifications of the gestion of the I/O : remove of MPP_IO).
Write a paragraph on Buildbot for the deliverable of OASIS4 number 7.2 in the frame of the European project IS-ENES, July 2011.
Interactions with Mike Hobson of the Met Office on the user-defined and debug oasis4 for multiple transients.

Compilation of OASIS4 and run of toys on CURIE (Prace project).
Compilation of OASIS4 and run of toys on JADE at CINES.

Tests with files of weights for OASIS4 (userdef) :
Creation of files of weights for interpolation with OASIS4 from files created by OASIS3 :

• creation of the files grids.nc, masks.nc, areas.nc for OASIS3 with the grids orca and arpegered at low and high resolution using NCL (on a PC Linux and on Corail)
  
• creation of all the files of weights thanks to OASIS3 for the interpolations that will be tested with OASIS4 at low and high resolution (with an adaptation of the the environment testNONE of OASIS3)
• adaptation of the files with the weights of OASIS3 to OASIS4, removing links equal to 0 and weights equal to 0
• test of different partitions for bilinear and conservative interpolations on CORAIL and CURIE
  
• creation of web pages with the results obtained (see http://www.cerfacs.fr/oa4web/projet_test_weights_dev/RESULTS_MONO/projet_weights.html, see http://www.cerfacs.fr/oa4web/projet_test_weights_dev/RESULTS_PARA/projet_weights.html)
• adaptation of the toy to run on massivelly parallel computer : read only the local grid to each processe and not the entire grid (there was a problem of increasing memory)
  

Use of Redmine to manage project OASIS4 -Userdefined to reference bugs (see https://inle.cerfacs.fr/projects)

Reference toys on Builbot :
Test and debug of toyoa4, create_restart, gridless, user3d-auto, toy_echam_mpiom_partition_multiblock on different platforms with Buildbot every night, using and adapting the scripts already existing for all the examples (see http://memphis.cerfacs.fr:8011/waterfall) :

• change the structure of the repositories of the different toys for homogenization
• add the examples to the scripts used by Buildbot
• update all scripts for the trunk version
  

Test all the low resolution interpolations (old project CICLE) on Linux every night (see http://memphis.cerfacs.fr:8013/waterfall).

OASIS3
Personal documentation of all the problems encountered.
Creation of the doxygen files for the oasis3 code (http://www.cerfacs.fr/oa4web/oasis3/doc-doxygen/html/index.html), update of the sources of oasis3, of the documentation every day with a cron.
Creation of personal documentation on OASIS3 (http://www.cerfacs.fr/oa4web/oasis3/personal_html_doc/oasis3_index.html)
Use of Redmine to manage projects (OASIS3, OASIS3-MCT) to reference bugs, developments (see https://inle.cerfacs.fr/projects)
Clean and update the OASIS web site (https://verc.enes.org/models/software-tools/oasis)

Test and debug a high resolution tutorial HR_tutorial on CORAIL and a PC Linux machine and check in it on SVN.
Verify and test (debug) what did Li Yan (February 2011 - February 2012) on the conservative interpolation using the true areas given in areas.nc:

• Creation of a toy  test_interp_conservative_truearea to calculate the interpolated field using OASIS3 in interpolation mode, the error of the interpolation and also the global integral of the initial field and the interpolated field (field*surface)
  
• Test the conservation of the global integral on simple grids with shape that is  varying (and so are the areas)
• Open and update corresponding ticket on redmine
  

Interactions with Marie Piazza, a PhD Student of Globc, on OASIS3 to interpolate her fields on the computer Napali at CERFACS (compile OASIS3 on Napali, test interpolation for one field).
Interactions with David Byrne from ETHZ in Switzerland working on the coupling of the Roms and Cosmo  models with OASIS3 : compilation of OASIS3 with pgi+openmpi and running of some toys on my computer.
Interactions with Baoqiang Xiang form the University of Hawai on the conservative interpolation : use the environment testNONE, which allows to interpolate files of data between different grids, to test the quality of the conservative interpolation with its grids.
Interactions with R. Doescher from SMHI who wanted to perform a simulation of 10 years. After different tests, it is not possible without making deep modifications in the code.
Modification of OASIS3 to have points at 1E20 for bilinear and bicubic interpolation for non masked points when no extra search is asked for the target points that have only masked source neighbour points

Calculation of the remapping files for the conservative and bilinear interpolations for t799-Orca025 (see https://inle.cerfacs.fr/projects/oasis3/wiki/Page_on_OASIS_User_Support)

• Creation of the grids
• Creation of an environment tests_oasis3_interpolations (close to the one developped for test_interp_conservative_truearea) to do interpolations with OASIS3 when remapping files are present or not, and also when remapping files come from ESMF.
  

ESMF
Download, compile and install ESMF on CORAIL to test this library of interpolation and compare it to SCRIP for conservative interpolation (ESMF is much faster than SCRIP)

• Creation of an environment to create the source and target grids for ESMF from grids.nc, masks.nc and areas.nc (files of OASIS3) and rename the variables of  the remapping file coming from ESMF to OASIS3 format
  
• Creation of an environment to execute ESMF on Corail
• Compare results obtained with OASIS3 and ESMF for conservative interpolation
• Run ESMF in parallel

OASIS3-MCT
17 January 2012 : meeting MORCE at Jussieu LMD, Paris
12-13 April 2012 : meeting at CERFACS, Toulouse with Joel Chavas and Edouard Audit from the "Maison de la Simulation" at CEA Paris about conservative interpolation in OASIS3-MCT
Interactions witth  J. Freeman (BOM): compilation of OASIS3-MCT with intel+openmpi and running of some toys on my computer.
Work with Eric Maisonnave on the couplde model WRF/OASIS3-MCT/NEMO:

• Creation of a toy toy_eric_pulsation to reproduce the coupled model that is used in the PULSATION project: WRF/OASIS3-MCT/NEMO. Compile and run it on my Linux machine with 1 proc for each model and on CURIE with 4000 procs for each model.
  
• Modification of the sources of oasis3-mct because to many files are written during the run and this is a problem when the number of processors is very high as on CURIE.

DATA OF GLOBC
Cleaning and update of Reanalysis and Observation of the Globc team:
- Reanalysis ERAI, NOAA_20CR, NCEP:

• Creation of a generic README for each reanalysis explaining the data downloaded, the processing done on them.
• Creation of the scripts to download the data if necessary.
  
• Creation of the scripts to process the data downloaded (monthly and daily), with CDO and NCO.
  

Trainings :
28 september 2011 morning : presentation of DDT and OPT by Allinea at CERFACS
20 January 2012 morning : training on the software GIT at CERFACS
Supervisor during the training on OASIS3 at CERFACS in Toulouse, 7-8-9 September 2011, write a summary of these days.
Organisation of the training on OASIS3 at CERFACS in Toulouse, 30-31 May 2012 and 1 June 2012. Supervisor during the training.

Administrative :
Management of the internal seminars SIESTES and put the presentations on the web (see http://www.cerfacs.fr/3-25803-Presentation.php ).
Management of the web where are located other internal shorter scientific presentations (CLUCS) of the group (see http://shou.cerfacs.fr/globc/index.html )