wiki:SourceCode

Version 84 (modified by jgipsl, 9 years ago) (diff)

--

Development Activities

Documentation

Source Code

Reference Simulations

Group Activities

Source Code


trunk

The trunk is the main version of ORCHIDEE where all developements are integreted before a tag is created. See the trunk here. New branches will, from now on, be created as a copy from the trunk. See below the main changes in the trunk. All modifications that change the results should be mentioned in this list. The complete list of all modifications at the trunk can be seen in the svn log. Follow also modifications on the offline configurations trunk/ORCHIDEE_OL.

News :

  • While compiling LMDZ with the trunk ORCHIDEE, preprocessor flag ORCHIDEE_NOOPENMP must be removed from main makefile of LMDZ, see below in section download.
  • Now compiling is done in modeles/ORCHIDEE using "gmake driver". See below how to extract and compile the trunk
  • The trunk can not use tagged version v2_2_1 or earlier of IOIPSL(since rev 511 ?). Revision 2020 or later of IOIPSL/trunk/src must be used.

tag

Different tags of ORCHIDEE are done as a copy of the trunk at a specific moment. See here all existing tags : SourceCode/Tags

Validation

Step 1: techincal regression tests

The trusting is a bench of tests done at nearly each new revision of ORCHIDEE trunk. All simulations are compared with previous version of ORCHIDEE. A set of tests coupled and offline are preformed at curie/TGCC, ada/IDRIS and obelix/LSCE.

See results here http://webservices.ipsl.fr/trusting (click on tab TGCC or IDRIS or LSCE and the underlying tab ORCHIDEE)

Step 2: more technical tests

More tests are performed on selected revision on the trunk of ORCHIDEE. See here test done on trunk revision

Step 3 : simulations

For a scientific validation, a set of simulations are done on some revision on the trunk. See the page ReferenceSimulations or go directly to the summary of simulations done:

Main changes on the trunk

  • [2580] : Correction in routing_main related to restart writing. See ticket #169. Change in results.
  • [2575] : Now read DT_SECHIBA from run.def instead of SPLIT_DT. ROUTING_TIMESTEP changed name to DT_ROUTING. DT_TURNOVER_TIME parameter was removed.
  • [2501] : Correct bug existing since 2441, in initialization of variables depending on t2m in stomate, see ticket #158
  • [2481] : Changed netcdf library at obelix, see ticket #160
  • [2452] : Corrected bug introduced in 2441
  • [2441] : Introduced 2 bugs, resolved in 2452 and 2501. New default value REFTEMP_FILE=NONE : t2m_longterm is calculated instead of read rom file, see ticket #158
  • [2425] : Added beta factor, important for coupled simulations, see ticket #157
  • [2422] : Bug correction for daily forcing, see ticket #156
  • [2419] : Corrected case SOILTYPE_CLASSIF=usda, removed case fao
  • [2360], [2361] : See ticket #136 : Limitation of snow mass. Might change results
  • [2293] : Homogenization of parameters controling the level of text output (add WRITELEVEL). Updated orchidee.default.
  • [2291] : Add new parameters for optimization: NBUFF=1 and OK_HISTSYNC=n by default
  • [2222] [2223] [2224] : Merge branch ORCHIDEE-MICT into the trunk : Soil freezing and explicit snow model. Revision 2223 contains ORCHIDEE-MICT except a bug correction in hydrol. This revision reproduces exactly the same results as the trunk before the merge. Revision 2224 on the trunk and the branch ORCHIDEE-MICT are now the same.
  • [2183] :
  • [2149] : Compile options and documentation of makeorchidee_fcm has been improved during several commit. Now compile options corresponds to AA_make.gdef.
  • [2093] : Bug correction in enerbil, see ticket #137
  • [2053] : The externalized parameter MAXMASS_GLACIER has changed name. New name MAXMASS_SNOW.
  • [2031] : Merge the Photosynthesis branch primarily based on the work of Yin et al (2009). Some change in results.
  • [1987] : Changed default value for DOWNREGULATAION_CO2 from TRUE to FALSE
  • [1957] : Change formula for snow albedo, ticket #74 . Change in results are excpected. But also introduced a bug. This revision do not compile!! Use rev [1959].
  • [1925] : Correction for OpenMP. Adapted use of XIOS also for OpenMP mode
  • [1914] : Corrected error introduced in rev 1907
  • [1907], [1913] : Do not use. Introduced compile error
  • [1907] : Add possiblity to use XIOS in coupled mode with LMDZ. Note: LMDZ is not yet available to run ORCHIDEE with XIOS.
  • [1882] : Modification of parameters for tropical forests and activation of CO2 downregulation, more details here trunk/DownRegulation
  • [1881] : Correct time axis with XIOS. Initialization in XIOS of calendar and date information directly done by the model(not in xml)
  • [1852] : Corrected bug on evap_bare_lim_ns in hydrol.f90, some change in results for CWRR only, see ticket #121
  • [1843] : Revert back (to rev 1843) arch file for curie using default netcdf
  • [1840] : Correction in xml file. Introduced by mistake modification in arch-X64_CURIE.path (revert in next commit)
  • [1814] : Corrected bug introduced in rev 1788. The bug causes probleme if using gfortran and CWRR.
  • [1791] and rev [1792] : Corrected variable emis in sechiba_out_2.nc output file from IOIPSL(rev 1791) and in output file from XIOS (rev 1792)
  • [1789] : Corrected variable ECanop in sechiba_out_2.nc file, output file from IOIPSL
  • [1788] : Added interface to XIOS
  • [1772] : Corrected bug for variable dss in sechiba_out_2.nc file
  • [1615] : double all OMP_BARRIER to fixe a bug with openmp parallelisation
  • [1383] : Changed default values for vc_max_opt and vj_max_opt parameters. Results change.
  • [1280] : Bug correction for spinup analytic for case forcing cycle less than 7 years.
  • [1263] and rev 1262: Updated teststomate due to commit 1078. Teststomate can now run in sequential and parallel MPI mode.
  • [1229] : Bug fix : forcesoil can now again run in parallele MPI mode.
  • [1224] : Added possibility to read the vertical structure of the forcing data from the forcing file itself. No change in results.
  • [1172] : Added writting of evapot_corr in restart - and initialisation by reading the restart or parameter file - Little change in the results
  • [1170] : Added extra dimension. No change in results.
  • [1102] : Merged new phenology branch into the trunk. Use with caution. Some changes are expected in stomate outputs (tropics especially). Please, look at Branches/ORCHIDEE-NEWPHENO for more details.
  • [1100] : Merged Spinup_analytical branch into the trunk. The analytical spin-up is optional, so no changes in results. See Branches/AccelerationSpinup for more information.
  • [1091] : Correction of the carbon CMOR output C3PftFrac. Changes in stomate IPCC outputs.
  • [1082] : Added externalization for CWRR parameters. Some obsolete parameters were removed and some equations in Choisnel have been simplified. Implies some little changes in the results.
  • [1078] : Added modifications for OpenMP. The interface intersurf_gathered was changed. The interface intersurf_gathered_2m was removed and variables q2m and t2m are integrated in the first interface as optional variables. The new interface was already added in LMDZ but the preprossor key ORCHIDEE_NOOPENMP must be removed to use this new interface. MPI parallelization is guaranteed but OpenMP is still under validation and debugging phase. Compiling and running is still done in mode MPI only. Teststomate has not been adapted to run with these modifications and cannot be used before update.
  • [1042] : Re-organization 1 : moved driver source code from ORCHIDEE_OL into ORCHIDEE/src_driver
  • [1028] : Corrected bug for routing in coupled and offline mode. The bug was introduced in rev 947
  • [949] : Small correction needed to compile at Vargas/IDRIS
  • [947] : Inclusion of the developements from branches/Hydrology => Change in results. The routing does not work in coupled mode as the case was at branche Hydrology. See notes for validation here.
  • [945] : Bug correction due to inverse loop
  • [890] :
    • Start inclusion of COV
    • Modification in initialization of lai if it is not in restart file and for the cas read_lai. Ligne 1051 in slowproc. => Change in results.
    • This revision do not run in coupled mode with currrent LMDZ due to modification in interface.
  • [875] : Created the tag 1.9.6 as a copy of this revision Tag 1.9.6
  • [864] : Specific leaf area (SLA) is prescribed => Change in results.
  • [861] : Correct solarang and time_zone functions. => May change results (not for LMDZ-OR at vargas)
  • Change in diagnostics (history) files in some revisions between 720-740.
  • [511] : Included externalization branch in the trunk => Change in results.
  • [392] : Created the tag 1.9.5.2 as a copy of this revision Tag 1.9.5.2

Download and compile the trunk

Offline

Download the trunk using the script model in modipsl :

> svn co http://forge.ipsl.jussieu.fr/igcmg/svn/modipsl/trunk modipsl
> cd modipsl/util
> ./model ORCHIDEE_trunk
> cd ../modeles/ORCHIDEE
> gmake driver

If using an older version than rev 1042(version without ORCHIDEE/src_driver), then use the script recup_my_ORCHIDEE. This script is stored in modipsl/util :

> svn co http://forge.ipsl.jussieu.fr/igcmg/svn/modipsl/trunk modipsl
> cd modipsl/util
> ./recup_my_ORCHIDEE your_svn_login trunk [optional_REV]

LMDZOR

A new configuration LMDZOR_v5.2 has been created to use ORCHIDEE trunk coupled to LMDZ5. This configuration contains by default the latest revision of the ORCHIDEE trunk. In this set up, also orchidee drivers will be compiled by default. The directory config/ORCHIDEE_OL is also extracted so that both experiences in coupled and offline mode can be performed with the same model sources.

There are no differences in the run set up between LMDZOR_v5 and LMDZOR_v5.2.

> svn co http://forge.ipsl.jussieu.fr/igcmg/svn/modipsl/trunk modipsl
> cd modipsl/util
> ./model LMDZOR_v5.2
> cd ../config/LMDZOR_v5.2
> gmake

LMDZOR and OFFLINE in same modipsl

Use new configuration LMDZOR_v5.2 as desribe above.

You can choose to do expements in offline mode in modipsl/config/ORCHIDEE_OL or in coupled mode in modipsl/config/LMDZOR_v5.2.

Commit on the trunk

Only some people have the permission to commit on the trunk. See here some guidelines for commit on the trunk?.
These guidelines can also be applied to branches.