Lines of Code
[root]/trunk/libf/dynlonlat_phylonlat/phylmd
| Author | Changes | Lines of Code | Lines per Change |
|---|---|---|---|
| Totals | 46 (100.0%) | 3786 (100.0%) | 82.3 |
| emillour | 32 (69.6%) | 3190 (84.3%) | 99.6 |
| dcugnet | 6 (13.0%) | 410 (10.8%) | 68.3 |
| lguez | 5 (10.9%) | 149 (3.9%) | 29.8 |
| acozic | 1 (2.2%) | 33 (0.9%) | 33.0 |
| fairhead | 2 (4.3%) | 4 (0.1%) | 2.0 |
Change dynamics/physics interface directory name from dynlonlat_phylonlat to dynphy_lonlat.
EM
0 lines of code changed in 8 files:
Follow-up from commit 2395: get rid of rlon and rlat, longitude_deg and latitude_deg (from module geometry_mod) should be used instead. Longitudes and latitudes are no longer loaded from startphy.nc but inherited from dynamics (and compatibility with values in startphy.nc is checked). This will change bench results because of roundoffs differences between the two.
EM
17 lines of code changed in 3 files:
Modifications from 2274 that were somehow lost
LF
2 lines of code changed in 1 file:
Change call to inca initialisation to fit with new sections dynamique/physic
33 lines of code changed in 1 file:
In etat0dyn: removed few useless lines: "masque" is always known because etat0dyn is called after etat0phys.
In grid_atob: shape error in routine fine2coarse now fixed: "msk" argument and local variable mask must have the dimensions of the output grid. Working unit of dist_sphe is no longer degree, but radian.
1 lines of code changed in 1 file:
Bug fix on bug introduced in previous bug fix (klon_glo is not set before call to init_physics_distribution, but has a correct value after).
EM
1 lines of code changed in 1 file:
Corrections for the 1D model to fix initializations broken with previous revision.
EM+MPL
8 lines of code changed in 1 file:
Correction for ce0l to compile and work in serial mode.
EM
5 lines of code changed in 1 file:
More on physics/dynamics separation and cleanup:
- Set things up so that all physics-related initializations are done via iniphysiq.
- Created a "geometry_mod.F90" module in phy_common to store information on the loacl grid (i.e. replaces comgeomphy) and moreover give these variables more obvious names (e.g.: rlond => longitude, rlatd => latitude, airephy => cell_area).
- removed obsolete comgeomphy.h and comgeomphy.F90
EM
135 lines of code changed in 2 files:
Fix call to iniphysiq in ce0l.
EM
2 lines of code changed in 1 file:
Make iniphysiq a module.
Fix call to iniphysiq in lmdz1d (missing arguments and arrays of wrong sizes).
EM
225 lines of code changed in 1 file:
Physics/dynamics separation:
- remove all references to dimensions.h from physics. nbp_lon (==iim) , nbp_lat (==jjm+1) and nbp_lev (==llm) from mod_grid_phy_lmdz should be used instead.
- added module regular_lonlat_mod in phy_common to store information about the global (lon-lat) grid cell boundaries and centers.
EM
34 lines of code changed in 1 file:
Physics/dynamics separation:
- move test_disvert_m to dynlonlat_phylonlat/phylmd since it is only used by ce0l and relies on dynamics.
- put "config_inca" in tracinca_mod so physics routines can get the info from there rather than from control_mod.
- get rid of references to "control_mod" from within the physics.
EM
75 lines of code changed in 1 file:
Another step towards a clean separation between physics and dynamics: adapted read_pstoke.F90, read_pstoke0.F90 initphysto.F90 and phystokenc.F90 (now module phystokenc_mod.F90) to not explicitely include/use file/modules from the dynamics.
In the process, added module "time_phylmdz_mod.F90" in the physics, which contains the information otherwise found in "temps.h" (which is in the dynamics) and should be used instead.
EM
11 lines of code changed in 1 file:
Follow-up for a cleaner separation between dynamics and physics:
- "write_field" is called from physics and dynamics but has no dependence on either so it should be in "misc"
- "write_field_phy" is common to all physics, so it goes in "phy_common"
- "init_ssrf_m" and "limit_netcdf" are only used by ce0l, so these should be in "dynlonlat_phylonlat/phylmd"
- "q_sat" is called from physics and dynamics but has no dependence on either so it should be in "misc"
EM
980 lines of code changed in 2 files:
2 lines of code changed in 1 file:
* Add parallel capability for ce0l.
* Small bug in grid_noro fixed (smoothed topography was used instead of unsmoothed one for geopotential computation at north pole).
* Removed average of mass at poles in etat0dyn_netcdf after start_init_dyn => different results in the zoomed grid case.
* ok_etat0=n and ok_limit=y combination now works fine (if no initial state is needed, but only limit.nc file). This required:
- to move grid_noro0 and start_init_noro0 subroutines from etat0dyn_netcdf.F90 to limit_netcdf.F90
- to create init_ssrf_m.F90 file, so that sub-surfaces can be initialized from limit_netcdf.F90 without any etat0*_netcdf routines call).
* Simplified somehow the corresponding code, in particular:
1) removed obsolete flags "oldice".
2) removed flag "ibar": barycentric interpolation is used everywhere (except in start_init_subsurf, still calling grille_m - to be changed soon).
3) removed useless CPP_PHY precompilation directives, considering the possibility to run ce0l without physics is useless (ce0l is dedicated to Earth physics).
251 lines of code changed in 3 files:
Bug fix. maase, pbaru and pbarv cannot be intent(in) in caldyn0
because they are modified in caldyn0 (masse is computed by massdair
and pbaru, pbarv are computed by flumass). Also masse is already
computed in etat0dyn_netcdf, and the difference with the computation
in caldyn0 is the averaging at the poles. If we do not want to lose
this averaging, we should not overwrite masse with an output argument
from caldyn0. So the simplest thing to do is to downgrade masse from
argument to local variable of caldyn0. We also simply downgrade pbaru,
pbarv from arguments to local variables of caldyn0.
2 lines of code changed in 1 file:
Fixed regression from revision 2315: comvert.h was replaced by
vertical_layers_mod in test_disvert, but variables ap, bp, preff of
vertical_layers_mod were not defined. So, in main program ce0l, moved
call to test_disvert after call to Init_Phys_lmdz, and inserted in
between them calls to infotrac_init and iniphysiq (required). Had then
to remove the call to infotrac_init in etat0dyn_netcdf. In main
program ce0l, had to remove the call to InitComgeomphy since this is
done in iniphysiq.
In main program ce0l: no need to use indice_sol_mod; removed
preprocessor tests on CPP_PHYS in ce0l.
133 lines of code changed in 3 files:
Physics/dynamics separation: make an infotrac_phy module, which should be used from within the physics, and is initialized from infotrac (dynamics) via iniphysiq.
EM
21 lines of code changed in 2 files:
More on physics/dynamics separation: make a vertical_layers_mod module to contain information on the vertical discretization. This module should be used from within the physics (instead of including comvert.h from dynamics).
EM
29 lines of code changed in 1 file:
Further modifications to enforce physics/dynamics separation:
- moved iniprint.h and misc_mod back to dyn3d_common, as these should only be used by dynamics.
- created print_control_mod in the physics to store flags prt_level, lunout, debug to be local to physics (should be used rather than iniprint.h)
- created abort_physic.F90 , which does the same job as abort_gcm() did, but should be used instead when in physics.
- reactivated inifis (turned it into a module, inifis_mod.F90) to initialize physical constants and print_control_mod flags.
EM
5 lines of code changed in 1 file:
Move etat0phys_netcdf.F90 to "dynlonlat_phylonlat/phylmd" as it relies on "phylmd" routines.
Some cleanup to remove obsolete and unecessary CPP_EARTH preprocessing condition.
EM
545 lines of code changed in 2 files:
Moved ioipsl_getin_p_mod.F90 to "phylmd" beacause files in "misc" should rely on neither physics nor dynamics modules (or include) files.
Moved etat0dyn_netcdf.F90 to "dynlonlat_phylonlat/phylmd" as it is only used by ce0l and moreover requires "grid_atob_m" which is in "dynlonlat_phylonlat".
With these adjustements, compiling with '-nophys' works again.
EM
770 lines of code changed in 1 file:
Initial states creation routines have been reorganized and simplified.
As far as possible, dynamics and physics related routines have been
separated.
Some routines have been converted to fortran 90 and repeated codes sections
have been "factorized".
Array/vector arguments have become implicit in some routines to avoid usage
of "dimensions.h" ; possible for routines with explicit interfaces and if
iim and jjm can be deduced from arguments sizes.
* dynlonlat_phylonlat/ce0l.F90 calls now phylmd/etat0phys_netcdf.F90 and
dyn3d/etat0dyn_netcdf.F90 that replace phylmd/etat0_netcdf.F90.
start.nc and startphy.nc creations are now independant.
* startvar.F90 has been suppressed ; corresponding operations have been
simplified and embedded in etat0*_netcdf.F90 routines as internal
procedures.
* Routines converted to fortran 90 and "factorized":
- dyn3d_common/conf_dat_m.F90 (replaces dyn3d_common/conf_dat2d.F
and dyn3d_common/conf_dat3d.F)
- dyn3d/dynredem.F90 (replaces dyn3d/dynredem.F)
- dyn3d/dynetat0.F90 (replaces dyn3d/dynetat0.F)
- phylmd/grid_noro_m.F90 (replaces dyn3d_common/grid_noro.F)
- dynlonlat_phylonlat/grid_atob_m.F90 (replaces dyn3d_common/grid_atob.F)
- dyn3d_common/caldyn0.F90 (replaces dyn3d_common/caldyn0.F)
- dyn3d_common/covcont.F90 (replaces dyn3d_common/covcont.F)
- dyn3d_common/pression.F90 (replaces dyn3d_common/pression.F)
* phylmd/phyredem.F90 and phylmd/limit_netcdf.F90 have been slightly
factorized.
TO DO:
* little fix needed in grid_noro_m.F90 ; untouched yet to ensure results
are exactly the same as before. Unsmoothed orography is used to compute
"zphi", but smoothed (should be unsmoothed) one is used at poles.
* add the dyn3dmem versions of dynredem.F90 and dynetat0.F90
(dynredem_loc.F90 and dynetat0_loc.F90, untested yet).
* test compilation in parallel mode for a single processor.
156 lines of code changed in 1 file:
Ooooops
Correction bug introduit dans r2225 qui menait au plantage du routage dans ORCHIDEE
----------------------------------------------------------------
Correcting bug introduced in r2225 that lead to ORCHIDEE crashing in initialisation
of the routing
2 lines of code changed in 1 file:
Correcting an imbroglio from revision 2247. Directory dynlmdz_phylmd
does not exist any longer, and ce0l.F90 should have been modified in
directory dynlonlat_phylonlat/phylmd.
14 lines of code changed in 1 file:
Further reorganizing of the physics-dynamics interface.
Things related to dynamics grid and physics grid go in "dynlonlat_phylonlat".
Any physics phy*** related stuff (e.g. iniphysiq, ce0l) goes in a phy*** subdirectory of dynlonlat_phylonlat.
EM
327 lines of code changed in 2 files: