| 3 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 4 |
============================== |
============================== |
| 5 |
|
|
| 6 |
|
|
| 7 |
|
untagged |
| 8 |
|
o some bug fixes for #undef REAL4_IS_SLOW |
| 9 |
|
|
| 10 |
|
checkpoint52 = ecco_c52_e35 |
| 11 |
|
o after testing adjoints of checkpoint51u_post |
| 12 |
|
|
| 13 |
|
checkpoint51u_post |
| 14 |
|
o this is the post-merge tag of the ecco-branch |
| 15 |
|
(needs another clean checkout and re-testing) |
| 16 |
|
o various natl_box_adjoint configs and results |
| 17 |
|
but no input fields |
| 18 |
|
o obcs adjoint not yet merged! |
| 19 |
|
o adding ECCO global 1x1 and 2x2 degree configs |
| 20 |
|
o updated/extended some store directives for seaice |
| 21 |
|
o pkg/ecco now containes ecco-specific part of cost function |
| 22 |
|
o top level routines the_main_loop, forward_step |
| 23 |
|
supersede those in model/src/ |
| 24 |
|
previous input data.cost now in data.ecco |
| 25 |
|
(new namelist ecco_cost_nml) |
| 26 |
|
o in ctrl/ cleaned some cross-dependencies and updated CPP options |
| 27 |
|
|
| 28 |
|
checkpoint51t_post |
| 29 |
|
o this is the pre-merge tag of the ecco-branch |
| 30 |
|
|
| 31 |
|
checkpoint51s_post |
| 32 |
|
o cleanup: removed various obsolete #define-s |
| 33 |
|
o pkg/grdchk |
| 34 |
|
- renamed 'ph-grd 3' to 'grad-res' |
| 35 |
|
- merged from ecco-branch |
| 36 |
|
(remaining bug fixes for obcs gradient checks) |
| 37 |
|
- additional high-precision output for testreport |
| 38 |
|
(grep for 'precision_grdchk_result') |
| 39 |
|
|
| 40 |
|
checkpoint51r_post |
| 41 |
|
o cleaned up or removed CPP_OPTIONS.h files |
| 42 |
|
o updated results for some of the tests to reflect the new |
| 43 |
|
optfiles (-ieee: -O0 instead of -O3) |
| 44 |
|
|
| 45 |
|
checkpoint51q_post |
| 46 |
|
o cd_code: |
| 47 |
|
- moved cd_scheme.F -> cd_code_scheme.F |
| 48 |
|
- separate read_checkpoint from cd_code_ini_vars.F |
| 49 |
|
- separated cd_code part from write_checkpoint |
| 50 |
|
- updated AD_SOURCE, generated .flow |
| 51 |
|
- added CD_CODE_VARS.h to the_main_loop |
| 52 |
|
- Moving cd_code specific store dir.s to pkg/cd_code |
| 53 |
|
o ctrl package totally restructured |
| 54 |
|
- pack/unpack now optional and decoupled from |
| 55 |
|
xx_/adxx_ I/O |
| 56 |
|
- ctrl_pack/unpack cleaned |
| 57 |
|
(new routines ctrl_init_ctrlvar.F, pkg/ctrl/ctrl_init_wet.F) |
| 58 |
|
- confined inclusion of AD_CONFIG.h to where necessary. |
| 59 |
|
o genmake2: changed AD dependencies from SRCFILES to AD_FILES |
| 60 |
|
|
| 61 |
|
checkpoint51p_post |
| 62 |
|
o create initial cd_code package |
| 63 |
|
- WARNING: the initialization of fields within packages |
| 64 |
|
is broken because INI_FIELDS (where READ_CHECKPOINT is |
| 65 |
|
called) is called *before* PACKAGES_INIT_VARIABLES. |
| 66 |
|
This seems to be backwards since |
| 67 |
|
1) its important to initially zero the entire field including |
| 68 |
|
the "halo" region |
| 69 |
|
2) zeroing needs to be done *before* reading pickup info |
| 70 |
|
and the pickup call might or might not happen in any |
| 71 |
|
particular run |
| 72 |
|
- this needs an overhaul for R2 |
| 73 |
|
o cleanup verification/*/code/GMREDI_OPTIONS.h |
| 74 |
|
o cleanup testreport |
| 75 |
|
|
| 76 |
|
checkpoint51o_post |
| 77 |
|
o bringing AD-related verif's up to date with genmake2 |
| 78 |
|
- carbon/ |
| 79 |
|
- global_ocean.90x40x15/ |
| 80 |
|
- hs.94.1x64x5/ |
| 81 |
|
- (front_relax/ ) |
| 82 |
|
o (more) consistent directory structure for AD-related verif.'s |
| 83 |
|
o cleaning ALLOW_GRADIENT_CHECK -> ALLOW_GRDCHK |
| 84 |
|
o cleaning some ALLOW_TANGENTLINEAR_RUN -> ALLOW_AUTODIFF |
| 85 |
|
o bug fix in find_alpha.F for MDJWF: |
| 86 |
|
- modif. to alpha = 1/D*( dN/dT - rho*dD/Dt) to account for |
| 87 |
|
change rho -> rho-rhoConst |
| 88 |
|
- replace call find_rho to find_rhonum |
| 89 |
|
o initialisation of rFlx extended to full array (required by TAF) |
| 90 |
|
and shifted to thermodynamics |
| 91 |
|
o removed PTRACERS.h in ptracers routine |
| 92 |
|
o added surfacetendencyPtr to S/R parameter list pracers_forcing |
| 93 |
|
o Correcting initialisations for fVerU, fVerV. |
| 94 |
|
o helping TAF to recognize dependency on kUp for |
| 95 |
|
fVerU, fVerV |
| 96 |
|
o Replacing delZ by delR in pkg/ctrl/ |
| 97 |
|
checkpoint51o_pre |
| 98 |
|
|
| 99 |
|
o undid all of cp51 pending some cleanups and discussion |
| 100 |
|
o added the [#include "AD_CONFIG.h"] statement to all files that need |
| 101 |
|
it for adjoint/tl #defines |
| 102 |
|
o re-worked the build logic in genmake2 to support AD_CONFIG.h |
| 103 |
|
o removed tools/genmake since it no longer works |
| 104 |
|
|
| 105 |
|
checkpoint51n_post |
| 106 |
|
o modifications to make FREEZE flux visible to pkg/kpp |
| 107 |
|
- moved surfaceTendencyTice from pkg/seaice to main code |
| 108 |
|
- FREEZE & EXTERNAL_FORCING_SURF moved to FORWARD_STEP |
| 109 |
|
- subroutine FREEZE now limits only surface temperature |
| 110 |
|
(this means new output.txt for global_ocean.90x40x15, |
| 111 |
|
global_ocean.cs32x15, and global_with_exf) |
| 112 |
|
o added surface flux output variables to TIMEAVE_STATVARS |
| 113 |
|
|
| 114 |
|
checkpoint51n_pre |
| 115 |
|
o added "#undef ALLOW_USE_MPI" support to eesupp/src/gather_2d.F |
| 116 |
|
o added [#include "AD_CONFIG.h"] to all files that use the following |
| 117 |
|
defines: |
| 118 |
|
- ALLOW_ADJOINT_RUN |
| 119 |
|
- ALLOW_TANGENTLINEAR_RUN |
| 120 |
|
- ALLOW_ECCO_OPTIMIZATION |
| 121 |
|
|
| 122 |
|
checkpoint51m_post |
| 123 |
|
o incorporating changes to get TAMC/TAF working in genmake2 |
| 124 |
|
- added new TAMC/TAF targets to the Makefile |
| 125 |
|
- various cleanups related to package handling |
| 126 |
|
- these changes should have no impact on the forward model |
| 127 |
|
|
| 128 |
|
checkpoint51l_post |
| 129 |
|
o changes to pkg/cal and pkg/exf that allow and speed-up long integrations |
| 130 |
|
- faster algorithm for cal_addtime.F |
| 131 |
|
- higher precision arithmetic, where needed |
| 132 |
|
- pkg/exf *startdate variables changed from calendar date to |
| 133 |
|
model integration time |
| 134 |
|
|
| 135 |
|
checkpoint51l_pre |
| 136 |
|
o add salt flux to forcing fields (used when salty sea-ice forms or melts) |
| 137 |
|
o enable to apply restoring on S & T only in a latitude band. |
| 138 |
|
o update output (external_forcing_surf modif. change truncation errors) |
| 139 |
|
global_ocean.cs32x15, global_ocean_pressure, lab_sea, natl_box |
| 140 |
|
|
| 141 |
|
checkpoint51k_post |
| 142 |
|
o change default to: useNHMTerms=.FALSE. |
| 143 |
|
- new output.txt for atmos. exp (NHMTerms were wrong but now turned off): |
| 144 |
|
adjustment.128x64x1, aim.5l_Eq_Ch, aim.5l_LatLon, hs94.128x64x5, hs94.1x64x5 |
| 145 |
|
- new output.txt for idealized oceanic exp (NHMTerms are now turned off): |
| 146 |
|
exp1, ideal_2D_oce |
| 147 |
|
- change data file of oceanic exp (leave the results unchanged for now) |
| 148 |
|
exp2, global_ocean[90x40x15,pressure,with_exf], lab_sea, natl_box |
| 149 |
|
o set gravitySign automatically according to the vertical coordinate used |
| 150 |
|
(and change sign: back to "logical" value: +1 when R increases downward) |
| 151 |
|
o added an include barrier to the default CPP_OPTIONS.h and |
| 152 |
|
removed the #include "PACKAGES_CONFIG.h" from it |
| 153 |
|
o fix a bug in mom_fluxform (from chk51j_post) |
| 154 |
|
|
| 155 |
|
checkpoint51j_post |
| 156 |
|
o Added some AD-related initialisations in mom_vecinv/ mom_fluxform/ |
| 157 |
|
o adjusted some flow directives (dynamics, thermodynamics) |
| 158 |
|
o added cubed-sphere case to test cost function |
| 159 |
|
o heckpoint_lev?_directives.h now include |
| 160 |
|
pkg-specific directives, whenever those are enabled. |
| 161 |
|
o Moving pkg-specific flow directives to corresponding pkg |
| 162 |
|
Moving pkg-specific differentiation list to corresponding pkg |
| 163 |
|
Moving pkg-specific checkpoint-lev storing to corresponding pkg |
| 164 |
|
---> conventions are |
| 165 |
|
- pkg_ad.flow |
| 166 |
|
- pkg_ad_diff.list |
| 167 |
|
- pkg_ad_check_lev?_dir.h |
| 168 |
|
|
| 169 |
|
checkpoint51i_post |
| 170 |
|
o merge of the branch-genmake2 |
| 171 |
|
- knowingly breaks the older genmake |
| 172 |
|
- automatic generation of all ALLOW_${PKG_NAME} defines |
| 173 |
|
- new testreport supersedes older testscript |
| 174 |
|
o large number of package inclusion and header cleanups |
| 175 |
|
checkpoint51i_pre |
| 176 |
|
|
| 177 |
|
o new output for adjoin of global ocean vector invariant |
| 178 |
|
o fix for checkpoint of thermodynamic seaice and bulk_force |
| 179 |
|
o Steph gchem various cleanups |
| 180 |
|
|
| 181 |
|
checkpoint 51h_pre |
| 182 |
|
o fixed some single quotes |
| 183 |
|
o modify gchem pkg to be more versatile |
| 184 |
|
o modifications to clean up dic pkg |
| 185 |
|
o add pkg/cfc |
| 186 |
|
|
| 187 |
|
checkpoint51g_post |
| 188 |
|
o provide links for ARPACK and ADM*TLM |
| 189 |
|
o start removing verification/vero/ |
| 190 |
|
o update AD-related stuff |
| 191 |
|
(in particular some IF-statements in model/src/ routines) |
| 192 |
|
o adding to verification/global_ocean.90x40x15/ adjoint of |
| 193 |
|
vector invariant setup |
| 194 |
|
|
| 195 |
|
checkpoint51f_post |
| 196 |
|
o update advect_xz/output.txt (left from chekpoint50f_pre, ini_vel modif) |
| 197 |
|
o Mods and bug fixes to pkg/cal, pkg/exf, etc., needed for computation |
| 198 |
|
of tracer Green's fucntions for ocean inversion project. |
| 199 |
|
|
| 200 |
|
checkpoint51f_pre |
| 201 |
|
o Added on-the-fly spatial interpolation capability |
| 202 |
|
"USE_EXF_INTERPOLATION" to pkg/exf. |
| 203 |
|
- This is a temporary Cartesian-grid hack until |
| 204 |
|
the super-duper ESMF coupler becomes available. |
| 205 |
|
- See verification/global_with_exf/README for usage example. |
| 206 |
|
- Removed obsolete EXFwindOnBgrid and SEAICEwindOnCgrid |
| 207 |
|
flags and modified pkg/seaice accordingly. |
| 208 |
|
o Bug fix to pkg/ptracers, pkg/generic_advdiff/gad_calc_rhs.F, |
| 209 |
|
and pkg/kpp/kpp_transport_ptr.F for dealing with tracer |
| 210 |
|
non-local transport term. |
| 211 |
|
|
| 212 |
|
checkpoint51e_post |
| 213 |
|
o pkg/mom_vecinv: |
| 214 |
|
- changes related to hFac: use S/R from aim.5l_cs/code (coriolis & vort.advec) |
| 215 |
|
- add time-ave diagnostic of vorticity advection. |
| 216 |
|
o add diagnostic (instantaneous) of Shapiro Filter effect for T,S & UV. |
| 217 |
|
o atmospheric geopotential: - include water vapor effect; |
| 218 |
|
- use p^kappa in p* scaling. |
| 219 |
|
o land pkg: read initial state from files. |
| 220 |
|
o AIM : add stratospheric drag in the upper level |
| 221 |
|
|
| 222 |
|
checkpoint51d_post |
| 223 |
|
o change function ILNBLNK (=> last non-blank char.) to return zero |
| 224 |
|
(instead of the string length) when the string. char is empty |
| 225 |
|
o fix bug in dxV,dyU definition of cubed-sphere grid (ini_curvilinear_grid) |
| 226 |
|
o modif's to restore TAMC compatibility |
| 227 |
|
- avoid "_d" constructs in PARAMETER declarations |
| 228 |
|
- avoid interrupting lines by comment lines |
| 229 |
|
o Use cluster local disks for purely local I/O |
| 230 |
|
vs. globally visible disks needed for ctrl stuff |
| 231 |
|
(merged from ecco-branch) |
| 232 |
|
- new active_file_loc... and mdsio_..._loc routines |
| 233 |
|
- affected packages: autodiff, ctrl, grdchk, mdsio |
| 234 |
|
|
| 235 |
|
checkpoint51c_post |
| 236 |
|
o introducing integer flag debugLevel |
| 237 |
|
- debLevZero: suppress all output |
| 238 |
|
- debLevA: standard |
| 239 |
|
- debLevB: equivalent to current debugMode=.TRUE. |
| 240 |
|
o introducing pathname variable mdsioLocalDir for mdsio |
| 241 |
|
that can be specified at runtime |
| 242 |
|
(relevant for cluster I/O to local disk) |
| 243 |
|
` |
| 244 |
|
>>>checkpoint51b_post |
| 245 |
|
adjoint of ptracers |
| 246 |
|
o disentangled ALLOW_PTRACERS using new ALLOW_GCHEM |
| 247 |
|
o extended state in checkpoint_lev?, etc. |
| 248 |
|
o alternative ctrl/cost with tr1 or ptracer |
| 249 |
|
o new comlev1_bibj_ptracers |
| 250 |
|
o new ptracers.flow |
| 251 |
|
o made ptracers_forcing same as external_forcing_s |
| 252 |
|
o added verification for adjoint with ptracers |
| 253 |
|
- adm. grdchk fail so far |
| 254 |
|
- tlm. grdchk OK |
| 255 |
|
- checked tapelev/comlev recomputations: OK |
| 256 |
|
<<<checkpoint51b_pre |
| 257 |
|
|
| 258 |
|
checkpoint51a_post |
| 259 |
|
o added code for biogeochemistry package |
| 260 |
|
- new pkgs: dic, gchem |
| 261 |
|
- new subroutine in ptracers |
| 262 |
|
o modify code for biogeo packages |
| 263 |
|
- in model/src and pkg/ptracers |
| 264 |
|
|
| 265 |
|
checkpoint51 |
| 266 |
|
o merged improved packages from ecco-branch (ecco_c51_e34) to MAIN |
| 267 |
|
- autodiff, ctrl, exf, grdchk, seaice |
| 268 |
|
kept separate packages |
| 269 |
|
- cost, ecco, obcs |
| 270 |
|
o updating verification/internal_wave, verification/lab_sea |
| 271 |
|
(all verif's pass on faulks.lcs.mit.edu RH7.3, g77) |
| 272 |
|
|
| 273 |
|
checkpoint50i_post |
| 274 |
|
o Preparing next differentiable checkpoint and sync |
| 275 |
|
of MAIN vs. ecco-branch |
| 276 |
|
- 1st step: this one is differentiable |
| 277 |
|
- 2nd step: next(! not this) checkpoint will have updated packages |
| 278 |
|
from ecco-branch) |
| 279 |
|
|
| 280 |
|
checkpoint50h_post |
| 281 |
|
o new package "land": provide surface BC for AIM. |
| 282 |
|
o pkg/ptracers: apply exchange to all tracers after reading a pickup. |
| 283 |
|
o AIM: use true log-P extrapolation for near surface temp (= as in |
| 284 |
|
F.Molteni paper) ; update aim.5l_cs output. |
| 285 |
|
o New variable in PARM03 for more flexible pickups |
| 286 |
|
o somewhat cleaned package initialisation sequence for |
| 287 |
|
ctrl/ cost/ ecco/ in packages_... routines |
| 288 |
|
|
| 289 |
|
checkpoint50g_post |
| 290 |
|
o merged with release1_p17 (pkg/seaice and verification/lab_sea) |
| 291 |
|
- added SEAICE_MULTILEVEL for 8-category sea-ice thermodynamics |
| 292 |
|
- LSR sea-ice dynamic solver moved to SouthWest B-grid location and |
| 293 |
|
made the default because of faster convergence than ADI |
| 294 |
|
|
| 295 |
|
checkpoint50f_post |
| 296 |
|
o exf totally restructured |
| 297 |
|
- replaced all exf_set_ routines by one generic exf_set_gen.F |
| 298 |
|
and made sure it's still differentiable |
| 299 |
|
- split exf_getffields.F into new exf_bulkformulae.F |
| 300 |
|
- exf_inscal_/exf_outscal_ suffixes consistent with field names |
| 301 |
|
- updated adjoint/makefile and relevant verif. accordingly |
| 302 |
|
|
| 303 |
|
checkpoint50f_pre |
| 304 |
|
o AIM: near surface temp evaluated from Pot.Temp: fixed for Part Cell ; |
| 305 |
|
+ change Evap over land. (update aim.5l_cs output) |
| 306 |
|
o added missing .cvsignores and made uniform across expts |
| 307 |
|
o changed configuration of tidal_basin_2d expt |
| 308 |
|
o added pkg/debug functions for tracing code in fatal crashes |
| 309 |
|
o tidied pkg/monitor to use MONITOR_OPTIONS.h |
| 310 |
|
o new fn in pkg/monitor to stop model if solution grows too big |
| 311 |
|
o removed lots of single quotes from comments |
| 312 |
|
|
| 313 |
|
checkpoint50e_post |
| 314 |
|
o make KPP work with PTRACERS |
| 315 |
|
- fix gad_calc_rhs to call new routine kpp_transport_ptr, which is |
| 316 |
|
nearly a copy of kpp_transport_s |
| 317 |
|
- there is no analogue to SurfaceTendencyS, so I have to use |
| 318 |
|
gPtr(of the surface layer) instead |
| 319 |
|
o add a new platform SunFire+mpi (SunFire 15000) to genmake |
| 320 |
|
checkpoint50e_pre |
| 321 |
|
|
| 322 |
|
checkpoint50d_post |
| 323 |
|
o change kpp output from multiple-record state files to single-record state |
| 324 |
|
files analogous to write_state.F |
| 325 |
|
o reduce the output frequency of cg3d-related stuff to the monitor frequency, |
| 326 |
|
analogous to the cg2d-related output. |
| 327 |
|
o fix small problem with in ptracers_write_checkpoint.F: len(suff)=512, |
| 328 |
|
so that writing to internal file fn (with length 512) fails. |
| 329 |
|
checkpoint50d_pre |
| 330 |
|
|
| 331 |
|
checkpoint50c_post |
| 332 |
|
o fix problem in ini_procs.F when using MPI (was there since checkpoint48f) |
| 333 |
|
o add monitor diagnostic of vorticity (mon_vort3.F) and potential energy. |
| 334 |
|
o add time-average diagnostic of U*V and vertical diffusive flux for theta. |
| 335 |
|
o bug fixed in convective adjustment diagnostic (multi-tile set-up). |
| 336 |
|
o 4th Order Adv scheme: return to previous (c40pre7) masking near boundaries. |
| 337 |
|
|
| 338 |
|
checkpoint50c_pre |
| 339 |
|
Merging from release1_p13: |
| 340 |
|
o bug fix for pkg/seaice dynamic solver |
| 341 |
|
o Added SEAICE_initialHEFF to pkg/seaice |
| 342 |
|
|
| 343 |
|
checkpoint50b_post |
| 344 |
|
o store u*,v* in gU,V instead of in guNm1,gvNm1 |
| 345 |
|
o in order to put the momForcing out of the Adams-Bashforth: |
| 346 |
|
move forcing & CD-scheme calls from mom_fluxform & mom_vecinv |
| 347 |
|
to timestep.F |
| 348 |
|
o new flag "useCDscheme" (default=F); replace guCD,gvCD by local arrays |
| 349 |
|
o re-generate output for highly sensitive test-exp: |
| 350 |
|
global_ocean_pressure, ideal_2D_oce, lab_sea, natl_box |
| 351 |
|
|
| 352 |
|
checkpoint50b_pre |
| 353 |
|
o restore the use of momAdvection & useCoriolis flags (mom_vecinv) |
| 354 |
|
o corrected CPP option mismatch in exf_Bulk... functions for case |
| 355 |
|
#define ALLOW_ATM_WIND && #undef ALLOW_ATM_TEMP |
| 356 |
|
|
| 357 |
|
checkpoint50a_post |
| 358 |
|
o Bug fix for merging between c50 and KPP. |
| 359 |
|
ikey was passed from thermodynamics to kpp_calc via |
| 360 |
|
common block rather than being recomputed in kpp_calc, |
| 361 |
|
in contradiction with new key itdkey. |
| 362 |
|
New key ikppkey created, and tamc.h headers updated. |
| 363 |
|
|
| 364 |
|
checkpoint50 |
| 365 |
|
- 2nd step of merge with ecco-branch: |
| 366 |
|
merge changes in ecco-branch back into main trunk |
| 367 |
|
- bug fix in find_alpha |
| 368 |
|
- more storing in gmredi_calc_tensor.F, kpp_routines.F |
| 369 |
|
- avoid iterative loop in ini_pressure for ALLOW_AUTODIFF_TAMC |
| 370 |
|
- merged some sliced AD support routines |
| 371 |
|
|
| 372 |
|
checkpoint49 |
| 373 |
|
merged from ecco-branch: |
| 374 |
|
o exf: |
| 375 |
|
- Enable initialisation of forcing fields to constant |
| 376 |
|
(runtime) values. |
| 377 |
|
- in exf_getffields.F |
| 378 |
|
Reduce i-/j-loop to interior domain, discarding overlaps. |
| 379 |
|
That also fixes wrong TAF-key computations for key_1, key_2 |
| 380 |
|
with bulf formulae. |
| 381 |
|
- exf_init.F modify #ifdef for exf_init_evap |
| 382 |
|
- exf_getffieldrec.F, ctrl_getrec.F |
| 383 |
|
The following INT-usages are not safe: |
| 384 |
|
fldsecs = int(fldsecs/fldperiod)*fldperiod |
| 385 |
|
fldcount = int(fldsecs/fldperiod) + 1 |
| 386 |
|
and were modified. |
| 387 |
|
o autodiff: checkpoint_lev... |
| 388 |
|
- updated storing for bulk formulae (some arrays were missing) |
| 389 |
|
- added state for seaice package |
| 390 |
|
- synchronized with main branch |
| 391 |
|
o DIVA: |
| 392 |
|
Modif.s in |
| 393 |
|
- the_main_loop.F |
| 394 |
|
- the_model_main.F |
| 395 |
|
and added cost_final_restore.F cost_final_store.F |
| 396 |
|
o With these changes, it is possible to fully separate |
| 397 |
|
the first forward run to compute the cost function (mdthe_main_loop) |
| 398 |
|
and adjoint runs over individual DIVA intervals (adthe_main_loop) |
| 399 |
|
into separate model executions. |
| 400 |
|
This enables better tuning of DIVA to given queue sizes. |
| 401 |
|
o Loss of cost function value in consecutive DIVA runs is |
| 402 |
|
fixed by storing to 'costfinal' file. |
| 403 |
|
o Appropriate call of ctrl_unpack/_pack,grdchk_main at |
| 404 |
|
beginning/end of adjoint integration is accounted for. |
| 405 |
|
The current changes apply to TAF-1.4.26. |
| 406 |
|
TAF has one remaining bug, requiring one manual change |
| 407 |
|
(see MITgcm manual). |
| 408 |
|
o cal: synchronized between MAIN and ecco-branch |
| 409 |
|
|
| 410 |
|
checkpoint48i_post |
| 411 |
|
o Changes to restore differentiability of code w.r.t. previous tag |
| 412 |
|
(mostly adding new routines to make list and replacing |
| 413 |
|
pressure by totPhiHyd). |
| 414 |
|
o Updated and merged grdchk package |
| 415 |
|
- has both ADM and TLM checks |
| 416 |
|
- works for single- and multi-proc. |
| 417 |
|
- output cleaned |
| 418 |
|
- worked successfully for parallel DIVA |
| 419 |
|
o genmake: |
| 420 |
|
- Adding -platform=SP4 |
| 421 |
|
- Modif. for DEC Alpha (many don't have makedepend; use mkdep instead). |
| 422 |
|
o Modif. for DIVA: |
| 423 |
|
ctrl_unpack should not be called in intermediate DIVA interval. |
| 424 |
|
(similar thing would apply for active files such as |
| 425 |
|
tbar, sbar, psbar, which should not be reinitialised; |
| 426 |
|
only relevant for ecco-branch) |
| 427 |
|
|
| 428 |
|
checkpoint48h_post |
| 429 |
|
o improve "ini_p_ground" when using finite difference form for calc_phi_hyd |
| 430 |
|
o Added net flux to downward flux conversion to pkg/exf/exf_getffields.F |
| 431 |
|
o Added SEAICE_initialHEFF to pkg/seaice |
| 432 |
|
o ecco_the_model_main.F: SEAICE_WRITE_PICKUP must precede WRITE_CHECKPOINT |
| 433 |
|
|
| 434 |
|
checkpoint48g_post |
| 435 |
|
--- EOS: |
| 436 |
|
o compute locally the pressure for use in EOS: UNESCO, JMD95P, MDJWF |
| 437 |
|
store total Potential in totPhyHyd for diagnostics & EOS. |
| 438 |
|
o fix restart and overlap Pb when using Z-coord and EOS funct. of P |
| 439 |
|
--- Pressure/geopotential gradient: |
| 440 |
|
o use Ro_surf & R_low instead of hFac in cal_phi_hyd: |
| 441 |
|
o NLFS + p-coord.: eta*Alpha' is not longer lagging 1.time-step behind |
| 442 |
|
( changes results of global_ocean_pressure ) |
| 443 |
|
o NLFS + z-coord.: add eta*Rho' contribution. |
| 444 |
|
o accurate phiHyd when using interface-W at the middle between 2 centers |
| 445 |
|
( changes results of ideal_2D_oce ) |
| 446 |
|
o includes r* 2nd term in potential gradient. |
| 447 |
|
|
| 448 |
|
checkpoint48f_post |
| 449 |
|
Merging from release1_p12: |
| 450 |
|
o Modifications for using pkg/exf with pkg/seaice |
| 451 |
|
- improved description of the various forcing configurations |
| 452 |
|
- added basic radiation bulk formulae to pkg/exf |
| 453 |
|
- units/sign fix for evap computation in exf_getffields.F |
| 454 |
|
- updated verification/global_with_exf/results/output.txt |
| 455 |
|
o Added pkg/sbo for computing IERS Special Bureau for the Oceans |
| 456 |
|
(SBO) core products, including oceanic mass, center-of-mass, |
| 457 |
|
angular, and bottom pressure (see pkg/sbo/README.sbo). |
| 458 |
|
o Lower bound for viscosity/diffusivity in pkg/kpp/kpp_routines.F |
| 459 |
|
to avoid negative values in shallow regions. |
| 460 |
|
- updated verification/natl_box/results/output.txt |
| 461 |
|
- updated verification/lab_sea/results/output.txt |
| 462 |
|
o MPI gather, scatter: eesupp/src/gather_2d.F and scatter_2d.F |
| 463 |
|
o Added useSingleCpuIO option (see PARAMS.h). |
| 464 |
|
o Updated useSingleCpuIO option in mdsio_writefield.F to |
| 465 |
|
work with multi-field files, e.g., for single-file pickup. |
| 466 |
|
o pkg/seaice: |
| 467 |
|
- bug fix in growth.F: QNET for no shortwave case |
| 468 |
|
- added HeffFile for specifying initial sea-ice thickness |
| 469 |
|
- changed SEAICE_EXTERNAL_FLUXES wind stress implementation |
| 470 |
|
o Added missing /* */ to CPP comments in pkg/seaice, pkg/exf, |
| 471 |
|
kpp_transport_t.F, forward_step.F, and the_main_loop.F |
| 472 |
|
o pkg/seaice: |
| 473 |
|
- adjoint-friendly modifications |
| 474 |
|
- added a SEAICE_WRITE_PICKUP at end of the_model_main.F |
| 475 |
|
|
| 476 |
|
checkpoint48e_post |
| 477 |
|
o New S/R for diagnostic of bottom pressure (phi0surf was missing in c48d_post) |
| 478 |
|
o dynamics: change definition of computational domain & adapt mom_fluxform |
| 479 |
|
accordingly ; when viscA4=0, allows to run the dynamics with Olx=Oly=2. |
| 480 |
|
o following exp works with Olx=Oly=2 (and pass through checkbounds option): |
| 481 |
|
adjust_nlfs.cs-32x32x1 adjustment.128x64x1 adjustment.cs-32x32x1 |
| 482 |
|
exp0 exp1 exp2 exp4 exp5 hs94.cs-32x32x5 hs94.128x64x5 hs94.1x64x5 |
| 483 |
|
global_ocean.90x40x15 global_with_exf global_ocean_pressure |
| 484 |
|
inverted_barometer solid-body.cs-32x32x1 |
| 485 |
|
(high order AdvScheme and/or multiDimAd works with Olx=Oly=2 ? not sure) |
| 486 |
|
|
| 487 |
|
checkpoint48d_post |
| 488 |
|
o in preparation for r*: |
| 489 |
|
a) use pre-computed gradient of hydrostatic potential: |
| 490 |
|
changes in timestep.F & mom_cdscheme.F affects results of ideal_2D_oce |
| 491 |
|
b) move phi0surf from calc_phi_hyd to calc_grad_phi_hyd: |
| 492 |
|
=> affects results of glob_oce_pressure (different truncation error) |
| 493 |
|
update output.txt of ideal_2D_oce & global_ocean_pressure |
| 494 |
|
|
| 495 |
|
checkpoint48d_pre |
| 496 |
|
o in preparation for r*: |
| 497 |
|
new S/R (calc_grad_phi_hyd.F) to compute Hydrostatic potential gradient. |
| 498 |
|
pass the 2 comp. of the grad. as arguments to momentum S/R. |
| 499 |
|
but for the moment, only used if it does not change the results. |
| 500 |
|
o add finite volume form of calc_phi_hyd for buoyancyRelation OCEANIC & OCEANICP |
| 501 |
|
|
| 502 |
|
checkpoint48c_post |
| 503 |
|
Modifications in treatment of initial flux forcing controls |
| 504 |
|
to avoid extra storing of flux fields. |
| 505 |
|
|
| 506 |
|
checkpoint48c_pre |
| 507 |
|
o Added tidal_basin_2d experiment (currently 1D!) |
| 508 |
|
o can use r* Coordinate with OBC (but Orlanski Rad. BC not yet implemented) |
| 509 |
|
|
| 510 |
|
checkpoint48b_post |
| 511 |
|
o r* coordinate added in #ifdef NONLIN_FRSURF block. |
| 512 |
|
(modification to pressure gradient not yet implemented) |
| 513 |
|
o mom_fluxform: new S/R mom_calc_rtrans.F computes vertical transports |
| 514 |
|
(for U & V) that are used for vertical advection of momentum. |
| 515 |
|
|
| 516 |
|
checkpoint48a_post |
| 517 |
|
o In preparation for r*, use temp. array to compute vertical transport |
| 518 |
|
in flux-form momentum pkg. => change results of few test-exp. |
| 519 |
|
Update output of: exp4, exp5, front_relax, global_ocean_pressure, |
| 520 |
|
lab_sea, natl_box & plume_on_slope |
| 521 |
|
|
| 522 |
|
checkpoint48 |
| 523 |
|
o fix small bug that was causing lot of warning (TIMER_STOP 'UPDATE_CG2D') |
| 524 |
|
|
| 525 |
|
checkpoint47j_post |
| 526 |
|
o add new exp. (ideal_2D_oce) to test "exotic" parameters and options. |
| 527 |
|
o moved adexch from forward_step to addummy_in_stepping |
| 528 |
|
o added exch in ctrl_map_... |
| 529 |
|
o aveFreq.=0. at end of cost_final to avoid overwrite in |
| 530 |
|
reverse checkpointing loop |
| 531 |
|
o Yet more changes in gmredi: |
| 532 |
|
* adgmredi_calc_tensor |
| 533 |
|
avoiding all recomputation of gmredi_slope_limit |
| 534 |
|
* adgmredi_x/y/rtransport |
| 535 |
|
added flag for excessive storing to avoid recomp. of |
| 536 |
|
u/v/rtans, dTdx/y/z |
| 537 |
|
-> this is not really necessary and very memory-consuming |
| 538 |
|
* adgmredi_slope_psi: |
| 539 |
|
consistency with gmredi_slope_limit in treatment of GM_slopeSqCutoff |
| 540 |
|
* gmredi_slope_limit |
| 541 |
|
re-activated full calculation of taperfct for case 'ac02' |
| 542 |
|
o updating verification/ adjoints of carbon, global_ocean.90x40x15 |
| 543 |
|
|
| 544 |
|
checkpoint47i_post |
| 545 |
|
o GM_EXCLUDE_() replace GM_TAPER_(): by default, all the code is compiled. |
| 546 |
|
o add GM_Small_Number & GM_slopeSqCutoff as run-time parameters (data.gmredi) |
| 547 |
|
default values = 1.D-12 & 1.D+48 (=> give same results as checkpoint47f_post) |
| 548 |
|
o cg2dTargetResWunit: define tolerance using deltaTmom (and not deltaTfreesurf) |
| 549 |
|
o gmredi: fix few bugs. |
| 550 |
|
|
| 551 |
|
checkpoint47h_post (duplicate tag for PH's c48 tag, in order to move c48) |
| 552 |
|
This will be a good checkpoint to plug into the ecco-branch. |
| 553 |
|
o few modif.'s to get latest version adjointed |
| 554 |
|
(mainly kick out code in ini_linear_phisurf) |
| 555 |
|
o modif's to run adjoint with exactConserv |
| 556 |
|
o case GM_BOLUS_ADVEC should be cleaned |
| 557 |
|
S/R gmredi_slope_psi should be cleaned |
| 558 |
|
o verification/carbon now has exactConserv=.TRUE. |
| 559 |
|
|
| 560 |
|
checkpoint47g_post |
| 561 |
|
o time-average AIM output: becomes independent from statvars time-average |
| 562 |
|
(can be used with #undef ALLOW_TIMEAVE) ; add new variables (e.g., RH) |
| 563 |
|
o After TAF bug fix: |
| 564 |
|
Tested adjoint of gmredi package. |
| 565 |
|
Tests were done with checkpoint47 rather than with latest |
| 566 |
|
checkpoint47f_post since the latter is broken for the adjoint |
| 567 |
|
verification/carbon/ contains both ADM and TLM gradient checks. |
| 568 |
|
Here they are, before they get changed/lost/stolen. |
| 569 |
|
Mostly modif.'s to fix numerical sensitivities. |
| 570 |
|
Gradient checks OK for |
| 571 |
|
- GM_taper_scheme: |
| 572 |
|
* clipping |
| 573 |
|
* ac02 |
| 574 |
|
* linear |
| 575 |
|
* glw91 |
| 576 |
|
* dm95 |
| 577 |
|
* ldd97 |
| 578 |
|
- GMREDI_OPTIONS: |
| 579 |
|
* GM_VISBECK_VARIABLE_K |
| 580 |
|
* GM_NON_UNITY_DIAGONAL |
| 581 |
|
* GM_EXTRA_DIAGONAL |
| 582 |
|
* GM_BOLUS_ADVEC |
| 583 |
|
in conjunction with data.gmredi parameters to be checked in |
| 584 |
|
in a few minutes under verification/carbon/code/ |
| 585 |
|
|
| 586 |
|
checkpoint47f_post |
| 587 |
|
Merging from release1_p10: |
| 588 |
|
o modifications for using pkg/exf with pkg/seaice |
| 589 |
|
- pkg/seaice CPP options SEAICE_EXTERNAL_FORCING |
| 590 |
|
and SEAICE_EXTERNAL_FLUXES |
| 591 |
|
- pkg/exf CPP options EXF_READ_EVAP and |
| 592 |
|
EXF_NO_BULK_COMPUTATIONS |
| 593 |
|
- usage examples are Experiments 8 and 9 in |
| 594 |
|
verification/lab_sea/README |
| 595 |
|
- verification/lab_sea default experiment now uses |
| 596 |
|
pkg/gmredi, pkg/kpp, pkg/seaice, and pkg/exf |
| 597 |
|
|
| 598 |
|
checkpoint47e_post |
| 599 |
|
o created a tag prior to creating a branch |
| 600 |
|
|
| 601 |
|
checkpoint47d_post |
| 602 |
|
o fixed bug in the oceanic pressure coordinates code: vertical viscosity |
| 603 |
|
at the bottom boundary had an erroneous half slip boundary condition |
| 604 |
|
o added bottom drag and no slip boundary condition capabilities to |
| 605 |
|
oceanic pressure coordinates code |
| 606 |
|
o added new verification experiment global_ocean_pressure, this experiment is |
| 607 |
|
also described in the tutorial section of the manual |
| 608 |
|
checkpoint47d_pre |
| 609 |
|
|
| 610 |
|
checkpoint47c_post |
| 611 |
|
o Added a verification experiment for a global cubed-sphere ocean run |
| 612 |
|
- verification/global_ocean.cubed32x32x30 |
| 613 |
|
- uses pkg/therm_seaice and pkg/bulk_forcing |
| 614 |
|
|
| 615 |
|
o allow to run AIM physics with SPEEDY input files (from Franco Molteni) |
| 616 |
|
o allow a more accurate definition of Ro_Surf (selectFindRoSurf=1) |
| 617 |
|
when using P-coordinate; only implemented for atmospheric config. |
| 618 |
|
o OCEANICP & realFreshWater: include P-E direct effect on wVel ; |
| 619 |
|
NOTES: requires option NONLIN_FRSURF to be "#define". |
| 620 |
|
o update advect_xz/results/output.txt (left from checkpoint44g_post) |
| 621 |
|
|
| 622 |
|
checkpoint47b_post |
| 623 |
|
Merging from release1_p9: |
| 624 |
|
o pkg/seaice |
| 625 |
|
- removed GOTO's and added taf directives |
| 626 |
|
- double precision constants to reduce the g77 (Linux) |
| 627 |
|
to F77 (SGI) differences reported in release1_p8 |
| 628 |
|
o tools/genmake |
| 629 |
|
- added SGI options |
| 630 |
|
o verification/testscript |
| 631 |
|
- updated to that of checkpoint47a_post |
| 632 |
|
o verification/global_ocean.90x40x15/input/eedata |
| 633 |
|
- modified for SGI f77 compatibility |
| 634 |
|
o verification/lab_sea |
| 635 |
|
- added description of sea-ice model |
| 636 |
|
- added missing matlab routines |
| 637 |
|
- added test of thermodynamics parallelization |
| 638 |
|
|
| 639 |
|
checkpoint47a_post |
| 640 |
|
o new pkg: aim_v23 |
| 641 |
|
= F. Molteni atmos.physics (SPEEDY, ver23) adapted to MITgcm. |
| 642 |
|
for now, keep the same surface forcing as before. |
| 643 |
|
- Part-Cell implemented into AIM; check that Heat & Water are conserved |
| 644 |
|
- aim.5l_cs: use new aim pkg (run 10yr & get better results than before) |
| 645 |
|
- aim.5l_LatLon & aim.5l_Equatorial_Channel are still using old aim pkg |
| 646 |
|
o add diagnostic of surface correction term in monitor output |
| 647 |
|
|
| 648 |
|
o bulk_force and therm_seaice |
| 649 |
|
- Two packages: bulk_force (Bulk forcing) |
| 650 |
|
and therm_seaice (thermodynamic_seaice) - adopted from LANL CICE.v2.0.2 |
| 651 |
|
- Earlier integration from Stephaine Dutkiewicz |
| 652 |
|
and Patrick Heimbach. |
| 653 |
|
- Two ifdef statements for compile time, |
| 654 |
|
ALLOW_THERM_SEAICE and ALLOW_BULK_FORCE |
| 655 |
|
|
| 656 |
|
checkpoint47 |
| 657 |
|
differentiable version of checkpoint46n_post |
| 658 |
|
o bug fix in quasihydrostaticterms |
| 659 |
|
o * "clean" adjoint code (in terms of extensive recomputations) |
| 660 |
|
can now be obtained for all GMREDI options (i.e. for |
| 661 |
|
- GM_VISBECK_VARIABLE_K |
| 662 |
|
- GM_NON_UNITY_DIAGONAL |
| 663 |
|
- GM_EXTRA_DIAGONAL |
| 664 |
|
- GM_BOLUS_ADVEC ) |
| 665 |
|
* However, wrong gradient check problem remains unsolved. |
| 666 |
|
* New CPP options have been introduced for different |
| 667 |
|
tapering schemes |
| 668 |
|
o external_fields_load now part of differentiation list |
| 669 |
|
o pressure needs multiple storing; |
| 670 |
|
would be nice to have store_pressure at beginning or |
| 671 |
|
end of forward_step, e.g. by having phiHyd global (5-dim.) |
| 672 |
|
(NB: pressure is needed for certain cases in find_rho, |
| 673 |
|
which is also invoked through convective_adjustment). |
| 674 |
|
o recomputations in find_rho for cases |
| 675 |
|
'JMD95'/'UNESCO' or 'MDJWF' are OK. |
| 676 |
|
o #define ATMOSPHERIC_LOADING should be differentiable |
| 677 |
|
o ini_forcing shifted to begining of initialise_varia |
| 678 |
|
o Incorporating QNVS line search routines into MITgcm |
| 679 |
|
(this is separate code, and therefore not under pkg) |
| 680 |
|
lsopt/, optim/ |
| 681 |
|
o Updated verification/carbon/ |
| 682 |
|
- #define GM_VISBECK_VARIABLE_K |
| 683 |
|
- GM_taper_scheme = 'ldd97' |
| 684 |
|
|
| 685 |
|
checkpoint46n_post |
| 686 |
|
Merging from release1_p8: |
| 687 |
|
o verification/natl_box: |
| 688 |
|
updating new external_fields_load routine |
| 689 |
|
o New package: pkg/seaice |
| 690 |
|
Sea ice model by D. Menemenlis (JPL) and Jinlun Zhang (Seattle). |
| 691 |
|
The sea-ice code is based on Hibler (1979-1980). |
| 692 |
|
Two sea-ice dynamic solvers, ADI and LSR, are included. |
| 693 |
|
In addition to computing prognostic sea-ice variables and diagnosing |
| 694 |
|
the forcing/external data fields that drive the ocean model, |
| 695 |
|
SEAICE_MODEL also sets theta to the freezing point under sea-ice. |
| 696 |
|
The implied surface heat flux is then stored in variable |
| 697 |
|
surfaceTendencyTice, which is needed by KPP package (kpp_calc.F and |
| 698 |
|
kpp_transport_t.F) to diagnose surface buoyancy fluxes and for the |
| 699 |
|
non-local transport term. Because this call precedes model |
| 700 |
|
thermodynamics, temperature under sea-ice may not be "exactly" at |
| 701 |
|
the freezing point by the time theta is dumped or time-averaged. |
| 702 |
|
N.B.: fairly large differences in accuracy occur across |
| 703 |
|
different platforms/compilers; comparison between |
| 704 |
|
g77 (Linux) and F77 (SGI Origin 2000/MIPSpro) |
| 705 |
|
generated output gives: |
| 706 |
|
T S U V |
| 707 |
|
C D M c m s m s m s m s |
| 708 |
|
n p a R g m m e . m m e . m m e . m m e . |
| 709 |
|
f n k u 2 i a a d i a a d i a a d i a a d |
| 710 |
|
g d e n d n x n . n x n . n x n . n x n . |
| 711 |
|
|
| 712 |
|
Y Y Y Y 5 5 7 7 7 8 10 9 6 6 6 6 7 5 7 5 7 FAIL lab_sea |
| 713 |
|
o GAD: |
| 714 |
|
- generated new common blocks to account for call of |
| 715 |
|
same gad routines with differing traceridentities |
| 716 |
|
(needed to modify tracerIdentity indices in GAD.h) |
| 717 |
|
- generated separate common blocks for case useCubedSphereExchange |
| 718 |
|
(Department of Futurology) |
| 719 |
|
- parameter lists to gmredi_?transport: added tracerIdentity |
| 720 |
|
- added new key indices to tamc.h |
| 721 |
|
o external_fields_load: |
| 722 |
|
- added this routine to TAF list |
| 723 |
|
- needed to make some common blocks global and additional storing |
| 724 |
|
along the same lines as exf package (checkpoint_lev?_directives.h) |
| 725 |
|
o exf: |
| 726 |
|
updated external forcing package |
| 727 |
|
- bug fixes carried over from ecco-branch |
| 728 |
|
(missing OBCS_OPTIONS.h in two routines) |
| 729 |
|
- enable easy to use "no forcing". |
| 730 |
|
- added exf I/O for atmospheric loading |
| 731 |
|
- added exf I/O for runoff data |
| 732 |
|
- transfered scaling between exf <-> MITgcm to exf namelist |
| 733 |
|
o Adding new verification experiment global_with_exf. |
| 734 |
|
This verification experiment is almost identical to the |
| 735 |
|
global_ocean.90x40x15 experiment, except that it uses |
| 736 |
|
the exf (and cal) package instead of the |
| 737 |
|
external_fields_load routine (cf README file). |
| 738 |
|
|
| 739 |
|
checkpoint46m_post |
| 740 |
|
o modified testscript to return -- when there are no lines of "%MON" output |
| 741 |
|
o added new flag "quasiHydrostatic" for QH mode |
| 742 |
|
and new routine quasihydrostaticterms() |
| 743 |
|
o added 2*Omega*Cos(phi)*W to u equations (non-hydrostatic Coriolis term) |
| 744 |
|
New routine: mom_u_coriolis_nh() |
| 745 |
|
o fixed sign of non-hydrostatic metric terms in u and v equations. |
| 746 |
|
Depend on correct sign of gravitySign. |
| 747 |
|
o added work.pc* to .cvsignore in aim directories using ifc |
| 748 |
|
o introduced flag for controllin non-hydrostatic metric terms |
| 749 |
o removed gravitySign from PARM04 |
o removed gravitySign from PARM04 |
| 750 |
o switched to g77/ifc from pgf77 in aim experiments |
o switched to g77/ifc from pgf77 in aim experiments |
| 751 |
|
|
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch