1 |
$Header: /u/gcmpack/MITgcm/doc/tag-index,v 1.53.4.29 2003/03/02 23:51:34 dimitri Exp $ |
2 |
|
3 |
Notes on tags used in MITgcmUV |
4 |
============================== |
5 |
|
6 |
ecco_c44_e27 |
7 |
o exf: |
8 |
- Enable initialisation of forcing fields to constant |
9 |
(runtime) values. |
10 |
- in exf_getffields.F |
11 |
Reduce i-/j-loop to interior domain, discarding overlaps. |
12 |
That also fixes wrong TAF-key computations for key_1, key_2 |
13 |
with bulf formulae. |
14 |
- exf_init.F modify #ifdef for exf_init_evap |
15 |
- exf_getffieldrec.F, ctrl_getrec.F |
16 |
The following INT-usages are not safe: |
17 |
fldsecs = int(fldsecs/fldperiod)*fldperiod |
18 |
fldcount = int(fldsecs/fldperiod) + 1 |
19 |
and were modified. |
20 |
o grdchk: updated and merged |
21 |
- has both ADM and TLM checks |
22 |
- works for single- and multi-proc. |
23 |
- output cleaned |
24 |
- worked successfully for parallel DIVA |
25 |
Needed one update in EEPARAMS.h and updates in ctrl package |
26 |
More modif's may be necessary for ECCO code to support TLM. |
27 |
o autodiff: checkpoint_lev... |
28 |
- updated storing for bulk formulae (some arrays were missing) |
29 |
- added state for seaice package |
30 |
o DIVA: |
31 |
Modif.s in |
32 |
- cost_init.F |
33 |
- cost_averagesfields.F |
34 |
- ecco_the_main_loop.F |
35 |
- ecco_the_model_main.F |
36 |
and added cost_final_restore.F cost_final_store.F |
37 |
o With these changes, it is possible to fully separate |
38 |
the first forward run to compute the cost function (mdthe_main_loop) |
39 |
and adjoint runs over individual DIVA intervals (adthe_main_loop) |
40 |
into separate model executions. |
41 |
This enables better tuning of DIVA to given queue sizes. |
42 |
o Loss of cost function value in consecutive DIVA runs is |
43 |
fixed by storing to 'costfinal' file. |
44 |
o Appropriate call of ctrl_unpack/_pack,grdchk_main at |
45 |
beginning/end of adjoint integration is accounted for. |
46 |
The current changes apply to TAF-1.4.26. |
47 |
TAF has one remaining bug, requiring one manual change |
48 |
(see MITgcm manual). |
49 |
|
50 |
icebear4 |
51 |
o Added cost_ice for sea-ice volume sensitivity experiment. |
52 |
o ecco_the_model_main.F: SEAICE_WRITE_PICKUP must precede WRITE_CHECKPOINT |
53 |
o Reduced derivative singularities in pkg/seaice. |
54 |
o Updated verification/natl_box_adjoint |
55 |
|
56 |
icebear3 |
57 |
o Removed spurious iikey and idkey |
58 |
o pkg/exf |
59 |
- bug fix in treatment of downward radiation |
60 |
o pkg/seaice |
61 |
- added initial sea-ice thickness parameter SEAICE_initialHEFF |
62 |
o updated verification/natl_box_adjoint |
63 |
|
64 |
icebear2 |
65 |
o Added missing /* */ to CPP comments in pkg/seaice, pkg/exf, |
66 |
kpp_transport_t.F,and ecco_the_main_loop.F |
67 |
o adjoint-friendly modifications to pkg/seaice |
68 |
o added verification/lab_sea as per release1_p12_pre |
69 |
|
70 |
ecco_c44_e26 |
71 |
o Merging pkg/exf and pkg/seaice from cvs tag ecco_ice2 |
72 |
o exf: Modifications to correctly handle evap field for adjoint |
73 |
o grdchk: |
74 |
Added code to perform gradient checks for bulk formulae/atmos. state |
75 |
(atemp, aqh, uwind, vwind). |
76 |
NOTE: |
77 |
This package is out of synch with c48 package. |
78 |
The latter also has the tangent linear gradient checks. |
79 |
Need to be merged. |
80 |
o verification/natl_box_adjoint: |
81 |
* added new setup: code_bulk/, input_bulk, results_bulk |
82 |
(Experiment 2, described in README) |
83 |
uses bulk formulae and atmos. state as control instead of |
84 |
air-sea fluxes. |
85 |
Performs gradient checks for first 4 elements of atemp control. |
86 |
* changed standard verification to perform |
87 |
gradient checks on first 4 elements on hflux control. |
88 |
* gmredi switched off (will come after merge with c48). |
89 |
|
90 |
ecco_c44_e25 |
91 |
o Merging some GM changes from main trunk (c47) to ecco-branch. |
92 |
o Updating verif. natl_box_adjoint with GMRedi. |
93 |
o Adding setup with explicit differentiaton of pressure solver. |
94 |
o Bug fix: missing OBCS_OPTIONS.h for sponge layer code. |
95 |
o (Re-?) Commiting old bug fixes in kpp to ecco-branch |
96 |
(kpp_calc.F, kpp_routines.F) |
97 |
o exf: |
98 |
updated external forcing package |
99 |
- cleaned old exfa stuff |
100 |
- bug fixes |
101 |
(missing OBCS_OPTIONS.h in two routines) |
102 |
- enable easy to use "no forcing". |
103 |
- added exf I/O for atmospheric loading |
104 |
- added exf I/O for runoff data |
105 |
- transfered scaling between exf <-> MITgcm to exf namelist |
106 |
|
107 |
ecco_c44_e24 |
108 |
o modifications to gradient check package (Martin Losch) |
109 |
- enable centered differences |
110 |
- modified format of standard output |
111 |
o bug fix for nwettile contents |
112 |
so far assumed global sum in first entry (1,1,k), |
113 |
needed for global control vector |
114 |
-> introduced new variable nwet[c/w/s]global(nr) |
115 |
o shifts of calls to cal/exf package to be accessible |
116 |
in MITgcm stand-alone |
117 |
o shifted packages_init_variables before convective_adjustment_ini |
118 |
(corresponding update of TAF tape initialisation). |
119 |
o removed auxiliary field in impldiff |
120 |
(caused numerical differences for GMRedi cases) |
121 |
o updated obcs control part (G. Gebbie) |
122 |
o removed some f90 comments ("!") |
123 |
o Bug fixes and updates in cost package |
124 |
* cost_final: correct multiplier mult_tauv, mult_sflux |
125 |
* cost_hyd: added call cost_ssh |
126 |
* cost_init_barfiles.F: replaced arrays by tmpfld's |
127 |
* cost_readtopexmean.F: ersmask -> tpmeanmask |
128 |
* cost_weights.F: corrected wtheta[2], wsalt[2] |
129 |
o cleaned exchanges in cg2d.F |
130 |
|
131 |
ecco_c44_e23 |
132 |
merged patches between release1_p1 and release1_p2: |
133 |
o fixed array boundary problem for adjoint |
134 |
(related to partial recomputation within subroutine |
135 |
related to dynamics -> calc_viscosity -> mom_... |
136 |
o fixed recomputation in thermodynamics |
137 |
o prepare (but not complete) recomputation problem in impldiff |
138 |
o fixed problem for adjoint of global_max |
139 |
o switched off sponge layer code in external_forcing |
140 |
Bug fix in ctrl_mask_set_yz (spotted by G. Gebbie) |
141 |
|
142 |
ecco_c44_e22 |
143 |
o Finally, the merges with the Scripps cost function terms: |
144 |
Various additions, modifications, quality checks. |
145 |
o Plus: Included switch ALLOW_AUTODIFF_MONITOR to handle |
146 |
monitoring of adjoint state. |
147 |
o Plus: removed comm_stats, ini_communication_patterns from diff. list |
148 |
(no more need to "keep" MPI headers for adjoint generation) |
149 |
o Plus: exf changes: |
150 |
o added run-off routine |
151 |
o additional storing needed for ustress,vstess |
152 |
because of state-dependent quality check |
153 |
o added missing initialisations |
154 |
o Plus: genmake with MAKEDEPEND variable was still missing in ecco-branch |
155 |
NB: - verification/natl_box_adjoint/ 240 steps run |
156 |
* looks ok under Linux (twain.lcs.mit.edu) |
157 |
* looks ok on MIT SUN's (fjord.mit.edu) |
158 |
* NaN in reverse mode on Scripps SUN's at iter=214 |
159 |
* NaN in reverse mode on IBM SP3 (BlueHorizon) at iter=204 |
160 |
- 1 year forward int. global 2x2 deg. produces similar cost functions |
161 |
* all cost terms of new code are slightly smaller (0.5 to 3%). |
162 |
|
163 |
ecco_c44_e21 |
164 |
o reversed order of ctrl/cost initialisation |
165 |
o gmredi_slope_psi.F: Inconsistent types in MIN/MAX expressions |
166 |
o modified to enable ctrl field swapping |
167 |
o bug fix for global sum of ncvarlength in ctrl_init |
168 |
o no more call to ctrl_unpack from ctrl_init |
169 |
o no more write of ctrl vector in ctrl_pack |
170 |
however, init. of control vector from ctrl_unpack |
171 |
o various bug fixes & consistent nobcs handling in |
172 |
ctrl_pack/unpack routines (by G. Gebbie) |
173 |
o optim package: Added obcs control part for lsopt I/O |
174 |
(by G. Gebbie) |
175 |
o missing call to obcs_diags |
176 |
o moved calls in packages_init_fixed to packages_readparms |
177 |
because some quantities are needed earlier. |
178 |
|
179 |
ecco_c44_e20 |
180 |
o before major merges from the Scripps setup |
181 |
here are two small updates: |
182 |
- field swapping for surface flux controls enabled |
183 |
(remains t.b.d. for atmos. state plus obcs) |
184 |
- bug fix in active_file.F for xz-/yz- routines |
185 |
- added missing xz-/yz- routines to active_file_ad.F |
186 |
|
187 |
ecco_c44_e19 |
188 |
o merged necessary patches: |
189 |
- AD-related changes for GMRedi |
190 |
- fixes i KPP (delZ -> drF) |
191 |
- hook to OBCS songe layer code in external_forcing |
192 |
- errorMessageUnit non-zero in eeboot.F |
193 |
- included GMREDI, EXACT_CONSERV to natl_box_adjoint verif. |
194 |
|
195 |
ecco_c44_e18 |
196 |
o sponge layer code for obcs was missing |
197 |
- added hook in external_forcing.F |
198 |
- added storing for TAF |
199 |
o new routine cost_init_barfiles to initialise |
200 |
adjoint fields adtbar, adsbar, adpsbar |
201 |
(needed in conjunction with corrected ladinit for active I/O) |
202 |
o cleaned up and corrected order in packages_readparms, ecco_readparms |
203 |
o corrected and simplified weights for obcs controls |
204 |
- weight fields simplified |
205 |
- bug fix in ctrl_pack |
206 |
- no more write of control vector in ctrl_pack |
207 |
o i/j index bug fixes in cost_ssh_mean |
208 |
|
209 |
ecco_c44_e17 |
210 |
o bug fixes in cost package |
211 |
* wrong identifier for active I/O of monthly mean psbar |
212 |
(cost_averagesfields, cost_ssh) xx_psbar_dummy |
213 |
* cleaned all identifier for active I/O of monthly fields |
214 |
* wrong ladinit in various cost routines |
215 |
o merge with interval (chkpt44c_post - chkpt44d_post) |
216 |
* added missing EXCLUDE_MONITOR flags |
217 |
* changed "e" to "_d" in gmredi_slope_limit, gmredi_slope_psi |
218 |
(incompatible typ in MIN/MAX expressions caused problems |
219 |
on IBM SP3) |
220 |
* in genmake added variable MAKEDEPEND |
221 |
plus resetting for case SunOS |
222 |
added timer_stats.c routine for IBM SP3 |
223 |
* removed variables in dynamics |
224 |
|
225 |
ecco_c44_e16 |
226 |
o new ECCO checkpoint, the first since ecco_c43_e15 |
227 |
which is branched from MITgcm. |
228 |
All changes were merged in between (see ecco-branch-mod? tags). |
229 |
o last changes: |
230 |
* assuming path for TAF and TAMC are known (e.g. in $path) |
231 |
* genmake: |
232 |
- added tools/xmakedepend to circumvene shortcoming of |
233 |
SUN's makedepend (size too small) |
234 |
- invoking ../tools/xmakedepend instead of makedepend |
235 |
- changed compiler flags for case SunOS |
236 |
|
237 |
ecco-branch-mod5 |
238 |
o mod's to reproduce open boundary run |
239 |
o enable field swapping for OB's and climatology |
240 |
o update store directives for OB control |
241 |
o added option ALLOW_OBCS_PRESCRIBE |
242 |
o call of getobcs now out of obcs_calc (param. list modified) |
243 |
|
244 |
ecco-branch-mod4 |
245 |
o this is the second part of changes necessary to reconcile |
246 |
the natl_box_adjoint verification (with and without grdchk). |
247 |
>>>!!! CHANGES BETWEEN THE TAGS ecco-branch-mod3 !!!<<< |
248 |
>>>!!! AND ecco-branch-mod4 !!!<<< |
249 |
>>>!!! SHOULD BE MERGED INTO THE MAIN TRUNK !!!<<< |
250 |
- genmake: removed $S64 overwrite for case SunOS |
251 |
- pkg/exf: update and corrections for field swapping and obcs |
252 |
- pkg/ecco: parameter lists for the_model_main, the_main_loop |
253 |
harmonized between ECCO and MITgcm |
254 |
- pkg/autodiff: added flow directives for obcs, mdsio_gl_slice |
255 |
updated checkpointing_lev... lists for obcs |
256 |
- model/src: minor changes in forward_step, plot_field |
257 |
- pkg/mdsio: added mdsio_gl_slice |
258 |
|
259 |
ecco-branch-mod3 |
260 |
o this is the first part of changes necessary to reconcile |
261 |
natl_box_adjoint verification (with and without grdchk). |
262 |
This set of changes is relevant for ECCO only. |
263 |
The next set/tag of changes (ecco-branch-mod4) |
264 |
should be merged into the main trunk. |
265 |
- update some options |
266 |
- parameter lists of ecco_the_model_main and ecco_the_main_loop |
267 |
differed from those of the MITgcm counterparts. |
268 |
|
269 |
ecco-branch-mod2 |
270 |
o update ecco-branch-mod1 to checkpoint44 |
271 |
o exclude from updates: |
272 |
- verification/ |
273 |
- adjoint/ |
274 |
- pkg/cost |
275 |
- pkg/ctrl |
276 |
- pkg/ecco |
277 |
- pkg/obcs |
278 |
- model/src/the_model_main.F |
279 |
- model/src/the_main_loop.F |
280 |
|
281 |
checkpoint44 |
282 |
o apply Shapiro filter on T&S before calc_PhiHyd (stagger time stepping) |
283 |
o fix bug in mom_vi_del2uv |
284 |
o added some stuff for AD |
285 |
|
286 |
ecco-branch-mod1 |
287 |
o based on c43 |
288 |
o replaced packages cost, ctrl, ecco, obcs |
289 |
o removed MITgcm verif.'s, added ECCO verif.'s |
290 |
o added lsopt, optim for line search optimization. |
291 |
|
292 |
checkpoint43 |
293 |
o added two new advection tests |
294 |
o added global 4x4 degree ocean test |
295 |
o added test of H&S94 with free surface on cube |
296 |
o some final mods for gradient check, adjoint,... |
297 |
|
298 |
checkpoint42 |
299 |
Fixed AD-related problems: |
300 |
o Store directives up-to-date with re-arranged Adams-Bashforth |
301 |
(mainly thermodynamics.F) |
302 |
o New store directives for multi-dim. advection schemes |
303 |
* new CPP flag ALLOW_MULTI_DIM_ADVECTION |
304 |
* new common block and key passkey |
305 |
(mainly gad_advection.F) |
306 |
o Modified store directives for split of dynamics/thermodynamics |
307 |
for the case ALLOW_KPP |
308 |
o Cleaned argument list for timestep_tracer.F |
309 |
|
310 |
checkpoint41 |
311 |
o re-formatted/added comments in prologues of key subroutines |
312 |
to work with protex |
313 |
|
314 |
checkpoint40 |
315 |
o moved calc_gs, calc_gt and calc_tr1 back to model/src |
316 |
o added FLT (floats) package from Arne along with demo. experiment |
317 |
o re-arranged Adams-Bashforth scheme for tracers to allow easier |
318 |
implementation of forward-in-time forcing terms |
319 |
o more mods for non-linear free-surface including new variable |
320 |
o modified multi-dim method to work on cube |
321 |
o bug fix in KPP (from DM) |
322 |
|
323 |
checkpoint40pre9 |
324 |
o added DST advection schemes |
325 |
o fix MPI bug on SGI's |
326 |
o modifications for non-linear free-surface |
327 |
o fixed conditionals for fixed flow/passive tracer only mode |
328 |
o added tension/strain form of stress tensor |
329 |
o changes to commenting of arguments for TAF/TAMC |
330 |
o added/updated many matlab scripts |
331 |
o added multi-dimensional advection method for tracers |
332 |
|
333 |
checkpoint40pre8 |
334 |
o new package rw/ |
335 |
Added method for dumping intermediate local arrays: |
336 |
mdsio_writetile - same as mdsio_writefield except works from inside |
337 |
bi,bj loop |
338 |
mdsio_writelocal - same as mdsio_writetile except works for local |
339 |
arrays |
340 |
write_local_r? - higher-level wrapper for mdsio_writelocal |
341 |
Controlled by diagFreq. Defaults to zero (ie. no dumps) |
342 |
Example given at end of mom_vecinv.F that dumps some local arrays. |
343 |
o replaced calc_mom_rhs.F by mom_fluxform.F/mom_vecinc.F to |
344 |
control flux-form versus vector invariant code. |
345 |
Switch is runtime flag vectorInvariantMomentum (defaults to false) |
346 |
or genmake -DISABLE ... |
347 |
o TAMC-related update to cope with mom_fluxform.F/mom_vecinc.F |
348 |
(additional storing is now needed for case vecinv; |
349 |
currently nothing implemented to switch this storing off |
350 |
in case fluxform is used since no CPP flag available). |
351 |
|
352 |
checkpoint40pre7 |
353 |
o bug fix in dynamics.F |
354 |
calc_viscosity called before k=1,Nr loop. |
355 |
o added some #ifdef's ALLOW_PASSIVE_TRACER |
356 |
o cleaned initialisations in ctrl_map_ini |
357 |
related to TAF |
358 |
|
359 |
checkpoint40pre6 |
360 |
o split dynamics into thermodynamics & dynamics |
361 |
o corresponding split of calc_diffusivity into |
362 |
calc_diffusivity, calc_viscosity |
363 |
(kpp_calc_diff, kpp_calc_visc). |
364 |
o Added diffkr, kapgm to control variables |
365 |
o bug fix in gmredi_slope_limit for ldd97 |
366 |
|
367 |
checkpoint40pre5 |
368 |
o proper initialization of Adams-Bashforth time stepping |
369 |
+ uses forward step for very first step of model (iter=0) |
370 |
o re-generated *all* output since almost all output is modified by this |
371 |
change (sorry but we've been putting this one off for about 2 years) |
372 |
+ natl_box, aim.5l_Equatorial_Channel and aim.5l_zon-ave were |
373 |
*not* regenerated since they are already out of date and I don't |
374 |
want to create the impression that they are suddenly working. |
375 |
o removed KAP stuff for DEC's |
376 |
o modified/added many .cvsignore files in */code and */input |
377 |
o added new expt "solid-body.cs-32x32x1" |
378 |
+ this uses an alternative form of ini_curvilinear_grid.F which |
379 |
is not necessarily any better than the current one... |
380 |
|
381 |
checkpoint40pre4 |
382 |
o added re-initialisations of utrans,vtrans,rtrans in |
383 |
calc_common_factors which are crucial for TAF. |
384 |
o extended iMin,jMin range for calc_common_factors, |
385 |
calc_diffusivity |
386 |
o added option ALLOW_PASSIVE_TRACER |
387 |
o a few minor updates and fixes |
388 |
|
389 |
checkpoint40pre3 |
390 |
o created calc_gtr1.F for gad package |
391 |
o split the_main_loop and re-introduced forward_step |
392 |
o some recomputations of nIter0 for TAMC |
393 |
|
394 |
checkpoint40pre2 |
395 |
o Updated to latest branch of AJA, JMC |
396 |
o Added gradient check package |
397 |
o Added routines for passive tracer handling |
398 |
o Added carbon sequestration sensitivity experiment |
399 |
|
400 |
checkpoint39 |
401 |
o Modifications and fixes to support TAMC |
402 |
(mainly missing or wrong directives) |
403 |
o Corrected common blocks when fooling TAMC to produce adjoint I/O |
404 |
in addummy_in_stepping |
405 |
o Updated verification/vero/ (bug fixes in code/, correct pickups in input/) |
406 |
o Enabled "sliced" (xz-,yz-) exchanges, needed for obcs. |
407 |
o Included following packages: |
408 |
cal: calendar |
409 |
exf: external forcing |
410 |
ecco: essential routines to control use of cal, exf |
411 |
(NB: this is not the full ECCO environment) |
412 |
o disabled some packages by default in genmake |
413 |
|
414 |
checkpoint38 |
415 |
o see notes_c37_adj.txt |
416 |
o corrected wrong layer index in external_forcing_t |
417 |
o replaced the_main_loop/forward_step by the_main_loop |
418 |
o removed forceoutput from write_state |
419 |
(replaced criteria for initial/final state) |
420 |
o added runtime flag for pre-C35 pickups: usePickupBeforeC35 |
421 |
|
422 |
checkpoint37 |
423 |
o mdsio.F routine split and packaged in pkg/mdsio |
424 |
(mdsio.F and mdsio_gl.F still exist in eesupp/src but are not |
425 |
used unless genmake -disable=mdsio) |
426 |
o allow a more precise surface-geopotential gradient for the atmosphere |
427 |
taking into account the change in space of Alpha (=1/rho) at the surface. |
428 |
o variable "etaN" (replace cg2d_x) is the surface R-anomaly [r unit] |
429 |
(ocean: sea surface height [m] / atmos: surface pressure anomaly [Pa]) |
430 |
o Change units: Phi_Hydrostatic (=phiHyd), Phi_Surface(=cg2d_x) and |
431 |
Phi_NonHydrostatic (=cg3d_x), all have now the dimension of a potential |
432 |
[m2/s2] (pressure/rhoConst in the ocean ; Geo-Potential in the atmosphere). |
433 |
o Time-Average diagnostics defined as a pseudo package "timeave". |
434 |
o pickup files : old pickup (before c35) did not work with new code (c35 |
435 |
and after) unless activating 2 commented lines in initialise-varia. |
436 |
o Non-Hydrostatic pickup file Pb fixed. |
437 |
o bug with Rigid-lid version (since c35) fixed. |
438 |
|
439 |
checkpoint36 |
440 |
o implement Crank-Nickelson barotropic time-stepping |
441 |
o recover Time-Average diagnostics |
442 |
o fix Pb in checkpoint (PR#19) |
443 |
o AIM become a standard package ; clean unused CPP_kees |
444 |
|
445 |
checkpoint35 |
446 |
o subtantial rewrite of dynamics.f |
447 |
o allows staggered time-stepping |
448 |
o packaged obcs, aim (Atmospheric Physics), zonal_filt, shap_filt |
449 |
o added bottom drag (linear + quadratic) |
450 |
|
451 |
checkpoint34 |
452 |
o Modified exchange routines for correct reverse simulation. |
453 |
o Added verification output to comply with bug fix in impldiff.F |
454 |
as of c33 (exp2, natl_box). |
455 |
o Corrected store directives in dynamics.F |
456 |
o Corrected array boundaries in impldiff.F |
457 |
o Corrected array dimensions in gmredi. |
458 |
Added temporary scalars to avoid storage by TAMC (gmredi, kpp). |
459 |
o Added routine GLOBAL_SUM_INT to global_sum.F |
460 |
o Added mdsio_gl.F for ECCO purposes. |
461 |
|
462 |
checkpoint33 |
463 |
o Fixed bug in lower BC in impldiff.F |
464 |
o Fixed confusion about units of forcing arrays in FFIELDS.h |
465 |
namely Fu,Fv,Qnet,Qsw,EmPmR: |
466 |
- Removed verification/natl_box/code/external_fields_scale.F |
467 |
(did not differ from that in model/src) |
468 |
- Changed units of fu,fv,Qnet,Qsw,EmPmR back to proper units |
469 |
(see FFIELDS.h for description) |
470 |
- Scale fu,fv,Qnet,Qsw,EmPmR when used in external_forcing_surf.F, |
471 |
kpp_calc.F and kpp_transport_t.F |
472 |
- Removed model/src/external_fields_scale.F and calls to it |
473 |
- verification/natl_box uses flux data with "atmospheric" sign so |
474 |
a special version of external_fields_load.F is used to |
475 |
change the data as it's read in. This way, the arrays |
476 |
have the right units and signs at all times tha a user could |
477 |
possibly use them. |
478 |
o Corrected genmake to delete a temporary file during configuration |
479 |
|
480 |
checkpoint32 |
481 |
o Scaling of forcing fields moved from external_forcing.F to |
482 |
external_fields_scale.F, called directly after loading fields. |
483 |
o Surface relaxation terms added to surface tendency fields in |
484 |
extra subroutine. |
485 |
o impldiff.F changed to avoid storing by TAMC of huge intermediate |
486 |
fields. |
487 |
o Changes in GMRedi to comply with packaging structure. |
488 |
o Modifications and bug fixes in KPP package. |
489 |
o Verification experiment for KPP added (verification/nat_box/) |
490 |
|
491 |
checkpoint31 |
492 |
o Packages interface brought in line with defined standard |
493 |
-> new routines packages_...F |
494 |
-> changes in initialise_fixed.F, initialise_varia.F |
495 |
-> enables initialisation interface for ECCO package |
496 |
o Rescaling of external forcing fields immediately after read |
497 |
(removed from external_forcing.F) |
498 |
-> new: surface_fields_scale.F |
499 |
-> affects ini_foorcing.F, external_fields_load.F |
500 |
o Computation of surface tendencies before tendency updates. |
501 |
-> new: external_forcing_surf.F |
502 |
o Change of variable names useGMRedi, useKPP, useECCO |
503 |
o Further changes in dynamics.F, impldiff.F for TAMC |
504 |
o Tested for experiments 0, 2, 4. |
505 |
- not yet tested for GMRedi. |
506 |
- not yet tested for KPP. |
507 |
|
508 |
checkpoint30 |
509 |
o Updates for OBCs |
510 |
o New experiment for test OBCs (internal_wave) |
511 |
o New output.txt in verification/exp4/results/ |
512 |
o Updates for KPP: calculation of viscosity corrected, new variables |
513 |
kpp_dumpFreq and kpp_taveFreq |
514 |
o CPP'd calls to system routines for TARGET_CRAY_VECTOR (P.H.) |
515 |
and entry for T90 in genmake |
516 |
|
517 |
checkpoint29 |
518 |
o packaged KPP and GM/Redi |
519 |
|
520 |
checkpoint28 |
521 |
o Corrected initialization of hFacMin for difference vertical coordinates |
522 |
o Modified calculation of partial cell thickness to more robust algorithm |
523 |
o config_summary.F: replaced write_1D_... by write_0D_... |
524 |
read_write.F: added subroutines write_0D_... |
525 |
o routines calc_common_factors, calc_gs, calc_gt, calc_phi_hyd, find_rho: |
526 |
included some initialisations required for TAMC. |
527 |
o routine calc_isoslopes: modified sqrt computation to make it |
528 |
differentiable for zero argument. |
529 |
o routines dynamics, impldiff: |
530 |
included store directives and tape key computations required for TAMC. |
531 |
o ALL TAMC related changes are between ifdef ALLOW_AUTODIFF_TAMC |
532 |
|
533 |
checkpoint27 |
534 |
o fixed calc_isoslopes() to so that cancellation of terms in tensor |
535 |
formulation works properly with variable resolution. |
536 |
o restructured the calling tree between the_model_main() and dynamics() |
537 |
o split initialise() into initialise_fixed() and initialise_varia() |
538 |
o introduced initialization file pSurfInitFile for free surface (cg2d_x) |
539 |
in new routine ini_psurf() |
540 |
|
541 |
checkpoint26 |
542 |
o moved some #ifdef in optional routines to encompass everything |
543 |
except the argument declarations. This was using up some memory. |
544 |
o fixed macros like _ADT to work with both versions of Linux cpp |
545 |
o cleaned up some unused and uninitialized variables |
546 |
(helps when debugging with strict compile options) |
547 |
o split up CPP_EEOPTIONS.h into options and macros |
548 |
(macros are now in eesupp/inc/CPP_EEMACROS.h) |
549 |
o patch for 2 processor JAM mode |
550 |
|
551 |
checkpoint25 |
552 |
o updates for OBCs and NH |
553 |
o dramatically reduced memory usage |
554 |
o added "JAM" routines for use on Artic network |
555 |
o parameterization of convection by implicit vertical diffusion |
556 |
o added a diagnostic of convective index |
557 |
o moved call to calc_divg_hat() from dynamics() to solve_for_pressure() |
558 |
(this is was partly for OBCs and NH but also in anticipation of |
559 |
an explicit free-surface option) |
560 |
o initial state input files for U and V |
561 |
|
562 |
checkpoint24 |
563 |
o Removed some spurious "REAL"s |
564 |
o Updated KPP to latest version from JPL |
565 |
o Added a COS(latitude) in horizontal viscosity |
566 |
|
567 |
checkpoint23 |
568 |
o Added del^4 diffusion for salt and temperature |
569 |
o Multiple minor fixes (implicit none's, arguments to global_sum, etc.) |
570 |
|
571 |
checkpoint22 |
572 |
o Made many small changes: implicit none, format(i5), etc... |
573 |
o Introduced KPP mixing scheme (compiles and runs but *not* tested) |
574 |
|
575 |
checkpoint21 |
576 |
o Introduced new I/O package that supports direct-access records (mdsio.F) |
577 |
o Split-up mid-level I/O routines |
578 |
o Updated input data for experiments 0,1,2,4 and 5 |
579 |
o Finally removed the genmake.* scripts (everyone seems happy with genmake) |
580 |
|
581 |
checkpoint20 |
582 |
o Tidied up open-boundaries some more |
583 |
o Added non-hydrostatic code |
584 |
o Fixed some input data |
585 |
o Modified MPI start-up to work in "coupled" context. |
586 |
|
587 |
checkpoint19 |
588 |
o Tidied up directory structure for verification experiments. |
589 |
o Added new code for open boundaries. |
590 |
o Corrected advective terms near boundaries. |
591 |
o Added CPP flag for "old UV" geometry and corrected spherical geometry. |
592 |
o Implemented free-slip/no-slip BCs and del^4 dissipation. |
593 |
o Split default parameters out of ini_parms.F into set_defaults.F. |
594 |
|
595 |
checkpoint18 |
596 |
o Further changes for optional code generation with |
597 |
CPP - ugh! |
598 |
|
599 |
checkpoint17 |
600 |
o Beginning to incorporating modularity |
601 |
for supporting atmos. and ocean configs. |
602 |
o Small changes for TAMC complinace |
603 |
o New ultra simple test case. Barotropic box |
604 |
configuration "exp0". |
605 |
|
606 |
checkpoint16 |
607 |
o Minor changes to exchange routines to support |
608 |
accumulation in reverse mode. Adjoint of an |
609 |
assignment is an addition. |
610 |
o Changes to support compilation by TAMC (LR(1) |
611 |
syntax) and by g77 under Linux ( 72 character |
612 |
limit ). |
613 |
|
614 |
checkpoint15 |
615 |
Rewrote exchange routines to allow |
616 |
o General tile <-> tile connectivity. |
617 |
o TAMC reverse mode flag ( is this really needed? ) |
618 |
o Variable width overlap and exchange region widths. |
619 |
o Hooks for sharing data via DMA, Arctic, Memory Channel, |
620 |
shmput/shmget, VIA, SYSV shmem and every other cool |
621 |
communication method. |
622 |
|
623 |
checkpoint14 |
624 |
Consistent isomorphism chages made. These allow dynamical core |
625 |
to support both pressure and meters as a vertical coordinate. |
626 |
|
627 |
checkpoint13 |
628 |
Release which corrects global_max problem and a couple |
629 |
of KGM time-averaging diagnostic errors. Also contains |
630 |
a double-gyre experiment setup as well as the baseline |
631 |
4 degree global template. The double gyre template is |
632 |
organised to allow it to be applied automatically to |
633 |
the baseline configuration. |
634 |
|
635 |
branch-point-rdot |
636 |
A few tidy-ups have been made bt the real point of this |
637 |
check-point (excuse the pun) is to allow a branch to be |
638 |
made for the purposes of adding the rDot alterations. |
639 |
I don't want to commit the rDot to the main-trunk until |
640 |
we've actually tested it... |
641 |
|
642 |
checkpoint12 |
643 |
Tidy up and added generic genmake script (formerly |
644 |
configure script). |
645 |
Released to testers. |
646 |
|
647 |
checkpoint11 |
648 |
Re-arranged initialise() and introduced ini_vertical_grid() |
649 |
and ini_masks_etc(). |
650 |
|
651 |
checkpoint10 |
652 |
Introduced the configure script. |
653 |
Separated out the diags package. |
654 |
Added template for writing time-averages of tile/intermediate quantities. |
655 |
Updated SIZE.h in exp2 and made the data big-endian. |
656 |
|
657 |
checkpoint9 |
658 |
MPI fix. Filename changes (meta/data). salbin*y stuff. |
659 |
General tidy-up. |
660 |
SST.bin SSS.bin added to verification/exp2 |
661 |
|
662 |
checkpoint8 |
663 |
Inludes time-dependant forcing and time-averaging. |
664 |
Works at 4x4 and 2.8125x2.8125 global |
665 |
|
666 |
checkpoint7-4degree-ref |
667 |
Branch from checkpoint7. Released for application to |
668 |
global modeling and general modeling - process studies, regional |
669 |
model etc.. |
670 |
Configured for verification/exp2 - 4 degree global, two process. |
671 |
|
672 |
checkpoint7 |
673 |
Created by cnh |
674 |
Simple 4 degree global simulation |
675 |
configuration. |
676 |
Includes climatological wind, |
677 |
hydrography, bathymetry etc |
678 |
Does not include Heat Flux and E-P |
679 |
|
680 |
checkpoint6 |
681 |
Created by cnh |
682 |
merged checkpoint5 with CD and |
683 |
spherical polar code. |
684 |
Still some problems with GM Redi |
685 |
abd full topography. |
686 |
|
687 |
checkpoint5 |
688 |
Created by AJA |
689 |
contains updates to GM Redi |
690 |
and reordered timestepping loop |
691 |
for Implicit Diffusion. |
692 |
|
693 |
checkpoint4 |
694 |
Created by cnh |
695 |
Contains memory saving macros ( _rdxc etc..) for when |
696 |
grid variations are limited. |
697 |
Contains polynomial coeffs for EOS and knudsen program |
698 |
for generation. |
699 |
|
700 |
checkpoint3 |
701 |
Created by cnh |
702 |
Added extra flags for runtime options |
703 |
Added this file |
704 |
Added implicit free-surface |
705 |
Added pre-conditioner |
706 |
Put compare01 into repository ( compare01/... ) |
707 |
Put data and eedata into repository ( verification/exp1 ) |
708 |
|
709 |
checkpoint2 |
710 |
Created by AJA |
711 |
Added GM/Redi |
712 |
|
713 |
checkpoint1 |
714 |
Created by AJA |
715 |
Restructured kloop in dynamics.F |
716 |
Added non-linear EOS |
717 |
|
718 |
baseline |
719 |
Original checked in code |