4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
5 |
============================== |
============================== |
6 |
|
|
7 |
|
o model/src: modifications relatd to atmosphere p*: |
8 |
|
- Switch to more accurate p* coordinate slope term (calc_grad_phi_hyd.F) |
9 |
|
- With p* or Sigma-P, use constant reference Pot.Temp (thetaConst) instead |
10 |
|
of vertical profile tRef in geopotential background and anomaly. |
11 |
|
- Acount for true p* correction in theta <-> T conversion in pkg/aim_v23 |
12 |
|
and pkg/atm_phys. |
13 |
|
- Update FWD output of experiments: aim.5l_cs (x2), fizhi-cs-32x32x40, |
14 |
|
fizhi-cs-aqualev20, fizhi-gridalt-hs, tutorial_held_suarez_cs, cpl_aim+ocn. |
15 |
|
o pkg/seaice: fix a bug in seaice_prepare_ridging.F, forgot a sqrt |
16 |
|
o pkg/ptracers: compute ptracer source/sink due to linear free surface. |
17 |
|
not used right now, but can be used in BGC packages to fix conservation. |
18 |
|
o pkg/dic, pkg/gchem: move tracer_meanarea to gchem, so it can be used by |
19 |
|
other packages too. |
20 |
|
o pkg/seaice/seaice_calc_ice_strength.F: replace SEAICE_rhoSnow by |
21 |
|
SEAICE_rhoIce in the Rothrock (1975)-type strength calculations (bug fix) |
22 |
|
o model/src: |
23 |
|
- since we have several heat capacity (even in model/src), define "recip_Cp" |
24 |
|
as local variable (no longer in common block) to avoid any confusion. |
25 |
|
- add thetaConst in PARAMS.h (to be used with pStar or sigma-P coords). |
26 |
|
o pkg/exf: |
27 |
|
- from Jeff Scott: add option (#define EXF_CALC_ATMRHO) to calculate local |
28 |
|
air density as function of air Temp, Humidity and atm pressure. |
29 |
|
|
30 |
|
checkpoint64w (2014/05/02) |
31 |
|
o verification: |
32 |
|
- update output_adm.txt files for experiment global_ocean.cs32x15 (seaice & |
33 |
|
seaice_dynmix) and lab_sea (primary test & noseaicedyn): moving/changing |
34 |
|
code from seaice_growth.F & seaice_advdiff.F to new ridging routines |
35 |
|
seaice_reg_ridge.F seaice_do_ridging.F affects AD gradient |
36 |
|
(down to 8-7 matching digits for global_ocean and 5-6 for lab_sea). |
37 |
|
o model/src: for atmosphere in p* coords: |
38 |
|
- store factor rStarFacC**atm_kappa (= pStarFacK) in common block |
39 |
|
- implement more accurate p* coordinate slope term (in calc_grad_phi_hyd.F), |
40 |
|
but not yet activated. |
41 |
|
o pkg/seaice: fix restart problem for ITD with new ice strength parameterisation: |
42 |
|
- move computation of opnWtrFrac = 1-AREA to before computing ice strength |
43 |
|
- this changes the results (but not for any verification experiment, because |
44 |
|
none tests this), but avoids having an extra field in the restart files |
45 |
|
o pkg/seaice: remove global variable TICE (not needed) |
46 |
|
- replace by TICES or local variable where necessary (cost_seaice_test, |
47 |
|
seaice_diagnostics_state) |
48 |
|
- move copying TICE to TICES from seaice_check_pickup to seaice_read_pickup |
49 |
|
- use specific read/write routines for reading and writing individual leves |
50 |
|
of a multilevel field |
51 |
|
o pkg/ctrl & model/inc: |
52 |
|
- move additional anomaly fields of control vars (related to options: |
53 |
|
ALLOW_KAPGM_CONTROL, ALLOW_KAPREDI_CONTROL and ALLOW_BOTTOMDRAG_CONTROL) |
54 |
|
from DYNVARS.h into new header file: CTRL_FIELDS.h |
55 |
|
o pkg/seaice: some small modifications to the ridging code |
56 |
|
- make fw2ObyRidge a global variable and add the freshwater flux to |
57 |
|
seaice_growth |
58 |
|
- pass actual thickness from seaice_prepare_ridging, so that it is |
59 |
|
computed only in one place and compute it in the same way as in |
60 |
|
seaice_growth |
61 |
|
- initialise more fields in seaice_do_ridging that where cause floating |
62 |
|
point exceptions over land |
63 |
|
o pkg/seaice: preparations for ridging code and first real ridging code |
64 |
|
- move most of "PART 1" of S/R seaice_growth into a new and separate |
65 |
|
S/R seaice_reg_ridge that will both regularize the state variable after |
66 |
|
advection call the actual ridging routine; |
67 |
|
- requires to make 5 fields global fields: d_HEFFbyNEG,d_HSNWbyNEG, |
68 |
|
d_AREAbyRLX,d_HEFFbyRLX,saltFluxAdjust; |
69 |
|
- remove ridging code from seaice_advdiff |
70 |
|
- clean up remaining part 1 of seaice_growth in an attempt to make it |
71 |
|
more readable (this is probably subjective) |
72 |
|
- call seaice_itd_redist once in the initialisation phase to have the initial |
73 |
|
conditions in the correct category |
74 |
|
- finally: add new ridging routine for ITD code |
75 |
|
|
76 |
|
checkpoint64v (2014/04/11) |
77 |
|
o pkg/ecco: |
78 |
|
- add runtime parameters mdt[start,end]date[1,2] |
79 |
|
- rename variable "tpmean" as "mdt" and "topexmeanfile" as "mdtdatfile" |
80 |
|
o pkg/seaice: |
81 |
|
- make deltaC a global variable, so that it can be used in ridging code |
82 |
|
- first step towards ridging: introduce new routines that compute |
83 |
|
ice strength based on ridging parameters |
84 |
|
- add diagnostic for HSNOWITD |
85 |
|
- add opnWtrFrac as a global variable (required for ridging scheme) |
86 |
|
o pkg/salt_plume + pkg/seaice/seaice_growth.F |
87 |
|
- add SALT_PLUME_IN_LEADS flag to activate pkg/salt_plume only when ice |
88 |
|
AREA exceeds a threshold (code contributed by I.Fenty in 2011) |
89 |
|
o verification/hs94.1x64x5: |
90 |
|
- leave ALLOW_AUTODIFF_TAMC undefined in code_oad/AUTODIFF_OPTIONS.h |
91 |
|
- do not use ECCO_CPPOPTIONS.h anymore for TAF & OpenAD Adjoint tests. |
92 |
|
o autodiff: |
93 |
|
- Start to include explicitly AUTODIFF_OPTIONS.h, COST_OPTIONS.h, |
94 |
|
and CTRL_OPTIONS.h in src files (to enable to skip the ECCO_CPPOPTIONS.h) |
95 |
|
For now, only in pkgs used in verification/hs94.1x64x5. |
96 |
|
- Replace ALLOW_AUTODIFF_TAMC by ALLOW_AUTODIFF (except for tape/storage |
97 |
|
which are specific to TAF/TAMC). |
98 |
|
- Replace ALLOW_AUTODIFF_OPENAD by ALLOW_OPENAD: |
99 |
|
since ALLOW_OPENAD is defined in PACKAGES_CONFIG.h (any time pkg/openad |
100 |
|
is compiled), this simplifies/reduces which *_OPTIONS.h file to include. |
101 |
|
- OpenAD: change Makefile to generate the appropriate AD_CONFIG.h when |
102 |
|
building mitgcmuv_ad (make adAll). |
103 |
|
o pkg/ecco: |
104 |
|
- cost_gencost_seaicev4.F: |
105 |
|
reformulate proxy cost as function of sst and local param SEAICE_freeze |
106 |
|
change cost names from siv4-[sst,vol] to siv4-[deconc,exconc] |
107 |
|
- add stop statement if old names siv4-[sst,vol] are used |
108 |
|
- add snapshots + cost reports for atmospheric longwave |
109 |
|
- remove factor 0.1 in front of wtau[u,v] |
110 |
|
- replace never used flag ALLOW_TMI_COST_CONTRIBUTION with |
111 |
|
ALLOW_TMI_SST_COST_CONTRIBUTION |
112 |
|
o pkg/atm_phys: |
113 |
|
- add some diagnostics for Short-wave and for convection S/R. |
114 |
|
o pkg/seaice/seaice_growth: |
115 |
|
- fix a bug in the lateral melt parameterization for SEAICE_ITD |
116 |
|
in seaice_growth; tempFrz needs to be computed inside the i/j-loop |
117 |
|
- rearrange loop order for better vector code in the same code block |
118 |
|
- rewrite the floeDiameter formula to have fewer divisions |
119 |
|
- (miraculously) none of this changes the results of seaice_itd |
120 |
|
o pkg/smooth: |
121 |
|
- move param smooth2Ddiffnbt to pkg/ecco |
122 |
|
o pkg/salt_plume: |
123 |
|
- salt_plume_frac.F: |
124 |
|
clean up local param names |
125 |
|
reformulate (1-frac) to frac; original 1-frac was based on swfrac code |
126 |
|
o pkg/rw: |
127 |
|
- add RS version of S/R READ_MFLDS_LEV. |
128 |
|
o pkg/streamice: |
129 |
|
- fix the restart test by writing/reading pickup_streamice files. |
130 |
|
o pkg/seaice: |
131 |
|
- refine checks for jfnk-related parameters |
132 |
|
- comment out if statement with SEAICEuseIMEX parameter as long as there |
133 |
|
is no IMEX code to avoid confusion and surprises |
134 |
|
|
135 |
|
checkpoint64u (2014/03/07) |
136 |
|
o verification/MLAdjust: |
137 |
|
- rename secondary test suffix: |
138 |
|
1.leith -> A4FlxF + switch to flux-form & add side-drag |
139 |
|
0.leithD -> AhFlxF + switch to flux-form |
140 |
|
0.smag -> AhVrDv (use Vorticity-Div. visc form) + add side-drag |
141 |
|
0.leith -> AhStTn (use Strain-Tension visc form) + add side-drag |
142 |
|
o pkg/mom_vecinv: |
143 |
|
- remove unused arguments from S/R MOM_VI_HDISSIP & MOM_HDISSIP (as it used |
144 |
|
to be before mom_calc_visc.F) |
145 |
|
- skip the call to MOM_CALC_TENSION & MOM_CALC_STRAIN if not needed. |
146 |
|
- add 2nd copy of vort3 & strain that knows about lateral BC (free/no slip): |
147 |
|
this is simpler for the adjoint and for diagnostics. |
148 |
|
o model/src + pkg/momentum: |
149 |
|
- add input files for 2-D additional Western and Southern cell-edge wall |
150 |
|
(e.g., to add "thin-wall" where it is =1). |
151 |
|
- fix sideDrag option for thin-walls with Non-Lin Free-Surf |
152 |
|
using 2nd hFacZ that is computed from initial (fix domain) hFac |
153 |
|
- known limitations of thin wall implementation: not yet coded for sigma |
154 |
|
coords; not working with CD-scheme neither. |
155 |
|
o pkg/seaice: |
156 |
|
- rename recently introduced parameters from (wrong name) AB to |
157 |
|
BDF (backward difference time discretisation) |
158 |
|
- retire runtime parameters SEAICEuseAB2, SEAICE_abEps |
159 |
|
o tools/adjoint_options: |
160 |
|
- remove setting of "OPENAD=true" (now set by command line "genmake2 -oad") |
161 |
|
|
162 |
|
checkpoint64t (2014/02/01) |
163 |
|
o pkg/ecco |
164 |
|
- add compile flag ALLOW_GENCOST_TIMEVARY_WEIGHT in ECCO_OPTIONS.h and |
165 |
|
runtime logical flag gencost_timevaryweight to read daily sea-ice sigma |
166 |
|
- replace hard-coded record number and period which only work for monthly |
167 |
|
with read-in runtime param that works for other periods (month,day,year) |
168 |
|
o pkg/thsice |
169 |
|
- corrected store directive fixes thsice adjoint gradient checks |
170 |
|
o genmake2 & testreport: |
171 |
|
- add genmake2 option "-oad" to generate a Makefile for OpenAD built and |
172 |
|
in this case, use tools/adjoint_options/adjoint_oad as default AD optfile. |
173 |
|
Update testreport accordingly. No need to set "export OPENAD=true" |
174 |
|
in adjoint_options/adjoint_oad anymore (will be removed later). |
175 |
|
o pkg/shelfice: |
176 |
|
- put T & S shelfice forcing into model surfaceForcing_T,_S (for now, |
177 |
|
only if SHELFICEboundaryLayer=F); Change model/src/external_forcing.F |
178 |
|
so that surface forcing applies to k=kSurf if using pkg/shelfice. |
179 |
|
o pkg/ctrl, pkg/grdchk, pkg/ecco: |
180 |
|
- Changes to make GENARR more modular (i.e. invisible if not used) |
181 |
|
o model/src/temp_integrate.F and salt_integrate.F |
182 |
|
- re-include store directives that were removed on Dec 27th. |
183 |
|
o model/src/external_forcing_surf.F |
184 |
|
- add storage directives for EmPmR and PmEpR to avoid recomputation of |
185 |
|
seaice_model in the adjoint that is problematic when seaice_model is |
186 |
|
turned off in the adjoint. The resolved issue dated back to July when |
187 |
|
EXTERNAL_FORCING_SURF call was moved. |
188 |
|
Update global_ocean.cs32x15/*/output_adm.seaice_dynmix.txt accordingly. |
189 |
|
o pkg/exf/exf_check.F |
190 |
|
- fix a small bug again (originally found by Wentao Liu, now found again) |
191 |
|
- streamline error output |
192 |
|
o pkg/streamice and aim_v23: |
193 |
|
- replace a few "_EXCH_XY_RL (" with "_EXCH_XY_RL(", so that |
194 |
|
Apple's /usr/bin/cpp knows what to do: |
195 |
|
Apple LLVM version 5.0 (clang-500.2.79) (based on LLVM 3.3svn) |
196 |
|
Target: x86_64-apple-darwin13.0.0 |
197 |
|
|
198 |
|
checkpoint64s (2014/01/05) |
199 |
|
o tools/adjoint_options: |
200 |
|
- switch to fastopt.net TAF server (instead of default fastopt.de) for |
201 |
|
AD optfiles adjoint_diva & adjoint_f95 (similar to adjoint_default). |
202 |
|
o model/src: |
203 |
|
- remove some unnecessary TAF storage directives in forward_step.F and |
204 |
|
in dynamics.F (note: not always removed by TAF, e.g., double storage |
205 |
|
of kappaRU,kappaRV in former version of dynamics.F). |
206 |
|
o model/src & pkg/ptracers: |
207 |
|
- call directly GAD_IMPLICIT_R & IMPLDIFF + DWNSLP_APPLY from temp_integrate.F |
208 |
|
salt_integrate.F and ptracers_integrate.F (instead of from thermodynamics.F) |
209 |
|
and remove ptracers_dwnslp_apply.F & ptracers_implicit.F |
210 |
|
- call MATRIX_STORE_TENDENCY_EXP for each tracer (in ptracers_integrate.F) |
211 |
|
- add fZon & fMer as output argument of S/R GAD_CALC_RHS. |
212 |
|
- update TAF AD-output of experiments global_ocean.90x40x15 (13 for AD-grad). |
213 |
|
o useSingleCpuInput separates single-CPU input from single-CPU output; |
214 |
|
for now only implemented for in exf_interp_read.F |
215 |
|
o verification/advect_xy: |
216 |
|
- fix missing EXCH call in local version of ini_salt.F and update output. |
217 |
|
o momentum diagnostics: |
218 |
|
- fix which deltaT is used in Total-Tendency diagnostics. |
219 |
|
- fix typo in diagnostics for momentum tendency from AB. |
220 |
|
o pkg/seaice/seaice_lsr.F: fix an initialisation bug for the |
221 |
|
SEAICE_VECTORIZE_LSR_ZEBRA option that lead to NaNs on a NEC-SX8 |
222 |
|
|
223 |
|
checkpoint64r (2013/12/08) |
224 |
|
o verification/*/results: |
225 |
|
- Update reference output of experiments that were passing with just 13 digits |
226 |
|
(make it easier to detect future changes in results) |
227 |
|
o model/src & pkg/ptracers: |
228 |
|
- move calls to GAD_ADVECTION (Multi-Dim advection) from thermodynamics.F to |
229 |
|
inside temp_integrate.F and salt_integrate.F and from ptracers_advection.F |
230 |
|
to inside ptracers_integrate.F ; remove ptracers_advection.F |
231 |
|
o coupling interface pkgs: |
232 |
|
[atm_compon_interf, ocn_compon_interf & atm_ocn_coupler] |
233 |
|
- change the config check to test the agreement of Ocean Depth |
234 |
|
with ATM land/sea mask (instead of ATM orography); |
235 |
|
- send to coupler the number of coupler time-steps to do for this run: |
236 |
|
this allows to check for consistency and don't need anymore to be |
237 |
|
specified in parameter file "data"; |
238 |
|
in coupler: |
239 |
|
- check consistency of RunOff map versus Ocean Depth and ATM land/sea mask. |
240 |
|
- improve report to log-file of config setting (pass log-file unit to few S/R) |
241 |
|
o pkg/compon_communic: |
242 |
|
- add coupling send/receive for non-tiled integer vector; |
243 |
|
- add S/R to check and stop if any proc has an error (similar to |
244 |
|
eesupp/src/stop_if_error.F but single thread and across Comm_World). |
245 |
|
o pkg/exch2: |
246 |
|
- in exch2*_get_rx[1,2]: use new list "W2_tileIndex" to replace search |
247 |
|
through list of all tiles my proc owns. |
248 |
|
o model/src: |
249 |
|
- AUTODIFF only: in calc_adv_flow.F, recompute rTransKp (instead of copying |
250 |
|
from rTrans) to simplify dependencies (no need to store rTrans anymore in |
251 |
|
calling S/R); this affects AD results at machine truncation level: |
252 |
|
update AD-output (TAF) of experiments: 1D_ocean_ice_column, |
253 |
|
bottom_ctrl_5x5 & obcs_ctrl (12 digits for AD grad, same FWD grad). |
254 |
|
o tools/build_options: |
255 |
|
- simplify Warnings options (-ieee & -devel cases) in linux_amd64_gfortran |
256 |
|
o coupling interface pkgs: |
257 |
|
[compon_communic, atm_compon_interf & ocn_compon_interf] |
258 |
|
- move declaration of local array "ibuf" out of header file "CPLR_SIG.h" |
259 |
|
to each S/R that needs it; remove un-used variables. |
260 |
|
- change component tile registration (cpl_register.F in both compon_interf |
261 |
|
and mitcomponent_tile_register.F) to pass the full list of tiles |
262 |
|
information (including bi,bj) in 1 vector (instead of in a bi,bj matrix). |
263 |
|
- with pkg/exch2: adopt the same convention used for I/O to define mapping |
264 |
|
towards coupler global array (unless W2_useE2ioLayOut=F): |
265 |
|
a) can now use different tiling between atm & ocn comp. |
266 |
|
b) no needs to modify the run-Off map when changing tiling (was needed |
267 |
|
even if using the same tiling between atm & ocn). |
268 |
|
c) will allow to run with blank-tiles in 1 component (e.g., for ocean |
269 |
|
full land tiles); not yet tested. |
270 |
|
- check for valid mapping from comp. tile to coupler global array |
271 |
|
- improve report of tile registration results to log file. |
272 |
|
o pkg/exch2: |
273 |
|
- in eesupp/src/eeboot.F: move call to W2_EEBOOT before Coupler |
274 |
|
tile-registration (CPL_REGISTER); |
275 |
|
- activate (un-comment) definition and setting of "W2_tileIndex" list to |
276 |
|
store the local bi,bj indices of each tile; |
277 |
|
- also print bi,bj in local report to log file; |
278 |
|
- fix pkg/debug/debug_cs_corner_uv.F for the case nSy > 1 (left from |
279 |
|
Jun 2009 modif of pkg/exch2). |
280 |
|
o testing on baudelaire: |
281 |
|
- run 2 tests at a time: no significant slow down (since enough cpu) and |
282 |
|
get the full set of test done earlier |
283 |
|
- switch the ifort test to "-devel" ; run the OpenAD test everyday. |
284 |
|
o model/src/thermodynamics.F and routines called from it: |
285 |
|
- pkg/gmredi: gmredi_residual_flow.F (replacing gmredi_calc_uvflow.F & |
286 |
|
gmredi_calc_wflow.F) to update eulerian velocity (input) into residual |
287 |
|
velocity (outp) by adding bolus velocity. |
288 |
|
- model/src: calc_adv_flow.F (replacing calc_common_factors.F) to compute |
289 |
|
mass transport from velocity + lateral grid cell area |
290 |
|
- rename calc_gt.F to temp_integrate.F (includes call to TIMESTEP_TRACER) |
291 |
|
- rename calc_gs.F to salt_integrate.F (includes call to TIMESTEP_TRACER) |
292 |
|
in thermodynamics.F |
293 |
|
- remove k from thermodynamics.F and move it to temp_integrate.F |
294 |
|
salt_integrate.F and ptracers_integrate.F (now done inside the tracer |
295 |
|
loop). |
296 |
|
- compute locally (in thermodynamics.F) 3-D velocity field that is used to |
297 |
|
advect tracers; pass it as argument to GAD_ADVECTION, GAD_SOM_ADVECT, |
298 |
|
PTRACERS_ADVECTION, TEMP_INTEGRATE, SALT_INTEGRATE, PTRACERS_INTEGRATE, |
299 |
|
GAD_IMPLICIT_R and PTRACERS_IMPLICIT |
300 |
|
- update AD-output (TAF & OAD) of experiments: 1D_ocean_ice_column, |
301 |
|
global_ocean.90x40x15, tutorial_global_oce_biogeo and |
302 |
|
tutorial_tracer_adjsens (+ TLM output). |
303 |
|
o tools/OAD_support and (verification/*/) code_oad: |
304 |
|
- OAD_support/cb2mGetModules.csh: add PACKAGES_CONFIG.h in header file to |
305 |
|
module conversion ; remove unneeded CPP_EEOPTIONS.h. |
306 |
|
- code_oad/CPP_OPTIONS.h: remove PACKAGES_CONFIG.h and use a more standard |
307 |
|
version (closer to model/inc/CPP_OPTIONS.h). |
308 |
|
o model/src: |
309 |
|
- remove parameter "useOldFreezing" from PARAMS.h and push it to the |
310 |
|
retired list of parameters (routine FREEZE has been replace by surface |
311 |
|
freezing (10.yrs ago) and by pkg/frazil (interior freezing). |
312 |
|
o pkg/gmredi: |
313 |
|
- add option (GM_CALC_TRANSP_FROM_VEL) to compute transport directly from |
314 |
|
velocity (same scheme but affects machine truncation) for GM_AdvForm. |
315 |
|
- update output of FWD experiments: front_relax.bvp, |
316 |
|
global_ocean.90x40x15.dwnslp, global_ocean.cs32x15 and ideal_2D_oce. |
317 |
|
|
318 |
|
checkpoint64q (2013/11/17) |
319 |
|
o verification: |
320 |
|
- generate new output on baudelaire (gfortran, -ieee) for OpenAD experiments |
321 |
|
tutorial_global_oce_biogeo and global_ocean.90x40x15. |
322 |
|
- generate new output (growth & melt) for seaice_itd experiment after |
323 |
|
updating the input pickup_seaice files (previously wrong). |
324 |
|
o pkg/cost: |
325 |
|
- define a "final cost function" for each tile, and use GLOBAL_SUM_TILE_RL |
326 |
|
(instead of _GLOBAL_SUM_RL) to get the final cost from all procs & threads: |
327 |
|
o this fixes global-sum value if multi-threads (since GLOBAL_SUM_RL |
328 |
|
cannot be applied to var in common block). |
329 |
|
o allows to get results independent of tile-per-proc setting (if using |
330 |
|
slower GLOBAL_SUM_SEND_RECV option). |
331 |
|
- update accordingly: seaice_cost_final.F, shelfice_cost_final.F, |
332 |
|
thsice_cost_final.F and ecco_cost_final.F (this one affects lab_sea |
333 |
|
forward gradients: update output). |
334 |
|
- update */code_oad/externalDummies.F by adding dummy GLOBAL_SUM_TILE_RL |
335 |
|
S/R and, for most exp., also comment out GLOBAL_SUM_R8. |
336 |
|
o pkg/mom_common: |
337 |
|
- move to pkg/mom_common and model/src (previously in tutorial_deep_convection |
338 |
|
code) 2nd version of isotropic 3-D Smagorinsky code interface: strain and |
339 |
|
viscosity are locally declared in dynmics.F and pass as argument to CALC_GW; |
340 |
|
ensure that all field value that are used are set. |
341 |
|
o tutorial_deep_convection (smag3d): |
342 |
|
- add missing scaling in mom_calc_smag_3d.F; |
343 |
|
- set smag3D_coeff in input.smag3d such that results do not change. |
344 |
|
o pkg/dic: |
345 |
|
- change dic_cost.F to compute cost on each tile first and then apply |
346 |
|
GLOBAL_SUM_TILE_RL: this affects forward gradient of AD experiments |
347 |
|
tutorial_dic_adjoffline & tutorial_global_oce_biogeo; update output. |
348 |
|
- remove kLev argument and use GLOBAL_SUM_TILE_RL in tracer_meanarea. |
349 |
|
o eesupp/src & pkg/mdsio: |
350 |
|
- remove gather/scatter_vector.F (from eesupp) and mdsio_read/writevector.F |
351 |
|
(from pkg/mdsio) since they are no longer used for tape IO. |
352 |
|
- move simple C subroutine "setdir.c" from pkg/compon_communic to eesupp/src |
353 |
|
(+ keep a copy in pkg/atm_ocn_coupler). |
354 |
|
o verification/isomip: |
355 |
|
- split domain in 8 squared tiles (instead of just 2) for FWD, AD & OAD tests. |
356 |
|
o tools/build_options: |
357 |
|
- optfile linux_amd64_ifort11, fast (no ieee) case: |
358 |
|
to get a working restart for all experiments, add "-fp-model source" |
359 |
|
to FOPTIM (also remove "-opt-streaming-stores auto" <- this is the default). |
360 |
|
|
361 |
|
checkpoint64p (2013/10/20) |
362 |
|
o pkg/autodiff: |
363 |
|
- adread_adwrite.F: when type and length do not match ALLOW_AUTODIFF_WHTAPEIO |
364 |
|
criteria, return to simple (default) tape IO format. |
365 |
|
o pkg/mdsio (+ eesupp/src): |
366 |
|
- rewrite/simplify tape-IO S/R (new S/R MDS_READ/WRITE_TAPE replace |
367 |
|
previous MDSREAD/WRITEVECTOR) with 2 array argument from each type (R4/R8); |
368 |
|
- fix globalFile and singleCpuIO options using simpler global mapping. |
369 |
|
- new pairs of scatter/gather vector S/R (one for each type, generated from |
370 |
|
template) corresponding to simpler global mapping. |
371 |
|
- adread_write.F: switch to new S/R (but leave old calls commented) and write |
372 |
|
to globalFile when parameter globalFiles is set to True. |
373 |
|
- switch on useSingleCpuIO in AD experiments: global_ocean.90x40x15 & isomip. |
374 |
|
o pkg/autodiff: |
375 |
|
- adread_adwrite.F: add a hack for RS arrays (when RS=real*4); |
376 |
|
with ALLOW_AUTODIFF_WHTAPEIO defined, since corresponding (real*4) is |
377 |
|
missing, just add a STOP if array type do not match. |
378 |
|
o pkg/exf: |
379 |
|
- to allow to use seaice-fraction from pkg/exf in other pkgs than pkg/seaice: |
380 |
|
rename CPP option EXF_ALLOW_SEAICE_RELAX to EXF_SEAICE_FRACTION |
381 |
|
and rename array obsSIce to exf_iceFraction. |
382 |
|
o tutorial_deep_convection: |
383 |
|
- add 1rst version of isotropic 3-D Smagorinsky code (from Louis-Philippe), |
384 |
|
for now all in this exp. code dir. |
385 |
|
- add secondary test using Smag-3D viscosity (keep very low backgroud visc for |
386 |
|
testing) and flux-limter advection scheme for Temp. (no explicit diffusion). |
387 |
|
o eesupp/src: |
388 |
|
- added SINGLE_DISK_IO option because reading of parameter files on |
389 |
|
very large number of processors did not work with existing set-up |
390 |
|
|
391 |
|
checkpoint64o (2013/09/25) |
392 |
|
o pkg/atm_phys & atm_compon_interf: |
393 |
|
- allows for coupling of gray-atmosphere (atm_phys) with ocean. |
394 |
|
o tools/build_options: |
395 |
|
- switch from "-fPIC" to "-mcmodel=medium" in optfiles: linux_amd64_gfortran |
396 |
|
& linux_amd64_ifort11 and in linux_amd64_g77 & linux_amd64_open64 also |
397 |
|
remove CPP setting (was identical to genmake2 default). |
398 |
|
o pkg/gmredi & model/src: Some updating and work done on the residual model. |
399 |
|
- created new file model/src/calc_eddy_stress.F to calculate the eddy stress. |
400 |
|
- fixed up the calculation of the eddy stress and how it is added to the |
401 |
|
external forcing via taueddy_external_forcing.F |
402 |
|
- at the moment, it is very experimental and only works with the variable eddy |
403 |
|
diffusivity calculated in gmredi_k3d.F |
404 |
|
o pkg/seaice/seaice_check.F: add a check to catch the case of specifying |
405 |
|
lw/swfluxfile instead of lw/swdownfile in data.exf together with pkg/seaice |
406 |
|
o testreport (& tools/do_tst_2+2): |
407 |
|
- for MPI test, set default COMMAND to 'mpirun -v -np TR_NPROC ./{executable}' |
408 |
|
o tools/build_options: linux_amd64_gfortran & linux_amd64_ifort11: |
409 |
|
- remove CPP='cpp -traditional -P' (this is genmake2 default) |
410 |
|
- add switch (for big setups) to use 'medium' memory-model (but leave it |
411 |
|
commented out for now). Note: needed with FC 19 to compile big setups. |
412 |
|
o pkg/atm_compon_interf: |
413 |
|
- regroup per field origin (model & pkgs) the coupling storage routines |
414 |
|
(which accumalate in time each coupling field) |
415 |
|
- move bi,bj loops inside atm_store_my_data.F and store also wind-stress; |
416 |
|
update accordingly + simplify aim_do_physics.F |
417 |
|
|
418 |
|
checkpoint64n (2013/08/26) |
419 |
|
o pkg/bbl: |
420 |
|
- update formulation to reduce dispersion of tracer properties in bbl |
421 |
|
- replaced horizontal and vertical transport parameters with velocities |
422 |
|
- changes results for verification/global_with_exf.yearly |
423 |
|
o verification/halfpipe_streamice: |
424 |
|
- change to matrix construction; update both Fwd & AD output. |
425 |
|
o verification/tutorial_tracer_adjsens: |
426 |
|
- OpenAD test: switch to vector-invariant momentum since fluxform with |
427 |
|
rStar causes problems with OpenAD; update output. |
428 |
|
o tools: |
429 |
|
- genmake2: define a common prefix for all temp-file names ; move them |
430 |
|
to local /tmp disk (but would be easy to move them back); |
431 |
|
- xmakedepend: apply the same switch for temp-files. |
432 |
|
o pkg/offline: |
433 |
|
- changes for use with rStar (NLFS): assume that we continue to provide |
434 |
|
UVELMASS/hFacW_ini & VVELMASS/hFacS_ini (hFacWi,S_ini as written in grid |
435 |
|
output files) and apply r* scaling at each time-step (function of the time |
436 |
|
varying Eta field); use normal pickup files (needed for restart with NLFS). |
437 |
|
o adjoint with OpenAD: |
438 |
|
- move OpenAD-specific files from verification/OpenAD/code_oad_all to a new |
439 |
|
package openad. This package has to be compiled for all runs with OpenAD |
440 |
|
instead of specifying code_oad_all as a modification directory. |
441 |
|
- link the fortran templates from tools/OAD_support to the build directory |
442 |
|
for preprocessing to avoid conflicts between different setups compiled |
443 |
|
simultaneously and simplify cleanup. |
444 |
|
o pkg/exf: |
445 |
|
- in exf_getforcing.F: comment out most of the EXCH calls (fields used by |
446 |
|
main model or other pkgs are exchanged in EXF_MAPFIELDS). |
447 |
|
o model/src: |
448 |
|
- do_stagger_fields_exchanges.F: skip EXCH calls when using pkg/offline |
449 |
|
(similar to do_fields_blocking_exchanges.F). |
450 |
|
o tools/genmake2: |
451 |
|
- remove TAF output src code just before calling staf (when not doing a fresh |
452 |
|
built, will stop if things go wrong with staf). |
453 |
|
- fix target "adobj" (left from TAF version-2 transition) |
454 |
|
o pkg/diagnostics: |
455 |
|
- add a parameter to track the status of pkg/diagnostics activation; |
456 |
|
this replace/extend the use of logical param "settingDiags" (to check |
457 |
|
when adding diag to the list) to also check any DIAGNOSTICS_[]FILL* call. |
458 |
|
New S/R to print error message and stop when pkgStatus is not right. |
459 |
|
- to try: add a warning when "data.diagnostics" exist but useDiagnostics=F. |
460 |
|
|
461 |
|
checkpoint64m (2013/08/13) |
462 |
|
o model/inc: |
463 |
|
- move FCORI_MACROS.h from PARAMS(_MACROS).h to GRID_MACROS.h (left from |
464 |
|
8 yrs ago modif when Coriolis arrays were moved from PARAMS.h to GRID.h) |
465 |
|
o pkg/kpp & aim_v23: |
466 |
|
- only call DIAGNOSTICS_FILL when useDiagnostics=TRUE. |
467 |
|
o pkg/seaice: |
468 |
|
- restore (last time on 2011/12/24) ability to compile pkg/seaice with |
469 |
|
undefined pkg/exf ALLOW_ATM_TEMP or ALLOW_RUNOFF CPP-options. |
470 |
|
- only call DIAGNOSTICS_FILL when useDiagnostics=TRUE. |
471 |
|
o eesupp/src: |
472 |
|
- fix check for overlap-size argument (myOLn.GT.MAX_OLY_EXCH) |
473 |
|
in exch1_rx[_ad].template (thanks to Francois Lefeuvre report). |
474 |
|
o switch to TAF version-2: |
475 |
|
- add flow directives to "freeze" naming convention for ad-routines that |
476 |
|
are used in I/O routines of adjoint state and associated common blocks |
477 |
|
that are used in adcommon.h (backward compatible with version-1) |
478 |
|
- update all ad-optfiles (in tools/adjoint_options + few local to some |
479 |
|
verification exp) to switch to latest TAF version (currently 2.5.5); |
480 |
|
merge "*_mpi" adoptfiles into the corresponding non-mpi (but commented |
481 |
|
out except for diva). |
482 |
|
- the major change with TAF version-2 is that both the differentiated |
483 |
|
routines and the forward routines (new) are returned as part of the TAF |
484 |
|
output source code. Update genmake2 accordingly. |
485 |
|
- new output for AD experiment tutorial_global_oce_biogeo (better AD gradient |
486 |
|
with version-2 when compared to TLM or FWD gradient); other experiments |
487 |
|
output show minor changes (machine truncation level); update the few AD & |
488 |
|
TLM output with 12 matching digits only (some prior to version-2 switch). |
489 |
|
|
490 |
|
checkpoint64l (2013/08/05) |
491 |
|
o model/src: |
492 |
|
- in dynamics.F, skip the call to CALC_VISCOSITY if momViscosity=F |
493 |
|
- main_do_loop.F: no longer includes additional (AUTODIFF) header files |
494 |
|
when using TAF but only if using OpenAD. |
495 |
|
o pkg/mom_common+fluxform+vecinv: |
496 |
|
- always set horiz. viscosity arrays to background value before calling |
497 |
|
MOM_CALC_VISC (in MOM_FLUXFORM & MOM_VECINV) and call S/R MOM_CALC_VISC |
498 |
|
only when using variable horiz. viscosity (useVariableVisc=T); |
499 |
|
- simplify mom_vecinv.F (only 1 block for momViscosity). |
500 |
|
o pkg/mom_common: variable horiz. viscosity in NH code |
501 |
|
- move horiz-viscosity 3-D arrays viscAh_W & viscA4_W (used in calc_gw.F) |
502 |
|
from model/inc/NH_VARS.h to pkg/mom_common/MOM_VISC.h |
503 |
|
- store in common block (in MOM_VISC.h) logical parameters: useHarmonicVisc, |
504 |
|
useBiharmonicVisc & useVariableVisc. |
505 |
|
- fix old bug (already in 1rst implementation of variable horizontal viscosity |
506 |
|
in calc_gw.F, version 1.27, 2006/06/20): |
507 |
|
horizontal bi-harmonic viscosity in gW was missing when wiscA4W=0 |
508 |
|
even if variable bi-harmonic viscosity (e.g., viscA4Grid) was non-zero. |
509 |
|
- change background horiz viscosity in vertical momentum (gW) to always be |
510 |
|
viscAhW,viscA4W (instead of switching to viscAhD, viscA4D when using |
511 |
|
any type of variable viscosity, e.g., viscAh/4Grid) |
512 |
|
- change default value of viscAhW,viscA4W to viscAhD,viscA4D (instead of |
513 |
|
viscAh,viscA4) which default to viscAh,viscA4 : this avoid missing |
514 |
|
horiz. viscosity in W eq when just setting viscAhD/Z,viscA4D/Z (since |
515 |
|
default viscAh,viscA4 is zero). |
516 |
|
o model/src: |
517 |
|
- extend length of drC & recip_drC from Nr to Nr+1. |
518 |
o Divided Adjoint (DIVA) with MPI: |
o Divided Adjoint (DIVA) with MPI: |
519 |
- genmake2: do not create local ./mpi_headers dir when no file to copy into. |
- genmake2: do not create local ./mpi_headers dir when no file to copy into. |
520 |
- remove option ALLOW_DIVIDED_ADJOINT_MPI (replaced by ALLOW_DIVIDED_ADJOINT |
- remove option ALLOW_DIVIDED_ADJOINT_MPI (replaced by ALLOW_DIVIDED_ADJOINT |