| 1 |
$Header$ |
$Header$ |
| 2 |
|
$Name$ |
| 3 |
|
|
| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o add HAVE_SETRLSTK define and useSETRLSTK flag (eedata) which calls |
| 8 |
|
a C routine to unlimit the stack size |
| 9 |
|
o mnc-ify pkg/aim snapshot vars |
| 10 |
|
o mnc-ify pkg/land as requested by Daniel |
| 11 |
|
o various changes to mnc including: |
| 12 |
|
- all files use the new "BASENAME[[.ITER].{t|f}NUM].nc" format |
| 13 |
|
- output can now be grouped so that all files within a group |
| 14 |
|
change the ITER portion of their names in lock-step together |
| 15 |
|
- can now read ("global") PER-FACE (in addition to PER-TILE) files |
| 16 |
|
and works with both EXCH1 and EXCH2 (but needs more testing) |
| 17 |
|
- writing works for all verification test cases w/ g77 on Linux |
| 18 |
|
|
| 19 |
|
checkpoint57r_post |
| 20 |
|
o make MNC variable initialization happen earlier |
| 21 |
|
o add diagnostics for (almost) each term of the momentum Eq. |
| 22 |
|
o remove SST,SSS,UVEL_k2,VVEL_k2 diagnostics. |
| 23 |
|
o move pkg timeave out of pkg-groups "oceanic" & "atmospheric" (pkg/pkg_groups) |
| 24 |
|
o remove MINIMAL_TAVE_OUTPUT option ; |
| 25 |
|
add "dumpInitAndLast" to replace "nodumps" (and MINIMAL_TAVE_OUTPUT). |
| 26 |
|
o move some basic S/R (used elsewhere in the code) from pkg/mdsio to eesupp/src |
| 27 |
|
allows to compile without pkg/mdsio (& test successfully hs94.1x64x5) |
| 28 |
|
o add MNC output for adjoint of EXF (adexf), SEAICE (adseaice) |
| 29 |
|
o add time-dependent SST, SSS control |
| 30 |
|
o cost terms in ecco/ modularized (new cost_generic.F) |
| 31 |
|
o cost_averages routines are modularized (new cost_averagesgeneric.F) |
| 32 |
|
o new infrastructure for seaice model vs. obs misfits in seaice/ |
| 33 |
|
* driver from the_main_loop.F is seaice_cost_driver.F |
| 34 |
|
* first term is AREA vs. SMR observation misfit (ALLOW_SEAICE_COST_SMR_AREA) |
| 35 |
|
* gradient check in Exp. 6a of natl_box_adjoint/ |
| 36 |
|
o natl_box_adjoint input_seaice for Lab. Sea now available online via wget |
| 37 |
|
(see README file) |
| 38 |
|
|
| 39 |
|
checkpoint57q_post |
| 40 |
|
o fix initialisation Pb in calc_gw (get NANs in exp5 with g77) |
| 41 |
|
o fix annoying months-old bug in mnc/mdsio output for seaice pkg |
| 42 |
|
- bug was reported by Jean-Michel and Ian Fenty |
| 43 |
|
o fix initialization Pb in AIM-Ocean coupled model (aim_sice2aim.F) |
| 44 |
|
update exp. cpl_aim+ocn output. |
| 45 |
|
o Implement 4-level checkpointing: AUTODIFF_4_LEVEL_CHECKPOINT |
| 46 |
|
Needed for high-res. runs with large global files |
| 47 |
|
to be able to limit tapelev size |
| 48 |
|
o Changed some tests for adjoint: |
| 49 |
|
* carbon/ now tests AUTODIFF_2_LEVEL_CHECKPOINT |
| 50 |
|
* natl_box_adjoint/code_bulk/ tests AUTODIFF_4_LEVEL_CHECKPOINT |
| 51 |
|
* global_ocean.90x40x15/ back to default 3-level checkpointing |
| 52 |
|
o Modifications to enable scripted COMMON to MODULE conversion |
| 53 |
|
* verification/OpenAD/code/ has necessary scripts and description; |
| 54 |
|
* converts, compiles and runs successfully; |
| 55 |
|
|
| 56 |
|
checkpoint57p_post |
| 57 |
|
o reading of delXfile & delYfile moved from ini_parms.F to ini_grid.F |
| 58 |
|
o move ${PKG}_MNC_INIT from ${PKG}_READ_PARAMS to ${PKG}_INIT_FIXED |
| 59 |
|
(already the case for some pkgs, including recent MNC_init, e.g. thsice) |
| 60 |
|
o call CALC_GW from DYNAMICS (instead of from FORWARD_STEP) |
| 61 |
|
o Dissipation & phiHyd gradient are always added to gU,gV in timestep.F |
| 62 |
|
(was already the case for dissipation with mom_vecinv, |
| 63 |
|
and also the case for grad.PhiHyd when staggered-timeStep) |
| 64 |
|
This will allow to put dissipation out-off the AB time-stepping. |
| 65 |
|
Unfortunately, affects truncation error. Update output of exp. that "fails" |
| 66 |
|
(cg2d matching number) and leave the others for later: front_relax(10), |
| 67 |
|
global_ocean_pressure(10),ideal_2D_oce(10),internal_wave(9),lab_sea(12), |
| 68 |
|
matrix_example(7!),natl_box(12),plume_on_slope(11),vermix(12x3,11x2). |
| 69 |
|
|
| 70 |
|
checkpoint57o_post |
| 71 |
|
o fix MNC/MDSIO flags in PTRACERS so that (1) it works, and |
| 72 |
|
(2) its ready to be documented during the 20050801 DocFest |
| 73 |
|
o MNC-ify the gmredi package -- tested and works w/ and w/o mnc |
| 74 |
|
for ideal_2D_oce and lab_sea on linux_ia32_g77 |
| 75 |
|
o set to #undef CTRL_SET_OLD_MAXCVARS_30 |
| 76 |
|
in global1x1_tot/code_seaice_adj to access indices 32, 33 |
| 77 |
|
(worth another tag) |
| 78 |
|
|
| 79 |
|
checkpoint57n_post |
| 80 |
|
o adding precip control xx_precip 32 |
| 81 |
|
o adding swflux control xx_swflux 33 |
| 82 |
|
o Replacing 10 forcing penalty cost routines by 1 generic. |
| 83 |
|
o Implementing partitioning in time-mean and variable adjustment |
| 84 |
|
for atmos. state. |
| 85 |
|
o Time-mean constant error is now read in data.ecco (instead of hard-coded) |
| 86 |
|
o in natl_box_adjoint: |
| 87 |
|
* updated exp 1 to 10 |
| 88 |
|
(but does not have time-mean vs. variable split results yet) |
| 89 |
|
o updated global1x1/*_seaice_adj/ |
| 90 |
|
to incorporate latest changes |
| 91 |
|
o enable useSingleCpuIO for mdsio vector routines |
| 92 |
|
(applies in particular to tapelev I/O) |
| 93 |
|
o removed inefficient K-loop in all active_file*control*.F |
| 94 |
|
|
| 95 |
|
checkpoint57m_post |
| 96 |
|
o re-arrange exp. adjustment.cs-32x32x1 (adjust_nlfs.cs-32x32x1 merge into) |
| 97 |
|
and advect_cs (use EXCH-2 and standard cs-32 grid files). |
| 98 |
|
o pkg/exch2: filling of face-corner halo regions is optional |
| 99 |
|
(ifdef W2_FILL_NULL_REGIONS) and is turned off by default. |
| 100 |
|
|
| 101 |
|
checkpoint57l_post |
| 102 |
|
tag before updating tutorials (pre-release-2) |
| 103 |
|
o add cos & sin of grid-direction orientation angle (curvilinear grid only, |
| 104 |
|
for output and interpolation of input fields onto model grid) |
| 105 |
|
o fix radius in mom_v_metric_cylinder.F (wrong by dy/2). |
| 106 |
|
Affects results of exp. rotating_tank ; update output.txt. |
| 107 |
|
o pkg/diagnostics: 2 new S/R to fill a diagnostics using a scaling factor |
| 108 |
|
and taking the square quantity and/or using a fraction-weight field. |
| 109 |
|
o fix aim_diagnostics_init.F mate numbers. |
| 110 |
|
o add mnc output capability to diagnostics/diagstat and update |
| 111 |
|
our cvsignore files for the ACSII output generated |
| 112 |
|
|
| 113 |
|
checkpoint57k_post |
| 114 |
|
o lab_sea now uses diagnostics package and provides a suite of |
| 115 |
|
new EXF diagnostics output. |
| 116 |
|
|
| 117 |
|
checkpoint57j_post |
| 118 |
|
o exf |
| 119 |
|
* bug fix in exf_bulkformulae for psimh (spotted by Eliyahu Biton): |
| 120 |
|
Bracket missing in eqn. for psimh (eqn. 7 of Large & Pond, 1982); |
| 121 |
|
terms -2*ATAN(x) + pi/2 were added to both stable and unstable case. |
| 122 |
|
* extended diagnostics output |
| 123 |
|
- now includes all EXF variables |
| 124 |
|
- hs, hl moved to common block |
| 125 |
|
- separate S/R for diagnostics, called after ctrl vector is added |
| 126 |
|
- name changed for EXFsw -> EXFswnet, EXFlw -> EXFlwnet |
| 127 |
|
- EXFqnet contains true Qnet (i.e. includes qsw) |
| 128 |
|
* updated verif. lab_sea, global_with_exf |
| 129 |
|
* Rearranging hflux (exf_mapfields -> exf_getforcing) |
| 130 |
|
to ensure that diagnostics sees a true EXFqnet |
| 131 |
|
o add matlab scripts for CS-grid vectors (in utils/cs_grid) |
| 132 |
|
o diagnostics pkg: change pointers so that 1 diagnostic can be used |
| 133 |
|
several times with different frequencies. |
| 134 |
|
o mnc-ify the thsice package as requested by Daniel Enderton |
| 135 |
|
- the monitor--MNC output needs work |
| 136 |
|
- many attributes need to be added (most are currently blank) |
| 137 |
|
o mnc-ify theaim_v23 package as requested by Daniel Enderton |
| 138 |
|
- many attributes need to be added (most are currently blank) |
| 139 |
|
o remove "groundAtK1" & "rkFac" (replaced by usingPCoords & -rkSign) |
| 140 |
|
o add Sadourny like discretization to the 4th.O vorticity advection scheme |
| 141 |
|
o fix bug in rStar code (calc_r_star.F) when used with topo/bathy. |
| 142 |
|
update output.txt from: aim.5l_cs (& .thSI), cpl_aim+ocn, |
| 143 |
|
fizhi-gridalt-hs and global_ocean.cs32x15 (& .thsice). |
| 144 |
|
o fix few little things to avoid floating error with -ftrapuv (ifort v9) |
| 145 |
|
(ini_spherical_polar_grid.F, calc_r_star.F, monitor/mon_surfcor.F) |
| 146 |
|
|
| 147 |
|
checkpoint57i_post |
| 148 |
|
o Diagnostics package was changed: diagnostic allocation is now deactivated |
| 149 |
|
by setting an empty file name in the namelist (used to be by setting 0 frequency) |
| 150 |
|
o momentum advection (VI form) options: activate 4th.O vorticity Coriolis ; |
| 151 |
|
add 1rst Order upwind vertical Shear. |
| 152 |
|
o add a quick fix for A/B checkpointing in OFFLINE mode |
| 153 |
|
o add IGNORE_TIME define and -ignore_time option to genmake2 as an |
| 154 |
|
end-run around the C/Fortran linking horror in eesup/src/timers.F |
| 155 |
|
o added pkg/runclock: allows user to terminate run based on wall clock |
| 156 |
|
o added non-hydrostatic scaling parameter: nh_am2 |
| 157 |
|
o pkg/my82: futile attempt to make results compiler-independent by adding |
| 158 |
|
some "_d 0", etc. no effect. |
| 159 |
|
o for mnc output, fill the 'T' coordinate var with myTime and create a |
| 160 |
|
separate 'iter' variable for iteration count |
| 161 |
|
o pkg/diagnostics: |
| 162 |
|
add new capability: output of Global/Regional & per-level statistics ; |
| 163 |
|
tested against monitor output (including land & thSIce) on various exp. |
| 164 |
|
o make seaice-pkg I/O type flags (MNC/MDS) more standard. |
| 165 |
|
|
| 166 |
|
checkpoint57h_done |
| 167 |
|
tag after changing user customized files (DIAGNOSTICS_SIZE.h, |
| 168 |
|
data.diagnostics, data.gchem, data.ptracers) is done. |
| 169 |
|
|
| 170 |
|
o pkg/diagnostics: prepare for Global/Regional statistics. |
| 171 |
|
add 2nd namelist in data.diagnostics and update DIAGNOSTICS_SIZE.h |
| 172 |
|
o remove "baseTime" (no used) from arg. list of DIFF_BASE_MULTIPLE |
| 173 |
|
and rename it: DIFFERENT_MULTIPLE |
| 174 |
|
|
| 175 |
|
checkpoint57h_post |
| 176 |
|
o move tIter0 from data.gchem to PTRACERS_Iter0 in data.ptracers |
| 177 |
|
o pkg/diagnostics: change output frequency, now in seconds (instead |
| 178 |
|
of Nb of iterations). This requires to change data.diagnostics. |
| 179 |
|
checkpoint57h_pre |
| 180 |
|
|
| 181 |
|
o in mnc, let the user create the output dir if both mnc_outdir_date |
| 182 |
|
and mnc_outdir_num are false |
| 183 |
|
o add fflags() to pkg/diagnostics which allows one to specify, on a |
| 184 |
|
per-file basis, the file precision [default behavior is unchanged] |
| 185 |
|
o added diagnostics to pkg/kpp, including computation of mixed layer |
| 186 |
|
depth based on a temperature/density criterion |
| 187 |
|
o updated verification/natl_box to test the new pkg/kpp diagnostics |
| 188 |
|
o bugs fixed (with useAbsVorticity=T & with useJamartMomAdv=T) |
| 189 |
|
o test AB-3 code with 4th.Order advection scheme (advect_xy/input.ab3_c4) |
| 190 |
|
o fix mnc flags in pkg/seaice so that they are "owned" by seaice, |
| 191 |
|
not mnc -- this mistake was found by JMC and is now corrected |
| 192 |
|
|
| 193 |
|
checkpoint57g_post |
| 194 |
|
o add Adams-Bashforth 3rd Order time stepping: |
| 195 |
|
- default remains AB-2 (#undef ALLOW_ADAMSBASHFORTH_3) |
| 196 |
|
- not yet implemented with NH, CD_Scheme, pTracers, MNC_pickup ... |
| 197 |
|
checkpoint57g_pre |
| 198 |
|
|
| 199 |
|
o make lambda[Theta,Salt]ClimRelax spatially varying |
| 200 |
|
(default remains spatially uniform constant based on tau) |
| 201 |
|
o added new control variables |
| 202 |
|
* init. uVel, vVel, etanN |
| 203 |
|
* lambda[Theta,Salt]ClimRelax |
| 204 |
|
o separate masks used for ctrl_pack/unpack 'from write_grid' output |
| 205 |
|
(suggested by G. Forget) |
| 206 |
|
|
| 207 |
|
checkpoint57f_post |
| 208 |
|
o add parameter "baseTime" = model base time = time at iteration zero |
| 209 |
|
new function DIFF_BASE_MULTIPLE (replace DIFFERENT_MULTIPLE) to |
| 210 |
|
account for baseTime shift |
| 211 |
|
o move Coriolis arrays from PARAMS.h to GRID.h |
| 212 |
|
checkpoint57f_pre |
| 213 |
|
|
| 214 |
|
o mnc-ifying the seaice package |
| 215 |
|
o introduced an alternative length scale calculation for use in the limiting |
| 216 |
|
maximum viscosity (off by default, see mom_vi_hdissip.F for details) |
| 217 |
|
o S/R ptracers_calc_diff: argument list bug for CALL KPP_CALC_DIFF_S |
| 218 |
|
o added modified Leith scheme with viscosity proportional to grad(Hdiv) |
| 219 |
|
as well as grad(vort3). Made it compatible with both approximate form |
| 220 |
|
for the cubed sphere and full form of these gradients. Added flag |
| 221 |
|
useFullLeith to select between approximate and full form. |
| 222 |
|
o add "d0" in getcon.F to reduce differences between compilers/platforms |
| 223 |
|
update output.txt of exp. fizhi-gridalt-hs |
| 224 |
|
o change port_rand funtion to allow to set the seed. |
| 225 |
|
o fix obscure MNC bug involving conversion between reals and integers |
| 226 |
|
- it caused no problems with certain (eg. gcc, ifc) compilers but the |
| 227 |
|
new version is now more standards-compliant |
| 228 |
|
o adjoint-related modifications |
| 229 |
|
* gad_dst3_adv_?: to be able to use advscheme=30 in adjoint |
| 230 |
|
* gad_calc_rhs: fool adjoint, revert to advscheme=30 in reverse when using 33 |
| 231 |
|
* mom_vi_hdissip: to be able to use Leith scheme in adjoint |
| 232 |
|
|
| 233 |
|
checkpoint57e_post |
| 234 |
|
o Adding eddy stress controls a la Ferreira et al. |
| 235 |
|
checkpoint57e_pre (actually was tagged: eckpoint57e_pre) |
| 236 |
|
|
| 237 |
|
o fizhi pkg: fix AtoC and CtoA when called with same input & output array |
| 238 |
|
update output.txt of exp. fizhi-gridalt-hs |
| 239 |
|
o add more "Z" dimensions and associated coordinate vars to MNC |
| 240 |
|
o pkg/exf: added twoDigitYear capability for option useExfYearlyFields |
| 241 |
|
o parameter calendarDumps: when set, approximate months (30-31 days) and years |
| 242 |
|
(360-372 days) for parameters chkPtFreq, pChkPtFreq, taveFreq, SEAICE_taveFreq, |
| 243 |
|
KPP_taveFreq, and freq in pkg/diagnostics are converted to exact calendar |
| 244 |
|
months and years. Requires pkg/cal. |
| 245 |
|
o diagnostics: add pickup functionality |
| 246 |
|
- off by default for backwards compatibility |
| 247 |
|
- current version only uses MDSIO -- MNC will be added |
| 248 |
|
o allow to use pkg diagnostics for LAND pkg output & Shapiro Tendencies. |
| 249 |
|
o pkg/shap_filt: fix for multi-tiles per face (CS-grid with EXCH2) |
| 250 |
|
|
| 251 |
|
checkpoint57d_post |
| 252 |
|
o pkg/thsice used in the ocean: fix tile Pb related to sIceLoad forcing |
| 253 |
|
o modify verification/ideal_2D_oce so that it tests the creation of mnc |
| 254 |
|
output from the diagnostics package |
| 255 |
|
o temporary fix in EXCH2_UV: add a 3rd call to fix the 24.tiles CS-grid set-up |
| 256 |
|
o make genmake2 more portable by using ".$FS" instead of ".f" for more |
| 257 |
|
of the test compilations (eg. IBM xlf on Mac OSX w/HFS filesys) |
| 258 |
|
o add coordinate variables to the MNC output from the diagnostics package |
| 259 |
|
o change MNC-flags default value: true when useMNC=TRUE |
| 260 |
|
o add a "readgrid_mnc" flag to MNC and give model/src/ini_curvilinear_grid.F |
| 261 |
|
the ability (off by default) to read the grid information using MNC |
| 262 |
|
o lab_sea: use LSR_ERROR=1.E-12 for testing ; update output.txt |
| 263 |
|
o diagnostics: fix a bug (sub-set of levels); keep full precision when averaging |
| 264 |
|
o do snap-shot diagnostics at the middle of the time-interval (instead of |
| 265 |
|
at the end). |
| 266 |
|
o update exp. aim.5l_cs (use default LAND_OPTIONS.h file) and add |
| 267 |
|
new test for AIM with thSIce + slab-ocean |
| 268 |
|
o cpl_aim+ocn test: update output (optfile & default GMREDI_OPTIONS.h) |
| 269 |
|
o move state variable diagnostics to the beginning of the time step ; |
| 270 |
|
(note: still needs some work specially for snap-shot diagnostics) |
| 271 |
|
duplicate & replace UVELVVEL diagnostics with UV_VEL_C & UV_VEL_Z |
| 272 |
|
|
| 273 |
|
checkpoint57c_post |
| 274 |
|
o Change testing option (-ieee) to get closer agreement between g77 & ifort |
| 275 |
|
o generate new output.txt files on faulks with g77 & the new -ieee option |
| 276 |
|
checkpoint57c_pre |
| 277 |
|
|
| 278 |
|
o add "_d 0" in freeze ; change some "FLOAT" to "DFLOAT" and went back (TAF); |
| 279 |
|
update output.txt of exp. global_ocean.90x40x15 |
| 280 |
|
o fix in ini_eos.F (one "_d 0" was missing): now g77 & ifc give much closer |
| 281 |
|
results; update global_ocean_pressure & global_ocean.cs32x15 output.txt |
| 282 |
|
o add saltFlux(+File) forcing capability with periodicExternalForcing |
| 283 |
|
o allow to use pkg diagnostics for GM-Redi & ThSIce output |
| 284 |
|
o diagnostics_fill: fill in up to sNx+1 or sNy+1 (if U,V or Z pt) for mnc. |
| 285 |
|
o improvements for the MNC output from DIAGNOSTICS: |
| 286 |
|
- correct handling of the unlimitied dimension |
| 287 |
|
- all dims now have associated coordinate variables |
| 288 |
|
- variables located at U, V, and Corner ("Z") points are now written |
| 289 |
|
with the correct horzontal dimensions (eg. 'Xp1', 'Yp1') so that |
| 290 |
|
*all* data points are output |
| 291 |
|
o added seaice_summary.F and removed obsolete ALLOW_SEAICE's from pkg/seaice |
| 292 |
|
o added pkg/seaice/seaice_monitor.F |
| 293 |
|
o mdsio_readfield.F can now read files visible only to master MPI process. |
| 294 |
|
o pkg/seaice: some test code for ice-modified ocean stress |
| 295 |
|
o exf_getffields interpolation defaults to bilinear for all scalar forcing |
| 296 |
|
fields (remains bicubic for wind velocity and stress). This avoids, e.g., |
| 297 |
|
spurious negative numbers for precipitation and humidity. Will cause |
| 298 |
|
some small numerical differences for integrations using pkg/exf/exf_interp.F. |
| 299 |
|
o pkg/ctrl/ |
| 300 |
|
* Add ctrlvec diagnostics in pack/unpack for nondimensional I/O |
| 301 |
|
May be enabled via doPackDiag |
| 302 |
|
* introduce z-scaling of |
| 303 |
|
+ gradient (doZscalePack) and |
| 304 |
|
+ control (doZscaleUnpack) |
| 305 |
|
o pkg/mdsio/ |
| 306 |
|
* enable useSingleCpuIO for state estimation (ctrl_pack/unpack) mdsio_gl.F |
| 307 |
|
* N.B.: leave tiled I/O for tapelev (i.e. mds...vector) |
| 308 |
|
|
| 309 |
|
checkpoint57b_post |
| 310 |
|
o simplify options in diagnostics_fill arguments (remove weird combinations) |
| 311 |
|
o fix MNC coordinate variables so that they now contain per-face index |
| 312 |
|
values (instead of xC,xG,yC,yG values) when run with EXCH2 since, in |
| 313 |
|
general, EXCH2 grids are only aligned to a "local" coordinate system |
| 314 |
|
o test diagnostics pkg in dic_example (no output) and ideal_2D_oce (with ouput) |
| 315 |
|
o Implement instantaneous output within pkg diagnostics: |
| 316 |
|
setting a negative frequency in data.diagnostics produces |
| 317 |
|
"snap-shot" output (as opposed to time-average if freq > 0) |
| 318 |
|
o move out off EQUIVALENCE all diagnostics that fizhi does not use ; |
| 319 |
|
NOTE: change ptracer diagnostics name: use 2 characters to work with |
| 320 |
|
more than 9 tracers. Add also diagnostics of advect & diffus flux. |
| 321 |
|
o more CF compliance: convert all MNC 'iter' variables to 'T' so that it |
| 322 |
|
becomes the coordinate variable for the 'T' dimension |
| 323 |
|
o add CF-style coordinate variables to MNC |
| 324 |
|
- just a first cut: numbers are meaningless on the cubesphere and |
| 325 |
|
missing along the T axis but otherwise it works! |
| 326 |
|
o add surface flux diagnostics: TAUX, TAUY, TFLUX, TRELAX, TICE, SFLUX, SRELAX |
| 327 |
|
o pkg/thSIce: read initial state from 6 files ; allow monitor to use MNC. |
| 328 |
|
o add diagnostics of advective & diffusive flux for T & S (pTracers coming soon) |
| 329 |
|
this is based directly on what the model compute, so that the content of |
| 330 |
|
each diagnostics might be affected by switching options (e.g., implicitDiffusion) |
| 331 |
|
o small changes for diagnostics (gad_biharm_?.F, arguments of S/R impldiff.F) |
| 332 |
|
o clean up build options files for SunOS, in particular |
| 333 |
|
- remove -traditional from xmakedepend |
| 334 |
|
- change -xarch=v9 to -xarch=native for f77 |
| 335 |
|
- restore use of make (not gmake) |
| 336 |
|
- add CC=gcc if using g77 |
| 337 |
|
o changes per discussion of AM, JMC, & EH3 today: |
| 338 |
|
- move redundant "grid" variables out of "state" -- in preparation of |
| 339 |
|
creating proper CF-style coordinate variables and in recognition |
| 340 |
|
that any/all grid variables can be easily copied (eg. w/NCO) from |
| 341 |
|
the grid files into any of the other NetCDF files |
| 342 |
|
- *always* output grid variables with MDSIO (this will eventually be |
| 343 |
|
changed) |
| 344 |
|
- output MNC grid file if useMNC is true |
| 345 |
|
o every instance of _END_MASTER() has been found and replaced with |
| 346 |
|
_END_MASTER( myThid ) in order to satisfy certain picky Sun |
| 347 |
|
preprocessors |
| 348 |
|
o re-arrange diagnostics pkg initialisation: |
| 349 |
|
- allow each package to extend the available diagnostics list |
| 350 |
|
- add some checking and fix small problems (multi-threaded, ...) |
| 351 |
|
- new S/R diagnostics_fill (replace fill_diagnostics): |
| 352 |
|
look through the short list of active diag. (instead of the long list) |
| 353 |
|
create function DIAGNOSTICS_IS_ON to tell if a diagnostics is active |
| 354 |
|
|
| 355 |
|
checkpoint57a_post |
| 356 |
|
o this set of changes restores TAMC compatibility |
| 357 |
|
tested for global_ocean.90x40x15 adjoint on trough |
| 358 |
|
(what a pain) |
| 359 |
|
|
| 360 |
|
checkpoint57a_pre |
| 361 |
|
o some modifs in tools/build_options for sunos |
| 362 |
|
o call ctrl_pack for ctrl at end for optimcycle=0 |
| 363 |
|
|
| 364 |
|
checkpoint57 |
| 365 |
|
o add horizontal bi-harmonic viscosity for vertical velocity |
| 366 |
|
- new parameter viscA4W (defaults to viscA4) |
| 367 |
|
- affects only non-hydrostatic code |
| 368 |
|
- to be done: no-slip lateral boundary conditions |
| 369 |
|
o switch back to "sflux = sflux - runoff" in exf_getforcing.F |
| 370 |
|
o implement Implicit Vertical advection for pTracers |
| 371 |
|
o depth convergence accelerator: replace deltaTtracer by dTtracerLev(k) |
| 372 |
|
(implemented also for implicit vertical diffusion & advection) |
| 373 |
|
o OBCS: extend application of tracer open boundary conditions |
| 374 |
|
to a strip beyond the boundary of width Olx/Oly |
| 375 |
|
o add call monitor to TAF set of required calls (flow directives) |
| 376 |
|
and include in natl_box_adjoint |
| 377 |
|
o untangle a few ifdefs related to cost, ctrl, autodiff |
| 378 |
|
untangel a few ALLOW_ECCO_OPTIMIZATION |
| 379 |
|
o Apply Cooper & Haines representative estimates also at sub-surface (k=Nr) |
| 380 |
|
(theta2, salt2 in ecco_cost_weights) |
| 381 |
|
|
| 382 |
|
checkpoint56c_post |
| 383 |
|
o OBCS as control variables |
| 384 |
|
- update ctrl_ad_diff.list, obcs_ad_diff.list |
| 385 |
|
- remove balance of obcs controls from default |
| 386 |
|
- fix index bug nobcs in ctrl_init |
| 387 |
|
- fix dummy fields filen in ctrl_pack |
| 388 |
|
- add dummy weights for obcs |
| 389 |
|
o lsopt: comment out fort.94 output |
| 390 |
|
o optim: |
| 391 |
|
- remove unnecessary header files |
| 392 |
|
- adjusted namelists |
| 393 |
|
- add xerbla.F to Makefile |
| 394 |
|
o exf: |
| 395 |
|
- multiple modifications of exf_check_range |
| 396 |
|
- unit change in exf_init_runoff from [m/year] to [m/s] |
| 397 |
|
o mdsio: |
| 398 |
|
- i/o-vector change debugLevel from debLevA to debLevB |
| 399 |
|
|
| 400 |
|
checkpoint56b_post |
| 401 |
|
o collect all ptracers output in an new routine (PTRACERS_OUTPUT) that |
| 402 |
|
is called from DO_THE_MODEL_IO |
| 403 |
|
- rename PTRACERS_STATV_WRITE to PTRACERS_WRITE_TIMEAVE |
| 404 |
|
- introduce new parameter PTRACERS_dumpFreq (defaults to dumpFreq) |
| 405 |
|
- clean up PTRACERS_WRITE_TIMEAVE a little |
| 406 |
|
o add another verification experiment that tests both the cfc-pkg and |
| 407 |
|
the gchem-pkg with GCHEM_SEPARATE_FORCING undefined, that is with |
| 408 |
|
reactive tendencies treated inline with regular timestepping |
| 409 |
|
o testreport previously did a lousy job of comparing small numbers, so |
| 410 |
|
we (Ed and Martin) changed the way the cut-off of 1.e-12 is |
| 411 |
|
implemented in the little c-progam tmp_cmpnum.c which is embedded in |
| 412 |
|
testreport. |
| 413 |
|
o ptracers_monitor: bug fix for MPI |
| 414 |
|
|
| 415 |
|
checkpoint56a_post |
| 416 |
|
o GCHEM: finish reorganizating the package |
| 417 |
|
- forward_step calls GCHEM_CALC_TENDENDY, which computes gchemTendency |
| 418 |
|
(introduces another 3D-array for each passive tracer, but only if |
| 419 |
|
GCHEM_SEPARATE_FORCING is undefined. For GCHEM_SEPARATE_FORCING |
| 420 |
|
gchemTendency is not needed because the timestep is done separately) |
| 421 |
|
- gchemTendency is added to gPtr in GCHEM_ADD_TENDENCY |
| 422 |
|
- GCHEM is now---more or less cleanly---separate from PTRACERS |
| 423 |
|
- to bo done: handling of surfaceForcingPtr that is needed for KPP not |
| 424 |
|
so clear to me how that should be done in a general way. |
| 425 |
|
o CFC: improve formatting of output in CFC_ATMOS |
| 426 |
|
o PTRACERS: |
| 427 |
|
- rename GCHEM_MONITOR to PTRACERS_MONITOR and call it from MONITOR, so |
| 428 |
|
that ALL experiments with ptracers enable can be checked. This makes |
| 429 |
|
GCHEM_MONITOR obsolete. |
| 430 |
|
- include a runtime parameter PTRACERS_monitorFreq that defaults to |
| 431 |
|
monitorFreq |
| 432 |
|
- set default PTRACERS_write_mdsio to false if PTRACERS_write_mnc is true |
| 433 |
|
o add PTRACERS (PTR0[1-5]) to testreport output |
| 434 |
|
o exf_check_range: |
| 435 |
|
* default changed to .TRUE. |
| 436 |
|
* will check values for niter0 if debLevA, every timestep if debLevB |
| 437 |
|
* STOP in exf_check_range if 'out of range' |
| 438 |
|
* range chosen generous to allow for potential extema |
| 439 |
|
TBD: a version with stricter range |
| 440 |
|
permitted range would be runtime parameters |
| 441 |
|
o GMredi: |
| 442 |
|
- fix ldd97 slope limit ; extend valid domain of Psi-Bolus ; |
| 443 |
|
- change S/R GRAD_SIGMA to be valid in the overlap on CS-grid. |
| 444 |
|
- change the default GMREDI_OPTIONS.h: compile everything except Visbeck |
| 445 |
|
o GCHEM: |
| 446 |
|
- move call to GCHEM_FORCING_SEP, so that it is now called before |
| 447 |
|
the blocking exchanges to avoid computing halos in GCHEM_FORCING_SEP |
| 448 |
|
- introduced a useGCHEM runtime flag and applied it everywhere. |
| 449 |
|
- renamed PTRACERS_SEPARATE_FORCING to GCHEM_SEPARATE_FORCING; |
| 450 |
|
this flag is now only visible within GCHEM-pkg routines. |
| 451 |
|
- moved gchem_forcing_int temporarily into ptracers_forcing, this will |
| 452 |
|
be replaced by a 3D array that is computed before thermodynamics; for |
| 453 |
|
this, the tracer ID iTracer had to be included into the parameter list |
| 454 |
|
of ptracers_forcing |
| 455 |
|
- set up forward_step to call gchem_forcing, that will replace |
| 456 |
|
gchem_forcing_int in ptracers_forcing |
| 457 |
|
- cleaning up |
| 458 |
|
o small cleaning of exch2: remove files that are also in eesupp/src |
| 459 |
|
|
| 460 |
|
checkpoint56 |
| 461 |
|
o AIM: read land-Fraction earlier (from ini-fixed) and monitor land at nIter0 |
| 462 |
|
o mnc: cleanup "full" files--fixes out-of-memory problem |
| 463 |
|
o ECCO-GODAE |
| 464 |
|
* model/, pkg/, code/, input/ now complete w.r.t. ECCO-GODAE production setup |
| 465 |
|
* fully merged and updated from SIO code |
| 466 |
|
windstress sign change affects |
| 467 |
|
- exf_inscal_[u/v]stress |
| 468 |
|
- sign of xx_tau[u/v] adjustments |
| 469 |
|
- signs in cost_scat |
| 470 |
|
* ctrl_get_gen: avoid update of tau[u/v] for first few days |
| 471 |
|
* natl_box_adjoint updated and compressed |
| 472 |
|
* 2-level checkpoint switch added |
| 473 |
|
(tested in EXP. 1 of natl_box_adjoint) |
| 474 |
|
* new switch doInitXX to avoid init. of xx_... in optimcycle=0 |
| 475 |
|
o SVD |
| 476 |
|
* first complete, non-hacked version for SVD calculations |
| 477 |
|
* all interfaces clean w.r.t to handling of xx_..., adxx_..., g_xx_... |
| 478 |
|
performed via direct access files admtlm_vector.it???? |
| 479 |
|
(via ctrl_[un]pack) |
| 480 |
|
* optimcycle is used for ARPACK iteration counter |
| 481 |
|
(note: adxx_... is iterated, rather than xx_...; latter is constant!) |
| 482 |
|
* EV problem ok w.r.t. L2-norm in tangent space; need to check other norms |
| 483 |
|
and add linear transforms where necessary |
| 484 |
|
* controls are initial/final SST, SSS (can be extended; need to check |
| 485 |
|
ctrl_set_[un]pack_xyz.F according to ctrl_set_[un]pack_xy.F |
| 486 |
|
o mdsio: filename info to both stderr, stdout |
| 487 |
|
o exf: exf_inscal_runoff: separated default conversion [m/year] -> [m/sec] |
| 488 |
|
from sign change |
| 489 |
|
|
| 490 |
|
checkpoint55j_post |
| 491 |
|
o isolate dissipation tendency (allow to keep it out off AB) in V.I form. |
| 492 |
|
affects (truncation level) output.txt of global_ocean.cs32x15(&.thsice) |
| 493 |
|
o add consistent build directories and .cvsignore files to verification |
| 494 |
|
o added viscA4GridMin and viscA4GridMax for more control on biharmonic viscosity |
| 495 |
|
o make all subroutines 30 or less characters in length for certain older |
| 496 |
|
compilers to work (a thorough search only turns up two routines in |
| 497 |
|
pkg/ptracers) |
| 498 |
|
o allow (vecinv only) a different horiz. viscosity for Divergence and Vorticity; |
| 499 |
|
affects (truncation level) output.txt of global_ocean.cs32x15(&.thsice) |
| 500 |
|
|
| 501 |
|
checkpoint55i_post |
| 502 |
|
o change arrays for total vertical diffusivity (2.D if explicit, |
| 503 |
|
3.D if implicit) ; each ptracer uses its own background value. |
| 504 |
|
o added mnc/NetCDF output capability to ocn_compon_interf |
| 505 |
|
- added bits to verification/cpl_aim+ocn so that it can run on |
| 506 |
|
ITRDA and other clusters without the mpirun "-wd" option |
| 507 |
|
o fixed wayward endif in mom_fluxform.F that messed up cyclindrical |
| 508 |
|
configuration |
| 509 |
|
o give MNC the ability to create additional files rather than |
| 510 |
|
exceeding a specified file size limit "MNC_MAX_FILESIZE" |
| 511 |
|
- the default is just less than 2GB which is a practical limit |
| 512 |
|
on many systems |
| 513 |
|
o Compute the non-local term using either KPP vertical diffusivity (=default) |
| 514 |
|
or the total vertical diffusivity (when KPP_ghatUseTotalDiffus=T) |
| 515 |
|
|
| 516 |
|
checkpoint55h_post |
| 517 |
|
o add verification test "vermix" |
| 518 |
|
o remove all the *_ioinc flags and replace with the single global |
| 519 |
|
outputTypesInclusive flag |
| 520 |
|
o allow to set a vertical profile of vertical diffusivity for T & S |
| 521 |
|
o call PACKAGE_BOOT & PACKAGES_READPARMS just after INI_PARMS |
| 522 |
|
o useMNC moved from data.mnc to data.pkg |
| 523 |
|
|
| 524 |
|
checkpoint55g_post |
| 525 |
|
o Renaming carbon AD experiments to comply with testreport conventions |
| 526 |
|
o Updating adjoint of Held-Suarez hs94.1x64x5 |
| 527 |
|
o Fixing directives for non-ECCO forward_step, after ECCO directives |
| 528 |
|
have been fixed to avoid recomp. of do_oceanic_phys |
| 529 |
|
Works well for global_ocean.90x40x15, |
| 530 |
|
but recomp. issue remains for carbon |
| 531 |
|
o Updating adjoint output for natl_box_adjoint after |
| 532 |
|
pkg/ecco has been modified (see taga c55d_post) |
| 533 |
|
|
| 534 |
|
checkpoint55f_post |
| 535 |
|
o rename to timeave_init_fixed |
| 536 |
|
|
| 537 |
|
checkpoint55e_post |
| 538 |
|
o some delicate re-shuffle of store directives to avoid one |
| 539 |
|
extra call of do_oceanic_physics |
| 540 |
|
affects ecco_the_main_loop.F, do_oceanic_physics.F |
| 541 |
|
checkpoint_lev1_directives.h |
| 542 |
|
o NB: this may break global_ocean adjoint temporarily, |
| 543 |
|
but it is clear how to fix it. Will do later, need this now. |
| 544 |
|
|
| 545 |
|
checkpoint55d_post |
| 546 |
|
o ECCO specific cost function terms (up-to-date with 1x1 runs) |
| 547 |
|
o ecco_cost_weights is modified to 1x1 runs |
| 548 |
|
o modifs to allow observations to be read in as |
| 549 |
|
single file or yearly files |
| 550 |
|
o enable to read exf forcing fields as either |
| 551 |
|
single file or yearly files (flag useExfYearlyFields) |
| 552 |
|
checkpoint55d_pre |
| 553 |
|
|
| 554 |
|
o switch between 2-level and 3-level checkpointing |
| 555 |
|
o add ability of MNC to write local and "partial" (eg. 2D slices where |
| 556 |
|
the full 3D field is never actually stored) arrays to NetCDF files |
| 557 |
|
with the correct (that is, the complete multi-dimensional) set of |
| 558 |
|
array indicies |
| 559 |
|
- used in mom_vecinv() to write the diagFreq output |
| 560 |
|
- tested (demonstrated) in verification/aim.5l_cs |
| 561 |
|
o further mnc-ification |
| 562 |
|
- add verification/lab_sea to the list with mnc setups |
| 563 |
|
- fix undefined mnc v-names in cd_code |
| 564 |
|
- add kpp to mnc output (both snapshot and timeave) |
| 565 |
|
o non-hydrostatic code |
| 566 |
|
- added horizontal viscosity parameter viscAhW for w, that replaces viscAh |
| 567 |
|
in calc_gw.F |
| 568 |
|
- defaults to viscAh |
| 569 |
|
- this is useful, when I want to use viscAh=0 (e.g., only biharmonic |
| 570 |
|
friction or Leith) in non-hydrostatic mode, as long as there is |
| 571 |
|
no biharmonic friction and/or Leigth or equivalent for vertical velocity. |
| 572 |
|
o pkg/cost |
| 573 |
|
extend simple cost functions to treat time mean quantities |
| 574 |
|
(T,U,V) via cost_accumulate_state from cost_tile |
| 575 |
o multidimAdvection on CS-grid: implement the full 3 pass ; overlap needs to |
o multidimAdvection on CS-grid: implement the full 3 pass ; overlap needs to |
| 576 |
be 2 times larger than the minimum size the advection scheme requires. |
be 2 times larger than the minimum size the advection scheme requires. |
| 577 |
now get a perfect conservation of salt in advect_cs test-exp (updated) |
now get a perfect conservation of salt in advect_cs test-exp (updated) |
| 595 |
- changed sponge layer relaxation parameter units for days to seconds |
- changed sponge layer relaxation parameter units for days to seconds |
| 596 |
- turned off sponge layer by default |
- turned off sponge layer by default |
| 597 |
- changed argument list of OBCS_CALC to myTime+deltaTclock |
- changed argument list of OBCS_CALC to myTime+deltaTclock |
| 598 |
|
- fixed about 35 bugs in obcs_external_fields_load |
| 599 |
|
|
| 600 |
checkpoint55c_post |
checkpoint55c_post |
| 601 |
o obcs stuff |
o obcs stuff |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch