| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o pkg/obcs: |
| 8 |
|
- fix bug in initialisation (broken since Oct 1rst modification) |
| 9 |
|
for useOBCSprescribe without useEXF. |
| 10 |
|
|
| 11 |
|
checkpoint61y |
| 12 |
|
o pkg/seaice |
| 13 |
|
- change (once again) the discretization for the evp solver: rather than |
| 14 |
|
squaring the averages of strain rates, average the squares to give |
| 15 |
|
a better approximation to VP behavior (in simple test); this breaks |
| 16 |
|
verification experiments that use EVP (lab_sea and lab_sea.hb87) |
| 17 |
|
--> update outputs |
| 18 |
|
o add CPP-flag (in mom_calc_visc.F) and TAF directive (in gmredi_slope_limit) |
| 19 |
|
to improve vectorization of adjoint code on TARGET_NEC_SX |
| 20 |
|
o verification/vermix: |
| 21 |
|
- change vermix to a more "classical" 1d column with high vertical |
| 22 |
|
resolution, seasonal heat flux forcing and constant wind forcing |
| 23 |
|
--> update outputs |
| 24 |
|
o pkg/seaice |
| 25 |
|
- fix diagnostics of principal stress components for EVP |
| 26 |
|
- add optional diagnostics of non-linear convergence for LSOR solver |
| 27 |
|
- make maximum zeta a runtime parameter |
| 28 |
|
o model/src (3-d solver): |
| 29 |
|
Undo modif from Dec 12, 2005 (solve_for_pressure.F cvs version v.49), |
| 30 |
|
which consisted in removing the Surf. Non-Hydrostatic pressure from the |
| 31 |
|
whole column and adding it to Eta (inspired from Casualli, 1999): |
| 32 |
|
Only affects 3-D solver initial guess and only if using ExactConserv, |
| 33 |
|
but tests did show a degradation of solver convergence. For this reason, |
| 34 |
|
decide to disable this piece of code. |
| 35 |
|
to recover previous version, just set zeroPsNH = exactConserv (line 78). |
| 36 |
|
Update output of exp. global_ocean.cs32x15.viscA4 & hs94.cs-32x32x5.impIGW |
| 37 |
|
o pkg/ggl90: |
| 38 |
|
- fix previous modif (unfortunatly has not been documented here). |
| 39 |
|
o pkg/seaice/autodiff: |
| 40 |
|
- add CPP flag to include clipping of zeta in EVP code; |
| 41 |
|
although clipping of zeta is not necessary for EVP and not recommended |
| 42 |
|
it was done by default, now clipping is disabled by default |
| 43 |
|
- modify computation and store zetaC/Z in seaice_evp for TAF |
| 44 |
|
- add CPP-flag SEAICE_ALLOW_CLIPVELS (undefined by default) |
| 45 |
|
- change seaice_calc_viscosities/strainrates for better |
| 46 |
|
vectorization after AD by TAF |
| 47 |
|
- rearrange common blocks in SEAICE.h once again for easier dumps of |
| 48 |
|
adjoint output, adjust adcommon.h and g_common.h |
| 49 |
|
- change flux form of default advection scheme for better vectorization |
| 50 |
|
of adjoint code |
| 51 |
|
- modify lab_sea/input.lsr to test the flux form of default advection |
| 52 |
|
scheme |
| 53 |
|
o autodiff: fix seaice fields in g_common.h (left over from 3 to 1 level |
| 54 |
|
transition in checkpoint61r) |
| 55 |
|
|
| 56 |
|
checkpoint61x |
| 57 |
|
o adjoint: adding two new verification experiments |
| 58 |
|
* tutorial_global_biogeo: an online dic sensitivity experiment |
| 59 |
|
(work with G. Forget) |
| 60 |
|
* tutorial_dic_adjoffline: an offline dic sensitivity experiment |
| 61 |
|
(an update of the Dutkiewicz et al., 2006 paper) |
| 62 |
|
Both use the same CO2FLUX cost function, but different controls |
| 63 |
|
Gradient checks look good, but some recomp. could still be traightened out |
| 64 |
|
Should add xx_dic control variable to test at some point |
| 65 |
|
o model/src: |
| 66 |
|
add run-time parameters to check and/or mask initial Temp & Salt |
| 67 |
|
o pkg/ptracers: |
| 68 |
|
apply zonal filter (if used) to all passive tracers |
| 69 |
|
|
| 70 |
|
checkpoint61w |
| 71 |
|
o adjoint: |
| 72 |
|
* update and finish work on adjoint of dic package (G. Forget) |
| 73 |
|
* start adjoint of offline package |
| 74 |
|
* bug fix for cube-sphere TLM code (active_file_g.F) |
| 75 |
|
o tutorial_global_oce_optim: |
| 76 |
|
now tested with 4 tiles (change data.grdchk and update output_adm.txt) |
| 77 |
|
and add SIZE.h_mpi for MPI testing |
| 78 |
|
o pkg autodiff |
| 79 |
|
- fix bug in autodiff_restore |
| 80 |
|
- make genmake2 remove comments and empty lines from ad_input_code.f in |
| 81 |
|
order to reduce file size (new script file remove_comments_sed) |
| 82 |
|
o allows to specify vertical profile of vertical viscosity (viscArNr) |
| 83 |
|
o pkg autodiff |
| 84 |
|
autodiff_ini_model_io: comment out writing of grid information to netcdf |
| 85 |
|
files, because at this time the grid fields are a bunch of zeros |
| 86 |
|
(this is analogous to ini_mnc_vars.F) |
| 87 |
|
o pkg seaice/autodiff: |
| 88 |
|
left-over from removing the 3-time-level fields: |
| 89 |
|
- small rearrangement of common blocks (SEAICE.h) |
| 90 |
|
- fix field declaration in order to get meaningful output in |
| 91 |
|
adseaice.*.nc (adcommon.h) |
| 92 |
|
o pkg obcs: |
| 93 |
|
- go through NH code only if nonHydrostatic=T |
| 94 |
|
- fix restart for Orlanski (Hydrostatic; still to fix NH+Orlanski restart) |
| 95 |
|
- move call to OBCS_CALC before SEAICE_MODEL which needs seaice-obc fields |
| 96 |
|
=> Update output of experiment seaice_obcs |
| 97 |
|
- useOBCSbalance: code is wrong with multi-tile / proc: add a stop. |
| 98 |
|
- fix NH+Orlanski restart (add wvel OBC value in pickup file) |
| 99 |
|
- fix restart for useOBCSprescribe + useEXF |
| 100 |
|
o pkg ctrl |
| 101 |
|
- make obcs as control parameter work also with useSingleCPUio |
| 102 |
|
- replace a few sny and snx by Ny and Nx to be consistent with |
| 103 |
|
ctrl_set_globfld_x/yz.F |
| 104 |
|
o pkg seaice |
| 105 |
|
- re-introduce global field frWtrAtm for the case of |
| 106 |
|
ALLOW_MEAN_SFLUX_COST_CONTRIBUTION defined (SEAICE.h, seaice_growth.F) |
| 107 |
|
- catch the case of an empty filename in seaice_cost_init_fixed.F |
| 108 |
|
o pkg ecco: |
| 109 |
|
- small bug fixes in the_main_loop |
| 110 |
|
. fix order of monitor arguments |
| 111 |
|
. change SEAICE_MULTILEVEL to SEAICE_MULTICATEGORY |
| 112 |
|
. move include DYNVARS.h up outside the ALLOW_AUTODIFF_TAMC block so |
| 113 |
|
that ALLOW_ECCO_FORWARD_RUN works with NONLIN_FRSURF |
| 114 |
|
- replace cos(yc*deg2rad) by _rA in cost_mean_heat/saltflux.F, fix the |
| 115 |
|
imbalance diagnostics for wmean_h/sflux = 0 |
| 116 |
|
- add actual values to f_obcsn/s/w/e in ecco_cost_final.F |
| 117 |
|
- catch the case of empty filenames in ecco_cost_init_fixed.F |
| 118 |
|
and ecco_readparms.F |
| 119 |
|
- fix the logic of wsalt=0.and.wsalt2=0 to wsalt=0.or.wsalt2=0 (and theta) |
| 120 |
|
to make this statement meaningfull in ecco_cost_weights.F |
| 121 |
|
o pkg zonal_filt: |
| 122 |
|
- change ZONAL_FILTER S/R interface (allows to filter 2-D fields) |
| 123 |
|
- when using rStar, filter uVel*hFacW instead of uVel |
| 124 |
|
(tested without topography: fix a Pb of growing instability with NLFS) |
| 125 |
|
o pkg/aim/thsice: |
| 126 |
|
-Add capacity to read a Q-Flux to be applied in slab-mixed layer thsice_slab_ocean.F |
| 127 |
|
(works as SST restoring) |
| 128 |
|
o pkg/thsice: |
| 129 |
|
-add specific parameter for salt restoring in slab mixed-layer |
| 130 |
|
(by default equal to the temperature restoring) |
| 131 |
|
|
| 132 |
|
checkpoint61v |
| 133 |
|
o pkg/kpp: |
| 134 |
|
-add double diffusive contributions as a hack, turned off by default |
| 135 |
|
for now, and the code can be excluded with a CPP-flag |
| 136 |
|
EXCLUDE_KPP_DOUBLEDIFF just as EXCLUDE_KPP_SHEAR_MIX |
| 137 |
|
-add a verification experiments to "vermix" to test this code |
| 138 |
|
o pkg/mom_common: |
| 139 |
|
Move computation of length scales to mom_init_fixed.F (avoid recomputation |
| 140 |
|
at each time-steps*levels). |
| 141 |
|
o pkg/layers: |
| 142 |
|
Created a new package for computing volume fluxes in temperature or |
| 143 |
|
[not yet implemented] density layers, e.g. for calculating residual |
| 144 |
|
overturning circulations. |
| 145 |
|
o pkg/diagnostics: |
| 146 |
|
implement a RS type version for the set of DIAGNOSTICS-FILL subroutines |
| 147 |
|
o fix few RS/RL type mismatch (check with ifort 11.1 "-warn all" option) |
| 148 |
|
o pkg/mdsio (& pkg/rw): rework high level S/R interface: |
| 149 |
|
To avoid mixing type (RS/RL) of input/output array argument, |
| 150 |
|
replace single mixed-type array with a pair from each type (RS/RL). |
| 151 |
|
o setting pickupSuff force to start from a pickup (even if nIter0=0): this |
| 152 |
|
is now implemented for all pkgs (previously only done for the main pickup) |
| 153 |
|
|
| 154 |
|
checkpoint61u |
| 155 |
|
o pkg/obcs: |
| 156 |
|
- change OB?eta from _RS to _RL and put them into a separate commen block |
| 157 |
|
- add code to read OB?eta from a file via obcs_prescribe_read: the code |
| 158 |
|
compiles and does not destroy any other experiments, the data is read |
| 159 |
|
properly, but the code is not tested in actual applications |
| 160 |
|
o ini_mnc_vars.F: if available (ALLOW_CAL), use startTime_1 and 2 to |
| 161 |
|
create a meaningful unit attribute for the time coordinate variable |
| 162 |
|
o pkg/ebm: add plenty of missing "_d 0" ; |
| 163 |
|
update results of experiment global_ocean_ebm (fwd + adjoint tests) |
| 164 |
|
Notes: several pieces of code only works for (very) specific set up |
| 165 |
|
(domain size, lat range, etc ...); should be documented & stop if one |
| 166 |
|
of those assumptions is not meet. |
| 167 |
|
o pkg/seaice: add missing "_d 0" in HSNOW initialisation |
| 168 |
|
update results of experiments: lab_sea (3 fwd tests + 2 adjoint tests) |
| 169 |
|
and offline_exf_seaice (seaicetd + adjoint). |
| 170 |
|
o add missing "_d 0" in ShortWave flux penetration (S/R SWFRAC): |
| 171 |
|
update results of experiments: lab_sea (4 fwd tests + 3 adjoint), |
| 172 |
|
seaice_obcs, global_ocean.cs32x15 (icedyn & thsice) and natl_box (x2). |
| 173 |
|
o few changes to pass when compiling with strick checking of arguments |
| 174 |
|
across S/R (allow to find few bugs in S/R calls). |
| 175 |
|
o pkg/seaice: |
| 176 |
|
- seaice_evp: move if-statement out of loop (TEM) at the cost of 11 new |
| 177 |
|
2d-fields, for slightly better performance |
| 178 |
|
|
| 179 |
|
checkpoint61t |
| 180 |
|
o pkg/seaice: |
| 181 |
|
- remove unused variables in seaice_evp |
| 182 |
|
- reduce number of SQRT in seaice_evp and seaice_calc_viscosities |
| 183 |
|
o eeset_parms.F, open_copy_data_file.F & ini_parms.F: |
| 184 |
|
named file replacing SCRATCH files (for TARGET_BGL & TARGET_CRAYXT): |
| 185 |
|
each processor opens a different file (a fix from Matt). |
| 186 |
|
o cg2d: |
| 187 |
|
use simple EXCH (overlap size = 1 and ignore corners), like in cg3d.F : |
| 188 |
|
this reduces number of EXCH calls by 2 when using exch2. |
| 189 |
|
For now, keep previous size array for gc2d_nsa.F to avoid Pb with store |
| 190 |
|
directives and missing simple EXCH (exch_s3d_rx) adjoint code. |
| 191 |
|
o pkg/thsice: |
| 192 |
|
add code for Near Infra-Red albedo (from Jeff) |
| 193 |
|
(turned on with run-time flag: thSIce_calc_albNIR, in data.ice). |
| 194 |
|
o ini_model_io: |
| 195 |
|
move (from packages_readparms.F & initialise_fixed.F) to ini_model_io.F |
| 196 |
|
MNC & MONITOR initialisation calls; move MNC calls from {pkg}_readparms.F |
| 197 |
|
to {pkg}_init_fixed.F for {pkg}= kpp, seaice & shelfice. |
| 198 |
|
|
| 199 |
|
checkpoint61s |
| 200 |
|
o pkg/exch2 (exch2_rxN_cube*): |
| 201 |
|
remove last BARRIER (no need to synchronise after getting data from shared |
| 202 |
|
buffer (get) as long as any change to buffer (put,recv) is between BARRIER) |
| 203 |
|
o change tiling of natl_box (from 2 to 4 tiles) to get it tested with MPI+MTH |
| 204 |
|
and generate new output (for the 2 tests) using default rotationPeriod. |
| 205 |
|
o pkg/exch2: |
| 206 |
|
- always call exch2_*_cube, not exch-1 anymore, if useCubedSphereExchange=F |
| 207 |
|
(was already the case with AUTODIFF_EXCH2 defined) |
| 208 |
|
- add bj in exch2 arrays and S/R. |
| 209 |
|
- can now use pkg/exch2 for regular (non CS) set-up; switch test experiment |
| 210 |
|
MLAdjust to exch2 (identical results except with MPI -> truncation diff); |
| 211 |
|
o add pkg/longstep for ptracer timestep a multiple of U/V/T/S timestep |
| 212 |
|
- adds deltaT argument to the subroutines TIMESTEP_TRACER DWNSLP_APPLY |
| 213 |
|
GAD_ADVECTION GAD_CALC_RHS GAD_IMPLICIT_R GAD_SOM_ADVECT |
| 214 |
|
- verification experiment natl_box.longstep |
| 215 |
|
o build_options: ifort+mpi_aces optfile (following Constantinos advise): |
| 216 |
|
allow different versions of ifort (8 or 9) and different versions |
| 217 |
|
of mpi (mpich or mpich2) to be used with the same optfile. |
| 218 |
|
requires to load a netcdf module and set MPI_INC_DIR (see comments inside |
| 219 |
|
optfile). Change testing scripts (example_scripts/ACESgrid) accordingly. |
| 220 |
|
o pkg/seaice: |
| 221 |
|
fix previous modif (3-level field removed) for old-pickup reading |
| 222 |
|
|
| 223 |
|
checkpoint61r |
| 224 |
|
o pkg/seaice: |
| 225 |
|
- clean up the 3-time levels of UICE,VICE,HEFF,AREA in three steps: |
| 226 |
|
1. introduce fields for time step (N-1) |
| 227 |
|
2. do not use levels 2 and 3 anymore (but keep them) |
| 228 |
|
3. replace 3D versions of UICE,VICE,HEFF,AREA by 2D versions. This |
| 229 |
|
includes changes to pkg/thsice/thsice_get_velocity.F, |
| 230 |
|
model/src/external_forcing_surf.F, pkg/obcs/obcs_apply_seaice.F, and |
| 231 |
|
pkg/obcs/obcs_apply_uvice.F. |
| 232 |
|
- completely remove code for leap-frog time stepping for second order |
| 233 |
|
advection |
| 234 |
|
- adjusted adjoint accordingly; |
| 235 |
|
gradient checks have improved by an order of magnitude! |
| 236 |
|
|
| 237 |
|
checkpoint61q |
| 238 |
|
o pkg/seaice: fix bug in free slip boundary conditons for LSOR solver |
| 239 |
|
- the boundary conditions along open boundaries are unclear, because the |
| 240 |
|
associated masks are set to zero and are thus treated as closed when |
| 241 |
|
applying the slip-boundary conditions. |
| 242 |
|
- move initialisation of grid variables from seaice_init_varia to _fixed |
| 243 |
|
o adding doc/diags_changes.txt |
| 244 |
|
o model/src: |
| 245 |
|
do_the_model_io: rename few pkg S/R which are called from do_the_model_io, |
| 246 |
|
to avoid "_diags" naming if no relation to pkg/diagnostics. |
| 247 |
|
ini_theta,ini_salt: fix (at least compile the code) reading from NetCDF |
| 248 |
|
file and only call PLOT_FIELD_XYZ if debugMode=T. |
| 249 |
|
remove few PLOT_FIELD_XY calls outside initialisation. |
| 250 |
|
remove few unnecessary BARRIER. |
| 251 |
|
o pkg/salt_plume: add SaltPlumeSouthernOcean to allow disabling |
| 252 |
|
of salt_plume package in Southern Ocean |
| 253 |
|
o pkg/thsice: |
| 254 |
|
- fix hOceMxL in overlap (needed for sea-ice advection); |
| 255 |
|
update output of experiment global_ocean.cs32x15.icedyn. |
| 256 |
|
o pkg/seaice: |
| 257 |
|
- make the (placeholder-)tracer IceAge a little more meaningful |
| 258 |
|
by avoiding ridging of age due to convergent ice flow |
| 259 |
|
- add diagnostics for surface temperature, atmospheric and oceanic |
| 260 |
|
heat flux, and atmospheric and oceanic ice growth |
| 261 |
|
|
| 262 |
|
checkpoint61p |
| 263 |
|
o eesupp (global_max & global_sum): |
| 264 |
|
- add one more element to buffer (start at index 0) for GL_MAX/SUM output; |
| 265 |
|
- remove starting & endding barrier (no longer needed) |
| 266 |
|
o pkg/mdsio, writelocal: |
| 267 |
|
works also in multi-threaded when LOCBIN_IO_THREAD_SAFE is undefined |
| 268 |
|
(remove the STOP): uses shared buffer IO to store data from all threads |
| 269 |
|
and then let master-thread write nThreads tiles. |
| 270 |
|
o pkg/mdsio (read/write field): |
| 271 |
|
- tiled IO done in 1 piece (all levels at a time); |
| 272 |
|
- new header file "MDSIO_BUFF_3D.h" with 3-D buffers. |
| 273 |
|
- multi-threaded: allow to read/write local (non-shared) array |
| 274 |
|
(was already working with singleCpuIO ; now works also without); |
| 275 |
|
=> no longer needs barrier call after reading a file (added inside |
| 276 |
|
mds_read/write field). |
| 277 |
|
- move barrier calls outside gather/scatter_2d to mds_read/write_field |
| 278 |
|
o pkg/obcs/exf |
| 279 |
|
- exf_getffieldrec.F: add a few ". _d 0" and |
| 280 |
|
add code to catch the case fldperiod = 0. |
| 281 |
|
- obcs_precribe_read.F: change if statements so that they match |
| 282 |
|
the variable type (_RL) |
| 283 |
|
- add a separate exf_iprec_obcs to exf_set_obcs (by default equal to |
| 284 |
|
exf_iprec) as part of namelist EXF_NML_01 |
| 285 |
|
o pkg/mdsio: |
| 286 |
|
- read/write tiled (local) files: read/write 1-level tile chunk at a time |
| 287 |
|
(instead of segment of length sNx); expected to speed up tiled IO. |
| 288 |
|
o pkg/exch2: |
| 289 |
|
- take buffer copy from/to array out of S/R exch2_send/recv into new |
| 290 |
|
S/R exch2_put/get ; Exch of local variable now works with MPI+MTH |
| 291 |
|
(tested by removing commom block statement in SOLVE_FOR_PRESSURE.h). |
| 292 |
|
- ad version of send/recv no longer needed (but needs exch2_ad_put/ad_get) |
| 293 |
|
- simplify argument list of S/R exch2_get_uv_bounds and exch2_get_scal_bounds |
| 294 |
|
(which replaces exch2_get_recv_bounds & exch2_get_send_bounds) by |
| 295 |
|
including Topology header file (and common blocks). |
| 296 |
|
- implement EXCH_IGNORE_CORNERS in scalar exchange (rx1); |
| 297 |
|
remove 2nd exch call in exch2_s3d_rx (no longer needed). |
| 298 |
|
- document S/R arg. list (+ update some comments) |
| 299 |
|
o pkg/seaice: |
| 300 |
|
- added SEAICE_CLIM_AIR code, which permits modulation of surface air |
| 301 |
|
temperature and humidity over sea ice based on climatological values. |
| 302 |
|
- clean up computation of Hibler+Bryan (1987) stress coupling for the case |
| 303 |
|
of LSR (this change is expected to change the results slightly because now |
| 304 |
|
it uses slightly difference moduli from the second last LSR solution, but |
| 305 |
|
that is more consistent with the stress computations; this part of the code |
| 306 |
|
is not tested in the verification experiments) |
| 307 |
|
- fix diagnostics for flux of ice in case of B-grid |
| 308 |
|
- turn strain rates eij and press into global fields within global |
| 309 |
|
common block in SEAICE.h for more straighforward diagnostics |
| 310 |
|
- fix diagnostics for principal stress components, press, zeta, eta |
| 311 |
|
to give something meaning full |
| 312 |
|
o change tiling of global_ocean.cs32x15 (12 tiles, 32x16) and generate |
| 313 |
|
new output (all 4 tests) using default rotationPeriod. |
| 314 |
|
o pkg KPP: |
| 315 |
|
- add flag (LimitHblStable) to avoid limiting BL depth under stable conditions |
| 316 |
|
- fix vertical indices in RI_IWMIX (match between KPP grid from 0 to Nr+1 |
| 317 |
|
and normal grid from 1 to Nr) |
| 318 |
|
o Fix computations of L4rdt in mom_calc_visc.F |
| 319 |
|
o eesupp: |
| 320 |
|
new S/R ALL_PROC_DIE (like EEDIE, without irrelevant threads err msg): |
| 321 |
|
should be called before a "stop" if we know that all process are going |
| 322 |
|
to stop; if one Proc does call this S/R, will hang in MPI_FINALISE. |
| 323 |
|
|
| 324 |
checkpoint61o |
checkpoint61o |
| 325 |
o pkg/exch2: change check_depths to issue just a warning (instead of error+stop) |
o pkg/exch2: change check_depths to issue just a warning (instead of error+stop) |
| 326 |
o model/src/ini_curvilinear_grid.F: |
o model/src/ini_curvilinear_grid.F: |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch