| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o pkg/diagnostics/diagstats_calc.F: add special code for TARGET_NEC_SX |
| 8 |
|
that vectorises. This code uses 3 extra 2D fields, F90 array assignments |
| 9 |
|
and F90 intrinsic routines SUM, MINVAL, MAXVAL, but since TARGET_NEC_SX |
| 10 |
|
does not even come with an F77 compiler anymore, this should not be |
| 11 |
|
a limitation |
| 12 |
|
|
| 13 |
|
checkpoint63e (2011/11/06) |
| 14 |
|
o pkg/seaice adjoint of LSR: |
| 15 |
|
- substantially rework STORE dirs. |
| 16 |
|
- add new flag SEAICE_LSR_ADJOINT_ITER and parameter SOLV_MAX_FIXED |
| 17 |
|
this flag will do SOLV_MAX_ITERS but only store SOLV_MAX_FIXED fields |
| 18 |
|
and only perform SOLV_MAX_FIXED adjoint iters. |
| 19 |
|
If flag SEAICE_LSR_ADJOINT_ITER is undef-ined, the store in the iteration |
| 20 |
|
always goes to same record, meaning that the last forward field will be |
| 21 |
|
the one used in the adjoint - an approximation. |
| 22 |
|
- now produces non-zero LSR-related gradients, and acceptable gradients for |
| 23 |
|
An's zonal (x-z) test case (to be checked in) |
| 24 |
|
- lab_sea: produces acceptable gradients for lab_sea with forward/backward LSR |
| 25 |
|
(but these are screwed up by the thermodynamics in the first place); |
| 26 |
|
reduce from 4 to 3 timesteps to avoid t.d. issue as temporary solution; |
| 27 |
|
output_adm.txt also reflects changes to seaice_lsr |
| 28 |
|
- global_ocean.cs32x15: |
| 29 |
|
* the "default" input_ad.seaice has been re-defined to always exclude dynamics |
| 30 |
|
(output_adm.seaice.txt was updated) |
| 31 |
|
* new input_ad.seaice_dynmix tests a forward LSR / backward free-drift |
| 32 |
|
(output_adm.seaice_dynmix.txt) |
| 33 |
|
o test-exp. exp4+stevens: |
| 34 |
|
- switch off pTracers and nonHydrostatic (restart problems with NH); update output |
| 35 |
|
o pkg/seaice (lsr solver): |
| 36 |
|
- put coriolis term into FORCEX,Y (as in seaice_evp.F; previously in rhsU,V only); |
| 37 |
|
update output (change truncation) of FWD test exp lab_sea & global_ocean.cs32x15 |
| 38 |
|
- fix SEAICEuseFREEDRIFT for new OBCS version (ndef OBCS_UVICE_OLD) |
| 39 |
|
- new parameter "LSR_mixIniGuess" allows (if=2,4) to mix free-drift solution |
| 40 |
|
into initial guess for linearised solver; use weight ~ (1/err)^2,4. |
| 41 |
|
- compute (with new S/R) and print initial (+ free-driff sol. if available) |
| 42 |
|
and final residual of linearised system. |
| 43 |
|
o pkg/pp81: add diagnostics for vertical eddy viscosity and diffusivity |
| 44 |
|
(requires new subroutines pp81_init_fixed.F/pp81_diagnostics_init.F) |
| 45 |
|
o pkg/obcs: Stevens BCs (still only for east and west) |
| 46 |
|
- fix Stevens BCs by using intermediate velocities of previous |
| 47 |
|
time step as "baroclinic" contribution rather velocity of one grid cell |
| 48 |
|
inside the boundaries; more consistent but makes the entire treatment |
| 49 |
|
more complicated, requires new pickup files and two new S/Rs: |
| 50 |
|
obcs_copy_uv_n and obcs_save_uv_n |
| 51 |
|
keep old version with flag OBCS_STEVENS_USE_INTERIOR_VELOCITY |
| 52 |
|
(undefined by default) |
| 53 |
|
- silently move computation of some grid factors into separate loops to |
| 54 |
|
save some recomputations |
| 55 |
|
- add a hack for non-linear free surface |
| 56 |
|
- add a verification experiment to exp4 that tests this code |
| 57 |
|
o make gluemnc/xplodemnc work with (moron) non-GNU shell tools |
| 58 |
|
|
| 59 |
|
checkpoint63d (2011/10/25) |
| 60 |
|
o pkg/obcs + seaice-dynamics: |
| 61 |
|
- new implementation calls OBCS_APPLY_UVICE (in seaice_dynsolver.F) before |
| 62 |
|
solving for uIce,vIce and prevent SEAICE_LSR or SEAICE_EVP to modify OB |
| 63 |
|
values (consistent with an implicit method). |
| 64 |
|
- add option OBCS_UVICE_OLD (in OBCS_OPTIONS.h, defined by default) to |
| 65 |
|
return to older implementation (same results as before); note that none of |
| 66 |
|
the "experimental" OBCS_SEAICE options are implemented in the new version. |
| 67 |
|
- switch test exp. seaice_obcs to use new version (update output). |
| 68 |
|
o pkg/seaice: |
| 69 |
|
- add few stop (B-grid & CS-exch, B-grid & obcs, AUTODIFF & SEAICE_maskRHS) |
| 70 |
|
- add parenthesis in Water-Ice drag calculations (evp & lsr); |
| 71 |
|
- in seaice_lsr.F, simplify and fix AA3 setting in case sNx or sNy=1; |
| 72 |
|
and fix case SEAICE_waterDrag=0 (but unstable if no or little seaice) |
| 73 |
|
- update output of test exp lab_sea that uses evp or lsr (fwd+ad) |
| 74 |
|
and global_ocean.cs32x15 (icedyn & seaice fwd+ad). |
| 75 |
|
o pkg/layers: |
| 76 |
|
- Include potential density as new coordinate (Thanks to David Munday) |
| 77 |
|
o pkg/dic: |
| 78 |
|
- Remove subroutine CALC_PCO2_APPROX_CO3 from carbon_chem.F and add |
| 79 |
|
carbonate computation/output to CALC_PCO2_APPROX |
| 80 |
|
o pkg/generic_advdiff |
| 81 |
|
- make TAF generate code that vectorizes: |
| 82 |
|
gad_dst3fl_adv_*.F, gad_u3c4_impl_r.F |
| 83 |
|
o model/src: |
| 84 |
|
- improve vectorization also for adjoint code of |
| 85 |
|
solve_tridiagonal and solve_pentadiagonal. As this requires up to |
| 86 |
|
5 new 3D-fields, this code is only turned on for the combination of |
| 87 |
|
ALLOW_AUTODIFF_TAMC and TARGET_NEC_SX. Also remove error message |
| 88 |
|
for singular matrix check (only for TARGET_NEC_SX). |
| 89 |
|
- remove hard-wiring of CPP-flag ALLOW_SOLVERS_KLOOPINSIDE to |
| 90 |
|
ALLOW_AUTODIFF. This flag now needs to be set outside of |
| 91 |
|
solve_tri/pentadiagonal.F |
| 92 |
|
o pkg/exf: |
| 93 |
|
- fix previous modif in exf_readparms.F (revision 1.66: treatment of |
| 94 |
|
optional namelist EXF_NML_OBCS). |
| 95 |
|
- add many debug-print (to trace down where pkg/cal error msg comes from) |
| 96 |
|
o pkg/mom_vecinv/mom_vi_hdissip.F |
| 97 |
|
- rearrange loops so that vectorization is improved and in particular |
| 98 |
|
the adjoint routine vectorizes. The code is a little longer now, but I |
| 99 |
|
think easier to read. |
| 100 |
|
Note: CPP-flag MOM_VI_ORIGINAL_VISCA4 is not tested. |
| 101 |
|
o pkg/seaice: |
| 102 |
|
- allow to specify initial seaice velocity from files (uIceFile,vIceFile) |
| 103 |
|
o pkg/exch2: |
| 104 |
|
- alternative code (without tile x tile matrix) for cumulated (global) sum; |
| 105 |
|
used as default; keep using matrix version in adjustment.cs-32x32x1. |
| 106 |
|
o tools/build_options: |
| 107 |
|
- after merging some MPI and/or multi-threads specific optfiles so that |
| 108 |
|
the modified default optfile support both built (using genmake2 -mpi |
| 109 |
|
for the MPI built, using genmake2 -omp for OpenMP built), |
| 110 |
|
move now obsolete optfiles to tools/build_options/unsupported/ |
| 111 |
|
(add copy of email sent to support list in doc/optfiles_changes.txt) |
| 112 |
|
|
| 113 |
|
checkpoint63c (2011/09/28) |
| 114 |
|
o pkg/ebm: |
| 115 |
|
- fix ebm_zonalmean.F for several procs in Ydir (nPy > 1); update output of |
| 116 |
|
test exp. global_ocean_ebm (affects machine truncation, 12 digits for cg2d) |
| 117 |
|
o pkg/seaice (SItracer pickup) |
| 118 |
|
* Simple code for pickup of SItracer fields (#define ALLOW_SITRACER) |
| 119 |
|
To be done: |
| 120 |
|
- thorough testing of missing fields in S/R seaice_read_pickup |
| 121 |
|
- thorough matching between generic tracers and specific fields used |
| 122 |
|
(e.g. volume-weighted age, area-weighted age, salinity, ...) |
| 123 |
|
o pkg/grdchk: |
| 124 |
|
- all procs calculate and print finite-difference gradient (for testreport) |
| 125 |
|
- change description output used by testreport; |
| 126 |
|
- AD testing: also compare and report Finite-difference gradient; |
| 127 |
|
=> update all output_adm*.txt |
| 128 |
|
o test exp. global_ocean_ebm: |
| 129 |
|
- switch back useRealFreshWaterFlux (no problem anymore with run-off); |
| 130 |
|
and turn on exactConserve; update (fwd) output. |
| 131 |
|
o pkg/seaice (LSR): |
| 132 |
|
- print to STDOUT number of solver iterations at SEAICE_monFreq interval. |
| 133 |
|
- modify solver to deal cleanly with u/v switch at face edges (CS-grid): |
| 134 |
|
* identical results with non-CS topology (but save up to half of EXCH calls) |
| 135 |
|
* no much changes in cs-32 verification tests since no ice at face edges |
| 136 |
|
(differences due to convergence criteria, since nb of iterations changed) |
| 137 |
|
* expect some differences on lat-lon-cap grid. |
| 138 |
|
update FWD results of global_ocean.cs32x15 test exp. (icedyn & seaice) |
| 139 |
|
o pkg/exch2: |
| 140 |
|
- add new CPP option to disable Cumul-Sum matrix (to save storage when using |
| 141 |
|
many tiles); until alternative coding is added, just add a stop. |
| 142 |
|
o pkg/kpp: |
| 143 |
|
- move ptracer/salt correction to KPP diffusivity from calc_3d_diffusivity.F |
| 144 |
|
to S/R KPP_CALC_DIFF_PTR (+ add tracer-index to argument list). |
| 145 |
|
o optfiles: |
| 146 |
|
- merged few MPI specific (*+mpi*) and multi-threads specific (*+mth) optfiles |
| 147 |
|
into corresponding standard one, which now support different build: |
| 148 |
|
(when using "genmake2 -mpi" for MPI built, "genmake2 -omp" for OpenMP built) |
| 149 |
|
- add "darwin_amd64_gfortran" optfile (initially from Constantinos, tested by Gael) |
| 150 |
|
o pkg/obcs (seaice): |
| 151 |
|
- add missing EXCH for uIce,vIce after applying OBCS; |
| 152 |
|
- remove S/R EXF_SWAPFFIELDS_X/Y , disable S/R EXF_SET_OBCS_X/Y by adding |
| 153 |
|
argument = number of levels to S/R EXF_SWAPFFIELDS_XZ/YZ and S/R |
| 154 |
|
EXF_SET_OBCS_XZ/YZ ; update obcs_exf_load.F; |
| 155 |
|
- fix EXF_SET_OBCS_XZ/YZ for multi-threaded (assuming loaded arrays are shared) |
| 156 |
|
- verification exp. seaice_obcs: split domain in 2 tiles ; add SIZE.h_mpi |
| 157 |
|
and eedata.mth for MPI and mutlti-threaded testing. |
| 158 |
|
o pkg/flt: |
| 159 |
|
- make FLT EXCH working with multi-threads (and MPI + multi-threads); |
| 160 |
|
- use buffer for floats IO (trajectories & profiles); size set in FLT_SIZE.h; |
| 161 |
|
- fix pickup and trajectories + profiles output with multi-threads; |
| 162 |
|
- add parameter to select group of fields to write along trajectory and |
| 163 |
|
profile (default unchanged); |
| 164 |
|
- only call PORT_RAND* function when flt_noise <> 0; |
| 165 |
|
- add multi-threaded test for verification exp. flt_example. |
| 166 |
|
o eesupp/src: |
| 167 |
|
- fix S/R ALL_PROC_DIE for MPI+multi-threads. |
| 168 |
|
- USE_OMP_THREADING: check that number of threads (from eedata) matches |
| 169 |
|
actual number (from env.var. setting). |
| 170 |
|
o pkg/ebm: |
| 171 |
|
- fix multi-threading in ebm_area_t.F ; clean-up EBM.h header file. |
| 172 |
|
- add multi-threaded test for verification exp. global_ocean_ebm |
| 173 |
|
|
| 174 |
|
checkpoint63b (2011/08/29) |
| 175 |
|
o pkg/ebm: |
| 176 |
|
- fix sign of runoff; |
| 177 |
|
- update directly forcing fields (Qnet,EmPmR) to use standard |
| 178 |
|
external_forcing_surf.F (remove ebm_forcing_surf.F); |
| 179 |
|
- avoid loading forcing fields a 2nd time (1rst time from ini_forcing.F); |
| 180 |
|
- set either tauThetaClimRelax (in data) or tauThetaZonRelax (in data.ebm) |
| 181 |
|
and apply relaxation towards corresponding SST field (2-D map or ZonAver). |
| 182 |
|
- update global_ocean_ebm parameters file data and update output (fwd+ad). |
| 183 |
|
o verification/matrix_example: |
| 184 |
|
- split domain in 8 tiles (2x4); add SIZE.h_mpi and eedata.mth for |
| 185 |
|
MPI and mutlti-threaded testing. |
| 186 |
|
o pkg/matrix: |
| 187 |
|
- change name of S/R MATRIX_INIT to S/R MATRIX_INIT_VARIA ; |
| 188 |
|
- fix for multi-tiles (add bi,bj indices to exp/impMatrixCounter); |
| 189 |
|
- fix for multi-threaded (remove DATA statement); |
| 190 |
|
- use standard DIFFERENT_MULTIPLE function; |
| 191 |
|
- add MATRIX_OPTIONS.h |
| 192 |
|
o verification/matrix_example: |
| 193 |
|
To allow to split the domain in several tiles, change the domain size |
| 194 |
|
from 31x31 to 32x32 by adding 1 row (on Southern side) + 1 column (on |
| 195 |
|
Western side) of land points. |
| 196 |
|
o (*_)write_pickup: |
| 197 |
|
- fix few bugs in mypackage_read/write_pickup.F |
| 198 |
|
- in MDS_WR_METAFILES calls, replace "myTime" by a local copy into |
| 199 |
|
array of length 1 (was caught by compiler syntax-check). |
| 200 |
|
o pkg/diagnostics (velocity potential solver): |
| 201 |
|
- fix initialisation of main diagonal matrix (for overlap values) |
| 202 |
|
o tools/build_options: |
| 203 |
|
- towards single optfile for MPI and non-MPI built: merge few +mpi optfile |
| 204 |
|
into previously non MPI one (linux_am64 pgf77,open64,pathf90,ifort11) |
| 205 |
|
- linux_amd64_gfortran: merge +mpi_generic version into standard optfile: |
| 206 |
|
* move IEEE flags to DEVEL section; take IEEE flags from +mpi_generic ; |
| 207 |
|
* FAST case: for now, keep -O3 with ini_masks_etc.F in NOOPTFILES list; |
| 208 |
|
and use "-devel" testreport option for standard testing on baudelaire. |
| 209 |
|
o pkg/grdchk: |
| 210 |
|
- Assume that all gradient-check position are within the same tile |
| 211 |
|
corresponding to global indices (iGloTile,jGloTile) and that |
| 212 |
|
positions (i,j indices or nbeg) are local to this tile; |
| 213 |
|
- set tile local-indices and proc. number according to tile global-indices; |
| 214 |
|
- remove grdchkwhichproc from namelist (retired param); |
| 215 |
|
- send adjoint gradient to all procs so that all procs print it to STDOUT; |
| 216 |
|
- switch on MPI test for global_ocean.90x40x15 AD test experiment. |
| 217 |
|
o adding package bbl (Bottom Boundary Layer) |
| 218 |
|
description in MITgcm/pkg/bbl/bbl_description.tex |
| 219 |
|
example/test experiment in MITgcm_contrib/bbl |
| 220 |
|
o adjoint verifications |
| 221 |
|
- changed 5 experiments to multi-tiles |
| 222 |
|
* global_ocean.90x40x15 |
| 223 |
|
* obcs_ctrl |
| 224 |
|
* offline_exf_seaice |
| 225 |
|
* tutorial_dic_adjoffline |
| 226 |
|
* tutorial_tracer_adjsens |
| 227 |
|
changed reference output for those that crossed tile boundaries |
| 228 |
|
(changed choices of gradient check positions) |
| 229 |
|
- with same tiling, switch on MPI test for the last 4 (add SIZE.h_mpi) |
| 230 |
|
- temporary bug fix in grdchk_loc.F for obcs |
| 231 |
|
o genmake2 & testreport |
| 232 |
|
- add target "obj" (just for OBJFILES files) to Makefile; |
| 233 |
|
- add option to specify FEXTRAFLAGS on command line. |
| 234 |
|
o pkg/mypackage: |
| 235 |
|
- remove bi,bj arguments from argument list of S/R MYPACKAGE_CALC_RHS |
| 236 |
|
|
| 237 |
|
checkpoint63a (2011/08/02) |
| 238 |
o pkg/diagnostics (velocity potential): |
o pkg/diagnostics (velocity potential): |
| 239 |
- fix velocity-potential calculation when using OBCS: balance net inflow |
- fix velocity-potential calculation when using OBCS: balance net inflow |
| 240 |
at each level (new S/R OBCS_DIAG_BALANCE). |
at each level (new S/R OBCS_DIAG_BALANCE). |
| 285 |
Add Sub-Meso Eddies parameterisation from Baylor (for now, disabled with TAF), |
Add Sub-Meso Eddies parameterisation from Baylor (for now, disabled with TAF), |
| 286 |
and test it in verification exp. "front_relax.bvp" (update output.bvp.txt) |
and test it in verification exp. "front_relax.bvp" (update output.bvp.txt) |
| 287 |
|
|
| 288 |
checkpoint63 (2011/07/11) |
checkpoint63 (2011/07/11) |
| 289 |
o pkg/exch2: |
o pkg/exch2: |
| 290 |
- rename+move: nTiles in W2_EXCH2_PARAMS.h --> exch2_nTiles in W2_EXCH2_TOPOLOGY.h |
- rename+move: nTiles in W2_EXCH2_PARAMS.h --> exch2_nTiles in W2_EXCH2_TOPOLOGY.h |
| 291 |
- exch2 version of Cumul-Sum routine: S/R W2_SET_MAP_CUMSUM set up the |
- exch2 version of Cumul-Sum routine: S/R W2_SET_MAP_CUMSUM set up the |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch