| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o pkg/seaice/seaice_lsr.F |
| 8 |
|
- modify the computation of the residual so that the mass of sea-ice |
| 9 |
|
is not used twice; this simplyfies the calculations, but since |
| 10 |
|
the residual is never used except when LSR_mixIniGuess.ge.4---and this is |
| 11 |
|
not tested in testreport---this does not affect the testreport results. |
| 12 |
|
o verification/seaice_obcs: |
| 13 |
|
- disable SEAICE_GROWTH_LEGACY; reset input files to match those of |
| 14 |
|
verification/lab_sea/input.salt_plume; changes results/output.txt* |
| 15 |
|
o model/src: |
| 16 |
|
- adapt quasi-hydrostatic formulation to work with p or p* coordinate. |
| 17 |
|
o pkg/seaice: |
| 18 |
|
- improve description of parameters in seaice_fgmres.F and start |
| 19 |
|
cleaning up the code |
| 20 |
|
- change name of SCALPROD to SEAICE_SCALPROD and modify meaning of |
| 21 |
|
second argument |
| 22 |
|
- add a metric based on grid cell area to SEAICE_SCALPROD; update |
| 23 |
|
results of offline_exf_seaice.dyn_jfnk |
| 24 |
|
- compute initial residual norm in seaice_jfnk based on SEAICE_SCALPROD |
| 25 |
|
o pkg/aim_v23: |
| 26 |
|
- fix an other bug in dissipation heating (grid-cell area was missing). |
| 27 |
|
o model/src + pkg/seaice: |
| 28 |
|
- add global (multi-tile) linear solver for pair of tri-diagonal system |
| 29 |
|
along X and Y lines, respectively for U and V component. |
| 30 |
|
Note: 1) MPI and cube-exchange not yet coded. |
| 31 |
|
2) probably not accurate for poorly conditioned / large size problem. |
| 32 |
|
- add option for SEAICE_LSR to use global (multi-tile) tri-diagonal solver; |
| 33 |
|
default: not compiled and not used. |
| 34 |
|
o pkg/seaice: |
| 35 |
|
- first attempt at a multi-threaded fgmres solver for the JFNK solver: |
| 36 |
|
introduce bi/bj loops and use global_sum_tile_rl instead of |
| 37 |
|
mpi_allreduce in seaice_fgmres.F |
| 38 |
|
- improve diagnostics output for fgmres solver (seaice_fgmres.F) |
| 39 |
|
- replace a few unnecessary goto statements by if-blocks (seaice_fgmres.F) |
| 40 |
|
- add a verification experiment for JFNK in offline_exf_seaice |
| 41 |
|
o verification: |
| 42 |
|
- change (for now, only fwd test) offline_exf_seaice experiment: |
| 43 |
|
use a re-entrant channel with idealized costline (trapezoidal shape) |
| 44 |
|
at southern boundary to explore seaice dynamics and thermodynamics |
| 45 |
|
with simple forcing. Main set-up has dynamics+thermo and SST relaxation; |
| 46 |
|
secondary test (input.thsice) use thermodynamics only with SST relaxation. |
| 47 |
|
- add seaice-dynamics only secondary test using LSR and pkg/thsice advection. |
| 48 |
|
o pkg/exf: |
| 49 |
|
- disable "sstExtrapol" when Nr=1 |
| 50 |
|
o verification: |
| 51 |
|
- add new experiment "cheapAML_box" (simple 1 layer set-up from Bruno Deremble) |
| 52 |
|
to test pkg/cheapaml. |
| 53 |
|
o pkg/aim_v23 & pkg/shap_filt: |
| 54 |
|
- fix dissipation heating (erroneous gravity multiplication factor). |
| 55 |
|
o pkg/seaice: |
| 56 |
|
- remove unused field pressPre from jfnk solver |
| 57 |
|
- rename two subroutines in seaice_preconditioner.F |
| 58 |
|
o eesupp/src: |
| 59 |
|
- S/R MDSFINDUNIT avoids reserved unit numbers (scrUnit1, eeDataUnit ...) |
| 60 |
|
o pkg/seaice/seaice_preconditioner.F: fix preconditioner once again |
| 61 |
|
- remove wrong parts of the rhs and the lhs (were in there |
| 62 |
|
because I did not understand the concept of this preconditioner) |
| 63 |
|
- add an exchange between u- and v-sweep. This increases communication |
| 64 |
|
by a factor of two but improves the convergence of jfnk. Ideally, |
| 65 |
|
one would like the u- and v-sweep to be independent. So far within |
| 66 |
|
undocumented ifdefs |
| 67 |
|
- reorganize the extra computations for printing the residuals |
| 68 |
|
o tools/genmake2: |
| 69 |
|
- comment out the "mpif.h" include in NetCDF test programm (for MPI built) |
| 70 |
|
since it's not necessary (and was preventing to use MNC in some cases). |
| 71 |
|
|
| 72 |
|
checkpoint64a (2012/11/16) |
| 73 |
o pkg/autodiff: |
o pkg/autodiff: |
| 74 |
- improve pkg-internal logical parameter switch (inAd/inFwd mode) so that |
- improve pkg-internal logical parameter switch (inAd/inFwd mode) so that |
| 75 |
autodiff_readparms.F no longer depends on calling order (within |
autodiff_readparms.F no longer depends on calling order (within |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch