| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o pkg/icefront: fix bug and reorder loop in icefront_init_varia.F |
| 8 |
|
o model/src: move k-loop outside of i/j-loops when (re-)computing R_low |
| 9 |
|
in ini/update_masks_etc.F in order to avoid -O3 optimization problems with |
| 10 |
|
some compilers (gfortran 4.6.0 on MacOS) |
| 11 |
|
o momentum: |
| 12 |
|
- separate fVer?(:,:,kUp) & fVer?(:,:,kDown) in argument list of MOM_FLUXFORM |
| 13 |
|
& MOM_VECINV subroutines (to help TAF). This fixes Pb of uninitialised |
| 14 |
|
fVer?(kUp) in admom_fluxform with non-lin free-surf or bottom control. |
| 15 |
|
|
| 16 |
|
checkpoint63k (2012/03/17) |
| 17 |
|
o model/src: |
| 18 |
|
- selectAddFluid (ALLOW_ADDFLUID): add missing term in horizontal momentum |
| 19 |
|
flux-form (mom_fluxform.F) and vertical momentum (calc_gw.F) equations. |
| 20 |
|
- SOLVE_DIAGONAL: switch default to use adjoinable & vectorized version |
| 21 |
|
(#undef SOLVE_DIAGONAL_LOWMEMORY); update output of exp. front_relax.bvp. |
| 22 |
|
o model/src: solve_*diagonal.F |
| 23 |
|
- replace division with multiplication by inverse (similar to original version) |
| 24 |
|
update output of AD exp. global_ocean.90x40x15. |
| 25 |
|
- fix pkg/gmredi/gmredi_calc_psi_bvp.F & pkg/ggl90/ggl90_calc.F |
| 26 |
|
to enable the use of the vectorized & differentiable solve_tridiagonal.F |
| 27 |
|
- test exp. global_ocean.90x40x15/input_ad: switch to advection scheme 20 for |
| 28 |
|
vertical advection (to get solve_tridiagonal tested in AD mode). |
| 29 |
|
o pkg/seaice: ECCO-CLIVAR merge: |
| 30 |
|
1. seaice AREA relaxation |
| 31 |
|
* CPP option: SEAICE_ALLOW_AREA_RELAXATION |
| 32 |
|
* runtime flag: SEAICE_tauAreaObsRelax |
| 33 |
|
2. modified global mean cost imbalance code: CPP option |
| 34 |
|
ALLOW_SSH_GLOBMEAN_COST_CONTRIBUTION |
| 35 |
|
o another iteration on solve_*diagonal.F |
| 36 |
|
Make choices independent of AUTODIFF |
| 37 |
|
Default will be the adjoinable vectorized version. |
| 38 |
|
Alternative choices are: |
| 39 |
|
* non-adjoinable original version with low memory footprint: |
| 40 |
|
#define SOLVE_DIAGONAL_LOWMEMORY |
| 41 |
|
* still adjoinable but does not vectorize (k as inner-most loop): |
| 42 |
|
#define SOLVE_DIAGONAL_KINNER (keeping SOLVE_DIAGONAL_LOWMEMORY undef) |
| 43 |
|
o testreport |
| 44 |
|
- add option to use a home made wrapper around "make" |
| 45 |
|
- remove MITGCM_IEEE env. variable (no longer active) |
| 46 |
|
o pkg/shelfice |
| 47 |
|
- improve computation of friction velocity for velocity dependent |
| 48 |
|
exchange coefficients: include velocities into SHELFICEBoundaryLayer |
| 49 |
|
block to average velocities vertically |
| 50 |
|
- add TAF store directives for SHI_ALLOW_GAMMAFRICT |
| 51 |
|
o Modify diagonal solvers, now three versions (upon request): |
| 52 |
|
#ifndef SOLVE_DIAGONAL_AUTODIFF |
| 53 |
|
1. default version (dating back a while ago) |
| 54 |
|
#else /* def SOLVE_DIAGONAL_AUTODIFF */ |
| 55 |
|
# ifdef SOLVE_DIAGONAL_KINNER |
| 56 |
|
2. adjoinable version with k-innermost loop (G. Forget) |
| 57 |
|
# else |
| 58 |
|
3. adjoinable vectorized version of 2. |
| 59 |
|
# endif |
| 60 |
|
#endif /* SOLVE_DIAGONAL_AUTODIFF */ |
| 61 |
|
All flags are currently set within routines, but could/should be moved to CPP-options. |
| 62 |
|
Array bounds for 1. are as before, for 2.,3. cover the entire domain |
| 63 |
|
(could be changed for 1. too). |
| 64 |
o verification/lab_sea (fwd): |
o verification/lab_sea (fwd): |
| 65 |
- update the 4 fwd lab_sea test experiment: no longer SEAICE_GROWTH_LEGACY, |
- update the 4 fwd lab_sea test experiment: no longer SEAICE_GROWTH_LEGACY, |
| 66 |
use 7 categories (except hb87), new solve4temp params (except salt_plume), |
use 7 categories (except hb87), new solve4temp params (except salt_plume), |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch