| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
|
o pkg/seaice/autodiff: |
| 8 |
|
- rearrange common blocks in SEAICE.h once again for easier dumps of |
| 9 |
|
adjoint output, adjust adcommon.h and g_common.h |
| 10 |
|
- change flux form of default advection scheme for better vectorization |
| 11 |
|
of adjoint code |
| 12 |
|
- modify lab_sea/input.lsr to test the flux form of default advection |
| 13 |
|
scheme |
| 14 |
|
o autodiff: fix seaice fields in g_common.h (left over from 3 to 1 level |
| 15 |
|
transition in checkpoint61r) |
| 16 |
|
|
| 17 |
|
checkpoint61x |
| 18 |
|
o adjoint: adding two new verification experiments |
| 19 |
|
* tutorial_global_biogeo: an online dic sensitivity experiment |
| 20 |
|
(work with G. Forget) |
| 21 |
|
* tutorial_dic_adjoffline: an offline dic sensitivity experiment |
| 22 |
|
(an update of the Dutkiewicz et al., 2006 paper) |
| 23 |
|
Both use the same CO2FLUX cost function, but different controls |
| 24 |
|
Gradient checks look good, but some recomp. could still be traightened out |
| 25 |
|
Should add xx_dic control variable to test at some point |
| 26 |
|
o model/src: |
| 27 |
|
add run-time parameters to check and/or mask initial Temp & Salt |
| 28 |
|
o pkg/ptracers: |
| 29 |
|
apply zonal filter (if used) to all passive tracers |
| 30 |
|
|
| 31 |
|
checkpoint61w |
| 32 |
o adjoint: |
o adjoint: |
| 33 |
* update and finish work on adjoint of dic package (G. Forget) |
* update and finish work on adjoint of dic package (G. Forget) |
| 34 |
* start adjoint of offline package |
* start adjoint of offline package |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch