| 4 |
Notes on tags used in MITgcmUV |
Notes on tags used in MITgcmUV |
| 5 |
============================== |
============================== |
| 6 |
|
|
| 7 |
o model/src (rStar): |
o verification/lab_sea: |
| 8 |
use simple average (not area weighted) of rStarFac at U,V point when using |
- overhaul of the forward model experiments. The main motivation |
| 9 |
vectorInvariant and selectKEscheme=1,3 (consistent with mom. vertical advection) |
for this revision is that lab_sea has proven quite fragile (when |
| 10 |
update output of exp. fizhi-cs-32x32x40 and fizhi-cs-aqualev20. |
changing the sea-ice code, and switching platforms) to a point |
| 11 |
|
when failures were very hard to attribute to specific code problems. |
| 12 |
|
- the suspected reason for this fragility is noise amplification by evp so, |
| 13 |
|
to start with, we decided to replace evp with lsr in the main experiment. |
| 14 |
|
- the input.lsr experiment thus became obsolete and was removed. |
| 15 |
|
- it was replaced with the input.fd experiment that uses an |
| 16 |
|
analytical "free drift" solution as the ice velocity field. |
| 17 |
|
- Taken together, input (lsr), input.fd (free drift) and input.hb87 (evp) |
| 18 |
|
will hopefully allow us to discriminate problems due to dynamic solvers |
| 19 |
|
from problems due to the thermodynamic code. input.salt_plume (lsr) allows |
| 20 |
|
us to test additional options, along with the salt plume package. |
| 21 |
|
- Finally a pickup was generated (at the end of a one year run) to |
| 22 |
|
make the tests more representative of an actual model run. Previously |
| 23 |
|
the tests starting point was a 1m thickness of ice everywhere. |
| 24 |
|
o pkg/seaice: |
| 25 |
|
- introducing S\R seaice_freedrift.F that compute the |
| 26 |
|
analytical solution to an approximate ice momentum |
| 27 |
|
equation (i.e. the simple stress/coriolis balance). |
| 28 |
|
- define SEAICE_ALLOW_FREEDRIFT to trigger the computation |
| 29 |
|
of the free drift velocity field, and enable its use. |
| 30 |
|
To actually use the free drift solution as the model ice |
| 31 |
|
velocity field set SEAICEuseFREEDRIFT to .TRUE. in data.seaice. |
| 32 |
|
o move "useOASIS" from PARAMS.h to EEPARAMS.h and read it from "eedata" |
| 33 |
|
|
| 34 |
|
checkpoint62l (2010/10/04) |
| 35 |
|
o pkg/seaice: |
| 36 |
|
-fix the USE_ORIGINAL_SBI version (for variable freezing point); |
| 37 |
|
-use "celsius2K" in SEAICE_VARIABLE_FREEZING_POINT and in new |
| 38 |
|
version (with USE_ORIGINAL_SBI undef); |
| 39 |
|
o add hooks for new packages OASIS, the package will follow |
| 40 |
|
(hooks exclude the seaice pkg for now) |
| 41 |
|
o pkg/autodiff: ALLOW_AUTODIFF_WHTAPEIO |
| 42 |
|
Option to use mdsio_read_whalos.F/mdsio_write_whalos.F, |
| 43 |
|
rather than mdsio_readvector.F/mdsio_writevector.F, within |
| 44 |
|
adread_adwrite.F to read/write adjoint-related tapes to disk. |
| 45 |
|
o eesupp and pkg/mdsio: ALLOW_WHIO |
| 46 |
|
Option to write/read 2D fields to files including tiles halos. |
| 47 |
|
The main purpose is for adjoint related "tape I/O". |
| 48 |
|
The secondary purpose is for debugging phases. |
| 49 |
|
A word on those rather specifically intended purposes. |
| 50 |
|
The code was meant to be stripped down to the bare minimum. |
| 51 |
|
Typically, there is a unique and basic mapping to the global |
| 52 |
|
buffer (tile by tile, then proc by proc, then level by level), and |
| 53 |
|
byteswaps and meta-files are omitted. Indeed the (*_WH*) code is not |
| 54 |
|
meant to generate user friendly outputs of various shapes and forms. |
| 55 |
|
Rather it is intended, for the knowledgeabale developers, |
| 56 |
|
to be largely independent of the many mdsio options, to allow |
| 57 |
|
outputs of tiles+halos in debugging phases and in adjoint runs, |
| 58 |
|
and to allow some flexibility in tuning (adoint) I/O depending on each |
| 59 |
|
file system behavior. With regard to tuning, most of the potential knobs |
| 60 |
|
(to leave files open, to allocate the I/O indep. of usesinglecpuio, |
| 61 |
|
to specify directories, to specify the typical size of I/O calls, etc.) |
| 62 |
|
are not yet included, and will be added if they prove useful. |
| 63 |
|
o pkg/seaice: |
| 64 |
|
fix uninitialised arrays in USE_ORIGINAL_SBI version of seaice_solve4temp.F |
| 65 |
|
o pkg/seaice: |
| 66 |
|
new S/R seaice_solve4temp.F: this is the merged version of Ian Fenty |
| 67 |
|
code (seaice_budget_ice_if.F) which enable to recover old standard |
| 68 |
|
code (seaice_budget_ice.F) when option USE_ORIGINAL_SBI is defined. |
| 69 |
|
temporary: add "#define USE_ORIGINAL_SBI" at the top of seaice_solve4temp.F. |
| 70 |
|
o model/src (rStar): |
| 71 |
|
use simple average (not area weighted) of rStarFac at U,V point when using |
| 72 |
|
vectorInvariant and selectKEscheme=1,3 (consistent with mom. vertical advection) |
| 73 |
|
update output of exp. fizhi-cs-32x32x40 and fizhi-cs-aqualev20. |
| 74 |
|
|
| 75 |
checkpoint62k (2010/09/17) |
checkpoint62k (2010/09/17) |
| 76 |
o adjoint |
o adjoint |
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch