| 1 |
cnh |
1.1 |
ESMF setup that does cg on one PE set, several fg's on some other |
| 2 |
|
|
PE's and has cg2fg and fg2cg that distribute and aggregate needed |
| 3 |
|
|
fields respectively |
| 4 |
|
|
|
| 5 |
|
|
cg - runs on one PE |
| 6 |
|
|
fg's - run on 16 PE's |
| 7 |
|
|
cg2fg and fg2cg - run over all 17 PE's |
| 8 |
|
|
|
| 9 |
|
|
MITgcm setup for cg is a 20x20x20 box with 2km grid |
| 10 |
|
|
resolution |
| 11 |
|
|
|
| 12 |
|
|
MITgcm setup for fg's is a set of 25 20x1x20 non-hydrostatic slices |
| 13 |
|
|
100m grid resolution and a custom driver that switches between slices. |
| 14 |
|
|
For now all slices are assumed to have identical parameters and |
| 15 |
|
|
bathymetry and only the the "pickup" state information |
| 16 |
|
|
is switched by the custom driver. |
| 17 |
|
|
|
| 18 |
|
|
run sequence is |
| 19 |
|
|
|
| 20 |
|
|
initialization |
| 21 |
|
|
: |
| 22 |
|
|
: |
| 23 |
|
|
: |
| 24 |
|
|
|
| 25 |
|
|
timestepping |
| 26 |
|
|
|
| 27 |
|
|
cg & |
| 28 |
|
|
fg & |
| 29 |
|
|
BARRIER |
| 30 |
|
|
cg2fg & |
| 31 |
|
|
fg2cg & |
| 32 |
|
|
BARRIER |
| 33 |
|
|
|
| 34 |
|
|
finish |
Parent Directory
| 
Revision Log
| 
Revision Graph