| ... | @@ -59,24 +59,27 @@ Normally now, you have something like this on your directory: |
... | @@ -59,24 +59,27 @@ Normally now, you have something like this on your directory: |
|
|
```
|
|
```
|
|
|
on NEMO_CONFIGS/models/NEMO/dev_XXXXX_config_legos/cfgs/CONFIG_CASE/EXP00/ :
|
|
on NEMO_CONFIGS/models/NEMO/dev_XXXXX_config_legos/cfgs/CONFIG_CASE/EXP00/ :
|
|
|
|
|
|
|
|
namelist_top_ref
|
|
iodef.xml # Xios file that manage The I/O writing
|
|
|
namelist_ref
|
|
context_nemo.xml # for nemo modele
|
|
|
namelist_pisces_ref
|
|
|
|
|
iodef.xml
|
|
|
|
|
nemo.exe -> nemo
|
|
|
|
|
namelist_pisces_cfg
|
|
|
|
|
domain_def_nemo.xml
|
|
|
|
|
field_def_nemo-oce.xml
|
|
|
|
|
file_def_nemo-oce.xml
|
|
file_def_nemo-oce.xml
|
|
|
context_nemo.xml
|
|
field_def_nemo-oce.xml
|
|
|
grid_def_nemo.xml
|
|
grid_def_nemo.xml
|
|
|
|
domain_def_nemo.xml
|
|
|
|
|
|
|
|
file_def_nemo-pisces.xml # for pisces model
|
|
|
field_def_nemo-pisces.xml
|
|
field_def_nemo-pisces.xml
|
|
|
file_def_nemo-pisces.xml
|
|
|
|
|
namelist_cfg
|
|
namelist_cfg # nam nemo model
|
|
|
namelist_top_cfg
|
|
namelist_ref
|
|
|
|
namelist_top_cfg # nam tracer top level
|
|
|
|
namelist_top_ref
|
|
|
|
namelist_pisces_cfg # nam pisces model
|
|
|
|
namelist_pisces_ref
|
|
|
|
|
|
|
app.conf
|
|
app.conf
|
|
|
CONFIG_CASE.db
|
|
CONFIG_CASE.db
|
|
|
nemo -> $WRKDIR/NEMO_CONFIGS/models/NEMO/dev_XXXXX_config_legos/cfgs/CONFIG_CASE/BLD/bin/nemo.exe
|
|
nemo -> $WRKDIR/NEMO_CONFIGS/models/NEMO/dev_XXXXX_config_legos/cfgs/CONFIG_CASE/BLD/bin/nemo.exe
|
|
|
|
nemo.exe -> nemo
|
|
|
run_nemo.ksh
|
|
run_nemo.ksh
|
|
|
save_nemo.ksh
|
|
save_nemo.ksh
|
|
|
```
|
|
```
|
| ... | @@ -101,7 +104,10 @@ empty |
... | @@ -101,7 +104,10 @@ empty |
|
|
|
|
|
|
|
User needs the mesh mask file for preprocessing inputs. It contains all geographical and grid data from the configuration nemo. You need to launch the model during almost one time step, and then stop the job. The model will produce the mesh mask file. As it produce it for each mpi domains, user needs to use the rebuild_nemo tools to create it for the whole domain.
|
|
User needs the mesh mask file for preprocessing inputs. It contains all geographical and grid data from the configuration nemo. You need to launch the model during almost one time step, and then stop the job. The model will produce the mesh mask file. As it produce it for each mpi domains, user needs to use the rebuild_nemo tools to create it for the whole domain.
|
|
|
|
|
|
|
|
__prerequisites__
|
|
---
|
|
|
|
1. Prepare analytical run
|
|
|
|
|
|
|
|
_prerequisites_ :
|
|
|
|
|
|
|
|
Check if specific cpp key are disabled , go to the cpp file and remove key_top key
|
|
Check if specific cpp key are disabled , go to the cpp file and remove key_top key
|
|
|
|
|
|
| ... | @@ -116,20 +122,23 @@ __prerequisites__ |
... | @@ -116,20 +122,23 @@ __prerequisites__ |
|
|
> **Note** : The first step could be useless if you rebuild nemo in this way
|
|
> **Note** : The first step could be useless if you rebuild nemo in this way
|
|
|
./makenemo –m X64_IRENE –r TATL025BIO-XX -n CONFIG_CASE -j8 -v3 **del_key key_top** > log_build_TATL025BIO.txt 2>&1
|
|
./makenemo –m X64_IRENE –r TATL025BIO-XX -n CONFIG_CASE -j8 -v3 **del_key key_top** > log_build_TATL025BIO.txt 2>&1
|
|
|
|
|
|
|
|
save your namelist_cfg file with another name (it will be use later) and just modify these key on your namelist_cfg file
|
|
> **Note** : Save your namelist_cfg file with another name (it will be use later) and just modify these key on your namelist_cfg file
|
|
|
|
|
|
|
|
__modification of namelist_cfg file__
|
|
_modification of includefile_ :
|
|
|
|
|
|
|
|
|
In order to have small simulation (larger than the model time step could be better ....), in days unit, for instance, put :
|
|
|
|
```
|
|
|
|
DAYS=2
|
|
|
|
```
|
|
|
|
|
|
|
|
on NEMO_CONFIGS/models/NEMO/dev_XXXXX_config_legos/cfgs/CONFIG_CASE/EXP00/:
|
|
_modification of namelist_cfg file_ :
|
|
|
|
|
|
|
|
|
|
|
|
|
1. Modify your namelist_cfg (specific analytical without forcing input):
|
|
first look on includefileini the JPNI and JPNJ variable ($JPNI*$JPNJ=$NOCEAN) for non-optimization use, you need to turn off the optimized MPI tiles parameters, put on section &nammpp :
|
|
|
first look on includefileini the JPNI and JPNJ variable for non-optimization use, you need to turn off the optimized MPI tiles parameters, put on section &nammpp :
|
|
|
|
|
|
|
|
|
|
```
|
|
```
|
|
|
jpni = 6 ! $JPNI
|
|
jpni = $JPNI
|
|
|
jpnj = 23 ! $JPNJ
|
|
jpnj = $JPNJ
|
|
|
```
|
|
```
|
|
|
|
|
|
|
|
and on &namdom section
|
|
and on &namdom section
|
| ... | @@ -151,15 +160,15 @@ on NEMO_CONFIGS/models/NEMO/dev_XXXXX_config_legos/cfgs/CONFIG_CASE/EXP00/: |
... | @@ -151,15 +160,15 @@ on NEMO_CONFIGS/models/NEMO/dev_XXXXX_config_legos/cfgs/CONFIG_CASE/EXP00/: |
|
|
ln_rnf = .false. # No runoff
|
|
ln_rnf = .false. # No runoff
|
|
|
ln_bdy = .false. # No BC
|
|
ln_bdy = .false. # No BC
|
|
|
```
|
|
```
|
|
|
|
|
---
|
|
|
3. Launch job
|
|
2. Launch job
|
|
|
|
|
|
|
|
```
|
|
```
|
|
|
ccc_msub run_nemo.ksh
|
|
ccc_msub run_nemo.ksh
|
|
|
# sbatch run_nemo.ksh
|
|
# sbatch run_nemo.ksh
|
|
|
```
|
|
```
|
|
|
|
|
---
|
|
|
4. Look on your scratch the rundir path ($NOCEAN : number of procs for nemo) for the 1st simulation (X=1)
|
|
3. Look on your scratch the rundir path ($NOCEAN : number of procs for nemo) for the 1st simulation (X=1)
|
|
|
```
|
|
```
|
|
|
cp $TOOLS_REBUILDNEMO_DIR/rebuild_nemo* $SCRATCHDIR/TMP/${CONFIG_CASE}_X/.
|
|
cp $TOOLS_REBUILDNEMO_DIR/rebuild_nemo* $SCRATCHDIR/TMP/${CONFIG_CASE}_X/.
|
|
|
./rebuild_nemo mesh_mask $NOCEAN
|
|
./rebuild_nemo mesh_mask $NOCEAN
|
| ... | |
... | |
| ... | | ... | |