Changes between Version 8 and Version 9 of Users/Reference Configurations/GYRE_PISCES


Ignore:
Timestamp:
2017-05-17T11:53:05+02:00 (3 years ago)
Author:
nicolasmartin
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Users/Reference Configurations/GYRE_PISCES

    v8 v9  
    1111 
    12121. Choose the grid size 
    13 In `CONFIG/GYRE/EXP00`, edit your `namelist_cfg` file to change the `jp_cfg`, `jpi`, `jpj`, `jpk` variables in &namcfg: 
    14 ||JP_CFG        ||JPI   ||JPJ   ||JPK   ||Number of points      ||Equivalent to|| 
    15 ||1             ||  30  ||  20  ||101   ||60600                 ||GYRE 1°      || 
    16 ||25            || 750  || 500  ||101   ||37875000              ||ORCA 1/2°    || 
    17 ||50            ||1500  ||1000  ||101   ||151500000             ||ORCA 1/4°    || 
    18 ||150           ||4500  ||3000  ||101   ||1363500000            ||ORCA 1/12°   || 
    19 ||200           ||6000  ||4000  ||101   ||2424000000            ||ORCA 1/16°   || 
     13 
     14   In `CONFIG/GYRE/EXP00`, edit your `namelist_cfg` file to change the `jp_cfg`, `jpi`, `jpj`, `jpk` variables in &namcfg: 
     15   ||  `jp_cfg`  ||  `jpi`  ||  `jpj`  ||  `jpk`   ||  Number of points  ||  Equivalent to  || 
     16   ||           1||       30||       20||   101    ||               60600||GYRE 1°          || 
     17   ||          25||      750||      500||   101    ||            37875000||ORCA 1/2°        || 
     18   ||          50||     1500||     1000||   101    ||           151500000||ORCA 1/4°        || 
     19   ||         150||     4500||     3000||   101    ||          1363500000||ORCA 1/12°       || 
     20   ||         200||     6000||     4000||   101    ||          2424000000||ORCA 1/16°       || 
    2021 
    21222. In `namelist_cfg` again, avoid problems in the physics (and results will not be meaningful in terms of physics) by setting `nn_bench = 1` in &namctl 
    2223{{{#!f 
    23 nn_bench    =    1     !  Bench mode (1/0): CAUTION use zero except for bench 
     24   nn_bench    =    1     !  Bench mode (1/0): CAUTION use zero except for bench 
    2425}}} 
    2526 
    26273. If you increase domain size, you may need to decrease time-step (for stability) by changing `rn_rdt` value in &namdom (i.e. for `jp_cfg = 150`, ORCA12 equivalent, use `rn_rdt = 1200`) 
    2728{{{#!f 
    28 rn_rdt      = 1200.     !  time step for the dynamics 
     29   rn_rdt      = 1200.     !  time step for the dynamics 
    2930}}} 
    3031 
    31324. Optional, in order to increase the number of MPI communication for benchmark purposes: you can change the number of sub-timesteps computed in the time-splitting scheme each iteration. First change the list of active CPP keys for your experiment, in `CONFIG/"your configuration name"/cpp_"your configuration name".fcm` : replace key_dynspg_flt by key_dynspg_ts and recompile/create your executable again 
    3233{{{#!sh 
    33 makenemo [...] add_key "key_dynspg_ts" del_key "key_dynspg_flt" 
     34   makenemo [...] add_key "key_dynspg_ts" del_key "key_dynspg_flt" 
    3435}}} 
    3536In your `namelist_cfg` file, edit the &namsplit namelist by adding the following line:  
    3637{{{#!f 
    37 nn_baro       =    30               !  Number of iterations of barotropic mode/ 
     38   nn_baro       =    30               !  Number of iterations of barotropic mode/ 
    3839}}} 
    3940`nn_baro = 30` is a kind of minimum (we usually use 30 to 60).So than increasing the `nn_baro` value will increase the number of MPI communications.