Problems generating the MPAS-A static file

[thaisa]

I'm having trouble generating the static file with MPAS-A. The first time I tried to run it with the 15-3km resolution mesh x5.8060930.grid.nc (with refinement for South America) using x5.8060930.cvt.part.360 but it stopped on the way and didn't finish (log file attached). I thought it could be the nodes and processors I used, but I've tried using fewer nodes and more nodes (as 36 nodes and 10 processors). After a few attempts the problem persisted, including when using another partition x5.8060930.graph.info.part.512, which also didn't work. Then I tried using as a test other meshes (for example x1.10242.grid.nc) and the same problem occurred (log file below), it doesn't finish and doesn't generate the static file. Please note: The first time I followed the tutorial and it went perfectly, but when I tried using a higher-resolution mesh, I started having problems. if you have any suggestions on how I should proceed?

namelist.init_atmosphere and streams.init_atmosphere files are attached below.

 

[afernandezody]

@thaisa. I'll let the NCAR Team chime in with a more specific answer but, in the meantime, you might try generating the files one-by-one (firstly the static file only, then moving on to the init file and so forth) instead of all at-once. Cheers. Arturo

 

[Ming Chen]

Hi, thaisa,

There is no error message in your log.init_atmopshere files. The case seemed to simply stop. Did you look at your log.init_atmopshere,err files? I wonder whether you can find some error messages in those files that may be helpful for us to figure out what is wrong.

In your streams.init_atmopshere, you specify the grid file "SouthAmerica_elip.grid.nc". Can you clarify how you created this file?

I notice that you set

config_geog_data_path = '/glade/campaign/mmm/wmr/mpas_tutorial/mpas_static/

Did you run this case in derecho? If so, please let me know the directory where you run this case.

 

[thaisa]

Hello, sorry for the delay in replying, the problem has been solved. It was simply a matter of adding "mpiexec -n 360" to the job submission file. Thank you for your attention and help.