[mpas-developers] 1/10 degree problems

[Michael Duda]

Mat,

that's interesting that the output.nc file appeared for a while, 
then disapeared. Could the issue be related to the large size of
the output.nc file, and could there be a similar problem when 
reading the large grid.nc file? It might be necessary here to 
force the use of 64-bit file offsets in netCDF accesses; if you
get a chance, could you check whether uncommenting

FILE_OFFSET = -DOFFSET64BIT

in the Makefile resolves the issue of creating an output.nc file?

Cheers,
Michael


On Fri, Apr 16, 2010 at 03:40:34PM -0600, Mathew Maltrud wrote:
> 
> On Apr 16, 2010, at 3:22 PM, Xylar Asay-Davis wrote:
> 
> > It sounds like there isn't an output.nc to look at at all.
> 
> Xylar is correct.  the weird thing is that the output.nc file does get  
> created (i can ls it while the code is running), but disappears after  
> the run is done.
> 
> the graph.info.part.64 seems to look ok, though i only looked at a few  
> thousand lines out of 5.8 million.
> 
> i just wanted to know if it was anything obvious (like the 'F' being a  
> timer thing) that i might waste time on.  i'll run under totalview  
> next week to see if i can figure it out.
> 
> thanks for the help...
> -mat
> _______________________________________________
> mpas-developers mailing list
> mpas-developers at mailman.ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/mpas-developers