Re: ESMF_grid stalls when converting many fields

Forgot to add something.

[4c] below does reuire some understanding of netCDF's
      predefine mode.

--
You can use the 'simple' method, to write out a nc file
for *each* regridded variable; then use the netCDF operators
to merge the output files.
On 5/1/13 11:06 AM, Dennis Shea wrote:
> Not much info to determine the problem.
>
> [1] 200MB is not particularly large. Nor is 40-50 variables.
>
> [2] How did you generate the weight file?
>       Did you write the NCL script or did you use the template at
>          http://www.ncl.ucar.edu/Applications/ESMF.shtml
>          Example 25
>       This would generate a weight file which you could use for
>       all subsequent regridding via:
>
> http://www.ncl.ucar.edu/Document/Functions/ESMF/ESMF_regrid_with_weights.shtml
>
> [3] How are you writing the netCDF file? NCL allows 2 different ways
>       to create netCDF: one is the 'simple but, possibly, inefficient'
>       method; the other requires predefintion of the output
>       netCDF file's contents.
>
>                http://www.ncl.ucar.edu/Applications/o-netcdf.shtml
>
>       While 40-50 is not a large number of variables per se, When you
>       use the simple method this could be slow for assorted reasons.
>
> [4] The following may be useful to look at.
>       (a) generate weight file
>        http://www.cgd.ucar.edu/~shea/se2fv_esmf.genwgt.ncl
>
>       (b) use weight file generated in (a) to regrid multiple
>           variables and write to netCDF via simple* method
>
>        http://www.cgd.ucar.edu/~shea/se2fv_esmf.regrid2file.ncl
>
>       (c) same as (b) BUT uses 'efficient (predefine)' method to
>           write netCDF. This requires a double pass over the data
>           (i) predefine output file's contents (structure);
>           (ii)write regridded variables to predefined file.
>
>        http://www.cgd.ucar.edu/~shea/se2fv_esmf.regrid2file.ncl_predefine
>
> On 05/01/2013 08:54 AM, Patrick Kelly wrote:
>> Hello NCL community–
>>
>> I am trying to use the 'ESMF_grid'
>> to convert output from the CESM HOMME model to a rectilinear grid.
>>
>> The file is about 200mb large and contains maybe 40-50 variables.
>>
>> The procedure works fine at the start, but then stalls about half way
>> through without any error message.
>> NCL doesn't quit, it just "hangs" Indefinitely.
>>
>> IS this a memory limit or stack issue thing with NCL? I thought I'd
>> receive some error output if it was...
>>
>> Thanks for any leads,
>> Patrick
>>
>>
>> -----------------------------------------------
>> Patrick Kelly, Ph.D.
>> Postdoctoral Research Associate
>> University of Miami / RSMAS
>> email: pkelly@rsmas.miami.edu <mailto:pkelly@rsmas.miami.edu>
>>
>>
>> _______________________________________________
>> ncl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk