[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Mark is onboard..

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: Re: Mark is onboard..
  • Date: Tue, 02 Nov 2010 09:57:29 -0600
"Schuchardt, Karen L" <address@hidden> writes:

> Hi Ed, I'm having problems running my model after converting to the
> netcdf4 api.  In particular, I get the dimensions defined but then
> errors when I try to define the variables.  It seems to not like my
> dimensions but the value looks good to me.  The first variable I try
> to write is time which is dimensioned by the time dimension which has
> a value of 1.  I get NetCDF: Invalid dimension ID or name.

Sorry, it's not clear to me what is happening, but perhaps this example
will help. Your case sounds similar to this example:

http://www.unidata.ucar.edu/software/netcdf/docs/netcdf-tutorial.html#pres_005ftemp_005f4D-in-F90

> Is there a test program for the f90 interface with parallel IO?  I
> don't see one in the source distribution.  In particular I'm looking
> for code that uses methods like nf90_open_par, nf90_create_par and the
> nf90 methods in general.

Try nf_test/f90tst_parallel.f90.

>
>  I'm using nersc/cray installed 4.0.1.3 parallel with hdf5 1.8.3 since
>  its the default and I'm not clear exactly what requirements for
>  modules exist for 4.1.1.0.
>

Ah, that may be a problem. You really want to build hdf5-1.8.5-patch1,
and, for performance reasons, you certainly want to get the latest
netCDF snapshot. I have made recent improvements which dramatically
affect performance.

ftp://ftp.unidata.ucar.edu/pub/netcdf/snapshot/netcdf-4-daily.tar.gz

If you ask your local sysadmin, they can probably help you build HDF5,
it's not hard.

Thanks,

Ed
-- 
Ed Hartnett  -- address@hidden