On Mon, Feb 07, 2005 at 12:54:48PM -0600, Howard M. Salis wrote:
>
> Thank you for the advice! Your English is just fine (much better than my
> French). As you suggest, I think I will open up the file with a single
> processor, define any missing variables, and then close the file (flushing
> all data to disk). Then all of the processors can open the file, read it,
> and proceed as normal. I think that will work the best.
Since you are already using MPI, you might want to take a look at
our parallel netcdf project (www.mcs.anl.gov/parallel-netcdf). We use
the NetCDF file format but change the API slightly to be more
favorable to parallel programming.
It might be more work than you are willing to spend to adjust to our
programming model and API, however you will not have to worry as much
about processes getting a different view of the variables in a file.
It sounds like you've got one reasonable solution, but let me know if
you'd like more information about pnetcdf.
==rob
--
Rob Latham
Mathematics and Computer Science Division A215 0178 EA2D B059 8CDF
Argonne National Labs, IL USA B29D F333 664A 4280 315B
