non contiguous data IO

Robert Latham robl at mcs.anl.gov
Mon Oct 16 10:07:02 CDT 2006


On Mon, Oct 16, 2006 at 10:28:50AM +0200, Seiss, Guntram wrote:
> This is not a very elegant way, but it works, in principle. But I
> have recognized, it's a better solution to get all data of a process
> sorted contiguous inside one bloc of the record, before writing
> them, to avoid multiple calls of nfmpi_put_vara_real.

Hi

You definitely have the right idea.  In serial netcdf, I would suspect
one does not pay a very large penalty when making 3 small calls or one
large call. In parallel-netcdf, though, there are significant benefits
to passing as much information as possible into each call into the
API.  

Thanks for the update.

==rob

-- 
Rob Latham
Mathematics and Computer Science Division    A215 0178 EA2D B059 8CDF
Argonne National Labs, IL USA                B29D F333 664A 4280 315B




More information about the parallel-netcdf mailing list