pnf_test.F

[Robert Latham]

On Tue, Sep 23, 2003 at 05:14:55PM +0200, Reiner Vogelsang wrote:
> Hi Rob,
> I heard that the serial NetCDF has a feature which is called 'large file
> support' in order to overcome
> that restriction (?). Maybe, I misinterpreted a certain information.
> There was something mentioned on the NETCDF home page.

Oh, look at that :>  specifically, 

http://www.unidata.ucar.edu/packages/netcdf/faq.html#lfs

. If you don't use the unlimited dimension, only one variable can
  exceed 2 Gbytes in size, but it can be as large as the underlying
  file system permits. It must be the last variable in the dataset,
  and the offset to the beginning of this variable must be less than
  about 2 Gbytes.

. If you use the unlimited dimension, any number of record variables may
  exceed 2 Gbytes in size, as long as the offset of the start of each
  record variable within a record is less than about 2 Gbytes.

So you are right: with a few restrictions, people can write gigantic
files.  I don't know if the "offset to the beginning of the variable"
restrictions are a big restriction or just a minor annoyance.  

I'll check and make sure that netcdf defines the right flags (on linux
at least... ) to use large files, but i think since we rely on MPI-IO
for our file operations, we might be able to punt to the MPI-IO
implementation.  I bet we aren't that lucky though :>

==rob

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