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

Re: 20030915:File Offset questions related to 2GB dataset sizes (patch attached)

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 20030915:File Offset questions related to 2GB dataset sizes (patch attached)
  • Date: Mon, 13 Oct 2003 13:01:25 -0600
On Mon, 13 Oct 2003 13:08:18 -0500, Robert Latham wrote:

> Greg, thanks very much for sending along this patch.  This has been a
> showstopper for at least one site using our software.  
> 
> Russ, i'm not sure how to keep the parallel-netcdf and netcdf-3.5
> approaches in sync, but i'm not planning on making any changes to
> Greg's patch.  Should i just keep an eye on your 3.5 branch to make
> sure we continue to address the 64-bit filesize issue the same way?
> That sounds doable, but error-prone...

I hope to integrate Greg's changes into netcdf-3.5-beta13 soon to
create netcdf-3.6, but I'm currently in a crunch that will delay the
necessary testing until mid November, when the Unidata Program Center
will have moved to a different location for about six months.  We also
need to update out autoconf/configure to detect and handle large file
support properly.

You're right, just looking at our releases to stay in sync will be
error-prone, but we'll try to keep you in mind and not make any
gratuitous changes, like re-indenting everything to a different style
:-).  I'm not sure what the alternative is unless we were to try to
merge netcdf-3.6 with pnetcdf.  If that could be done without much
impact on non-parallel users, it would be worth considerable effort,
but we're also committed to netCDF-4 on HDF5, and currently have 1.25
FTE plus a student assistant for both efforts.  Once you start
optimizing pnetcdf, things may diverge even more.  If you have any
better ideas what we could do to help keep things in sync, please let
me know.

Meanwhile, I think it's great that your group is making such great
progress on pnetcdf.

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden          http://www.unidata.ucar.edu/staff/russ