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

Re: 20010306: NetCDF 2.4.3 on VMS

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 20010306: NetCDF 2.4.3 on VMS
  • Date: Tue, 06 Mar 2001 20:57:37 -0700
>To: address@hidden
>From: Zane AE <address@hidden>
>Subject: NetCDF 2.4.3 on VMS
>Organization: Caltech
>Keywords: 200103070020.f270KTL02326 netCDF 2.4.3 DEC VMS

Hi Zane,

>       Hi.  I'm using NetCDF 2.4.3 on a DEC Alpha box, and am having
> trouble using floating point data.

You don't say why you're using version 2.4.3 rather than some some
version such as 3.4, or whether you built the VMS version from source
and it passed the included tests.  But I'm guessing it was built
from source using the wrong floating-point format (as I recall, VMS
supported about 3 machine representations for floating-point), which
would have been caught with a "make test" after it was built.

>       I first wrote a simple test program that stored some data as
> floating point numbers, and the file that it created, when ncdump'ed had
> several problems.  First, only the first value of the record variable was
> meaningful, the other four apparently came out of other parts of memory.
> Similarly, all of the floating point data I had stored was gone, and
> replaced by outlandish figures like 2.99333721e-22, either that, or ncdump
> would crash when it got to displaying the data part of the file.
>       I changed the type of the data to be integer, and all worked well.
> The same program, using floating point data, compiled and run on a Unix
> machine, also runs fine.
>       I then wrote a CDL template for my floating point data, and ran
> ncgen on it, and ran that program to generate a NetCDF file, and then
> ncdump'ed that file with results similar to my own hand written program.
>       I've talked to one other person who has had a similar experience
> on VMS. Is this a known problem?  If so, is there a known work-around?
> I'd appriciate any information you can provide.  Thanks.  Zane AE

Here's a couple of possible work-arounds:

 - See <http://www.unidata.ucar.edu/glimpse/netcdf/3738> for a way to
   just install the binary for netCDF 2.4.3 from a VMS BACKUP save set
   for alpha-vms from
   <ftp://ftp.unidata.ucar.edu/pub/binary/vms-alpha/netcdf-2_4_3.bck>

 - See
   <ftp://sideshow.jpl.nasa.gov/pub/mjr/netcdf/netcdf-3.4_one_source.tar.gz>
   for a version of the netCDF 3.4 source that claims to be portable
   to Alpha VMS as well as lots of other platforms.

 - Try building from our latest source (which we haven't tested on
   Alpha VMS, because we don't have access to one of those anymore)
   from
     ftp://ftp.unidata.ucar.edu/pub/netcdf/netcdf-beta.tar.Z

All the versions, 2.4.3, 3.4, and 3.5beta use the same file format, so
any of them you can get to work should be suitable for most uses.

--Russ

_____________________________________________________________________

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