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

Re: 20040330:unaligned access pid=##### va =0x

>To: <address@hidden>,
>To: <address@hidden>
>From: "Joseph Benckert" <address@hidden>
>Subject: netcdf troubles
>Organization: UCAR/Unidata
>Keywords: 200403291443.i2TEhsrV014206

Hi Joseph,

> Sorry for the newb questions but I cannot figure out what I'm doing wrong
> here.   I'm trying to write multiple variables to one dataset (fortran code
> attached).  I'm using two utitlites to view my data: ncBrowse, and ncDump.
> When using ncdump, I'm getting a lot of errors like 'unaligned access
> pid=##### va =0x....' but I get a listing of all the variables.  When using
> ncBrowse the only variable I can see is the first variable.  Do you see
> anything that I'm doing obviously wrong?
> 
> Also, I did a search on the netcdf archives and a support person posted that
> this might be an issue with telling the CPU to ignore unaligned access
> errors.  I still don't understand why I can't see the variables in both
> viewing programs.
> 
> This e-mail has gone to PSC support and netCDF support.  Those at PSC can
> see the errors by doing an ncdump on any file in the $SCRATCH directory that
> begins with: UG5sf_
> 
> Joseph A. Benckert
> Department of Physics
> North Carolina State University
> address@hidden

Was the ncdump you tried built for the platform on which you're
running it?

Please try the ncdump on a known good netCDF file to make sure you're
not just seeing a problem with ncdump.  There are sample netCDF files
in the netCDF source distribution, or you can create them with the
ncgen utility, or you can grab any .nc file from here:

 ftp://ftp.unidata.ucar.edu/pub/netcdf/sample_data/

and run ncdump on it.

The program you sent is not complete enough to compile, so I can't
duplicate the problem you're seeing.  If you want to make one of your
netCDF files available to us, we can test if our ncdump can read
them.

Just looking at your program, I see a few problems.  NF90_STRERROR is
a function, not a subroutine, so a more appropriate way to use it than
calling it would be something like:

   if(status /= nf90_noerr) print *, nf90_strerror(status)

But all your calls to netCDF functions when ndim == 1 look OK to me.

--Russ

_____________________________________________________________________

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