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

20040225: Problem with f90 test for netCDF: pgf90 compiler

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 20040225: Problem with f90 test for netCDF: pgf90 compiler
  • Date: Thu, 26 Feb 2004 09:19:48 -0700
Jeff,

>Date: Thu, 26 Feb 2004 10:35:56 -0500
>From: "Jeff Konz" <address@hidden>
>Organization: Linux Networx
>To: "'Steve Emmerson'" <address@hidden>
>Subject: RE: 20040225: Problem with f90 test for netCDF: pgf90 compiler
> Keywords: 200402252015.i1PKFrrV004061 netCDF F90

The above message contained the following:

> I really need the f90 interface. I tried with the Intel compiler and it has
> even more problems.
> 
> I set the optimization level on pgi to -O1 and still had the same error.
> Can you give me a hint on which routines to look at for this error.

I would start debugging at line 97 of the file "f90/netcdf_test.f90":

    call check(nf90_put_var(ncFileID, pressVarID, pressure(:, :, 2), start = (/ 
1,
   1, 2 /)) )

This is where the incorrect values are written to the output netCDF
file.  From within the nf90_put_var routine, I would look at the first
pressure value; it should be 962 (but is probably 950).

> Do you think it's a Fortran problem or in the C routines in libnetcdf that
> the test program links in.

I think it's a problem with the pgf90 compiler.  I think the compiler
isn't correctly handling the "pressure(:, :, 2)" expression.

> Thanks,
> 
> -Jeff
> 
> Jeff Konz
> Linux Networx
> Applications Specialist
> Mobile: 248.219.2814

Regards,
Steve Emmerson