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

Re: 19980818: building netcdf on CRAY J90: ncvgt1 failure

George,

>Date: Wed, 19 Aug 1998 16:00:35 -0400
>From: "George S. Lai" <address@hidden>
>Organization: NASA/GSFC
>To: address@hidden
>Subject: Re: 19980818: building netcdf on CRAY J90: ncvgt1 failure
>Keywords: 199808181657.KAA22331

In the above message, you wrote:

>   >      1.  Send me the output from the command "uname -a".
> 
>      sn9609 charney 9.2.0.05 roo.5 CRAY J90
> 
>   >      2.  Tell me what versions of the C and Fortran compilers you're
>   >   using.
> 
>      f90:    Cray CF90 Version 3.0.1.3 08/18/98 17:15:09
>      cc :    Cray Standard C Version 6.0.1.3 08/18/98 17:15:30
> 
>   >      3.  Go to the top-level, netCDF source directory.
> 
>   >      4.  Execute the command "make distclean".
> 
>   >      5.  Remove the file "config.cache" if it exists.
> 
>   >      6.  Set your environment variables as before -- but don't use the 
>   >   "-DUSE_IEG" option and remove the space in the "CFLAGS= -O3"
>   >   setting (it's a typographical error).
> 
>   >      7.  Execute the configure script.  Redirect standard output and
>   >   standard error to the file "configure.log".  Send me the file.
> 
>   >      8.  Send me the file "config.log".
> 
>   >      9.  Execute the command "make".  Redirect standard output and
>   >   standard error to the file "make.log".  Send me the file.
> 
>   >      10. Execute the command "make test".  Redirect standard output and
>   >   standard error to the file "test.log".  Send me the file.
> 
>      After I re-did the above, now it passed the test!

Great!

>      I think I must have
>      done something wrong at the first try and the config.cache kept having
>      its effects. Thanks for the help. One last question is, would the package
>      work without "-dp" on FFLAGS?

The "-dp" option is necessary for all Fortran code that passes a
DOUBLEPRECISION argument to any netCDF routine.  It equivalences
Fortran's DOUBLEPRECISION datatype with C's "double" datatype, which is
necessary because the netCDF library is written in C.

--------
Steve Emmerson   <http://www.unidata.ucar.edu>