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

Re: 20050511: netcdf in double precision?

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 20050511: netcdf in double precision?
  • Date: Wed, 11 May 2005 11:46:00 -0600
>To: address@hidden
>From: Christine Delire <address@hidden>
>Organization: UCAR/Unidata
>Keywords: 200505111434.j4BEYkP3003223

Hi Christine,

> I run a  model that uses netcdf as format for its restart files (files  
> that are read by the model when restarted). For the restarts to be  
> seemless, I need netcdf to be compiled in double precision.
> As I never had to do that my-self, I am not 100% sure of what that  
> means. I am now trying to use my model on an IBM SP4 with the xlf  
> compiler. How should I compile the netcdf library ? with the -q64 flag  
> ?  -qrealsize=r8 ? both ?
> Any help would be greatly appreciated. I've been tangled in this issue  
> for a long time now.

You might want to check this support reply:

  
http://my.unidata.ucar.edu/cgi-bin/getfile?file=/content/support/help/MailArchives/netcdf/msg01602.html

I'm not sure what you mean by "I need netcdf to be compiled in double
precision" either.

If your model is compiled with -qrealsize=r8 and -q64, your netCDF
library should be compiled the same way.  That is, if your model is
Fortran and uses "-qrealsize=r8" as an xlf compiler flag, then in
building the netCDF library you should include "-qrealsize=r8" in
your FFLAGS environment variable or even in your compiler environment
variable FC before invoking the "configure" script to build the
library.  

--Russ