[MITgcm-devel] Re: MITgcm at gfdl

[Stephanie Dutkiewicz]

I come across a similar problem. Particularly when the bathymetry
fell exactly in the mid-depth. Some compilers round up,
and some round down. This hasn't been fixed (as far as I
know). 
My "fix" was to use the Depth.data file created on the
machine that the pickups came from...

steph

On Wed, 9 Jun 2004, Patrick Heimbach wrote:

> Tom,
> 
> that does indeed sound worrying, I haven't heard about this issue.
> I have been switching between SGI, IBM, and Beowulf,
> and can't remember having seen this kind of thing.
> Are you using identical parameters in your 'data' file
> (e.g. parameter hFacMin, hFacMindr)?
> 
> -Patrick
> 
> Quoting THOMAS HAINE <thomas.haine at jhu.edu>:
> 
> > Hi Patrick, 
> > 
> > Thanks a lot for your help. I've managed to build an mpi executable
> > successfully for my 9km resolution regional Irminger Sea model (Dan Lea's
> > forward code).
> > 
> > I had 1 small problem (unrelated to mpi): hFacC is not identical on the GFDL
> > SGIs to the hFacC I get under pIII linux g77. So, when I try and restart on
> > the SGI from my linux pickup files I bomb out in ini_theta.F with theta
> > identically zero in a couple of gridcells. Presumably, everything will be OK
> > once I write out a new SGI pickup file. But it's a bit worrying that the
> > topography masks are machine dependent. I'm committed to an ecco variant
> > (inherited from Jake Gebbie), so I can't simply update to the latest release.
> > But perhaps this hFacC issue has now been fixed?
> > 
> > Cheers, Tom.
> > 
> > 
> > ----- Original Message -----
> > From: Patrick Heimbach <heimbach at MIT.EDU>
> > Date: Tuesday, June 8, 2004 5:42 pm
> > Subject: Re: MITgcm at gfdl
> > 
> > > Hi Tom,
> > > 
> > > I've placed a README_FOR_TOM file under
> > > ~psh/ecco/latest/verification/global_ocean.90x40x15/
> > > which walks you through the steps on how download,
> > > compile and run the model at GFDL for
> > > the verification/global_ocean.90x40x15/
> > > experiment.
> > > 
> > > Let me know if you have problems.
> > > Cheers
> > > -Patrick
> > > 
> > > 
> > > 
> > > Quoting THOMAS HAINE <thomas.haine at jhu.edu>:
> > > 
> > > > Hi Patrick, 
> > > > 
> > > > I'm spending some time at GFDL this summer and I plan to run 
> > > MITgcm. I
> > > > understand you have been doing this. Would you mind telling me 
> > > how to
> > > > configure MITgcm for their cluster please?  You could share your 
> > > makefile> etc., for example. I'd really appreciate it!
> > > > 
> > > > Cheers, Tom.
> > > > 
> > > > 
> > > 
> > > 
> > > _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
> > > Patrick Heimbach     Massachusetts Institute of Technology
> > > FON: +1/617/253-5259                    EAPS, Room 54-1518
> > > FAX: +1/617/253-4464               77 Massachusetts Avenue
> > > heimbach at mit.edu                 Cambridge MA 02139
> > > http://www.mit.edu/~heimbach/                          USA
> > > 
> > > 
> > 
> > 
> 
> 
> _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
> Patrick Heimbach     Massachusetts Institute of Technology
> FON: +1/617/253-5259                    EAPS, Room 54-1518
> FAX: +1/617/253-4464               77 Massachusetts Avenue
> mailto:heimbach at mit.edu                 Cambridge MA 02139
> http://www.mit.edu/~heimbach/                          USA
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-devel
>