[MITgcm-support] RE: cube halo

[Chris Hill]

 P.S. you should use XG and YG directly from the tileNNN files. Halo'ing
these is not good, but XC and YC are fine. You should only need XG and YG to
+1 in X and Y respectively. Thesse numbers are in the tileNNN files.

Chris

> -----Original Message-----
> From: mitgcm-support-bounces at mitgcm.org 
> [mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Chris Hill
> Sent: Friday, January 30, 2004 4:06 PM
> To: 'Benny Cheng'
> Cc: 'Ichiro Fukumori'; mitgcm-support at mitgcm.org
> Subject: [MITgcm-support] RE: cube halo
> 
> Yes 
> 
> > -----Original Message-----
> > From: Benny Cheng [mailto:bcheng at jpl.nasa.gov]
> > Sent: Friday, January 30, 2004 3:38 PM
> > To: cnh at mit.edu
> > Cc: 'Ichiro Fukumori'; 'Ben Cheng'; 'Ed Hill'
> > Subject: RE: cube halo
> > 
> > Ok.. I'll check it out... in the meantime, was there an 
> answer already 
> > to Ichiro's question about the XC YC grid ?
> > ie. do the grid points on a face halo matches exactly with 
> the XC YC 
> > grid points on the corresponding faces ?
> > 
> > - Benny
> > 
> > 
> > 
> > On Fri, 2004-01-30 at 09:23, Chris Hill wrote:
> > > Hi All,
> > > 
> > > The grid files ( call tileNNN.mitgrid ) already contain the
> > XC and YC
> > > for an overlap of one.
> > > They a multirecord files with XC and YC as the first two
> > records, the
> > > record size should be 511x511.
> > > Example code that reads them is in ini_curvilinear_grid.F
> > > 
> > > Chris
> > > 
> > > > -----Original Message-----
> > > > From: Ichiro Fukumori [mailto:if at pacific.jpl.nasa.gov]
> > > > Sent: Friday, January 30, 2004 12:16 PM
> > > > To: Chris Hill; 'Ben Cheng'
> > > > Cc: 'Benny Cheng'; 'Ed Hill'
> > > > Subject: RE: kalpana
> > > > 
> > > > Chris,
> > > > 
> > > > Benny needs Xc and Yc, i.e., the grid.
> > > > 
> > > > -- Ichiro
> > > > 
> > > > At 11:33 AM 1/30/2004 -0500, Chris Hill wrote:
> > > > >Hi Benny,
> > > > >
> > > > >  it would be offline. it would just be a call to a function
> > > > halouv()
> > > > >that would sit after
> > > > >  the
> > > > >  read(16,rec=(j-1)*nz+k) u
> > > > >  read(17,rec=(j-1)*nz+k) v
> > > > >
> > > > >  in your code that returns uhaloed and vhaloed
> > > > >
> > > > >Chris
> > > > >
> > > > > > -----Original Message-----
> > > > > > From: Ben Cheng [mailto:bnc at pacific.jpl.nasa.gov]
> > > > > > Sent: Friday, January 30, 2004 2:03 AM
> > > > > > To: Chris Hill
> > > > > > Cc: 'Ichiro Fukumori'; 'Benny Cheng'; 'Ed Hill'
> > > > > > Subject: RE: kalpana
> > > > > >
> > > > > >
> > > > > > Chris,
> > > > > > I do agree it can be somewhat easier to do within the
> > > > code itself,
> > > > > > since the halos are already built-in with the inputs.
> > > > > > However, I still need to do interpolation offline,
> > since we're
> > > > > > converting cubesphere u,v output fields to lat lon grid for 
> > > > > > comparison. What this means is that I still need to
> > get the halo
> > > > > > info for the XC,YC grid offline.
> > > > > >
> > > > > >  -Benny
> > > > > >
> > > > > > ps. got a doctor's appt. tomorrow morning , so won't
> > be in the
> > > > > > office till after noon (PST).
> > > > > >
> > > > > >
> > > > > > On Thu, 29 Jan 2004, Chris Hill wrote:
> > > > > >
> > > > > > > OK - lets hope that's not an omen.
> > > > > > >
> > > > > > > I have some paperwork to do for the INS tonight 
> and several
> > > > > > meetings
> > > > > > > tomorrow morning. However, thinking about what Benny is
> > > > > > trying to do
> > > > > > > it seems like it ought to be fairly easy to cut and
> > > > paste his code
> > > > > > > together with some pieces of the mitgcm fortran 
> code (to do 
> > > > > > > the
> > > > > > > exchanges) to give a code that reads, exchanges and then
> > > > > > interpolates.
> > > > > > > I plan to take a look at this tomorrow afternoon unless
> > > > > > Benny e-mails
> > > > > > > to say he is done. Benny does that make sense to you? The
> > > > > > mitgcm code
> > > > > > > can do exchanges without needing MPI etc...., the
> > > > > > configuration I have
> > > > > > > in mind would have just six tiles each 510x510 + overlaps.
> > > > > > Let me know if you think this makes sense.
> > > > > > >
> > > > > > > Chris
> > > > > > >
> > > > > > > > -----Original Message-----
> > > > > > > > From: Ichiro Fukumori [mailto:if at pacific.jpl.nasa.gov]
> > > > > > > > Sent: Thursday, January 29, 2004 8:01 PM
> > > > > > > > To: Chris Hill; 'Benny Cheng'; 'Ed Hill'
> > > > > > > > Subject: kalpana
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > FYI, I think the Ames altix was renamed after 
> one of the 
> > > > > > > > Columbia astronauts.
> > > > > > > >
> > > > > > > > -- Ichiro
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > 
> > > > ===============================================================
> > > > Ichiro Fukumori               | Jet Propulsion Laboratory
> > > > e-mail: fukumori at jpl.nasa.gov | Mail Stop 300-323
> > > > phone: +1 (818) 354-6965      | 4800 Oak Grove Drive
> > > > fax: +1 (818) 393-6720        | Pasadena, California 
> 91109, U.S.A.
> > > > ===============================================================
> > > > 
> > > 
> > > 
> > 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support
>