[MITgcm-devel] cube-sphere question

[Chris Hill]

J-M/A,

I did have a thought on some tests yesterday, but they are quite a pain to
set up. It might be good to create some symmetry tests that do flow around
the sphere with two cube faces land. There are three directions that could
be done - corresponding to my infamous nr, ng, nb. If you change coriolis
appropriately you should get the identical results for the same experiment
in each direction, even though the topology and which points are at nx+1 v.
1 etc... would be changing. An experiment along these lines with topography
could be useful to figure out whether there is a coding issue or a numerical
issue.

At the R2 meeting this coming week I want us to go over various "software
engineering" issues about tests and checkpoints etc..., so we can discuss
this then.

Chris  

> -----Original Message-----
> From: mitgcm-devel-bounces at mitgcm.org 
> [mailto:mitgcm-devel-bounces at mitgcm.org] On Behalf Of 
> Jean-Michel Campin
> Sent: Saturday, February 21, 2004 2:06 PM
> To: menemenlis at jpl.nasa.gov
> Cc: mitgcm-devel at mitgcm.org
> Subject: Re: [MITgcm-devel] cube-sphere question
> 
> Dimitri,
> 
> No problems. Everything is worth to check.
> 
> I don't know the reason for the edge-crash.
> Could still be a well hidden bug.
> But the thing that is particular to the edges is the 
> discontinuity in the grid-spacing derivatives, that can make 
> the high order viscosity (more generally, high order scheme) 
> less stable or less efficient in removing the grid noise. 
> This could become even more difficult when you have 
> topography gradients (or hFac gradients) at the edges. But I 
> don't know what can be tested.
> 
> See you,
> 
> Jean-Michel
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-devel
>