[MITgcm-devel] Re: [MITgcm-cvs] MITgcm/pkg/mom_vecinv CVS Commit

[Alistair Adcroft]

I've finally debugged my re-implementation of Leith in the context of 
the latest code (it was working in October and is now working again).

The big difference is in the calculation of length scales and in my new 
implementation I take into account the notional stencil so I have 
factors of 6,8,16,20 in various places and it knows about the aspect 
ratio. The bottom-line is that a coefficient of 1 is techically the 
formal stable limit and in practice it differs from this only because of 
Adams-Bashforth and other linear terms (e.g. bottom drag).

Before checking this in I want to provide an indication of how to 
translate coefficients used in the present schemes to coefficients used 
in the new scheme (which will be parallel to the old).

Rather than run loads of tests find the stable range of coefficients for 
the old scheme can we have an informal survey of usuable values for 
viscAhGridMax and viscA4GridMax that seem to work? R.S.V.P. A.S.A.P.

Thanks,

Alistair.

Baylor Fox-Kemper wrote:
> Hi Alistair,
>   No, I thought you did OK.  I only meant that I expected you to change 
> things in future, at which point the bitwise comparision would be invalid.
>   However, adapting your suggestion allowed me to come up with a version 
> that is bitwise identical if viscleithd=0...
>   Cheers,
>    -Baylor
> 
> On Mar 25, 2005, at 1:46 PM, Alistair Adcroft wrote:
> 
>> Baylor,
>>
>> Can you clarify for me whether you meant that I broke the bit-wise 
>> backward compatibility with my tidy up?
>>
>> A.
>> -- 
>> Dr Alistair Adcroft              email: Alistair.Adcroft at noaa.gov
>> NOAA/GFDL, Princeton University               Tel: (609) 987-5073
>> Forrestal Campus, U.S. Route 1, P.O. Box 308, Princeton, NJ 08542
> 
> 


-- 
Dr Alistair Adcroft              email: Alistair.Adcroft at noaa.gov
NOAA/GFDL, Princeton University               Tel: (609) 987-5073
Forrestal Campus, U.S. Route 1, P.O. Box 308, Princeton, NJ 08542