[Mitgcm-support] Re: numbers

[mitgcm-support at dev.mitgcm.org]

A., 

Depends on your definition of perfect! I guess
I was expecting to be able to get the exact same numbers, as 
though running the exact same test on the exact smae machine.
The converged solver numbers are only the same to 7 digits,
do you understand why that it is?

Chris
Alistair Adcroft wrote:
> 
> I followed your instructions, and the numbers look perfect to me.
> (in /scratch3/adcroft/qq/...)
> 
> soyinka{input}% ./mitgcmuv | & grep CG2D
>  CG2D iters, err =      0          1.98852420248921E+00
>  CG2D iters, err =     18          4.55046795998284E-13
> soyinka{input}% grep CG2D ../results/output.txt | tail -2
>  CG2D iters, err =      0          1.98852420248921E+00
>  CG2D iters, err =     18          4.55046796001293E-13