[Getdp] Magnetodynamic 3D - Electromagnetic circuit element

[Christophe Geuzaine]

Michael J. Rivard wrote:
> Hi Christophe,
> 
> Is the GetDP direct solver (Algorithm = 11) optimized for sparse matrices?
> 

Unfortunately, no: it's a dense LU.

It's in the code for testing purposes--for example to investigate
problems like Florin's.

Christophe

> Mike
> 
> On Monday 14 April 2003 09:59 am, Christophe Geuzaine wrote:
> 
>>Florin CIUPRINA wrote:
>>
>>>Is this correct? If yes, probably something else is wrong because
>>>unfortunately with this the problem does not converge:
>>>
>>>GetDP :  998  1.0825175e+06  9.4973969e-01
>>>GetDP :  999  1.0559604e+06  9.2643995e-01
>>>GetDP : Warning   : Iterative solver has iterated too many times
>>>GetDP : 1000  1.0548821e+06  9.2549394e-01
>>>GetDP : Solver    : 1000 Iterations / Residual: 1.21005e+06
>>>
>>>What can I do?
>>
>>Did you try to solve the linear system with a direct solver
>>(Algorithm=11 in solver.par)?
>>
>>Christophe
> 
> 
> 

-- 
Christophe Geuzaine

Tel: (626) 395-4552    http://www.geuz.org
Fax: (626) 578-0124    mailto:geuzaine at acm.caltech.edu