Hi,

the iteration numbers look like hypre is not used. I get similar high
iteration numbers with a preconditioner like "./step-40 -pc_type
bjacobi".

Are you running in debug mode? Can you run with "./step -log_summary"
and look if the output says something about HYPRE (like in the blob
after "Option Table entries")?

On Wed, Feb 16, 2011 at 10:54 AM, Svyatoslav Gladkov <[email protected]> wrote:
> Dear dealers,
>
>  I just wanted to try the new technology for distributed meshes under
>  deal.II  7.0.0 thus I've complied all the stuff with p4est and PETSC
>  with  Hypre.  But  the  results  I  get  from  computation are quite
>  different  from the ones shown in tutorial (number of iterations for
>  linear  solver is much much larger than reference values). Output is
>  below. Any clue where might be a problem (I run it with 1 or 4 cores
>  but the results are more or less the same)?
>
> Regards,
> Slawa
>
> Cycle 0:
>   Number of active cells:       1024
>   Number of degrees of freedom: 4225
>   Solved in 99 iterations.
>
> Cycle 1:
>   Number of active cells:       1966
>   Number of degrees of freedom: 8453
>   Solved in 154 iterations.
>
> Cycle 2:
>   Number of active cells:       3676
>   Number of degrees of freedom: 16211
>   Solved in 216 iterations.
>
> Cycle 3:
>   Number of active cells:       7114
>   Number of degrees of freedom: 31743
>   Solved in 321 iterations.
>
> Cycle 4:
>   Number of active cells:       13846
>   Number of degrees of freedom: 62405
>   Solved in 467 iterations.
>
> Cycle 5:
>   Number of active cells:       27646
>   Number of degrees of freedom: 125401
>   Solved in 643 iterations.
>
> Cycle 6:
>   Number of active cells:       55726
>   Number of degrees of freedom: 253931
>   Solved in 956 iterations.
>
> Cycle 7:
>   Number of active cells:       115618
>   Number of degrees of freedom: 527203
>   Solved in 1319 iterations.
>
> _______________________________________________
> dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
>



-- 
Timo Heister
http://num.math.uni-goettingen.de/~heister
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to