[petsc-users] I expected the Laplace run with a smaller delta to have a more accurate solution, yet they come out almost exactly the same.

[Matthew Knepley]

On Wed, Sep 24, 2014 at 5:03 PM, Alletto, John M <john.m.alletto at lmco.com>
wrote:

>  I have set up a test for running a Laplacian solver with 2 sets of data.
>
> One has twice as many points as the other.
>
> Both cover the same range, I input the X, Y and Z variables set using
> SetCoordinates.
>
>
>
> I compare the results with an analytical model.
>
>
>
> I expected the Laplace run with a smaller delta to have a more accurate
> solution, yet they come out almost exactly the same.
>
>
>
> Any ideas?
>
>
I think you may have the wrong idea of accuracy. If you are expecting to
converge in the L_2 norm, you must
do an integral to get the error, rather than just take the difference of
vertex values (that is the l2 norm). You could
be seeing superconvergence at the vertices, but I do not know what
discretization you are using.

  Matt

-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20140924/63f090f0/attachment-0001.html>