[Dune] [#778] MatrixHelper::detAAT is unstable

Dune flyspray at dune-project.org
Thu Apr 22 09:52:44 CEST 2010


THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY.

The following task has a new comment added:

FS#778 - MatrixHelper::detAAT is unstable
User who did this - Oliver Sander (sander)

----------
Hi Martin!
The element that belongs to the matrix is not an element that I do finite elements on.  All I need to do is integrate over it.  Hence the extreme ill-conditioning is not a problem, and I don't even have to invert the matrix.  If the computed determinant was zero or nearly zero that would be fine.  I don't care much about precision here, I just cannot have the code abort.

Regarding your questions:
a) you can have separate algorithms for quadratic and non-quadratic matrices. Actually I wonder why you don't have those anyways.  Quadratic matrices are a common case and a dedicated algorithm for them may be a lot faster
b) It would really be a pity if we couldn't use generic geometries in dune-grid-glue (which is what I am talking about). They are perfect for the job.

I would much like to avoid the use of higher-precision types, because those extremely degenerate intersections really make up only a very small fraction.  In the vast majority of cases everything works fine with double precision.
----------

More information can be found at the following URL:
http://www.dune-project.org/flyspray/index.php?do=details&task_id=778#comment1947

You are receiving this message because you have requested it from the Flyspray bugtracking system.  If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.




More information about the Dune mailing list