sieve-dev Using the Sieve library

[Matthew Knepley]

On Tue, Aug 3, 2010 at 5:29 AM, Stefan Kurzbach <stefan.kurzbach at tuhh.de>wrote:

> Dear Mr. Knepley,
>
> I would like to adapt our finite element code to use Sieve for distributed
> data management and as an interface to the PETSc solvers. Unfortunately the
> examples are currently broken, so I am having problems getting started. From
> the sieve-dev mailing list I could see that there is some development
> activity, however I am not sure if there is anything planned to be fixed or
> released that would be a starting point. Can you give me any additional
> information?
>

Sure. Most examples are currently broken. I should probably take them out.
The reason is that I spend all my
time supporting the specific needs of applications that use Sieve rather
than the basics, like the examples. Not
optimal, but its only me so far.

So what hope is there of using it?

  1) There is an extensive tutorial about its integration with PETSc solvers
(the first link in Previous):


http://www.mcs.anl.gov/petsc/petsc-as/documentation/tutorials/index.html

      which comes with a repository with a fully working example for FEM.
The slides tell you what revision
      of the repository to update to for each section.

  2) There is a fully working, visco-elastic-plastic crustal
deformation+faults simulation code that uses it for FEM:

      http://www.geodynamics.org/cig/software/pylith

      which is extensively documented and tested. If you want to see how to
do anything in unstructured grid FEM
      with Sieve, it is probably already done in PyLith.

  3) There is a paper on our approach. It gives an idea of our philosophy
concerning mesh representation:

      http://iospress.metapress.com/index/K5647WT076773465.pdf

  Hope this helps,

    Matt

Best regards
> Stefan Kurzbach
>
> Hamburg University of Technology
> Department of River and Coastal Engineering
>



-- 
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/sieve-dev/attachments/20100803/21f0cb87/attachment.htm>