[petsc-dev] About the refine/coarsen in DM

[Matthew Knepley]

On Wed, Jan 15, 2025 at 11:51 AM Ye Changqing <Ye_Changqing at outlook.com>
wrote:

> Dear PETSc developers,
>
> I am curious about the mystery behind ``refine'' and ``coarsen'' routines
> of DM.
>
> - As mentioned on the webpage, these routines appear to neglect the
> coordinate information associated with the current DM. I suspect that some
> graph-based algebraic operations might be involved. Could you kindly share
> the underlying design principles or implementation strategies behind these
> routines?
>

Some of the confusion is because DM can represent lots of kinds of meshes,
but the refine/coarsen algorithms are mesh-type specific. For example,

1) DMForest can refine and coarsen using quadtree algorithms

2) DMPlex meshes can be regularly refined, and adaptively refined using the
SBR algorithm

3) DMPlex meshes can be refined and coarsened using metric-based adaptation
through MMG.

  Thanks,

     Matt


> Best,
> Changqing



-- 
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

https://urldefense.us/v3/__https://www.cse.buffalo.edu/*knepley/__;fg!!G_uCfscf7eWS!bGEN7gIKzUtDZ3Tll_Kl1xvRrW0jG9BWKCDtHx4xjvxwQXHnMRqN6VyzStaEMQuZeecebRnNMLftf4D2f6FT$  <https://urldefense.us/v3/__http://www.cse.buffalo.edu/*knepley/__;fg!!G_uCfscf7eWS!bGEN7gIKzUtDZ3Tll_Kl1xvRrW0jG9BWKCDtHx4xjvxwQXHnMRqN6VyzStaEMQuZeecebRnNMLftf7IK_ng8$ >
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20250115/6cb40830/attachment.html>