[petsc-dev] PetscSection

[Jed Brown]

On Sat, Nov 10, 2012 at 5:03 PM, Matthew Knepley <knepley at gmail.com> wrote:

> Interpolation. I do not have a good, graph theoretic way to phrase it.


So if you know the topology of the n-cell, it implies a structure to its
cone-closure. (You could store that as a cache of single-cell-rooted
complex.) To interpolate the mesh, you graft in the single-cell, matching
the intermediate levels (fast query with a vertex index). I've written some
code for this, though not in full generality (I didn't explicitly store the
reference single-cells).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20121110/7d5a0c1e/attachment.html>