[petsc-dev] Jed will not be happy

[Aron Ahmadia]

I'm pretty sure our needs are simple for image coarsening, but I'm going to
cc Manuel and David because I seem to recall that the levels of linear
interpolation available on PETSc grids are only suitable for image
rendering, not, for example, obtaining an accurate low-resolution coarsening
of a solution that you could do numerical analysis on, and I don't remember
if they had grander ambitions for DA interpolation before they discovered
petsc-3.1's current limitations.

A

On Sun, Aug 14, 2011 at 9:29 AM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> On Sun, Aug 14, 2011 at 01:21, Aron Ahmadia <aron.ahmadia at kaust.edu.sa>wrote:
>
>> For what it's worth, one of the PetClaw features (grid coarsening for 2D
>> image rendering), relies on interpolation working for 1 and 2-dimensional
>> uniform and periodic grids, so having this code ready would be useful to us
>> as well.
>
>
> Cool, we all agree that the simple periodic case has to work for the
> release.
>
> Is this application using a linear interpolation procedure or is it based
> on some nonlinear reconstruction?
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20110814/2fd4f564/attachment.html>