> On 28 Aug 2019, at 15:02, Matthew Knepley <knepley@gmail.com> wrote:
>
> On Wed, Aug 28, 2019 at 8:48 AM Sander Rhebergen <srhebergen@uwaterloo.ca> wrote:
> Thanks Matt, Patrick. That indeed took care of that error. There are some other errors now - the array of nodes that are fixed is empty so that I still have a zero pivot. Instead of fixing [0, 0] I fixed a point which I know lies in the mesh. This gives me a non-empty array of indices now (just the one point), but set.getDof(i)=0 for that point and so the nodes array is empty. I'll play around with the code a bit more, but for now I'll use Lawrence's suggestion.
>
> If sec.getDof(p) == 0 for your point p, there are no dofs associate with that point for the space V. Should this
> be true in your problem? If not, are you passing the correct V? You could always use sec.view() to see what
> is going on.
>
For periodic meshes, the coordinate field we have is DG, i.e. with no dofs associated with vertices.
But we are checking the solution space, not coordinate space. Right?
Thanks,
Matt
Cheers,
Lawrence
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