On Fri, 16 Nov 2007, Roy Stogner wrote:
> I assure you, I haven't merely been breaking libMesh... except for the
> commit where I broke adaptive coarsening in find_neighbors()... and
> the bit where I broke MeshTools::bounding_box...
And the bit where I broke DofMap::enforce_constraints_exactly()
On Fri, 16 Nov 2007, Benjamin Kirk wrote:
> What are your thoughts on how to best to avoid node duplication when
> performing refinement in parallel?
At the moment we do it by comparing the physical positions of each
node that's newly created on a ghost element, and changing its
temporary index t
What are your thoughts on how to best to avoid node duplication when
performing refinement in parallel?
The reason I ask is because I wonder if the partition-agnostic global
indexing I just checked in could be used. This computes a global, unique
index for each node/element based on its Hilbert k
