On Thu, 15 Sep 2011, Vasilis Vavourakis wrote:
was just wondering if it would be easy to add in the libMesh library
the feature of doing the domain decomposition for nodes, in addition
to elements as it does right now.
Hmm... we currently *do* do domain decomposition for nodes. There's
two catches to it:
Each node currently gets put in the lowest-numbered domain available
among elements containing that node. (I don't recall for certain
whether "containing" here is in the geometric or graph sense in the
hanging nodes case, but I think graph.) There's some flexibility with
this, but:
Each node needs to get put in *some* domain that is shared among
elements containing that node. If that's not valid then a lot of
library code probably breaks.
the reason of my inquiry is that this new feature would be handy in case one
implements libMesh for Meshless methods...or any
other techniques (don't have something else in mind) for which partitioning is
done for the mesh nodes instead for mesh
elements.
That's a lot trickier. It's impossible to come up with an elements
partitioning for which no valid nodes partitioning exists, and once
you've got the elements partitioning it's easy to create a valid nodes
partitioning. However, it is possible to come up with a nodes
partitioning for which no valid elements partitioning exists, and even
when a valid elements partitioning does exist it's not always trivial
to construct.
i guess that this will need to provide a few additional functions to the
MeshBase class, like:
Node::active() currently doesn't mean much, just "was given a node
id". In an AMR mesh, I believe *every* node which is attached to an
element is also attached to an active element. So with that in mind:
n_active_local_nodes()
Probably just tells you n_local_nodes()
active_local_node_begin()
active_local_node_end()
would match local_nodes_begin/end()
active_not_local_elements_begin()
active_not_local_elements_end()
already exist.
---
Roy
------------------------------------------------------------------------------
Doing More with Less: The Next Generation Virtual Desktop
What are the key obstacles that have prevented many mid-market businesses
from deploying virtual desktops? How do next-generation virtual desktops
provide companies an easier-to-deploy, easier-to-manage and more affordable
virtual desktop model.http://www.accelacomm.com/jaw/sfnl/114/51426474/
_______________________________________________
Libmesh-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
