![]() |
|
This file is part of OpenFOAM. OpenFOAM is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. OpenFOAM is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with OpenFOAM; if not, write to the Free Software Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
Calculates points shared by more than two processor patches or cyclic patches. Is used in parallelInfo. (this info is needed for point-edge communication where you do all but these shared points with patch to patch communication but need to do a reduce on these shared points) Works purely topological and using local communication only. Needs: - domain to be one single domain (i.e. all faces can be reached through face-cell walk). - patch face ordering to be ok - f[0] ordering on patch faces to be ok. Works by constructing equivalence lists for all the points on processor patches. These list are procPointList and give processor and meshPoint label on that processor. E.g. ((7 93)(4 731)(3 114)) means point 93 on proc7 is connected to point 731 on proc4 and 114 on proc3. It then gets the lowest numbered processor (the 'master') to request a sharedPoint label from processor0 and it redistributes this label back to the other processors in the equivalence list. Algorithm: - get meshPoints of all my points on processor patches and initialize equivalence lists to this. loop - send to all neighbours in relative form: - patchFace - index in face - receive and convert into meshPoints. Add to to my equivalence lists. - mark meshPoints for which information changed. - send data for these meshPoints again endloop until nothing changes At this point one will have complete point-point connectivity for all points on processor patches. Now - remove point equivalences of size 2. These are just normal points shared between two neighbouring procPatches. - collect on each processor points for which it is the master - request number of sharedPointLabels from the Pstream::master. This information gets redistributed to all processors in a similar way as that in which the equivalence lists were collected: - initialize the indices of shared points I am the master for loop - send my known sharedPoints + meshPoints to all neighbours - receive from all neighbour. Find which meshPoint on my processor the sharedpoint is connected to - mark indices for which information has changed endloop until nothing changes.
Definition in file globalPoints.H.
Go to the source code of this file.
Namespaces | |
namespace | Foam |
Classes | |
class | globalPoints |