Ok. I'll find you on Skype if necessary. My existing routine does separate inter-processor syncs from prolongation operations; they run at different steps in the algorithm. However, it may not be a problem to do them all at once. I'll try it and see...
-- Scott H. Hawley, Ph.D. Assistant Professor of Physics Belmont University. 615-460-6206. Sent from my iPad
On Apr 9, 2012, at 9:39 AM, "Erik Schnetter" schnetter@cct.lsu.edu wrote:
The routine that Roland describes determines what kind of boundary a boundary is -- inter-processor, symmetry boundary, outer boundary, etc., but unfortunately does not distinguish between mesh refinement boundaries and inter-process ghost zones. The reason is that a given boundary may be both. Consider an L-shaped refined region consisting of two rectangular component. Where they touch, ghost zones are filled by copying from the other component, elsewhere, they are filled via prolongation from the next coarser grid.
Therefore, the distinction between inter-process and mesh refinement boundaries is hidden from the application, and synchronization applies both kinds of boundary conditions.
To determine the details, you can query Carpet about which region of a component are filled by synchronization and which are filled by prolongation; however, there is no one-to-one connection to what Cactus calls a "face" of a grid.
Let's discuss on Skype, which should be easier than email.
-erik
On Mon, Apr 9, 2012 at 10:12 AM, Roland Haas roland.haas@physics.gatech.edu wrote:
Hello Scott,
For my elliptic solver, I need to know not only if a given grid boundary is an outer boundary or not, I need to know if it's an inter-processor boundary or a mesh refinement boundary. cctk_bbox does not provide the latter distinction. Is there something analogous provided by Carpet?
GenericFD's GenericFD_GetBoundaryInfo routine (in Kranc, see https://github.com/ianhinder/Kranc/blob/6bed8e74aac5bd1e76f2e7a4d8efcf166a60...) will do this for you. I believe there was discussion at one point to move this functionality into the flesh but cannot find the thread right now.
Yours, Roland
-- My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from http://keys.gnupg.net.
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
-- Erik Schnetter schnetter@cct.lsu.edu http://www.perimeterinstitute.ca/personal/eschnetter/ _______________________________________________ Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users