I have a question about the whole mesh bootstrap procedure as I track down
an issue that Derek and I bumped into when tiling and stitching meshes a
few weeks back.
We are starting to grow utilities to programmatically build meshes for our
simulations on the fly out of several primitives like build_cube and
stitch_mesh. Is it valid or even wise to call "build_cube" and
"stitch_mesh" repeatedly when running in parallel? I'm concerned about
what might happen with the partitioner that gets called in each of the
functions as I'm building up my final mesh. I seems like we might end up
partitioning our mesh in some terrible way since the individual pieces will
get partitioned before each stitch call. In fact if we tile enough, we
might not even have any elements on some processors since individual pieces
might be too small on their own to partition.
First off, is my understanding correct? If it is, is there something
simple we can do to mitigate this problem like passing a Boolean to either,
or both of these utilities to avoid "preparing" the mesh prematurely?
Thanks,
Cody
------------------------------------------------------------------------------
How ServiceNow helps IT people transform IT departments:
1. A cloud service to automate IT design, transition and operations
2. Dashboards that offer high-level views of enterprise services
3. A single system of record for all IT processes
http://p.sf.net/sfu/servicenow-d2d-j
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
