Guys,

The performance of ParmetisPartitioner::assign_partitioning degrades 
substantially as mesh size increases.  If I were to guess, I would say the 
count() method and n_active_local_elem() method in the asserts are the culprits 
since they appear to enumerate a significant portion of the mesh at each 
iteration.  While a 1600 element mesh reads lickety-split, a 90,000 element 
mesh takes longer than 1/2 hour to get through this method.  I say longer, 
because I never saw the code exit this method.

Thanks,
Bill
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to