Ben: I am assuming that "not necessarily" means that the SerialMesh can be somehow disabled?
A few other questions: * Can you describe briefly how libMesh handles refinement-dependent rebalancing of the mesh partitions? * Per the primary reference, it seems that the design of libMesh is such that it deals with assembling the system of equations, as well as handling the mesh operations. Is it possible to separate these two functionalities, i.e., to use the mesh manipulation functionality, but not perform the element operations? Thanks for the info. Neil -----Original Message----- From: <Kirk>, "Benjamin (JSC-EG311)" <[email protected]> Date: Monday, July 14, 2014 at 10:35 AM To: Neil Hodge <[email protected]> Cc: "[email protected]" <[email protected]> Subject: Re: [Libmesh-users] A quick question about current libMesh parallel mesh representation >In a word, no. > >In more words, not necessarily. > >We continue to carry around our original mesh data structure as a >"SerialMesh" object but have implemented a "ParallelMesh" flavor where >all data structures are parallelized, which has been used successfully to >thousands of cores. I don't want to quote a stale number and enrage the >INL guys, so I'll let them comment from here. > >-Ben > > > >On Jul 14, 2014, at 12:32 PM, Hodge, Neil E. <[email protected]> wrote: > >> I recently sat down and read >> >> Kirk, "libMesh: a C++ library for adaptive mesh refinement/coarsening >> simulations", Engineering with Computers, 2006. >> >> In that paper, I found the following: >> >> "On distributed memory machines, such as PC clusters, a complete copy of >> the mesh is maintained independently on each processor. This design >> decision limits practical 3D applications to on the order of 128 >> processors because of the overhead associated with storing the global >> mesh." >> >> Is it still the case that the complete mesh is represented in every >> process? Just curious, since the paper is from 2006 . . . Thanks. >> >> >> Neil >> >> ====================================================================== >> Neil Hodge, Ph.D. >> Methods Development Group >> Lawrence Livermore National Laboratory >> >> >> >> >> >> >>------------------------------------------------------------------------- >>----- >> Want fast and easy access to all the code in your enterprise? Index and >> search up to 200,000 lines of code with a free copy of Black Duck® >> Code Sight™ - the same software that powers the world's largest >>code >> search on Ohloh, the Black Duck Open Hub! Try it now. >> http://p.sf.net/sfu/bds >> _______________________________________________ >> Libmesh-users mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/libmesh-users > > ------------------------------------------------------------------------------ Want fast and easy access to all the code in your enterprise? Index and search up to 200,000 lines of code with a free copy of Black Duck® Code Sight™ - the same software that powers the world's largest code search on Ohloh, the Black Duck Open Hub! Try it now. http://p.sf.net/sfu/bds _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
