With 4.32 million dofs (432000 elements) in system_of_equations_ex6.C, I get memory usage of 12-13 GB. Is that what is expected? My code seems to use more than this , and the only major difference is that I have more systems in my code. Thanks, Subramanya
> Date: Thu, 22 Aug 2013 18:33:14 -0400 > From: [email protected] > To: [email protected] > Subject: Re: [Libmesh-users] memory usage by build_sparsity() > > Why don't you report memory usage based on systems_of_equations_ex4 > (uniformly refine the mesh a few times to make it "big enough"). Then > its easy for others to check if they see the same thing as you. > > > > > On 08/22/2013 06:30 PM, subramanya sadasiva wrote: > > Just out of curiosity, is the memory usage I see typical? or is there > > something wrong? Because the number of elements is pretty small - even > > though they are all densely connected. > > Thanks, Subramanya > > > >> From: [email protected] > >> To: [email protected] > >> Date: Thu, 22 Aug 2013 18:27:44 -0400 > >> CC: [email protected] > >> Subject: Re: [Libmesh-users] memory usage by build_sparsity() > >> > >> > >> It is 50*50*100, with 3 DOF per node.. From: [email protected] > >> Date: Thu, 22 Aug 2013 16:24:10 -0600 > >> Subject: Re: [Libmesh-users] memory usage by build_sparsity() > >> To: [email protected] > >> CC: [email protected] > >> > >> > >> > >> > >> On Thu, Aug 22, 2013 at 4:15 PM, subramanya sadasiva <[email protected]> > >> wrote: > >> > >> > >> > >> > >> > >> Hi John, I have around 250000 HEX 20 elements with 3 dof per node (3D > >> elasticity). I am running this in parallel with around 4 mpi processes. > >> In that case you might get some benefit from using ParallelMesh if you > >> aren't already tried that. > >> > >> > >> You have 250k elements so that's what, about 64^3? something like 128^3 > >> total DOFs? > >> -- > >> John > >> > >> ------------------------------------------------------------------------------ > >> Introducing Performance Central, a new site from SourceForge and > >> AppDynamics. Performance Central is your source for news, insights, > >> analysis and resources for efficient Application Performance Management. > >> Visit us today! > >> http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk > >> _______________________________________________ > >> Libmesh-users mailing list > >> [email protected] > >> https://lists.sourceforge.net/lists/listinfo/libmesh-users > > > > ------------------------------------------------------------------------------ > > Introducing Performance Central, a new site from SourceForge and > > AppDynamics. Performance Central is your source for news, insights, > > analysis and resources for efficient Application Performance Management. > > Visit us today! > > http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk > > _______________________________________________ > > Libmesh-users mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/libmesh-users > > > ------------------------------------------------------------------------------ > Introducing Performance Central, a new site from SourceForge and > AppDynamics. Performance Central is your source for news, insights, > analysis and resources for efficient Application Performance Management. > Visit us today! > http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk > _______________________________________________ > Libmesh-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/libmesh-users ------------------------------------------------------------------------------ Introducing Performance Central, a new site from SourceForge and AppDynamics. Performance Central is your source for news, insights, analysis and resources for efficient Application Performance Management. Visit us today! http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
