Hi all, I have a follow-up question about step 31. Is it possible to convert step 31 into a parallel code? I checked TrilinosWrappers module and there is only mpi::vector. No mpi::matrix/blockmatrix is found. So I am a little doubtful.
Thanks! Huan On Fri, 2010-04-02 at 10:16 -0400, Huan Sun wrote: > Oh, that's my fault. I assumed usage of Petsc or Trilinos just means > parallel computations by default. > > Thanks, > > Huan > > > On Fri, 2010-04-02 at 04:49 -0500, Wolfgang Bangerth wrote: > > Huan, > > > > > I have a question on the example of step-31. When I ran the program, I > > > found the processors worked pretty independently, or asynchronously -- > > > from what I read in the output. The time step numbers can vary quite a > > > lot on different processors. (I am assuming this is normal. please > > > correct me if not.) However, each time step is determined by the CFL > > > condition, which needs the global info of the velocity to get the > > > maximal velocity. This would require the faster processors to wait for > > > the slower ones to finish so that they can determine the next time step, > > > which contradicts my observation. > > > I would greatly appreciate your explanations or thoughts. > > > > I am not entirely sure what you refer to: step-31 is a sequential program > > and does not use multiple processors. Or what do I miss? > > > > Best > > W. > > > > ------------------------------------------------------------------------- > > Wolfgang Bangerth email: [email protected] > > www: http://www.math.tamu.edu/~bangerth/ > > > _______________________________________________ > dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
