hi Bin, I'm not a MPI fan, but I think you can use MPI_Gather function to gather the results from all processors and send them in root.
pan --- Bin <[EMAIL PROTECTED]> wrote: > Hi libmesh developers: > > I use libmesh to code a biot problem, which is > running well in sequential. > Now, I want to convert this code to parallel, I have > the following > questions: > > In my code, I use some std::vector values to store > the derivatives on gauss > points inside every elements. For example, I use > std::vector<strain> > elem_strain to store the strain tensor on gauss > points. I understand that in > parallel, each processor own a part of this > std::vector. In the final stage, > I want to use only processor 0 to write these value > to a data file. How can > I do it? > > By the way, does libmesh support the latest version > of PETSC (v2.3.3 p8) > now? > > Thank you very much. > > > Bin > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio > 2005. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/> _______________________________________________ > Libmesh-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/libmesh-users > ____________________________________________________________________________________ Never miss a thing. Make Yahoo your home page. http://www.yahoo.com/r/hs ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
