Hi, have you startet your program with
mpirun -n 4 ./program ? where after "-n" you set the number of cpus to be used (4 in this example). Thomas ------------------------------------------------- On Tue, 1 Dec 2009, Huan Sun wrote: > Hi Thomas, > > Thanks so much the help. Now I can compile and run both step-17 and 18. > > However deal.ii code finds only one mpi communicator so everything turns > out to be serial. I was running the program on a computer with multiple > cpus, not on a cluster. Is that the way it should be? > > Thanks, > > Huan > > > > On Mon, 2009-11-30 at 09:53 +0100, [email protected] > wrote: > > Hi, > > > > I think you should specify the variable LD_LIBRARY_PATH in your system. > > > > Something like (in your .bashrc or .profile) > > > > LD_LIBRARY_PATH=$LD_LIBRARY_PATH:path_to_libmpichcxx; export LD_LIBRARY_PATH > > > > or in your .tcshrc, if you have the tcsh shell > > > > setenv LD_PATH $LD_LIBRARY_PATH > > setenv LD_LIBRARY_PATH $LD_PATH:path_to_libmpichcxx > > > > where "path_to_libmpichcxx" is something like > > /ima/home/hsun/petsc/externalpackages/mpich2-1.0.8/lib > > > > > > Thomas > > > > ------------------------------------------------- > > > > On Mon, 30 Nov 2009, Huan Sun wrote: > > > > > Hi all, > > > > > > Thanks for the help. > > > > > > Somehow I got a new prblem -- I am new to the MPI area. > > > I enable the MPI download when configuring Petsc following the > > > instructions in FAQ on deal.ii webpage. Everything works fine. And I > > > configured deal.ii's compilers as Toby and Thomas suggested: > > > > > > ./configure --disable-threads --enable-shared > > > CC=/ima/home/hsun/petsc/externalpackages/mpich2-1.0.8/bin/mpicc > > > CXX=/ima/home/hsun/petsc/externalpackages/mpich2-1.0.8/bin/mpicxx > > > > > > deal.ii complained it cannot find -lmpichcxx. > > > > > > > > > ======================optimized========= Linking library: > > > libpetscall.so > > > ======================debug============= Linking library: > > > libpetscall.g.so > > > /usr/bin/ld: cannot find -lmpichcxx > > > collect2: ld returned 1 exit status > > > make[1]: *** [libpetscall.so] Error 1 > > > make[1]: *** Waiting for unfinished jobs.... > > > /usr/bin/ld: cannot find -lmpichcxx > > > collect2: ld returned 1 exit status > > > make[1]: *** [libpetscall.g.so] Error 1 > > > > > > I wonder if any more options I need to specify when configuring deal.ii. > > > > > > Thanks a lot! > > > > > > Huan > > > > > > > > > > > > > > > > > > > > > On Sun, 2009-11-29 at 14:55 +0100, Toby D. Young wrote: > > > > > [0]PETSC ERROR: Configure run at Fri Nov 27 23:13:43 2009 > > > > > [0]PETSC ERROR: Configure > > > > > options --with-cc=gcc --with-fc=gfortran --with-cxx=g++ > > > > > --download-f-blas-lapack=1 > > > > > --download-mpich=1 --with-shared=1 --with-dynamic=1 > > > > > --with-clanguage=C++ --with-x=0 > > > > > > > > In any case, to run step-17 on multi processors (which I think you are > > > > trying to do here), you will need to be compiling with MPI. This means > > > > configuring PETSc with: > > > > --with-mpi=1 --with-mpi-dir=path/to/mpi > > > > and configuringdealii with the same compilers by: > > > > CC=mpicc CXX=mpicxx > > > > > > > > As Thomas pointed out, there is some wierdness using the DataOut class > > > > with PETSc when threading is enabled. Try configuring > > > > dealii with: > > > > --disable-threads > > > > to be sure to get step-17 to run. Step-18 does not use DataOut in the > > > > same > > > > way and seems to somehow avoid these problems. > > > > > > > > There are other hints on the dealii FAQ which can help you get things > > > > working the way you really want them to. Maybe that helps? > > > > > > > > Best, > > > > Toby > > > > > > > > > > > > ----- > > > > > > > > Toby D. Young > > > > Assistant Professor > > > > Philosophy-Physics > > > > Polish Academy of Sciences > > > > Warszawa, Polska > > > > > > > > www: http://www.ippt.gov.pl/~tyoung > > > > skype: stenografia > > > > > > > > > _______________________________________________ > > > dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii > > > > > > _______________________________________________ > dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii > _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
