[petsc-users] Write binary to matrix
Florian Lindner
mailinglists at xgm.de
Fri Sep 23 04:15:13 CDT 2016
Am 22.09.2016 um 18:34 schrieb Hong:
> Florian:
> Would it work if replacing MATSBAIJ to MATAIJ or MATMPISBAIJ?
MATAIJ works, but is not an option for my actual application.
MATMPISBAIJ does not work. Not very suprisingly, since afaik setting it to MATSBAIJ and executing it on multiple MPI
ranks actually results in MATMPISBAIJ.
Best,
Florian
>
> Hong
>
> Hey,
>
> this code reproduces the error when run with 2 or more ranks.
>
> #include <petscmat.h>
> #include <petscsys.h>
>
> int main(int argc, char *argv[])
> {
> PetscInitialize(&argc, &argv, "", NULL);
>
> Mat matrix;
> MatCreate(PETSC_COMM_WORLD, &matrix);
> MatSetType(matrix, MATSBAIJ);
> MatSetSizes(matrix, 10, 10, PETSC_DETERMINE, PETSC_DETERMINE);
> MatSetFromOptions(matrix);
> MatSetUp(matrix);
>
> MatAssemblyBegin(matrix, MAT_FINAL_ASSEMBLY);
> MatAssemblyEnd(matrix, MAT_FINAL_ASSEMBLY);
>
> PetscViewer viewer;
> PetscViewerBinaryOpen(PETSC_COMM_WORLD, "test.mat", FILE_MODE_WRITE, &viewer);
> MatView(matrix, viewer);
> PetscViewerDestroy(&viewer);
> MatDestroy(&matrix);
>
> PetscFinalize();
> }
>
>
> The complete output is:
>
>
> lindnefn at neon /data/scratch/lindnefn/aste (git)-[master] % mpic++ petsc.cpp -lpetsc && mpirun -n 2 ./a.out
>
> [0]PETSC ERROR: --------------------- Error Message --------------------------------------------------------------
> [0]PETSC ERROR: No support for this operation for this object type
> [0]PETSC ERROR: Cannot get subcomm viewer for binary files or sockets unless SubViewer contains the rank 0 process
> [0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html
> <http://www.mcs.anl.gov/petsc/documentation/faq.html> for trouble shooting.
> [0]PETSC ERROR: Petsc Release Version 3.7.3, unknown
> [0]PETSC ERROR: ./a.out on a arch-linux2-c-debug named neon by lindnefn Thu Sep 22 16:10:34 2016
> [0]PETSC ERROR: Configure options --with-debugging=1 --download-petsc4py=yes --download-mpi4py=yes
> --download-superlu_dist --download-parmetis --download-metis
> [0]PETSC ERROR: #1 PetscViewerGetSubViewer_Binary() line 46 in
> /data/scratch/lindnefn/software/petsc/src/sys/classes/viewer/impls/binary/binv.c
> [0]PETSC ERROR: #2 PetscViewerGetSubViewer() line 43 in
> /data/scratch/lindnefn/software/petsc/src/sys/classes/viewer/interface/dupl.c
> [0]PETSC ERROR: #3 MatView_MPISBAIJ_ASCIIorDraworSocket() line 900 in
> /data/scratch/lindnefn/software/petsc/src/mat/impls/sbaij/mpi/mpisbaij.c
> [0]PETSC ERROR: #4 MatView_MPISBAIJ() line 926 in
> /data/scratch/lindnefn/software/petsc/src/mat/impls/sbaij/mpi/mpisbaij.c
> [0]PETSC ERROR: #5 MatView() line 901 in /data/scratch/lindnefn/software/petsc/src/mat/interface/matrix.c
> WARNING! There are options you set that were not used!
> WARNING! could be spelling mistake, etc!
> Option left: name:-ksp_converged_reason (no value)
> Option left: name:-ksp_final_residual (no value)
> Option left: name:-ksp_view (no value)
> [neon:113111] *** Process received signal ***
> [neon:113111] Signal: Aborted (6)
> [neon:113111] Signal code: (-6)
> [neon:113111] [ 0] /lib/x86_64-linux-gnu/libc.so.6(+0x36cb0) [0x7feed8958cb0]
> [neon:113111] [ 1] /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x37) [0x7feed8958c37]
> [neon:113111] [ 2] /lib/x86_64-linux-gnu/libc.so.6(abort+0x148) [0x7feed895c028]
> [neon:113111] [ 3]
> /data/scratch/lindnefn/software/petsc/arch-linux2-c-debug/lib/libpetsc.so.3.7(PetscTraceBackErrorHandler+0x563)
> [0x7feed8d8db31]
> [neon:113111] [ 4] /data/scratch/lindnefn/software/petsc/arch-linux2-c-debug/lib/libpetsc.so.3.7(PetscError+0x374)
> [0x7feed8d88750]
> [neon:113111] [ 5] /data/scratch/lindnefn/software/petsc/arch-linux2-c-debug/lib/libpetsc.so.3.7(+0x19b2f6)
> [0x7feed8e822f6]
> [neon:113111] [ 6]
> /data/scratch/lindnefn/software/petsc/arch-linux2-c-debug/lib/libpetsc.so.3.7(PetscViewerGetSubViewer+0x4f1)
> [0x7feed8e803cb]
> [neon:113111] [ 7] /data/scratch/lindnefn/software/petsc/arch-linux2-c-debug/lib/libpetsc.so.3.7(+0x860c95)
> [0x7feed9547c95]
> [neon:113111] [ 8] /data/scratch/lindnefn/software/petsc/arch-linux2-c-debug/lib/libpetsc.so.3.7(+0x861494)
> [0x7feed9548494]
> [neon:113111] [ 9] /data/scratch/lindnefn/software/petsc/arch-linux2-c-debug/lib/libpetsc.so.3.7(MatView+0x12b6)
> [0x7feed971c08f]
> [neon:113111] [10] ./a.out() [0x400b8b]
> [neon:113111] [11] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5) [0x7feed8943f45]
> [neon:113111] [12] ./a.out() [0x4009e9]
> [neon:113111] *** End of error message ***
> --------------------------------------------------------------------------
> mpirun noticed that process rank 1 with PID 113111 on node neon exited on signal 6 (Aborted).
> --------------------------------------------------------------------------
>
> Thanks,
> Florian
>
>
>
> Am 22.09.2016 um 13:32 schrieb Matthew Knepley:
> > On Thu, Sep 22, 2016 at 5:42 AM, Florian Lindner <mailinglists at xgm.de <mailto:mailinglists at xgm.de>
> <mailto:mailinglists at xgm.de <mailto:mailinglists at xgm.de>>> wrote:
> >
> > Hello,
> >
> > I want to write a MATSBAIJ to a file in binary, so that I can load it later using MatLoad.
> >
> > However, I keep getting the error:
> >
> > [5]PETSC ERROR: No support for this operation for this object type!
> > [5]PETSC ERROR: Cannot get subcomm viewer for binary files or sockets unless SubViewer contains the rank 0 process
> > [6]PETSC ERROR: PetscViewerGetSubViewer_Binary() line 46 in
> > /data/scratch/lindnefn/software/petsc/src/sys/classes/viewer/impls/binary/binv.c
> >
> >
> > Do not truncate the stack.
> >
> > Run under valgrind.
> >
> > Thanks,
> >
> > Matt
> >
> >
> > The rank 0 is included, as you can see below, I use PETSC_COMM_WORLD and the matrix is also created like that.
> >
> > The code looks like:
> >
> > PetscErrorCode ierr = 0;
> > PetscViewer viewer;
> > PetscViewerBinaryOpen(PETSC_COMM_WORLD, filename.c_str(), FILE_MODE_WRITE, &viewer); CHKERRV(ierr);
> > MatView(matrix, viewer); CHKERRV(ierr);
> > PetscViewerDestroy(&viewer);
> >
> > Thanks,
> > Florian
> >
> >
> >
> >
> > --
> > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any
> > results to which their experiments lead.
> > -- Norbert Wiener
>
>
More information about the petsc-users
mailing list