<html>
  <head>
    <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <br>
    <div class="moz-cite-prefix">Am 22.04.2014 13:57, schrieb Matthew
      Knepley:<br>
    </div>
    <blockquote
cite="mid:CAMYG4GkXGE5Vx6qTncw1z0xw+yvzxUCJS5Y6co_3K3UUZotRcw@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div class="gmail_extra">
          <div class="gmail_quote">On Tue, Apr 22, 2014 at 6:48 AM,
            Niklas Fischer <span dir="ltr"><<a
                moz-do-not-send="true" href="mailto:niklas@niklasfi.de"
                target="_blank">niklas@niklasfi.de</a>></span> wrote:<br>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">Am
              22.04.2014 13:08, schrieb Jed Brown:<br>
              <blockquote class="gmail_quote" style="margin:0 0 0
                .8ex;border-left:1px #ccc solid;padding-left:1ex">
                Niklas Fischer <<a moz-do-not-send="true"
                  href="mailto:niklas@niklasfi.de" target="_blank">niklas@niklasfi.de</a>>

                writes:<br>
                <br>
                <blockquote class="gmail_quote" style="margin:0 0 0
                  .8ex;border-left:1px #ccc solid;padding-left:1ex">
                  Hello,<br>
                  <br>
                  I have attached a small test case for a problem I am
                  experiencing. What<br>
                  this dummy program does is it reads a vector and a
                  matrix from a text<br>
                  file and then solves Ax=b. The same data is available
                  in two forms:<br>
                    - everything is in one file (matops.s.0 and
                  vops.s.0)<br>
                    - the matrix and vector are split between processes
                  (matops.0,<br>
                  matops.1, vops.0, vops.1)<br>
                  <br>
                  The serial version of the program works perfectly fine
                  but unfortunately<br>
                  errors occure, when running the parallel version:<br>
                  <br>
                  make && mpirun -n 2 a.out matops vops<br>
                  <br>
                  mpic++ -DPETSC_CLANGUAGE_CXX -isystem<br>
                  /home/data/fischer/libs/petsc-3.4.3/arch-linux2-c-debug/include

                  -isystem<br>
                  /home/data/fischer/libs/petsc-3.4.3/include
                  petsctest.cpp -Werror -Wall<br>
                  -Wpedantic -std=c++11 -L<br>
                  /home/data/fischer/libs/petsc-3.4.3/arch-linux2-c-debug/lib

                  -lpetsc<br>
                  /usr/bin/ld: warning: libmpi_cxx.so.0, needed by<br>
/home/data/fischer/libs/petsc-3.4.3/arch-linux2-c-debug/lib/libpetsc.so,<br>
                  may conflict with libmpi_cxx.so.1<br>
                  /usr/bin/ld: warning: libmpi.so.0, needed by<br>
/home/data/fischer/libs/petsc-3.4.3/arch-linux2-c-debug/lib/libpetsc.so,<br>
                  may conflict with libmpi.so.1<br>
                  librdmacm: couldn't read ABI version.<br>
                  librdmacm: assuming: 4<br>
                  CMA: unable to get RDMA device list<br>
--------------------------------------------------------------------------<br>
                  [[43019,1],0]: A high-performance Open MPI
                  point-to-point messaging module<br>
                  was unable to find any relevant network interfaces:<br>
                  <br>
                  Module: OpenFabrics (openib)<br>
                     Host: <a moz-do-not-send="true"
                    href="http://dornroeschen.igpm.rwth-aachen.de"
                    target="_blank">dornroeschen.igpm.rwth-aachen.de</a><br>
                  CMA: unable to get RDMA device list<br>
                </blockquote>
                It looks like your MPI is either broken or some of the
                code linked into<br>
                your application was compiled with a different MPI or
                different version.<br>
                Make sure you can compile and run simple MPI programs in
                parallel.<br>
              </blockquote>
              Hello Jed,<br>
              <br>
              thank you for your inputs. Unfortunately MPI does not seem
              to be the issue here. The attachment contains a simple MPI
              hello world program which runs flawlessly (I will append
              the output to this mail) and I have not encountered any
              problems with other MPI programs. My question still
              stands.<br>
            </blockquote>
            <div><br>
            </div>
            <div>This is a simple error. You created the matrix A using
              PETSC_COMM_WORLD, but you try to view it</div>
            <div>using PETSC_VIEWER_STDOUT_SELF. You need to use
              PETSC_VIEWER_STDOUT_WORLD in</div>
            <div>order to match.</div>
            <div><br>
            </div>
            <div>  Thanks,</div>
            <div><br>
            </div>
            <div>     Matt</div>
            <div> </div>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">
              Greetings,<br>
              Niklas Fischer<br>
              <br>
              mpirun -np 2 ./mpitest<br>
              <br>
              librdmacm: couldn't read ABI version.<br>
              librdmacm: assuming: 4<br>
              CMA: unable to get RDMA device list<br>
--------------------------------------------------------------------------<br>
              [[44086,1],0]: A high-performance Open MPI point-to-point
              messaging module<br>
              was unable to find any relevant network interfaces:<br>
              <br>
              Module: OpenFabrics (openib)<br>
                Host: <a moz-do-not-send="true"
                href="http://dornroeschen.igpm.rwth-aachen.de"
                target="_blank">dornroeschen.igpm.rwth-aachen.de</a><br>
              <br>
              Another transport will be used instead, although this may
              result in<br>
              lower performance.<br>
--------------------------------------------------------------------------<br>
              librdmacm: couldn't read ABI version.<br>
              librdmacm: assuming: 4<br>
              CMA: unable to get RDMA device list<br>
              Hello world from processor <a moz-do-not-send="true"
                href="http://dornroeschen.igpm.rwth-aachen.de"
                target="_blank">dornroeschen.igpm.rwth-aachen.de</a>,
              rank 0 out of 2 processors<br>
              Hello world from processor <a moz-do-not-send="true"
                href="http://dornroeschen.igpm.rwth-aachen.de"
                target="_blank">dornroeschen.igpm.rwth-aachen.de</a>,
              rank 1 out of 2 processors<br>
              [<a moz-do-not-send="true"
                href="http://dornroeschen.igpm.rwth-aachen.de:128141"
                target="_blank">dornroeschen.igpm.rwth-aachen.de:128141</a>]
              1 more process has sent help message help-mpi-btl-base.txt
              / btl:no-nics<br>
              [<a moz-do-not-send="true"
                href="http://dornroeschen.igpm.rwth-aachen.de:128141"
                target="_blank">dornroeschen.igpm.rwth-aachen.de:128141</a>]
              Set MCA parameter "orte_base_help_aggregate" to 0 to see
              all help / error messages<br>
            </blockquote>
          </div>
          <br>
        </div>
      </div>
    </blockquote>
    Thank you, Matthew, this solves my viewing problem. Am I doing
    something wrong when initializing the matrices as well? The matrix'
    viewing output starts with "Matrix Object: 1 MPI processes" and the
    Krylov solver does not converge.<br>
    <br>
    Your help is really appreciated,<br>
    Niklas Fischer<br>
    <blockquote
cite="mid:CAMYG4GkXGE5Vx6qTncw1z0xw+yvzxUCJS5Y6co_3K3UUZotRcw@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div class="gmail_extra"><br clear="all">
          <div><br>
          </div>
          -- <br>
          What most experimenters take for granted before they begin
          their experiments is infinitely more interesting than any
          results to which their experiments lead.<br>
          -- Norbert Wiener </div>
      </div>
    </blockquote>
    <br>
  </body>
</html>