<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>For us it clearly creates problems in real computations...<br>
    </p>
    <p>I understand the need to have clean test for PETSc, but for me,
      it reveals that hypre isn't usable with more than one thread for
      now...</p>
    <p>Another solution:  force single-threaded configuration for hypre
      until this is fixed?</p>
    <p>Eric<br>
    </p>
    On 2021-03-13 8:50 a.m., Pierre Jolivet wrote:<br>
    <blockquote type="cite"
      cite="mid:7932C4DD-11BB-4BBB-A171-08585D7ADADE@joliv.et">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div class="">-pc_hypre_boomeramg_relax_type_all Jacobi => </div>
      <div class="">  Linear solve did not converge due to
        DIVERGED_INDEFINITE_PC iterations 3</div>
      <div class="">
        <div class="">-pc_hypre_boomeramg_relax_type_all l1scaled-Jacobi
          => </div>
      </div>
      <div class="">OK, independently of the architecture it seems (Eric
        Docker image with 1 or 2 threads or my macOS), but contraction
        factor is higher</div>
      <div class="">  Linear solve converged due to CONVERGED_RTOL
        iterations 8</div>
      <div class="">  Linear solve converged due to CONVERGED_RTOL
        iterations 24</div>
      <div class="">  Linear solve converged due to CONVERGED_RTOL
        iterations 26</div>
      <div class="">v. currently</div>
      <div class="">  Linear solve converged due to CONVERGED_RTOL
        iterations 7</div>
      <div class="">  Linear solve converged due to CONVERGED_RTOL
        iterations 9</div>
      <div class="">  Linear solve converged due to CONVERGED_RTOL
        iterations 10</div>
      <div class=""><br class="">
      </div>
      <div class="">Do we change this? Or should we force
        OMP_NUM_THREADS=1 for make test?</div>
      <div class=""><br class="">
      </div>
      <div class="">Thanks,</div>
      <div class="">Pierre</div>
      <div class=""><br class="">
      </div>
      <div>
        <blockquote type="cite" class="">
          <div class="">On 13 Mar 2021, at 2:26 PM, Mark Adams <<a
              href="mailto:mfadams@lbl.gov" class=""
              moz-do-not-send="true">mfadams@lbl.gov</a>> wrote:</div>
          <br class="Apple-interchange-newline">
          <div class="">
            <div dir="ltr" class="">Hypre uses a multiplicative smoother
              by default. It has a chebyshev smoother. That with a
              Jacobi PC should be thread invariant.
              <div class="">
                <div class="">Mark</div>
              </div>
            </div>
            <br class="">
            <div class="gmail_quote">
              <div dir="ltr" class="gmail_attr">On Sat, Mar 13, 2021 at
                8:18 AM Pierre Jolivet <<a
                  href="mailto:pierre@joliv.et" class=""
                  moz-do-not-send="true">pierre@joliv.et</a>> wrote:<br
                  class="">
              </div>
              <blockquote class="gmail_quote" style="margin:0px 0px 0px
                0.8ex;border-left:1px solid
                rgb(204,204,204);padding-left:1ex">
                <div style="overflow-wrap: break-word;" class=""><br
                    class="">
                  <div class="">
                    <blockquote type="cite" class="">
                      <div class="">On 13 Mar 2021, at 9:17 AM, Pierre
                        Jolivet <<a href="mailto:pierre@joliv.et"
                          target="_blank" class=""
                          moz-do-not-send="true">pierre@joliv.et</a>>
                        wrote:</div>
                      <br class="">
                      <div class="">
                        <div style="overflow-wrap: break-word;" class="">Hello
                          Eric,
                          <div class="">I’ve made an “interesting”
                            discovery, so I’ll put back the list in c/c.</div>
                          <div class="">It appears the following snippet
                            of code which uses Allreduce() + lambda
                            function + MPI_IN_PLACE is:</div>
                          <div class="">- Valgrind-clean with MPICH;</div>
                          <div class="">- Valgrind-clean with OpenMPI
                            4.0.5;</div>
                          <div class="">- not Valgrind-clean with
                            OpenMPI 4.1.0.</div>
                          <div class="">I’m not sure who is to blame
                            here, I’ll need to look at the MPI
                            specification for what is required by the
                            implementors and users in that case.</div>
                          <div class=""><br class="">
                          </div>
                          <div class="">In the meantime, I’ll do the
                            following:</div>
                          <div class="">-
                            update config/BuildSystem/config/packages/OpenMPI.py
                            to use OpenMPI 4.1.0, see if any other error
                            appears;</div>
                          <div class="">- provide a hotfix to bypass the
                            segfaults;</div>
                        </div>
                      </div>
                    </blockquote>
                    <div class=""><br class="">
                    </div>
                    <div class="">I can confirm that splitting the
                      single Allreduce with my own MPI_Op into two
                      Allreduce with MAX and BAND fixes the segfaults
                      with OpenMPI (*).</div>
                    <br class="">
                    <blockquote type="cite" class="">
                      <div class="">
                        <div style="overflow-wrap: break-word;" class="">
                          <div class="">- look at the hypre issue and
                            whether they should be deferred to the hypre
                            team.</div>
                        </div>
                      </div>
                    </blockquote>
                    <div class=""><br class="">
                    </div>
                    <div class="">I don’t know if there is something
                      wrong in hypre threading or if it’s just a side
                      effect of threading, but it seems that the number
                      of threads has a drastic effect on the quality of
                      the PC.</div>
                    <div class="">By default, it looks that there are
                      two threads per process with your Docker image.</div>
                    <div class="">If I force OMP_NUM_THREADS=1, then I
                      get the same convergence as in the output file.</div>
                    <div class=""><br class="">
                    </div>
                    <div class="">Thanks,</div>
                    <div class="">Pierre</div>
                    <div class=""><br class="">
                    </div>
                    <div class="">(*) <a
                        href="https://gitlab.com/petsc/petsc/-/merge_requests/3712"
                        target="_blank" class="" moz-do-not-send="true">https://gitlab.com/petsc/petsc/-/merge_requests/3712</a></div>
                    <br class="">
                    <blockquote type="cite" class="">
                      <div class="">
                        <div style="overflow-wrap: break-word;" class="">
                          <div class="">Thank you for the Docker files,
                            they were really useful.</div>
                          <div class="">If you want to avoid
                            oversubscription failures, you can edit the
                            file
                            /opt/openmpi-4.1.0/etc/openmpi-default-hostfile
                            and append the line:</div>
                          <div class="">localhost slots=12</div>
                          <div class="">If you want to increase the
                            timeout limit of PETSc test suite for each
                            test, you can add the extra flag in your
                            command line TIMEOUT=180 (default is 60,
                            units are seconds).</div>
                          <div class=""><br class="">
                          </div>
                          <div class="">Thanks, I’ll ping you on GitLab
                            when I’ve got something ready for you to
                            try,</div>
                          <div class="">Pierre<br class="">
                            <div class=""><br class="">
                            </div>
                          </div>
                        </div>
                        <span
                          id="gmail-m_3567963440499379521cid:15B6BE6E-0C96-4CBA-9ADC-EFB1DE1BDFC3"
                          class=""><ompi.cxx></span>
                        <div style="overflow-wrap: break-word;" class="">
                          <div class="">
                            <div class=""><br class="">
                              <blockquote type="cite" class="">
                                <div class="">On 12 Mar 2021, at 8:54
                                  PM, Eric Chamberland <<a
                                    href="mailto:Eric.Chamberland@giref.ulaval.ca"
                                    target="_blank" class=""
                                    moz-do-not-send="true">Eric.Chamberland@giref.ulaval.ca</a>>
                                  wrote:</div>
                                <br class="">
                                <div class="">
                                  <div class="">
                                    <p class="">Hi Pierre,</p>
                                    <p class="">I now have a docker
                                      container reproducing the problems
                                      here.</p>
                                    <p class="">Actually, if I look at
                                      snes_tutorials-ex12_quad_singular_hpddm 
                                      it fails like this:</p>
                                    <p class="">not ok
                                      snes_tutorials-ex12_quad_singular_hpddm
                                      # Error code: 59<br class="">
                                      #       Initial guess<br class="">
                                      #       L_2 Error: 0.00803099<br
                                        class="">
                                      #       Initial Residual<br
                                        class="">
                                      #       L_2 Residual: 1.09057<br
                                        class="">
                                      #       Au - b = Au + F(0)<br
                                        class="">
                                      #       Linear L_2 Residual:
                                      1.09057<br class="">
                                      #       [d470c54ce086:14127] Read
                                      -1, expected 4096, errno = 1<br
                                        class="">
                                      #       [d470c54ce086:14128] Read
                                      -1, expected 4096, errno = 1<br
                                        class="">
                                      #       [d470c54ce086:14129] Read
                                      -1, expected 4096, errno = 1<br
                                        class="">
                                      #       [3]PETSC ERROR:
                                      ------------------------------------------------------------------------<br
                                        class="">
                                      #       [3]PETSC ERROR: Caught
                                      signal number 11 SEGV:
                                      Segmentation Violation, probably
                                      memory access out of range<br
                                        class="">
                                      #       [3]PETSC ERROR: Try option
                                      -start_in_debugger or
                                      -on_error_attach_debugger<br
                                        class="">
                                      #       [3]PETSC ERROR: or see <a
href="https://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind"
                                        target="_blank" class=""
                                        moz-do-not-send="true">https://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind</a><br
                                        class="">
                                      #       [3]PETSC ERROR: or try <a
                                        href="http://valgrind.org/"
                                        target="_blank" class=""
                                        moz-do-not-send="true">http://valgrind.org</a>
                                      on GNU/linux and Apple Mac OS X to
                                      find memory corruption errors<br
                                        class="">
                                      #       [3]PETSC ERROR: likely
                                      location of problem given in stack
                                      below<br class="">
                                      #       [3]PETSC ERROR:
                                      ---------------------  Stack
                                      Frames
                                      ------------------------------------<br
                                        class="">
                                      #       [3]PETSC ERROR: Note: The
                                      EXACT line numbers in the stack
                                      are not available,<br class="">
                                      #       [3]PETSC ERROR:      
                                      INSTEAD the line number of the
                                      start of the function<br class="">
                                      #       [3]PETSC ERROR:       is
                                      given.<br class="">
                                      #       [3]PETSC ERROR: [3]
                                      buildTwo line 987
                                      /opt/petsc-main/include/HPDDM_schwarz.hpp<br
                                        class="">
                                      #       [3]PETSC ERROR: [3] next
                                      line 1130
                                      /opt/petsc-main/include/HPDDM_schwarz.hpp<br
                                        class="">
                                      #       [3]PETSC ERROR:
                                      --------------------- Error
                                      Message
                                      --------------------------------------------------------------<br
                                        class="">
                                      #       [3]PETSC ERROR: Signal
                                      received<br class="">
                                      #       [3]PETSC ERROR: [0]PETSC
                                      ERROR:
                                      ------------------------------------------------------------------------</p>
                                    <p class="">also
                                      ex12_quad_hpddm_reuse_baij fails
                                      with a lot more "Read -1, expected
                                      ..." which I don't know where they
                                      come from...?</p>
                                    <p class="">Hypre (like in
                                      diff-snes_tutorials-ex56_hypre) 
                                      is also having
                                      DIVERGED_INDEFINITE_PC failures...</p>
                                    <p class="">Please see the 3
                                      attached docker files:</p>
                                    <p class="">1)
                                      fedora_mkl_and_devtools : the
                                      DockerFile which install fedore 33
                                      with gnu compilers and MKL and
                                      everything to develop.</p>
                                    <p class="">2) openmpi: the
                                      DockerFile to bluid OpenMPI</p>
                                    <p class="">3) petsc: The las
                                      DockerFile that build/install and
                                      test PETSc</p>
                                    <p class="">I build the 3 like this:</p>
                                    <p class="">docker build -t
                                      fedora_mkl_and_devtools -f
                                      fedora_mkl_and_devtools .</p>
                                    <p class="">docker build -t openmpi
                                      -f openmpi .</p>
                                    <p class="">docker build -t petsc -f
                                      petsc .</p>
                                    <p class="">Disclaimer: I am not a
                                      docker expert, so I may do things
                                      that are not
                                      docker-stat-of-the-art but I am
                                      opened to suggestions... ;)<br
                                        class="">
                                    </p>
                                    <p class="">I have just ran it on my
                                      portable (long) which have not
                                      enough cores, so many more tests
                                      failed (should force
                                      --oversubscribe but don't know how
                                      to).  I will relaunch on my
                                      workstation in a few minutes.</p>
                                    <p class="">I will now test your
                                      branch! (sorry for the delay).</p>
                                    <p class="">Thanks,</p>
                                    <p class="">Eric<br class="">
                                    </p>
                                    <div class="">On 2021-03-11 9:03
                                      a.m., Eric Chamberland wrote:<br
                                        class="">
                                    </div>
                                    <blockquote type="cite" class="">
                                      <p class="">Hi Pierre,</p>
                                      <p class="">ok, that's
                                        interesting!</p>
                                      <p class="">I will try to build a
                                        docker image until tomorrow and
                                        give you the exact recipe to
                                        reproduce the bugs.</p>
                                      <p class="">Eric</p>
                                      <p class=""><br class="">
                                      </p>
                                      <div class="">On 2021-03-11 2:46
                                        a.m., Pierre Jolivet wrote:<br
                                          class="">
                                      </div>
                                      <blockquote type="cite" class="">
                                        <br class="">
                                        <div class=""><br class="">
                                          <blockquote type="cite"
                                            class="">
                                            <div class="">On 11 Mar
                                              2021, at 6:16 AM, Barry
                                              Smith <<a
                                                href="mailto:bsmith@petsc.dev"
                                                target="_blank" class=""
                                                moz-do-not-send="true">bsmith@petsc.dev</a>>
                                              wrote:</div>
                                            <br class="">
                                            <div class="">
                                              <div style="overflow-wrap:
                                                break-word;" class="">
                                                <div class=""><br
                                                    class="">
                                                </div>
                                                  Eric,
                                                <div class=""><br
                                                    class="">
                                                </div>
                                                <div class="">   Sorry
                                                  about not being more
                                                  immediate. We still
                                                  have this in our
                                                  active email so you
                                                  don't need to submit
                                                  individual issues.
                                                  We'll try to get to
                                                  them as soon as we
                                                  can.</div>
                                              </div>
                                            </div>
                                          </blockquote>
                                          <div class=""><br class="">
                                          </div>
                                          <div class="">Indeed, I’m
                                            still trying to figure this
                                            out.</div>
                                          <div class="">I realized that
                                            some of my configure flags
                                            were different than yours,
                                            e.g., no --with-memalign.</div>
                                          <div class="">I’ve also added
                                            SuperLU_DIST to my
                                            installation.</div>
                                          <div class="">Still, I can’t
                                            reproduce any issue.</div>
                                          <div class="">I will continue
                                            looking into this, it
                                            appears I’m seeing some
                                            valgrind errors, but I don’t
                                            know if this is some side
                                            effect of OpenMPI not being
                                            valgrind-clean (last time I
                                            checked, there was no error
                                            with MPICH).</div>
                                          <div class=""><br class="">
                                          </div>
                                          <div class="">Thank you for
                                            your patience,</div>
                                          <div class="">Pierre</div>
                                          <div class=""><br class="">
                                          </div>
                                          <div class="">
                                            <div class="">/usr/bin/gmake
                                              -f gmakefile test
                                              test-fail=1</div>
                                            <div class="">Using
                                              MAKEFLAGS: test-fail=1</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex12_quad_hpddm_reuse_baij.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex12_quad_hpddm_reuse_baij</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex12_quad_hpddm_reuse_baij</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tests-ex33_superlu_dist_2.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex33_superlu_dist_2</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex33_superlu_dist_2</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tests-ex49_superlu_dist.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-0_conv-0</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-0_conv-0</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-0_conv-1</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-0_conv-1</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-1_conv-0</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-1_conv-0</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-1_conv-1</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-1_conv-1</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-0_conv-0</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-0_conv-0</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-0_conv-1</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-0_conv-1</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-1_conv-0</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-1_conv-0</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-1_conv-1</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-1_conv-1</div>
                                            <div class="">        TEST
                                              arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex50_tut_2.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex50_tut_2</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tutorials-ex50_tut_2</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tests-ex33_superlu_dist.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tests-ex33_superlu_dist</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tests-ex33_superlu_dist</div>
                                            <div class="">        TEST
                                              arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex56_hypre.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex56_hypre</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex56_hypre</div>
                                            <div class="">        TEST
                                              arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex56_2.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex56_2</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tutorials-ex56_2</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex17_3d_q3_trig_elas.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex17_3d_q3_trig_elas</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex17_3d_q3_trig_elas</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex12_quad_hpddm_reuse_threshold_baij.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex12_quad_hpddm_reuse_threshold_baij</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex12_quad_hpddm_reuse_threshold_baij</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex5_superlu_dist_3.counts</div>
                                            <div class="">not ok
                                              ksp_ksp_tutorials-ex5_superlu_dist_3
                                              # Error code: 1</div>
                                            <div class="">#<span style="white-space:pre-wrap" class="">       </span>srun:
                                              error: Unable to create
                                              step for job 1426755: More
                                              processors requested than
                                              permitted</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex5_superlu_dist_3
                                              # SKIP Command failed so
                                              no diff</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex5f_superlu_dist.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex5f_superlu_dist
                                              # SKIP Fortran required
                                              for this test</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex12_tri_parmetis_hpddm_baij.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex12_tri_parmetis_hpddm_baij</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex12_tri_parmetis_hpddm_baij</div>
                                            <div class="">        TEST
                                              arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex19_tut_3.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex19_tut_3</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex19_tut_3</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex17_3d_q3_trig_vlap.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex17_3d_q3_trig_vlap</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex17_3d_q3_trig_vlap</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex5f_superlu_dist_3.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex5f_superlu_dist_3
                                              # SKIP Fortran required
                                              for this test</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex19_superlu_dist.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex19_superlu_dist</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex19_superlu_dist</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex56_attach_mat_nearnullspace-1_bddc_approx_hypre.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex56_attach_mat_nearnullspace-1_bddc_approx_hypre</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex56_attach_mat_nearnullspace-1_bddc_approx_hypre</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex49_hypre_nullspace.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex49_hypre_nullspace</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tutorials-ex49_hypre_nullspace</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex19_superlu_dist_2.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex19_superlu_dist_2</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex19_superlu_dist_2</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex5_superlu_dist_2.counts</div>
                                            <div class="">not ok
                                              ksp_ksp_tutorials-ex5_superlu_dist_2
                                              # Error code: 1</div>
                                            <div class="">#<span style="white-space:pre-wrap" class="">       </span>srun:
                                              error: Unable to create
                                              step for job 1426755: More
                                              processors requested than
                                              permitted</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex5_superlu_dist_2
                                              # SKIP Command failed so
                                              no diff</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre.counts</div>
                                            <div class=""> ok
                                              snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre</div>
                                            <div class=""> ok
                                              diff-snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre</div>
                                            <div class="">        TEST
                                              arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex64_1.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex64_1</div>
                                            <div class=""> ok
                                              diff-ksp_ksp_tutorials-ex64_1</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex5_superlu_dist.counts</div>
                                            <div class="">not ok
                                              ksp_ksp_tutorials-ex5_superlu_dist
                                              # Error code: 1</div>
                                            <div class="">#<span style="white-space:pre-wrap" class="">       </span>srun:
                                              error: Unable to create
                                              step for job 1426755: More
                                              processors requested than
                                              permitted</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex5_superlu_dist
                                              # SKIP Command failed so
                                              no diff</div>
                                            <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex5f_superlu_dist_2.counts</div>
                                            <div class=""> ok
                                              ksp_ksp_tutorials-ex5f_superlu_dist_2
                                              # SKIP Fortran required
                                              for this test</div>
                                          </div>
                                          <br class="">
                                          <blockquote type="cite"
                                            class="">
                                            <div class="">
                                              <div style="overflow-wrap:
                                                break-word;" class="">
                                                <div class="">   Barry</div>
                                                <div class=""><br
                                                    class="">
                                                  <div class=""><br
                                                      class="">
                                                    <blockquote
                                                      type="cite"
                                                      class="">
                                                      <div class="">On
                                                        Mar 10, 2021, at
                                                        11:03 PM, Eric
                                                        Chamberland <<a
href="mailto:Eric.Chamberland@giref.ulaval.ca" target="_blank" class=""
moz-do-not-send="true">Eric.Chamberland@giref.ulaval.ca</a>> wrote:</div>
                                                      <br class="">
                                                      <div class="">
                                                        <div class="">
                                                          <p class="">Barry,</p>
                                                          <p class="">to
                                                          get a some
                                                          follow up on
                                                          --with-openmp=1
                                                          failures,
                                                          shall I open
                                                          gitlab issues
                                                          for:</p>
                                                          <p class="">a)
                                                          all hypre
                                                          failures
                                                          giving <span style="white-space:pre-wrap" class="">DIVERGED_INDEFINITE_PC</span></p>
                                                          <p class=""><span style="white-space:pre-wrap" class="">b) all superlu_dist failures giving different results with </span><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class="">initia and "Exceeded timeout limit of 60 s"</span></span></p>
                                                          <p class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class="">c) hpddm failures "free(): invalid next size (fast)" and "Segmentation Violation"
</span></span></p>
                                                          <p class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class="">d) all tao's </span></span><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class="">"Exceeded timeout limit of 60 s"</span></span></span></span></p>
                                                          <p class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class="">I don't see how I could do all these debugging by myself...</span></span></span></span></p>
                                                          <p class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class="">Thanks,</span></span></span></span></p>
                                                          <p class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class=""><span style="white-space:pre-wrap" class="">Eric
</span></span></span></span></p>
                                                          <br class="">
                                                        </div>
                                                      </div>
                                                    </blockquote>
                                                  </div>
                                                  <br class="">
                                                </div>
                                              </div>
                                            </div>
                                          </blockquote>
                                        </div>
                                        <br class="">
                                      </blockquote>
                                      <pre cols="72" class="">-- 
Eric Chamberland, ing., M. Ing
Professionnel de recherche
GIREF/Université Laval
(418) 656-2131 poste 41 22 42</pre>
                                    </blockquote>
                                    <pre cols="72" class="">-- 
Eric Chamberland, ing., M. Ing
Professionnel de recherche
GIREF/Université Laval
(418) 656-2131 poste 41 22 42</pre>
                                  </div>
                                  <span
                                    id="gmail-m_3567963440499379521cid:26EB476E-4C18-4B68-9DC9-6FBE92E94935"
                                    class=""><fedora_mkl_and_devtools.txt></span><span
id="gmail-m_3567963440499379521cid:EC379F4B-01BD-409E-8BBC-6FBA5A49236E"
                                    class=""><openmpi.txt></span><span
id="gmail-m_3567963440499379521cid:3CA91B7D-219A-4965-9FF3-D836488847A0"
                                    class=""><petsc.txt></span></div>
                              </blockquote>
                            </div>
                            <br class="">
                          </div>
                        </div>
                      </div>
                    </blockquote>
                  </div>
                  <br class="">
                </div>
              </blockquote>
            </div>
          </div>
        </blockquote>
      </div>
      <br class="">
    </blockquote>
    <pre class="moz-signature" cols="72">-- 
Eric Chamberland, ing., M. Ing
Professionnel de recherche
GIREF/Université Laval
(418) 656-2131 poste 41 22 42</pre>
  </body>
</html>