<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>Done:<br>
    </p>
    <p><a class="moz-txt-link-freetext" href="https://github.com/hypre-space/hypre/issues/303">https://github.com/hypre-space/hypre/issues/303</a></p>
    <p>Maybe I will need some help about PETSc to answer their
      questions...</p>
    <p>Eric<br>
    </p>
    <div class="moz-cite-prefix">On 2021-03-14 3:44 a.m., Stefano
      Zampini wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:5324E970-3CAF-4980-8387-31AA0E208087@gmail.com">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      Eric
      <div class=""><br class="">
      </div>
      <div class="">You should report these HYPRE issues upstream <a
          href="https://github.com/hypre-space/hypre/issues" class=""
          moz-do-not-send="true">https://github.com/hypre-space/hypre/issues</a></div>
      <div class=""><br class="">
        <div><br class="">
          <blockquote type="cite" class="">
            <div class="">On Mar 14, 2021, at 3:44 AM, Eric Chamberland
              <<a href="mailto:Eric.Chamberland@giref.ulaval.ca"
                class="" moz-do-not-send="true">Eric.Chamberland@giref.ulaval.ca</a>>
              wrote:</div>
            <br class="Apple-interchange-newline">
            <div class="">
              <meta http-equiv="Content-Type" content="text/html;
                charset=UTF-8" class="">
              <div class="">
                <p class="">For us it clearly creates problems in real
                  computations...<br class="">
                </p>
                <p class="">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 class="">Another solution:  force single-threaded
                  configuration for hypre until this is fixed?</p>
                <p class="">Eric<br class="">
                </p>
                On 2021-03-13 8:50 a.m., Pierre Jolivet wrote:<br
                  class="">
                <blockquote type="cite"
                  cite="mid:7932C4DD-11BB-4BBB-A171-08585D7ADADE@joliv.et"
                  class="">
                  <meta http-equiv="Content-Type" content="text/html;
                    charset=UTF-8" class="">
                  <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 class="">
                    <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>
              </div>
            </div>
          </blockquote>
        </div>
        <br class="">
      </div>
    </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>