<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>Hi,</p>
    <p>Just to add something I didn't knew before: I got an answer from
      intel and tried the mkl_link_tool script shipped with mkl (into
      OneAPI package):</p>
    <p>$MKLROOT/bin/intel64/mkl_link_tool -libs --compiler=gnu_c
      --arch=intel64 --linking=dynamic --parallel=yes --interface=lp64
      --threading-library=gomp --mpi=openmpi --cluster_library=scalapack</p>
    <p><br>
          Intel(R) oneAPI Math Kernel Library (oneMKL) Link Tool v6.1<br>
             ==========================================================<br>
      <br>
      Output<br>
      ======<br>
      <br>
      Linking line:<br>
       -L${MKLROOT}/lib/intel64 -lmkl_scalapack_lp64 -Wl,--no-as-needed
      -lmkl_intel_lp64 -lmkl_gnu_thread -lmkl_core
      -lmkl_blacs_openmpi_lp64 -lgomp -lpthread -lm -ldl<br>
    </p>
    <p>I don't know if this is intersting for PETSc configuration
      scripts to try to use this or not, but it exists!</p>
    <p>(see
<a class="moz-txt-link-freetext" href="https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Intel-MKL-Link-Line-Advisor-as-external-tool/m-p/1260895#M30974">https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Intel-MKL-Link-Line-Advisor-as-external-tool/m-p/1260895#M30974</a>
      )<br>
    </p>
    <p>Eric</p>
    <p><br>
    </p>
    <div class="moz-cite-prefix">On 2021-03-03 11:13 a.m., Barry Smith
      wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:70E8C846-F726-491B-A7A5-34F8C6DD8F66@petsc.dev">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div class=""><br class="">
      </div>
      <div>
        <div>
          <blockquote type="cite" class="">
            <div class="">
              <blockquote type="cite"
                cite="mid:260581B5-04AC-4B8D-8A91-236401709785@joliv.et"
                class="">
                <div class="">2) I can reproduce
                  the src/mat/tests/ex242.c error (which explicitly uses
                  ScaLAPACK, none of the above PC uses it explicitly,
                  except PCBDDC/PCHPDDM when using MUMPS on “big”
                  problems where root nodes are factorized using
                  ScaLAPACK, see -mat_mumps_icntl_13)</div>
                <div class="">3) I’m seeing that both on your machine
                  and mine, PETSc BuildSystem insist on linking
                  libmkl_blacs_intelmpi_lp64.so even though we supply
                  explicitly libmkl_blacs_openmpi_lp64.so</div>
                <div class="">This for example yields a wrong
                  Makefile.inc for MUMPS:</div>
                <div class="">$ cat
                  arch-linux2-c-opt-ompi/externalpackages/MUMPS_5.3.5/Makefile.inc|grep
                  blacs</div>
                <div class="">
                  <div class="">SCALAP  = […] -lmkl_blacs_openmpi_lp64</div>
                  <div class="">LIBBLAS = […] -lmkl_blacs_intelmpi_lp64
                    -lgomp -ldl -lpthread -lm […]</div>
                </div>
                <div class=""><br class="">
                </div>
                <div class="">Despite what Barry says, I think PETSc is
                  partially to blame as well (why use
                  libmkl_blacs_intelmpi_lp64.so even though BuildSystem
                  is capable of detecting we are using OpenMPI).</div>
                <div class="">I’ll try to fix this to see if it solves
                  2).</div>
              </blockquote>
              <p class="">Okay, that's a very nice finding!!!  Hope it
                will be "fixable" easily!</p>
              <div class=""><br class="">
              </div>
            </div>
          </blockquote>
          The knowledge is there but the information may not be
          trivially available to make the right decisions. Parts of the
          BLAS/LAPACK checks use the "check everything" approach. For
          example</div>
        <div><br class="">
        </div>
        <div>
          <div># Look for Multi-Threaded MKL for MKL_C/Pardiso</div>
          <div>      useCPardiso=0</div>
          <div>      usePardiso=0</div>
          <div>      if self.argDB['with-mkl_cpardiso'] or
            'with-mkl_cpardiso-dir' in self.argDB or
            'with-mkl_cpardiso-lib' in self.argDB:</div>
          <div>        useCPardiso=1</div>
          <div>       
mkl_blacs_64=[['mkl_blacs_intelmpi'+ILP64+''],['mkl_blacs_mpich'+ILP64+''],['mkl_blacs_sgimpt'+ILP64+''],['mkl_blacs_openmpi'+ILP64+'']]</div>
          <div>       
mkl_blacs_32=[['mkl_blacs_intelmpi'],['mkl_blacs_mpich'],['mkl_blacs_sgimpt'],['mkl_blacs_openmpi']]</div>
          <div>      elif self.argDB['with-mkl_pardiso'] or
            'with-mkl_pardiso-dir' in self.argDB or
            'with-mkl_pardiso-lib' in self.argDB:</div>
          <div>        usePardiso=1</div>
          <div>        mkl_blacs_64=[[]]</div>
          <div>        mkl_blacs_32=[[]]</div>
          <div>      if useCPardiso or usePardiso:</div>
          <div>        self.logPrintBox('BLASLAPACK: Looking for
            Multithreaded MKL for C/Pardiso')</div>
          <div>        for libdir in
[os.path.join('lib','64'),os.path.join('lib','ia64'),os.path.join('lib','em64t'),os.path.join('lib','intel64'),'lib','64','ia64','em64t','intel64',</div>
          <div>                     
             os.path.join('lib','32'),os.path.join('lib','ia32'),'32','ia32','']:</div>
          <div>          if not
            os.path.exists(os.path.join(dir,libdir)):</div>
          <div>            self.logPrint('MKL Path not found.. skipping:
            '+os.path.join(dir,libdir))</div>
          <div>          else:</div>
          <div>            self.log.write('Files and directories in that
directory:\n'+str(os.listdir(os.path.join(dir,libdir)))+'\n')</div>
          <div>            #  iomp5 is provided by the Intel compilers
            on MacOS. Run source /opt/intel/bin/compilervars.sh intel64
            to have it added to LIBRARY_PATH</div>
          <div>            #  then locate libimp5.dylib in the
            LIBRARY_PATH and copy it to os.path.join(dir,libdir)</div>
          <div>            for i in mkl_blacs_64:</div>
          <div>              yield ('User specified MKL-C/Pardiso
            Intel-Linux64', None,
[os.path.join(dir,libdir,'libmkl_intel'+ILP64+'.a'),'mkl_core','mkl_intel_thread']+i+['iomp5','dl','pthread'],known,'yes')</div>
          <div>              yield ('User specified MKL-C/Pardiso
            GNU-Linux64', None,
[os.path.join(dir,libdir,'libmkl_intel'+ILP64+'.a'),'mkl_core','mkl_gnu_thread']+i+['gomp','dl','pthread'],known,'yes')</div>
          <div>              yield ('User specified MKL-Pardiso
            Intel-Windows64', None,
[os.path.join(dir,libdir,'mkl_core.lib'),'mkl_intel'+ILP64+'.lib','mkl_intel_thread.lib']+i+['libiomp5md.lib'],known,'yes')</div>
          <div>            for i in mkl_blacs_32:</div>
          <div>              yield ('User specified MKL-C/Pardiso
            Intel-Linux32', None,
[os.path.join(dir,libdir,'libmkl_intel.a'),'mkl_core','mkl_intel_thread']+i+['iomp5','dl','pthread'],'32','yes')</div>
          <div>              yield ('User specified MKL-C/Pardiso
            GNU-Linux32', None,
[os.path.join(dir,libdir,'libmkl_intel.a'),'mkl_core','mkl_gnu_thread']+i+['gomp','dl','pthread'],'32','yes')</div>
          <div>              yield ('User specified MKL-Pardiso
            Intel-Windows32', None,
[os.path.join(dir,libdir,'mkl_core.lib'),'mkl_intel_c.lib','mkl_intel_thread.lib']+i+['libiomp5md.lib'],'32','yes')</div>
          <div>        return</div>
          <div><br class="">
          </div>
          <div>The assumption is that the link will fail unless the
            correct libraries are in the list. But apparently this is
            not the case; it returns the first case that links but that
            case does not run which is why it appears to be producing
            "silly" results.</div>
          <div><br class="">
          </div>
          <div>If you set the right MPI and threading library, at these
            locations instead of trying all of them it might resolve the
            problems. </div>
          <div><br class="">
          </div>
          <div>
            <div>    if self.openmp.found:</div>
            <div>      ITHREAD='intel_thread'</div>
            <div>      ITHREADGNU='gnu_thread'</div>
            <div>      ompthread = 'yes'</div>
            <div>    else:</div>
            <div>      ITHREAD='sequential'</div>
            <div>      ITHREADGNU='sequential'</div>
            <div>      ompthread = 'no'</div>
            <div><br class="">
            </div>
            <div>
              <div>       
mkl_blacs_64=[['mkl_blacs_intelmpi'+ILP64+''],['mkl_blacs_mpich'+ILP64+''],['mkl_blacs_sgimpt'+ILP64+''],['mkl_blacs_openmpi'+ILP64+'']]</div>
              <div>       
mkl_blacs_32=[['mkl_blacs_intelmpi'],['mkl_blacs_mpich'],['mkl_blacs_sgimpt'],['mkl_blacs_openmpi']]</div>
            </div>
          </div>
          <div><br class="">
          </div>
          <div><br class="">
          </div>
          <div><br class="">
          </div>
          <div><br class="">
          </div>
          <blockquote type="cite" class="">
            <div class=""><br class="webkit-block-placeholder">
            </div>
          </blockquote>
          <br class="">
        </div>
      </div>
      <div><br class="">
        <blockquote type="cite" class="">
          <div class="">On Mar 3, 2021, at 8:22 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="">Hi Pierre,</p>
              <div class="moz-cite-prefix">On 2021-03-03 2:42 a.m.,
                Pierre Jolivet wrote:<br class="">
              </div>
              <blockquote type="cite"
                cite="mid:260581B5-04AC-4B8D-8A91-236401709785@joliv.et"
                class="">
                <meta http-equiv="Content-Type" content="text/html;
                  charset=UTF-8" class="">
                <blockquote type="cite" class="">
                  <div class="">
                    <p class="">If it ends that there is a problem
                      combining MKL + openMP that relies on linking
                      configuration for example, should it be a good
                      thing to have this (--with-openmp=1) tested into
                      the pipelines (with external packages of course)?</p>
                  </div>
                </blockquote>
                <div class="">As Barry said, there is not much (if any)
                  OpenMP in PETSc.</div>
                <div class="">There is however some workers with the MKL
                  (+ Intel compilers) turned on, but I don’t think we
                  test MKL + GNU compilers (which I feel like is a very
                  niche combination, hence not really worth testing,
                  IMHO).</div>
              </blockquote>
              <p class="">Ouch, this is my almost my personal working
                configuration and for most of our users too... and it
                worked well until I activated the OpenMP thing...<br
                  class="">
              </p>
              <p class="">We had good reasons to work with g++ or
                clang++ instead of intel compilers:</p>
              <p class="">- It is mandatory to pay to work with an intel
                compiler (didn't looked at OneAPI licensing yet, but it
                may have changed?)<br class="">
              </p>
              <p class="">- No support of Intel compilers with iceccd
                (slow recompilation)<br class="">
              </p>
              <p class="">- MKL was freely distributed, so it can be
                used with any compiler</p>
              <p class="">That doesn't mean we don't want to use intel
                compiler, but maybe we just want to to a specific
                delivery with it but continue to develop with g++ or
                clang++ (my personal choice).</p>
              <p class="">But I understand it is less straightforward to
                combine gcc and MKL than using native Intel
                tool-chain....<br class="">
              </p>
            </div>
          </div>
        </blockquote>
        <div><br class="">
        </div>
            I agree the MKL + GNU compilers is commonly used and should
        be tested and maintained in PETSc. <br class="">
        <blockquote type="cite" class="">
          <div class="">
            <div class="">
              <p class=""> </p>
              <blockquote type="cite"
                cite="mid:260581B5-04AC-4B8D-8A91-236401709785@joliv.et"
                class="">
                <blockquote type="cite" class="">
                  <div class="">
                    <p class="">Does the guys who maintain all these
                      libs are reading petsc-dev? ;)</p>
                  </div>
                </blockquote>
                <div class="">I don’t think they are, but don’t worry,
                  we do forward the appropriate messages to them :)</div>
              </blockquote>
              :)<br class="">
              <blockquote type="cite"
                cite="mid:260581B5-04AC-4B8D-8A91-236401709785@joliv.et"
                class="">
                <div class=""><br class="">
                </div>
                About yesterday’s failures…
                <div class="">1) I cannot reproduce any of the
                  PCHYPRE/PCBDDC/PCHPDDM errors (sorry I didn’t bother
                  putting the SuperLU_DIST tarball on my cluster)</div>
              </blockquote>
              <p class="">Hmmm, maybe my environment variables may play
                a role into this?</p>
              <p class="">for comparisons considerations, we explicitly
                set:</p>
              <p class="">export MKL_CBWR=COMPATIBLE<br class="">
                export MKL_NUM_THREADS=1<br class="">
              </p>
              <p class="">but it would be surprising it helps reproduce
                a problem: they usually stabilize results...<br class="">
              </p>
            </div>
          </div>
        </blockquote>
      </div>
      <div>
        <blockquote type="cite" class="">
          <div class="">
            <div class="">
              <p class="">Merci,</p>
              <p class="">Eric<br class="">
              </p>
              <blockquote type="cite"
                cite="mid:260581B5-04AC-4B8D-8A91-236401709785@joliv.et"
                class="">
                <div class=""><br class="">
                </div>
                <div class="">Thanks,</div>
                <div class="">Pierre</div>
                <div class=""><br class="">
                </div>
                <div class=""><a
                    href="http://joliv.et/irene-rome-configure.log"
                    class="" moz-do-not-send="true">http://joliv.et/irene-rome-configure.log</a></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_tutorials-ex50_tut_2.counts</div>
                  <div class=""> ok ksp_ksp_tutorials-ex50_tut_2 # SKIP
                    PETSC_HAVE_SUPERLU_DIST requirement not met</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/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/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/mat_tests-ex242_3.counts</div>
                  <div class="">not ok mat_tests-ex242_3 # Error code:
                    137</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR:
                    ------------------------------------------------------------------------</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: Caught signal number 11 SEGV: Segmentation
                    Violation, probably memory access out of range</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: Try option -start_in_debugger or
                    -on_error_attach_debugger</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: or see <a
                      href="https://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind"
                      class="" moz-do-not-send="true">https://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind</a></div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: or try <a href="http://valgrind.org/"
                      class="" moz-do-not-send="true">http://valgrind.org</a>
                    on GNU/linux and Apple Mac OS X to find memory
                    corruption errors</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: configure using --with-debugging=yes,
                    recompile, link, and run</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: to get more information on the crash.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: --------------------- Error Message
                    --------------------------------------------------------------</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: Signal received</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: See <a
                      href="https://www.mcs.anl.gov/petsc/documentation/faq.html"
                      class="" moz-do-not-send="true">https://www.mcs.anl.gov/petsc/documentation/faq.html</a>
                    for trouble shooting.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: Petsc Development GIT revision:
                    v3.14.4-733-g7ab9467ef9  GIT Date: 2021-03-02
                    16:15:11 +0000</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR:
                    ------------------------------------------------------------------------</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: Caught signal number 11 SEGV: Segmentation
                    Violation, probably memory access out of range</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: Try option -start_in_debugger or
                    -on_error_attach_debugger</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: or see <a
                      href="https://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind"
                      class="" moz-do-not-send="true">https://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind</a></div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: or try <a href="http://valgrind.org/"
                      class="" moz-do-not-send="true">http://valgrind.org</a>
                    on GNU/linux and Apple Mac OS X to find memory
                    corruption errors</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: configure using --with-debugging=yes,
                    recompile, link, and run</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: to get more information on the crash.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: --------------------- Error Message
                    --------------------------------------------------------------</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: Signal received</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: See <a
                      href="https://www.mcs.anl.gov/petsc/documentation/faq.html"
                      class="" moz-do-not-send="true">https://www.mcs.anl.gov/petsc/documentation/faq.html</a>
                    for trouble shooting.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: Petsc Development GIT revision:
                    v3.14.4-733-g7ab9467ef9  GIT Date: 2021-03-02
                    16:15:11 +0000</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR:
/ccc/work/cont003/rndm/rndm/petsc/arch-linux2-c-opt-ompi/tests/mat/tests/runex242_3/../ex242
                    on a arch-linux2-c-opt-ompi named irene4047 by
                    jolivetp Wed Mar  3 08:21:20 2021</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: Configure options --download-hpddm
                    --download-hpddm-commit=origin/main --download-hypre
                    --download-metis --download-mumps
                    --download-parmetis --download-ptscotch
                    --download-slepc --download-slepc-commit=origin/main
                    --download-tetgen --known-mpi-c-double-complex
                    --known-mpi-int64_t --known-mpi-long-double
                    --with-avx512-kernels=1
--with-blaslapack-dir=/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl/lib/intel64
                    --with-cc=mpicc --with-cxx=mpicxx --with-debugging=0
                    --with-fc=mpifort --with-fortran-bindings=0
                    --with-make-np=40
--with-mkl_cpardiso-dir=/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281
                    --with-mkl_cpardiso=1
--with-mkl_pardiso-dir=/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl
                    --with-mkl_pardiso=1 --with-mpiexec=ccc_mprun
                    --with-openmp=1
--with-packages-download-dir=/ccc/cont003/home/enseeiht/jolivetp/Dude/externalpackages/
--with-scalapack-include=/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl/include
--with-scalapack-lib="[/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl/lib/intel64/libmkl_scalapack_lp64.so,/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl/lib/intel64/libmkl_blacs_openmpi_lp64.so]"
                    --with-scalar-type=real --with-x=0 COPTFLAGS="-O3
                    -fp-model fast -mavx2" CXXOPTFLAGS="-O3 -fp-model
                    fast -mavx2" FOPTFLAGS="-O3 -fp-model fast -mavx2"
                    PETSC_ARCH=arch-linux2-c-opt-ompi</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: #1 User provided function() line 0 in
                     unknown file</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[2]PETSC
                    ERROR: Run with -malloc_debug to check if memory
                    corruption is causing the crash.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR:
/ccc/work/cont003/rndm/rndm/petsc/arch-linux2-c-opt-ompi/tests/mat/tests/runex242_3/../ex242
                    on a arch-linux2-c-opt-ompi named irene4047 by
                    jolivetp Wed Mar  3 08:21:20 2021</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: Configure options --download-hpddm
                    --download-hpddm-commit=origin/main --download-hypre
                    --download-metis --download-mumps
                    --download-parmetis --download-ptscotch
                    --download-slepc --download-slepc-commit=origin/main
                    --download-tetgen --known-mpi-c-double-complex
                    --known-mpi-int64_t --known-mpi-long-double
                    --with-avx512-kernels=1
--with-blaslapack-dir=/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl/lib/intel64
                    --with-cc=mpicc --with-cxx=mpicxx --with-debugging=0
                    --with-fc=mpifort --with-fortran-bindings=0
                    --with-make-np=40
--with-mkl_cpardiso-dir=/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281
                    --with-mkl_cpardiso=1
--with-mkl_pardiso-dir=/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl
                    --with-mkl_pardiso=1 --with-mpiexec=ccc_mprun
                    --with-openmp=1
--with-packages-download-dir=/ccc/cont003/home/enseeiht/jolivetp/Dude/externalpackages/
--with-scalapack-include=/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl/include
--with-scalapack-lib="[/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl/lib/intel64/libmkl_scalapack_lp64.so,/ccc/products/mkl-19.0.5.281/intel--19.0.5.281__openmpi--4.0.1/default/19.0.5.281/mkl/lib/intel64/libmkl_blacs_openmpi_lp64.so]"
                    --with-scalar-type=real --with-x=0 COPTFLAGS="-O3
                    -fp-model fast -mavx2" CXXOPTFLAGS="-O3 -fp-model
                    fast -mavx2" FOPTFLAGS="-O3 -fp-model fast -mavx2"
                    PETSC_ARCH=arch-linux2-c-opt-ompi</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: #1 User provided function() line 0 in
                     unknown file</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>[1]PETSC
                    ERROR: Run with -malloc_debug to check if memory
                    corruption is causing the crash.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>--------------------------------------------------------------------------</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>MPI_ABORT
                    was invoked on rank 2 in communicator MPI_COMM_WORLD</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>with
                    errorcode 50176059.</div>
                  <div class="">#</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>NOTE:
                    invoking MPI_ABORT causes Open MPI to kill all MPI
                    processes.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>You
                    may or may not see output from other processes,
                    depending on</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>exactly
                    when Open MPI kills them.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>--------------------------------------------------------------------------</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>--------------------------------------------------------------------------</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>MPI_ABORT
                    was invoked on rank 1 in communicator MPI_COMM_WORLD</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>with
                    errorcode 50176059.</div>
                  <div class="">#</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>NOTE:
                    invoking MPI_ABORT causes Open MPI to kill all MPI
                    processes.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>You
                    may or may not see output from other processes,
                    depending on</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>exactly
                    when Open MPI kills them.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>--------------------------------------------------------------------------</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>srun:
                    Job step aborted: Waiting up to 302 seconds for job
                    step to finish.</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>slurmstepd-irene4047:
                    error: *** STEP 1374176.36 ON irene4047 CANCELLED AT
                    2021-03-03T08:21:20 ***</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>srun:
                    error: irene4047: task 0: Killed</div>
                  <div class="">#<span class="Apple-tab-span" style="white-space:pre">        </span>srun:
                    error: irene4047: tasks 1-2: Exited with exit code
                    16</div>
                  <div class=""> ok mat_tests-ex242_3 # SKIP Command
                    failed so no diff</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/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/ts_tutorials-ex18_p1p1_xper_ref.counts</div>
                  <div class=""> ok ts_tutorials-ex18_p1p1_xper_ref</div>
                  <div class=""> ok diff-ts_tutorials-ex18_p1p1_xper_ref</div>
                  <div class="">        TEST
arch-linux2-c-opt-ompi/tests/counts/ts_tutorials-ex18_p1p1_xyper_ref.counts</div>
                  <div class=""> ok ts_tutorials-ex18_p1p1_xyper_ref</div>
                  <div class=""> ok
                    diff-ts_tutorials-ex18_p1p1_xyper_ref</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 # SKIP
                    PETSC_HAVE_SUPERLU_DIST requirement not met</div>
                  <div class=""><br class="">
                    <blockquote type="cite" class="">
                      <div class="">On 3 Mar 2021, at 6:21 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="">Just started a discussion on the
                            side:</p>
                          <p class=""><a class="moz-txt-link-freetext"
href="https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Intel-MKL-Link-Line-Advisor-as-external-tool/m-p/1260895#M30974"
                              moz-do-not-send="true">https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Intel-MKL-Link-Line-Advisor-as-external-tool/m-p/1260895#M30974</a></p>
                          <p class="">Eric<br class="">
                          </p>
                          <div class="moz-cite-prefix">On 2021-03-02
                            3:50 p.m., Pierre Jolivet wrote:<br class="">
                          </div>
                          <blockquote type="cite"
                            cite="mid:FF313618-3465-474B-B13E-A7FDA1FB074D@joliv.et"
                            class="">
                            <meta http-equiv="Content-Type"
                              content="text/html; charset=UTF-8"
                              class="">
                            Hello Eric,
                            <div class="">
                              <div class=""><span style="white-space: pre-wrap;" class="">src/mat/tests/ex237.c is a recent test with some code paths that should be disabled for “old” MKL versions. It’s tricky to check directly in the source (we do check in BuildSystem) because there is no such thing as PETSC_PKG_MKL_VERSION_LT, but I guess we can change if defined(PETSC_HAVE_MKL) to if defined(PETSC_HAVE_MKL) && defined(PETSC_HAVE_MKL_SPARSE_OPTIMIZE), I’ll make a MR, thanks for reporting this.</span></div>
                              <div class="">
                                <div class=""><span style="white-space: pre-wrap;" class="">
</span></div>
                                <div class="">
                                  <div class="">For the other issues,
                                    I’m sensing this is a problem with
                                    gomp + intel_gnu_thread, but this is
                                    pure speculation… sorry.</div>
                                  <div class="">I’ll try to reproduce
                                    some of these problems if you are
                                    not given a more meaningful answer.</div>
                                </div>
                                <div class=""><span style="white-space: pre-wrap;" class="">
</span></div>
                                <div class=""><span style="white-space: pre-wrap;" class="">Thanks,</span></div>
                                <div class=""><span style="white-space: pre-wrap;" class="">Pierre</span></div>
                              </div>
                              <div class="">
                                <div class=""><span style="white-space: pre-wrap;" class="">
</span></div>
                                <div class="">
                                  <blockquote type="cite" class="">
                                    <div class="">On 2 Mar 2021, at 9:14
                                      PM, 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="">Hi,</p>
                                        <p class="">It all started when
                                          I wanted to test PETSC/CUDA
                                          compatibility for our code.</p>
                                        <p class="">I had to activate
                                          --with-openmp to configure
                                          with --with-cuda=1
                                          successfully.</p>
                                        <p class="">I then saw that
                                          PETSC_HAVE_OPENMP  is used at
                                          least in MUMPS (and some other
                                          places).</p>
                                        <p class="">So, I configured and
                                          tested petsc with openmp
                                          activated, without CUDA.<br
                                            class="">
                                        </p>
                                        <p class="">The first thing I
                                          see is that our code CI
                                          pipelines now fails for many
                                          tests.</p>
                                        <p class="">After looking
                                          deeper, it seems that PETSc
                                          itself fails many tests when I
                                          activate openmp!</p>
                                        Here are all the configurations
                                        I have results for, after/before
                                        activating OpenMP for PETSc:<br
                                          class="">
                                        <p class="">==============================================================================</p>
                                        <p class="">==============================================================================</p>
                                        <p class="">For petsc/master +
                                          OpenMPI 4.0.4 + MKL
                                          2019.4.243:</p>
                                        <p class="">With OpenMP=1</p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/petsc-master-debug/2021.03.02.02h00m02s_make_test.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/petsc-master-debug/2021.03.02.02h00m02s_make_test.log</a></p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/petsc-master-debug/2021.03.02.02h00m02s_configure.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/petsc-master-debug/2021.03.02.02h00m02s_configure.log</a></p>
                                        <pre style="font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-thickness: initial; overflow-wrap: break-word; white-space: pre-wrap;" class=""># -------------
#   Summary    
# -------------
# FAILED snes_tutorials-ex12_quad_hpddm_reuse_baij diff-ksp_ksp_tests-ex33_superlu_dist_2 diff-ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-0_conv-0 diff-ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-0_conv-1 diff-ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-1_conv-0 diff-ksp_ksp_tests-ex49_superlu_dist+nsize-1herm-1_conv-1 diff-ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-0_conv-0 diff-ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-0_conv-1 diff-ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-1_conv-0 diff-ksp_ksp_tests-ex49_superlu_dist+nsize-4herm-1_conv-1 ksp_ksp_tutorials-ex50_tut_2 diff-ksp_ksp_tests-ex33_superlu_dist diff-snes_tutorials-ex56_hypre snes_tutorials-ex17_3d_q3_trig_elas snes_tutorials-ex12_quad_hpddm_reuse_threshold_baij ksp_ksp_tutorials-ex5_superlu_dist_3 ksp_ksp_tutorials-ex5f_superlu_dist snes_tutorials-ex12_tri_parmetis_hpddm_baij diff-snes_tutorials-ex19_tut_3 mat_tests-ex242_3 snes_tutorials-ex17_3d_q3_trig_vlap ksp_ksp_tutorials-ex5f_superlu_dist_3 snes_tutorials-ex19_superlu_dist diff-snes_tutorials-ex56_attach_mat_nearnullspace-1_bddc_approx_hypre diff-ksp_ksp_tutorials-ex49_hypre_nullspace ts_tutorials-ex18_p1p1_xper_ref ts_tutorials-ex18_p1p1_xyper_ref snes_tutorials-ex19_superlu_dist_2 ksp_ksp_tutorials-ex5_superlu_dist_2 diff-snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre ksp_ksp_tutorials-ex64_1 ksp_ksp_tutorials-ex5_superlu_dist ksp_ksp_tutorials-ex5f_superlu_dist_2
# success 8275/10003 tests (82.7%)
# <b class="">failed 33/10003</b> tests (0.3%)</pre>
                                        <p class="">With OpenMP=0</p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/petsc-master-debug/2021.02.26.02h00m16s_make_test.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/petsc-master-debug/2021.02.26.02h00m16s_make_test.log</a></p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/petsc-master-debug/2021.02.26.02h00m16s_configure.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/petsc-master-debug/2021.02.26.02h00m16s_configure.log</a><br
                                            class="">
                                        </p>
                                        <pre style="font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-thickness: initial; overflow-wrap: break-word; white-space: pre-wrap;" class=""># -------------
#   Summary    
# -------------
# FAILED tao_constrained_tutorials-tomographyADMM_6 snes_tutorials-ex17_3d_q3_trig_elas mat_tests-ex242_3 snes_tutorials-ex17_3d_q3_trig_vlap tao_leastsquares_tutorials-tomography_1 tao_constrained_tutorials-tomographyADMM_5
# success 8262/9983 tests (82.8%)
# <b class="">failed 6/9983</b> tests (0.1%)</pre>
                                        <p class="">==============================================================================</p>
                                        <p class="">==============================================================================</p>
                                        <p class="">For OpenMPI
                                          3.1.x/master:<br class="">
                                        </p>
                                        <p class="">With OpenMP=1:</p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/ompi_3.x/2021.03.01.22h00m01s_make_test.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/ompi_3.x/2021.03.01.22h00m01s_make_test.log</a></p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/ompi_3.x/2021.03.01.22h00m01s_configure.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/ompi_3.x/2021.03.01.22h00m01s_configure.log</a></p>
                                        <pre style="font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-thickness: initial; overflow-wrap: break-word; white-space: pre-wrap;" class=""># -------------
#   Summary    
# -------------
# FAILED mat_tests-ex242_3 mat_tests-ex242_2 diff-mat_tests-ex219f_1 diff-dm_tutorials-ex11f90_1 ksp_ksp_tutorials-ex5_superlu_dist_3 diff-ksp_ksp_tutorials-ex49_hypre_nullspace ksp_ksp_tutorials-ex5f_superlu_dist_3 snes_tutorials-ex17_3d_q3_trig_vlap diff-snes_tutorials-ex56_attach_mat_nearnullspace-1_bddc_approx_hypre diff-snes_tutorials-ex19_tut_3 diff-snes_tutorials-ex56_hypre diff-snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre tao_leastsquares_tutorials-tomography_1 tao_constrained_tutorials-tomographyADMM_4 tao_constrained_tutorials-tomographyADMM_6 diff-tao_constrained_tutorials-toyf_1
# success 8142/9765 tests (83.4%)
# <b class="">failed 16/9765</b> tests (0.2%)</pre>
                                        <p class="">With OpenMP=0:</p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/ompi_3.x/2021.02.28.22h00m02s_make_test.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/ompi_3.x/2021.02.28.22h00m02s_make_test.log</a></p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/ompi_3.x/2021.02.28.22h00m02s_configure.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/ompi_3.x/2021.02.28.22h00m02s_configure.log</a><br
                                            class="">
                                        </p>
                                        <pre style="font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-thickness: initial; overflow-wrap: break-word; white-space: pre-wrap;" class=""># -------------
#   Summary    
# -------------
# FAILED mat_tests-ex242_3 mat_tests-ex242_2 diff-mat_tests-ex219f_1 diff-dm_tutorials-ex11f90_1 ksp_ksp_tutorials-ex56_2 snes_tutorials-ex17_3d_q3_trig_vlap tao_leastsquares_tutorials-tomography_1 tao_constrained_tutorials-tomographyADMM_4 diff-tao_constrained_tutorials-toyf_1
# success 8151/9767 tests (83.5%)
# <b class="">failed 9/9767</b> tests (0.1%)
</pre>
                                        <p class="">==============================================================================</p>
                                        <p class="">==============================================================================</p>
                                        <p class="">For OpenMPI
                                          4.0.x/master:</p>
                                        <p class="">With OpenMP=1:</p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/ompi_4.x/2021.03.01.20h00m01s_make_test.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/ompi_4.x/2021.03.01.20h00m01s_make_test.log</a></p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/ompi_4.x/2021.03.01.20h00m01s_configure.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/ompi_4.x/2021.03.01.20h00m01s_configure.log</a><br
                                            class="">
                                        </p>
                                        <pre style="font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-thickness: initial; overflow-wrap: break-word; white-space: pre-wrap;" class=""># FAILED snes_tutorials-ex17_3d_q3_trig_elas snes_tutorials-ex19_hypre ksp_ksp_tutorials-ex56_2 tao_leastsquares_tutorials-tomography_1 tao_constrained_tutorials-tomographyADMM_5 mat_tests-ex242_3 ksp_ksp_tutorials-ex55_hypre ksp_ksp_tutorials-ex5_superlu_dist_2 tao_constrained_tutorials-tomographyADMM_6 snes_tutorials-ex56_hypre snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre ksp_ksp_tutorials-ex5f_superlu_dist_3 ksp_ksp_tutorials-ex34_hyprestruct diff-ksp_ksp_tutorials-ex49_hypre_nullspace snes_tutorials-ex56_attach_mat_nearnullspace-1_bddc_approx_hypre ksp_ksp_tutorials-ex5f_superlu_dist ksp_ksp_tutorials-ex5f_superlu_dist_2 ksp_ksp_tutorials-ex5_superlu_dist snes_tutorials-ex19_tut_3 snes_tutorials-ex19_superlu_dist ksp_ksp_tutorials-ex50_tut_2 snes_tutorials-ex17_3d_q3_trig_vlap ksp_ksp_tutorials-ex5_superlu_dist_3 snes_tutorials-ex19_superlu_dist_2 tao_constrained_tutorials-tomographyADMM_4 ts_tutorials-ex26_2
# success 8125/9753 tests (83.3%)
# <b class="">failed 26/9753</b> tests (0.3%)</pre>
                                        <p class="">With OpenMP=0</p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/ompi_4.x/2021.02.28.20h00m04s_make_test.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/ompi_4.x/2021.02.28.20h00m04s_make_test.log</a></p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/ompi_4.x/2021.02.28.20h00m04s_configure.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/ompi_4.x/2021.02.28.20h00m04s_configure.log</a><br
                                            class="">
                                        </p>
                                        <pre style="font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-thickness: initial; overflow-wrap: break-word; white-space: pre-wrap;" class=""># FAILED mat_tests-ex242_3
# success 8174/9777 tests (83.6%)
# <b class="">failed 1/9777</b> tests (0.0%)

</pre>
                                        <p class="">==============================================================================</p>
                                        <p class="">==============================================================================</p>
                                        <p class="">Is that known and
                                          normal?</p>
                                        <p class="">In all cases, I am
                                          using MKL and I suspect it 
                                          may come from there... :/</p>
                                        <p class="">I also saw a second
                                          problem, "make test" fails to
                                          compile petsc examples on
                                          older versions of MKL (but
                                          that's less important for me,
                                          I just upgraded to OneAPI to
                                          avoid this, but you may want
                                          to know):<br class="">
                                        </p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/dernier_ompi/2021.03.02.02h16m01s_make_test.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/dernier_ompi/2021.03.02.02h16m01s_make_test.log</a></p>
                                        <p class=""><a
                                            class="moz-txt-link-freetext"
href="https://giref.ulaval.ca/~cmpgiref/dernier_ompi/2021.03.02.02h16m01s_configure.log"
                                            moz-do-not-send="true">https://giref.ulaval.ca/~cmpgiref/dernier_ompi/2021.03.02.02h16m01s_configure.log</a><br
                                            class="">
                                        </p>
                                        <p class=""> Thanks,</p>
                                        <p class="">Eric<br class="">
                                        </p>
                                        <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>
                            </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>
                        </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>
            </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>