<div dir="ltr">On Sun, Dec 23, 2012 at 8:08 PM, Sanjay Govindjee <span dir="ltr"><<a href="mailto:s_g@berkeley.edu" target="_blank">s_g@berkeley.edu</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
  
    
  
  <div bgcolor="#FFFFFF" text="#000000">
    Not sure what you mean by where is your matrix?  I am simply running
    ex6 in the ksp/examples/tests directory.<br>
    <br>
    The reason I ran this test is because I was seeing the same behavior
    with my finite element code (on perfectly benign problems).<br>
    <br>
    Is there a built-in test that you use to check that superlu_dist is
    working properly with petsc?<br>
    i.e. something you know that works with with petsc 3.3-p5?<br></div></blockquote><div><br></div><div style>1) Run it on a SNES ex5 (or KSP ex2), which is a nice Laplacian</div><div style><br></div><div style>2) Compare with MUMPS</div>
<div style><br></div><div style>   Matt</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div bgcolor="#FFFFFF" text="#000000">
    -sanjay<div><div class="h5"><br>
    <br>
    <br>
    <div>On 12/23/12 4:56 PM, Jed Brown wrote:<br>
    </div>
    <blockquote type="cite">
      <div dir="ltr">Where is your matrix? It might be ending up with a
        very bad pivot. If the problem can be reproduced, it should be
        reported to the SuperLU_DIST developers to fix. (Note that we do
        not see this with other matrices.) You can also try MUMPS.</div>
      <div class="gmail_extra"><br>
        <br>
        <div class="gmail_quote">On Sun, Dec 23, 2012 at 6:48 PM, Sanjay
          Govindjee <span dir="ltr"><<a href="mailto:s_g@berkeley.edu" target="_blank">s_g@berkeley.edu</a>></span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
            <div bgcolor="#FFFFFF" text="#000000"> I wanted to use
              SuperLU Dist to perform a direct solve but seem to be
              encountering<br>
              a problem.  I was wonder if this is a know issue and if
              there is a solution for it.<br>
              <br>
              The problem is easily observed using ex6.c in
              src/ksp/ksp/examples/tests.<br>
              <br>
              Out of the box: make runex6 produces a residual error of
              O(1e-11), all is well.<br>
              <br>
              I then changed the run to run on two processors and add
              the flag<br>
              -pc_factor_mat_solver_package spooles  this produces a
              residual error of O(1e-11), all is still well.<br>
              <br>
              I then switch over to -pc_factor_mat_solver_package
              superlu_dist and the<br>
              residual error comes back as 22.6637!  Something seems
              very wrong.<br>
              <br>
              My build is perfectly vanilla:<br>
              <br>
              export PETSC_DIR=/Users/sg/petsc-3.3-p5/<br>
              export PETSC_ARCH=intel<br>
              <br>
              ./configure --with-cc=icc --with-fc=ifort  \
-download-{spooles,parmetis,superlu_dist,prometheus,mpich,ml,hypre,metis}<br>
              <br>
              make PETSC_DIR=/Users/sg/petsc-3.3-p5/ PETSC_ARCH=intel
              all<br>
              make PETSC_DIR=/Users/sg/petsc-3.3-p5/ PETSC_ARCH=intel
              test<span><font color="#888888"><br>
                  <br>
                  -sanjay<br>
                </font></span></div>
          </blockquote>
        </div>
        <br>
      </div>
    </blockquote>
    <br>
    </div></div><pre cols="72">-- 
-----------------------------------------------
Sanjay Govindjee, PhD, PE
Professor of Civil Engineering
Vice Chair for Academic Affairs

779 Davis Hall
Structural Engineering, Mechanics and Materials
Department of Civil Engineering
University of California
Berkeley, CA 94720-1710

Voice:  <a href="tel:%2B1%20510%20642%206060" value="+15106426060" target="_blank">+1 510 642 6060</a>
FAX:    <a href="tel:%2B1%20510%20643%205264" value="+15106435264" target="_blank">+1 510 643 5264</a>
<a href="mailto:s_g@berkeley.edu" target="_blank">s_g@berkeley.edu</a>
<a href="http://www.ce.berkeley.edu/~sanjay" target="_blank">http://www.ce.berkeley.edu/~sanjay</a>
-----------------------------------------------

New Books:  

Engineering Mechanics of Deformable 
Solids: A Presentation with Exercises
<a href="http://www.oup.com/us/catalog/general/subject/Physics/MaterialsScience/?view=usa&ci=9780199651641" target="_blank">http://www.oup.com/us/catalog/general/subject/Physics/MaterialsScience/?view=usa&ci=9780199651641</a>
<a href="http://ukcatalogue.oup.com/product/9780199651641.do" target="_blank">http://ukcatalogue.oup.com/product/9780199651641.do</a>
<a href="http://amzn.com/0199651647" target="_blank">http://amzn.com/0199651647</a>


Engineering Mechanics 3 (Dynamics)
<a href="http://www.springer.com/materials/mechanics/book/978-3-642-14018-1" target="_blank">http://www.springer.com/materials/mechanics/book/978-3-642-14018-1</a>
<a href="http://amzn.com/3642140181" target="_blank">http://amzn.com/3642140181</a>

-----------------------------------------------
</pre>
  </div>

</blockquote></div><br><br clear="all"><div><br></div>-- <br>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
-- Norbert Wiener
</div></div>