<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <br>
    <div class="moz-cite-prefix">On 09/23/2013 09:48 AM, Mark F. Adams
      wrote:<br>
    </div>
    <blockquote cite="mid:D6E1D3CE-5365-42AF-9668-F20085EEF6F6@lbl.gov"
      type="cite">
      <meta http-equiv="Content-Type" content="text/html;
        charset=ISO-8859-1">
      <br>
      <div>
        <div>On Sep 23, 2013, at 12:27 PM, Michele Rosso <<a
            moz-do-not-send="true" href="mailto:mrosso@uci.edu">mrosso@uci.edu</a>>
          wrote:</div>
        <br class="Apple-interchange-newline">
        <blockquote type="cite">
          <meta content="text/html; charset=ISO-8859-1"
            http-equiv="Content-Type">
          <div bgcolor="#FFFFFF" text="#000000"> The boundary conditions
            are periodic. <br>
            The equation I am solving is:<br>
            <br>
                   div(beta*grad(u))= f<br>
            <br>
            where beta is 1 inside the gas phase, 0.001 inside the
            liquid phase and a value in between for the nodes close to
            the interface.<br>
          </div>
        </blockquote>
        <div><br>
        </div>
        <div>This is a pretty big jump for geometric MG.  You might try
          AMG.  I suspect that the geometry is getting more complex as
          the simulation progresses.  Does the simulation start with
          both phases?  Also this problem is singular.  You might try
          projecting out the constant.  It could be that as the geometry
          gets more complex floating point errors are creeping in and
          you are getting an effective constant component to your RHS.</div>
        <br>
      </div>
    </blockquote>
    The simulation does start with both phases and the geometry is
    supposed to become more complex as the simulation progresses.<br>
    But so far the run is stopped before there are significant changes
    in the shape of the droplet. <br>
    I can give a shot to AMG: which options would you suggest to use.<br>
    Also, how can I project out the constant from the rhs? Thanks a lot!<br>
    <br>
    Michele<br>
    <blockquote cite="mid:D6E1D3CE-5365-42AF-9668-F20085EEF6F6@lbl.gov"
      type="cite">
      <div>
        <blockquote type="cite">
          <div bgcolor="#FFFFFF" text="#000000"> The system matrix is
            built so to remain symmetric positive defined despite the
            coefficients.<br>
            <br>
            Michele<br>
            <br>
            <br>
            <div class="moz-cite-prefix">On 09/23/2013 09:11 AM, Matthew
              Knepley wrote:<br>
            </div>
            <blockquote
cite="mid:CAMYG4G=iy3OLnFsFx33afmA5KmAyoj9aJBxuHvnh-mRrk7KUKA@mail.gmail.com"
              type="cite">
              <div dir="ltr">On Mon, Sep 23, 2013 at 8:55 AM, Michele
                Rosso <span dir="ltr"><<a moz-do-not-send="true"
                    href="mailto:mrosso@uci.edu" target="_blank">mrosso@uci.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"> Hi,<br>
                        <br>
                        <font face="Ubuntu">I am successfully using
                          PETSc to solve a 3D Poisson's equation with CG
                          + MG </font>.  Such equation arises from a
                        projection algorithm for a multiphase
                        incompressible flow simulation.<br>
                        I set up the solver  <font face="Ubuntu">as I
                          was suggested to do in a previous thread</font>
                        (title: "GAMG speed") and run a test case
                        (liquid droplet with surface tension falling
                        under the effect of gravity in a quiescent
                        fluid). <br>
                        The solution of the Poisson Equation via
                        multigrid is correct but it becomes
                        progressively slower and slower as the
                        simulation progresses (I am performing
                        successive solves) due to an increase in the
                        number of iterations.<br>
                        Since the solution of the Poisson equation is
                        mission-critical, I need to speed it up as much
                        as I can.<br>
                        Could you please help me out with this?<br>
                      </div>
                    </blockquote>
                    <div><br>
                    </div>
                    <div>First, what does the coefficient look like?</div>
                    <div><br>
                    </div>
                    <div>Second, what are the boundary conditions?</div>
                    <div><br>
                    </div>
                    <div>    Matt</div>
                    <div> </div>
                    <blockquote class="gmail_quote" style="margin:0 0 0
                      .8ex;border-left:1px #ccc solid;padding-left:1ex">
                      <div bgcolor="#FFFFFF" text="#000000"> I run the
                        test case with the following options: <br>
                        <br>
                        -pc_type mg  -pc_mg_galerkin  -pc_mg_levels 5  
                        -mg_levels_ksp_type richardson
                        -mg_levels_ksp_max_it 1 <br>
                        -mg_coarse_pc_type lu  
                        -mg_coarse_pc_factor_mat_solver_package
                        superlu_dist <br>
                        -log_summary -ksp_view 
                        -ksp_monitor_true_residual  -options_left <br>
                        <br>
                        Please find the diagnostic for the final solve
                        in the attached file "final.txt'. <br>
                        Thank you, <br>
                        <br>
                        Michele<br>
                      </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>
            </blockquote>
            <br>
          </div>
        </blockquote>
      </div>
      <br>
    </blockquote>
    <br>
  </body>
</html>