<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#ffffff" text="#000000">
    Thanks Matt, that was my guess actually.<br>
    <br>
    On 09.01.2012 17:46, Matthew Knepley wrote:
    <blockquote
cite="mid:CAMYG4GkWbMuE6hNWEEQOqtkZzjUdzmg28wzrpm4u7wp1hDTwzg@mail.gmail.com"
      type="cite">On Mon, Jan 9, 2012 at 10:37 AM, Alexander Grayver <span
        dir="ltr">&lt;<a moz-do-not-send="true"
          href="mailto:agrayver@gfz-potsdam.de">agrayver@gfz-potsdam.de</a>&gt;</span>
      wrote:<br>
      <div class="gmail_quote">
        <blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt
          0.8ex; border-left: 1px solid rgb(204, 204, 204);
          padding-left: 1ex;">
          <div bgcolor="#ffffff" text="#000000"> I see this is due to
            preconditioner, but could you explain shortly what
            particularly causes this problem? It's not obvious to me.<br>
          </div>
        </blockquote>
        <div><br>
        </div>
        <div>The size of the block being factorized by ILU(0) decreases
          as p increases. Thus, it is</div>
        <div>a weaker preconditioner. I recommend Yousef Saad's book
          which explains all this clearly.</div>
        <div><br>
        </div>
        <div>&nbsp; &nbsp;Matt</div>
        <div>&nbsp;</div>
        <blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt
          0.8ex; border-left: 1px solid rgb(204, 204, 204);
          padding-left: 1ex;">
          <div bgcolor="#ffffff" text="#000000"> Thanks.<br>
            <br>
            On 09.01.2012 17:36, Matthew Knepley wrote:
            <blockquote type="cite">On Mon, Jan 9, 2012 at 10:31 AM,
              Alexander Grayver <span dir="ltr">&lt;<a
                  moz-do-not-send="true"
                  href="mailto:agrayver@gfz-potsdam.de" target="_blank">agrayver@gfz-potsdam.de</a>&gt;</span>
              wrote:<br>
              <div class="gmail_quote">
                <blockquote class="gmail_quote" style="margin: 0pt 0pt
                  0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204);
                  padding-left: 1ex;"> I have tested default
                  GMRES+ILU(0) solver for problem of size ~200000 and I
                  observed that both convergence rate and accuracy
                  degrade when I increase number of processes.<br>
                  E.g., number of iterations almost doubles when going
                  from 1 to 4 processors.<br>
                  <br>
                  I noticed that ILU(0) is somehow represented by
                  sequential matrices internally and feel that this
                  might be a reason?<br>
                </blockquote>
                <div><br>
                </div>
                <div>The default is Block Jacobi+ILU(0) and GMRES(30)
                  which will definitely degrade as the number of</div>
                <div>processes is increased. You would need to use an
                  optimal preconditioner like Multigrid if you want</div>
                <div>a constant number of iterates.</div>
                <div><br>
                </div>
                <div>&nbsp; &nbsp;Matt</div>
                <div>&nbsp;</div>
                <blockquote class="gmail_quote" style="margin: 0pt 0pt
                  0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204);
                  padding-left: 1ex;"> Regards,<br>
                  Alexander<br>
                </blockquote>
              </div>
              <br>
              <br clear="all">
              <span class="HOEnZb"><font color="#888888">
                  <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<br>
                </font></span></blockquote>
            <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<br>
    </blockquote>
    <br>
  </body>
</html>