<html><head><meta http-equiv="Content-Type" content="text/html charset=iso-8859-1"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Well this is probably my "bizarre coarse grid solver". I explicitly create the solver after reducing the processor to one proc. PETSc seems to make it a GMRES solver in PCSetUp_MG and I explicitly set it back to PREONLY but it looks like GMRES slipped in anyway.<div><br></div><div>Mark</div><div><br><div><div>On Sep 14, 2012, at 7:17 PM, Matthew Knepley <<a href="mailto:knepley@gmail.com">knepley@gmail.com</a>> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite">On Fri, Sep 14, 2012 at 6:14 PM, Jed Brown <span dir="ltr"><<a href="mailto:jedbrown@mcs.anl.gov" target="_blank">jedbrown@mcs.anl.gov</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Also, add -pc_gamg_agg_nsmooths 1 as Mark suggested. It will reduce the iteration count significantly at the expense of somewhat higher setup cost.</blockquote><div><br></div><div>Unless I am reading this wrong, you have a bizarre coarse grid solver. It should be redundant LU but instead</div>
<div>you have GMRES/BJacobi?</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 class="HOEnZb"><div class="h5"><br>
<div class="gmail_quote">On Fri, Sep 14, 2012 at 6:11 PM, Jed Brown <span dir="ltr"><<a href="mailto:jedbrown@mcs.anl.gov" target="_blank">jedbrown@mcs.anl.gov</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><font size="4" face="courier new, monospace" color="#cc0000"><b><br></b></font></div><div><font size="4" face="courier new, monospace" color="#cc0000"><b> ##########################################################</b></font></div>
<div><font size="4" face="courier new, monospace" color="#cc0000"><b> # #</b></font></div><div><font size="4" face="courier new, monospace" color="#cc0000"><b> # WARNING!!! #</b></font></div>
<div><font size="4" face="courier new, monospace" color="#cc0000"><b> # #</b></font></div><div><font size="4" face="courier new, monospace" color="#cc0000"><b> # This code was compiled with a debugging option, #</b></font></div>
<div><font size="4" face="courier new, monospace" color="#cc0000"><b> # To get timing results run ./configure #</b></font></div><div><font size="4" face="courier new, monospace" color="#cc0000"><b> # using --with-debugging=no, the performance will #</b></font></div>
<div><font size="4" face="courier new, monospace" color="#cc0000"><b> # be generally two or three times faster. #</b></font></div><div><font size="4" face="courier new, monospace" color="#cc0000"><b> # #</b></font></div>
<div><font size="4" face="courier new, monospace" color="#cc0000"><b> ##########################################################</b></font></div><div><br><div class="gmail_quote">On Fri, Sep 14, 2012 at 6:08 PM, Zhenglun (Alan) Wei <span dir="ltr"><<a href="mailto:zhenglun.wei@gmail.com" target="_blank">zhenglun.wei@gmail.com</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">
<div>I'm sorry about that. I attached the
output files here with ' -ksp_monitor -ksp_view -log_summary'.
They are named after the grid size and pc-type. <br>
<br>
cheers,<br>
Alan<div><br>
On 9/14/2012 5:51 PM, Jed Brown wrote:<br>
</div></div><div>
<blockquote type="cite">
<div class="gmail_quote">On Fri, Sep 14, 2012 at 5:49 PM, Matthew
Knepley <span dir="ltr"><<a href="mailto:knepley@gmail.com" target="_blank">knepley@gmail.com</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div>On Fri, Sep 14, 2012 at 5:40 PM, Zhenglun
(Alan) Wei <span dir="ltr"><<a href="mailto:zhenglun.wei@gmail.com" target="_blank">zhenglun.wei@gmail.com</a>></span>
wrote:<br>
</div>
<div class="gmail_quote">
<div>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF" text="#000000">
<div>Dear folks,<br>
I did some test with -pc_type gamg with
/src/ksp/ksp/example/tutorial/ex45.c. It is not as
good as default -pc_type when my mesh (Cartisian) is
100*50*50; while it is a little bit better than the
default one when the mesh is 200*100*100. Therefore,
I guess this type of pc is good for larger problem.
Is that ture? or is there any rule of thumb for this
type of preconditioner? BTW, I tested it with 8
processes. <br>
</div>
</div>
</blockquote>
<div><br>
</div>
</div>
<div>When asking questions about convergence, always always
ALWAYS send the output of -ksp_monitor -ksp_view. If</div>
<div>you don't, we are just guessing blindly.</div>
</div>
</blockquote>
<div><br>
</div>
<div>And -log_summary because this is about performance.</div>
</div>
</blockquote>
<br>
</div></div>
</blockquote></div><br>
</div></blockquote></div><br>
</div></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></div><br></div></body></html>