We can't say anything without at least seeing<br><br> a) The output of -log_summary<br><br> b) The output -ksp_monitor and -ksp_view for a few solves<br><br> Matt<br><br><div class="gmail_quote">On Tue, Sep 1, 2009 at 6:25 PM, xiaoyin ji <span dir="ltr"><<a href="mailto:sapphire.jxy@gmail.com">sapphire.jxy@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Hi,<br>
<br>
I'm embedding a PETSc solver inside a molecular dynamics program. I<br>
used MATMPIAIJ matrix format and set ksp solver with KSPBICG and<br>
preconditioer as PCBJACOBI. Now the problem is ksp solver is getting<br>
slower and slower for each MD step, at the very beginning it takes<br>
about 1sec, and after 5000 steps the solving part takes up to<br>
10sec.(increasing gradually) However, I've also have an identical<br>
program in Fortran version, which did exactly the same thing and did<br>
not have such problem for over 1000000 steps, th Fortran version<br>
always solve within 1sec. The MPI_Wtime() shows that ksp solver is<br>
exactly the slowing down part. Thanks for any hints.<br>
<br>
Best,<br>
Xiaoyin Ji<br>
<br>
Department of Material Science and Engineering<br>
North Carolina State University<br>
</blockquote></div><br><br clear="all"><br>-- <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>