<div class="gmail_quote">On Wed, Jun 22, 2011 at 17:32, Qian Zhu <span dir="ltr"><<a href="mailto:qzhu@mcs.anl.gov">qzhu@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;">
Thanks for the reply, Jed. I worked on ex27 so I know setting certain values for the input parameters can make the case converge/non-converge...I am just wondering whether this is also true with ex5 and ex20 (We picked them for other interesting characteristics).</blockquote>
<div><br></div><div>Well, ex5 is a really simple problem which converges pretty easily when lambda < 6.81 (the bifurcation point). I don't know how hard ex20 is, but you could certainly increase the power beta.</div>
<div><br></div><div>What characteristics are you looking for? Asking for a solver that doesn't converge just seems like an odd request considering that we spend lots of time figuring out how to make them converge robustly.</div>
<div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"> Which examples are solving the Stokers problem?</blockquote></div><br><div>src/ksp/ksp/examples/tutorials/ex43.c</div>
<div><br></div><div>This will happily not converge:</div><div><br></div><div>$ ./ex43 -mx 40 -my 40 -c_str 2 -sinker_eta0 1e-8 -sinker_eta1 1 -stokes_ksp_monitor_true_residual -stokes_ksp_type fgmres</div><div><br></div><div>
Run this to see some configurations that do work:</div><div><br></div><div>$ make -n runex43 runex43_2</div><div><br></div><div><br></div><div>The discretization above is not particularly robust, but is relatively easy to solve. This one is harder [1]</div>
<div><br></div><div>src/ksp/ksp/examples/tests/ex11.c</div><div><br></div><div>This will pretend to converge, but the true residuals are nonsense.</div><div><br></div><div>./ex11 -f $DATAFILESPATH/matrices/underworld32.gz -fc_ksp_monitor_true_residual</div>
<div><br></div><div>If you add -fc_ksp_type fgmres to the above, it will stop pretending (and just not converge). To see some solver configurations that work, use</div><div><br></div><div>$ make -n runex11 runex11_2</div>
<div><br></div><div><br></div><div>[1] Follow the directions here to get the test matrix: <a href="http://www.mcs.anl.gov/petsc/petsc-2/documentation/faq.html#datafiles">http://www.mcs.anl.gov/petsc/petsc-2/documentation/faq.html#datafiles</a></div>