That makes sense. Is there a reasonably easy way of doing that in PETSc currently for reasonably large systems? <div><br></div><div>Gaetan<br><br><div class="gmail_quote">On Mon, Apr 29, 2013 at 4:32 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>Gaetan Kenway <<a href="mailto:gaetank@gmail.com" target="_blank">gaetank@gmail.com</a>> writes:<br>
<br>
> I would be very interested in looking at the problematic eigenmodes as<br>
> well. On my end I use Tecplot for all my visualization. What sort of<br>
> visualization technique are you thinking about? Is it the KSP subspace<br>
> vectors you want to look at?<br>
<br>
</div>No, it would be an eigensolve for either outliers or eigenvalues very<br>
close to zero. The cost to find the vector is like several solves, but<br>
it would tell us exactly what sort of functions are poorly approximated<br>
by the preconditioner. Then we would think about how we can change<br>
algorithms to make the preconditioner correct those functions better.<br>
</blockquote></div><br></div>