<div class="gmail_quote">On Wed, Feb 16, 2011 at 16:18, Juha Jäykkä <span dir="ltr"><<a href="mailto:juhaj@iki.fi">juhaj@iki.fi</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div id=":4z6">So the "solution" in the KSP should actually be identically zero for a<br>
converged result?<br></div></blockquote><div><br></div><div>Yes</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div id=":4z6">
<div class="im"><br>
> defect. If the initial guess is zero, then it would normally pick up your<br>
> Dirichlet boundary conditions on the first iteration and all subsequent<br>
> solves would have zero in those locations.<br>
<br>
</div>It is not zero initially. But, on the other hand, it has zeros at both ends<br>
even at the very first iteration. If I understood your reply correctly, this<br>
would be expected for an initial guess which has correct values at the<br>
boundaries and it should only pick them up on the first iteration if they were<br>
not correct to begin with.<br>
<br>
Having ruled out a possibility of a bug in KSP, I need to continue my hunt for<br>
DIVERGED_LINEAR_SOLVE... None of the convergence tolerances seem to make any<br>
difference, it always diverges. The funny thing is, it diverges even if I<br>
start with an *exact* *solution*...</div></blockquote></div><br><div>This is a problem. Run with -ksp_converged_reason to find out why it's diverging.</div>