<div class="gmail_quote">On Wed, Apr 25, 2012 at 16:06, Dominik Szczerba <span dir="ltr"><<a href="mailto:dominik@itis.ethz.ch">dominik@itis.ethz.ch</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div id=":eg1">I am happy to be guilty if I know how I can improve: why do I get the<br>
right results without this option, for several cases?<br></div></blockquote><div><br></div><div>How do you define "right results"? Is SNES converging quadratically when you do not use -snes_mf_operator?</div><div>
</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div id=":eg1">
<div class="im"><br>
> Run with -snes_type test and see what it reports.<br>
<br>
</div>The moment I specify this option I get:<br>
<br>
[0]PETSC ERROR: Invalid argument!<br>
[0]PETSC ERROR: Cannot test with alternative preconditioner!<br>
<br>
tracing back to SNESSolve()</div></blockquote></div><br><div>You either (a) set different matrices for the operator and preconditioning matrix or (b) passed -snes_mf. Just pass one matrix.</div>