<div class="gmail_quote">On Fri, Dec 23, 2011 at 04:11, Alexander Grayver <span dir="ltr"><<a href="mailto:agrayver@gfz-potsdam.de">agrayver@gfz-potsdam.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
I would also be interested more to know how it scales system? I mean
what is idea behind this scaling.</blockquote></div><div><br></div><a href="http://www.mcs.anl.gov/petsc/petsc-dev/docs/manualpages/KSP/KSPSetDiagonalScale.html">http://www.mcs.anl.gov/petsc/petsc-dev/docs/manualpages/KSP/KSPSetDiagonalScale.html</a><div>
<br></div><div>It takes the inverse square root of the diagonal and does symmetric scaling. If you used this with no additional preconditioner, it would be like preconditioning with Jacobi. The point is to correct some ill-scaling *before* sending the matrix to an algebraic preconditioner.</div>