[petsc-dev] Multigrid is confusing

[Mark F. Adams]

I like to compartmetnalize: deal with the math and computer science separately.

And you (or someone) decided to only do one comm for SSOR.  You could, and maybe should, communicate between SOR steps.

On May 25, 2012, at 9:33 AM, Jed Brown wrote:

> Fair point for large subdomains, but two additive have two comm steps instead of one.
> 
> On May 25, 2012 8:24 AM, "Mark F. Adams" <mark.adams at columbia.edu> wrote:
> 
> On May 25, 2012, at 8:59 AM, Jed Brown wrote:
> 
>> On Fri, May 25, 2012 at 1:07 AM, Mark F. Adams <mark.adams at columbia.edu> wrote:
>> And, I've never seen Gauss-Siedel used with Cheby because G-S has the correct damping properties, as is, for the Laplacian.
>> 
>> The point is to have something adequate for things that are not Laplacians. I tried running SOR without Cheby, but it was far less robust.
>> 
>> So I know it looks funny, but I don't have a similarly robust alternative. If we are living in a world where local work is cheap, we might as well do local SOR instead of pbjacobi. (Note that Cheby+pbjacobi is nearly as good as Cheby+SOR in some cases, but much worse in others.)
>> 
>> 
>> Note, G-S is not symmetric and Cheby for unsymmetric is a different can of worms.  So if A is symmetric then maybe try SSOR.
>> 
>> The default SOR is local_symmetric.
> 
> So it does a forward and backward pass.  So it is really two smoothing steps.  One should compare it with two additive smoothers.
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20120525/561cfd2a/attachment.html>