[petsc-dev] Parallel Gauss - Seidel

[Jed Brown]

On Sun, Jun 10, 2012 at 12:05 PM, Mark F. Adams <mark.adams at columbia.edu>wrote:

> As far as PETSc design, the current SOR is not really SOR, it is block
> Jacobi with SOR local solver.  So I think you are taking my name space!
>  THe current BJ/SOR is very practical so I don't like making  everyone
> change their command line args ... its up to you.


I would just add a bit to MatSORType
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20120610/e7f0c9fc/attachment.html>