On Oct 15, 2012, at 11:48 PM, Matthew Knepley <knepley at gmail.com> wrote:
> On Tue, Oct 16, 2012 at 3:24 PM, Barry Smith <bsmith at mcs.anl.gov> wrote: > > It is seriously misleading that TSType rosw uses SNES as the solver since > it is only using SNESKSP and the algorithm really is built around only linear > solves. Are you sure that using SNES is the right model for this family of > algorithms? Convince me. > > I realize one person among us thinks we should use SNES for both linear > and nonlinear solves, but he is wrong :-) > > Let me try a Venn Diagram: > _________________________________ > / _____________ \ > | / \ | > | Nonlinear problems | Linear problems | | > | \ _____________/ | > \_________________________________/ > > Also, there is no overhead using SNES, so I would say Do Not Multiply > Entities Beyond The Necessary. I am not concerned about overhead. I am concerned about things looking like they are doing one thing but that are actually doing something else. In this case, there is actually a nonlinear problem hanging around but the rosw algorithms avoid solving it, which is ok but I find the fact that it prints SNES is then misleading because given a nonlinear problem and SNES one would think it is actually solving a nonlinear problem with SNES, when it is not. > > Matt > > Matt > > > Barry > > > > > -- > What most experimenters take for granted before they begin their experiments > is infinitely more interesting than any results to which their experiments > lead. > -- Norbert Wiener