On Mon, Sep 17, 2012 at 4:35 PM, Roy Stogner <[email protected]> wrote: > > On Thu, 13 Sep 2012, John Peterson wrote: > >> You get the wrong solution with boomeramg on Stokes or Navier-Stokes >> because the preconditioner is singular. >> >> Try passing -ksp_monitor_true_residual when you run that way, and >> you'll see that it's not really converging. >> >> http://permalink.gmane.org/gmane.comp.mathematics.libmesh.user/2187 > > > Thanks for the update. This doesn't mesh with what I thought I > remembered (correct solutions on symmetric problems, failure with > asymmetry) but indeed I'm seeing failures on plain Stokes now. > > I don't understand the reason for the failures, though. In that > thread Jed said that multigrid breaks down for indefinite systems, > which sounded reasonable. (inability to invert an indefinite matrix > is basically the null hypothesis) But Stokes stops being indefinite > if you handle the unconstrained pressure mode, yet both the pinned > pressure and the lagrange multiplier pressure formulations in our > examples still fail with BoomerAMG.
Not sure what happens to constraints with AMG. In geometric multigrid, anyway, I think the constrained DOF can get coarsened out in the V-cycle, and then you are indefinite again? -- John ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
