I used -ksp_type cg -pc_type icc -pc_factor_shift_type positive_definite, but this takes ~300 iterations.
> On May 2, 2018, at 12:52 PM, Roy Stogner <royst...@ices.utexas.edu> wrote: > > > On Tue, 1 May 2018, Lee, Jae Ho wrote: > >> Actually, this issue is now resolved. The problem was that the >> default linear solver was not converging. This issue can be resolved >> by using a stronger linear solver. > > Thanks for the update! > > You had me worried at first - I don't think any of the most active > developers are currently doing anything with subdivision elements, > which is exactly the sort of situation where regressions are a risk. > Good to hear that wasn't the case. > > How far in the "stronger linear solver" direction did you have to go? > Stronger tolerance(s); stronger preconditioner; direct instead of > iterative? > --- > Roy ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users