[petsc-dev] defaulting to right preconditioning with GMRES?

Barry Smith bsmith at mcs.anl.gov
Thu Jul 10 22:43:39 CDT 2014


   Should we change GMRES to default to right preconditioning? The problem with left is if you get a funky preconditioning you may see “good” convergence in the preconditioned residual norm while the true residual norm goes no where. 

   Of course this will change the default behavior of a huge number of examples and user code.

   Barry




More information about the petsc-dev mailing list