Do not pass -snes_mf or -snes_mf_operator with the -snes_type test option
Barry On Apr 25, 2012, at 4:16 PM, Jed Brown wrote: > On Wed, Apr 25, 2012 at 16:06, Dominik Szczerba <dominik at itis.ethz.ch> > wrote: > I am happy to be guilty if I know how I can improve: why do I get the > right results without this option, for several cases? > > How do you define "right results"? Is SNES converging quadratically when you > do not use -snes_mf_operator? > > > > Run with -snes_type test and see what it reports. > > The moment I specify this option I get: > > [0]PETSC ERROR: Invalid argument! > [0]PETSC ERROR: Cannot test with alternative preconditioner! > > tracing back to SNESSolve() > > You either (a) set different matrices for the operator and preconditioning > matrix or (b) passed -snes_mf. Just pass one matrix.
