[petsc-users] "snes_type test" is gone?

[Smith, Barry F.]

   I think you need to write your own tester, possibly by taking code from the previous release of PETSc where we supported -snes_type test.

    Barry

  Note that the new code does a much more comprehensive testing of the Jacobian.


> On Jun 29, 2018, at 5:06 PM, Alexander Lindsay <alexlindsay239 at gmail.com> wrote:
> 
> Is there a way to mimic the behavior of `-snes_type test` with new PETSc? E.g. don't attempt to perform any solves? They're stupid but we have a bunch of tests in MOOSE that are set-up only to test the Jacobian, and any attempts to actually solve the system are disastrous. I can hack around this by doing things like setting the convergence reason to `SNES_CONVERGED_ITS` and similarly for ksp, making sure I use `-pc_type gamg` since there are missing Jacobian diagonal entries, etc. but this doesn't seem ideal.
> 
> Alex
> 
> On Mon, Jun 4, 2018 at 4:26 PM, Kong, Fande <fande.kong at inl.gov> wrote:
> Thanks, Hong,
> 
> I see. It is better if "-snes_type test" did not exist at the first place.  
> 
> 
> Fande,
> 
> On Mon, Jun 4, 2018 at 4:01 PM, Zhang, Hong <hongzhang at anl.gov> wrote:
> 
> 
> > On Jun 4, 2018, at 4:59 PM, Zhang, Hong <hongzhang at anl.gov> wrote:
> > 
> > -snes_type has been removed. We can just use -snes_test_jacobian instead. Note that the test is done every time the Jacobian is computed.
> 
> It was meant to be "-snes_type test".
> 
> > Hong (Mr.)
> > 
> >> On Jun 4, 2018, at 3:27 PM, Kong, Fande <fande.kong at inl.gov> wrote:
> >> 
> >> Hi PETSc Team,
> >> 
> >> I was wondering if "snes_type test" has been gone? Quite a few MOOSE users use this option to test their Jacobian matrices. 
> >> 
> >> If it is gone, any reason?
> >> 
> >> Fande,
> > 
> 
> 
>