If you are "throwing things" away in computing the Jacobian then any
expectation that Newton will converge is also thrown out the window.
If you used the PETSc SNES nonlinear solver you would have much more space to
experiment in. For example you could use -snes_mf_operator to apply the "
Hi,
For these problems, how large are they? And are they linear or nonlinear?
What I can do is use some fancier tools to help with what is going on with
the solvers in certain cases.
For Barry's question, the matrix in the SS solver is a diagonal matrix plus
a column scaling of the Jacobian.
On Thu, Oct 24, 2019 at 6:09 AM Thibaut Appel
wrote:
> Hi Matthew,
>
> Thanks for having a look, your example runs just like mine in Fortran.
>
> In serial, the value (0.0,0.0) was inserted whereas it shouldn't have.
>
I do not see that in serial.
> In parallel, you'll see that an error "Inserti
Hi Matthew,
Thanks for having a look, your example runs just like mine in Fortran.
In serial, the value (0.0,0.0) was inserted whereas it shouldn't have.
In parallel, you'll see that an error "Inserting a new nonzero at global
row/column" is triggered.
In both cases, MatSetValues tries to in
See bottom
> On Oct 14, 2019, at 1:12 PM, Justin Chang via petsc-users
> wrote:
>
> It might depend on your application, but for my stuff on maximum principles
> for advection-diffusion, I found RS to be much better than SS. Here’s the
> paper I wrote documenting the performance numbers I c
