[petsc-dev] ksp_error_if_not_converged in multilevel solvers

Mark Adams mfadams at lbl.gov
Sun Oct 20 08:43:00 CDT 2019

> If one just wants to run a fixed number of iterations, not checking for
> convergence, why would one set ksp->errorifnotconverged to true?
Good question. I can see not worrying too much about convergence on the
coarse grids, but to not allow it ... and now that I think about it, it
seems like we might want to error out with in indefinite PC. Maybe make
ksp->errorifnotconverged an int <1>:
0: no error
1: error if indefinite only
2: error if any error

> Thanks,
> Pierre
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20191020/ba877b8f/attachment-0001.html>

More information about the petsc-dev mailing list