Linear firedrake_0_fieldsplit_3_ solve did not converge due to DIVERGED_DTOL iterations 35 The option "fieldsplit_3_pc_factor_shift_type": "nonzero" does not help. On 02/12/16 15:55, Lawrence Mitchell wrote:
On 2 Dec 2016, at 15:48, Anna Kalogirou <a.kalogirou@leeds.ac.uk> wrote:
Hi David,
Indeed, the issue is due to the 3rd block. The mathematical operator on the 3rd block is basically a mass operator so I don't understand why it is indefinite? What was printed by fieldsplit_3_ksp_converged_reason?
Plausibly it is just that the default PETSc LU does not pivoting (they changed this default a while ago). Does it help to say "fieldsplit_3_pc_factor_shift_type": "nonzero"?
Lawrence _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake