<div dir="ltr"><div dir="ltr">On Thu, Oct 8, 2020 at 3:43 AM Yang Juntao <<a href="mailto:Y.Juntao@hotmail.com">Y.Juntao@hotmail.com</a>> wrote:<br></div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div lang="EN-US">
<div class="gmail-m_-7371589878672563102WordSection1">
<p class="MsoNormal">Hello, <u></u><u></u></p>
<p class="MsoNormal"><u></u> <u></u></p>
<p class="MsoNormal">I’m working on a nonlinear solver with SNES with handcoded jacobian and function. Each linear solver is solved with KSP solver.
<u></u><u></u></p>
<p class="MsoNormal">But sometimes I got issues with ksp solver convergence. I tried with finite difference approximated jacobian, but get the same error.
<u></u><u></u></p>
<p class="MsoNormal"><u></u> <u></u></p>
<p class="MsoNormal">From the iterations, the convergence seems ok at the beginning but suddenly diverged in the last iteration.<u></u><u></u></p>
<p class="MsoNormal">Hope anyone with experience on ksp solvers could direct me to a direction I can debug the problem.</p></div></div></blockquote><div><br></div><div>KSP Object: 1 MPI processes<br> type: gmres<br> restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement<br><br>The GMRES restarted at iteration 30. You can increase the subspace size using</div><div><br></div><div> -ksp_gmres_restart 100</div><div><br></div><div> Thanks,</div><div><br></div><div> Matt</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div lang="EN-US"><div class="gmail-m_-7371589878672563102WordSection1">
<p class="MsoNormal">iter = 0, SNES Function norm 2.94934e-06<u></u><u></u></p>
<p class="MsoNormal">iteration 0 KSP Residual norm 1.094600281831e-06<u></u><u></u></p>
<p class="MsoNormal">iteration 1 KSP Residual norm 1.264284474186e-08<u></u><u></u></p>
<p class="MsoNormal">iteration 2 KSP Residual norm 6.593269221816e-09<u></u><u></u></p>
<p class="MsoNormal">iteration 3 KSP Residual norm 1.689570779457e-09<u></u><u></u></p>
<p class="MsoNormal">iteration 4 KSP Residual norm 1.040661505932e-09<u></u><u></u></p>
<p class="MsoNormal">iteration 5 KSP Residual norm 5.422761817348e-10<u></u><u></u></p>
<p class="MsoNormal">iteration 6 KSP Residual norm 2.492867371369e-10<u></u><u></u></p>
<p class="MsoNormal">iteration 7 KSP Residual norm 8.261522376775e-11<u></u><u></u></p>
<p class="MsoNormal">iteration 8 KSP Residual norm 4.246401544245e-11<u></u><u></u></p>
<p class="MsoNormal">iteration 9 KSP Residual norm 2.514366787388e-11<u></u><u></u></p>
<p class="MsoNormal">iteration 10 KSP Residual norm 1.982940267051e-11<u></u><u></u></p>
<p class="MsoNormal">iteration 11 KSP Residual norm 1.586470414676e-11<u></u><u></u></p>
<p class="MsoNormal">iteration 12 KSP Residual norm 9.866392216207e-12<u></u><u></u></p>
<p class="MsoNormal">iteration 13 KSP Residual norm 4.951342176999e-12<u></u><u></u></p>
<p class="MsoNormal">iteration 14 KSP Residual norm 2.418292660318e-12<u></u><u></u></p>
<p class="MsoNormal">iteration 15 KSP Residual norm 1.747418526086e-12<u></u><u></u></p>
<p class="MsoNormal">iteration 16 KSP Residual norm 1.094150535809e-12<u></u><u></u></p>
<p class="MsoNormal">iteration 17 KSP Residual norm 4.464287492066e-13<u></u><u></u></p>
<p class="MsoNormal">iteration 18 KSP Residual norm 3.530090494462e-13<u></u><u></u></p>
<p class="MsoNormal">iteration 19 KSP Residual norm 2.825698091454e-13<u></u><u></u></p>
<p class="MsoNormal">iteration 20 KSP Residual norm 1.950568425807e-13<u></u><u></u></p>
<p class="MsoNormal">iteration 21 KSP Residual norm 1.227898091813e-13<u></u><u></u></p>
<p class="MsoNormal">iteration 22 KSP Residual norm 5.411106347374e-14<u></u><u></u></p>
<p class="MsoNormal">iteration 23 KSP Residual norm 4.511115848564e-14<u></u><u></u></p>
<p class="MsoNormal">iteration 24 KSP Residual norm 4.063546606691e-14<u></u><u></u></p>
<p class="MsoNormal">iteration 25 KSP Residual norm 3.677694771949e-14<u></u><u></u></p>
<p class="MsoNormal">iteration 26 KSP Residual norm 3.459244943466e-14<u></u><u></u></p>
<p class="MsoNormal">iteration 27 KSP Residual norm 3.263954971093e-14<u></u><u></u></p>
<p class="MsoNormal">iteration 28 KSP Residual norm 3.087344619079e-14<u></u><u></u></p>
<p class="MsoNormal">iteration 29 KSP Residual norm 2.809426925625e-14<u></u><u></u></p>
<p class="MsoNormal">iteration 30 KSP Residual norm 4.366149884754e-01<u></u><u></u></p>
<p class="MsoNormal"> Linear solve did not converge due to DIVERGED_DTOL iterations 30<u></u><u></u></p>
<p class="MsoNormal"><b><u></u> <u></u></b></p>
<div style="border-top:none;border-right:none;border-left:none;border-bottom:1pt solid windowtext;padding:0in 0in 1pt">
<p class="MsoNormal" style="border:none;padding:0in"><u></u> <u></u></p>
</div>
<p class="MsoNormal">SNES Object: 1 MPI processes<u></u><u></u></p>
<p class="MsoNormal"> type: newtonls<u></u><u></u></p>
<p class="MsoNormal"> SNES has not been set up so information may be incomplete<u></u><u></u></p>
<p class="MsoNormal"> maximum iterations=50, maximum function evaluations=10000<u></u><u></u></p>
<p class="MsoNormal"> tolerances: relative=1e-08, absolute=1e-50, solution=1e-08<u></u><u></u></p>
<p class="MsoNormal"> total number of linear solver iterations=0<u></u><u></u></p>
<p class="MsoNormal"> total number of function evaluations=0<u></u><u></u></p>
<p class="MsoNormal"> norm schedule ALWAYS<u></u><u></u></p>
<p class="MsoNormal"> SNESLineSearch Object: 1 MPI processes<u></u><u></u></p>
<p class="MsoNormal"> type: bt<u></u><u></u></p>
<p class="MsoNormal"> interpolation: cubic<u></u><u></u></p>
<p class="MsoNormal"> alpha=1.000000e-04<u></u><u></u></p>
<p class="MsoNormal"> maxstep=1.000000e+08, minlambda=1.000000e-12<u></u><u></u></p>
<p class="MsoNormal"> tolerances: relative=1.000000e-08, absolute=1.000000e-15, lambda=1.000000e-08<u></u><u></u></p>
<p class="MsoNormal"> maximum iterations=40<u></u><u></u></p>
<p class="MsoNormal"> KSP Object: 1 MPI processes<u></u><u></u></p>
<p class="MsoNormal"> type: gmres<u></u><u></u></p>
<p class="MsoNormal"> restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement<u></u><u></u></p>
<p class="MsoNormal"> happy breakdown tolerance 1e-30<u></u><u></u></p>
<p class="MsoNormal"> maximum iterations=10000, initial guess is zero<u></u><u></u></p>
<p class="MsoNormal"> tolerances: relative=1e-08, absolute=1e-50, divergence=10000.<u></u><u></u></p>
<p class="MsoNormal"> left preconditioning<u></u><u></u></p>
<p class="MsoNormal"> using DEFAULT norm type for convergence test<u></u><u></u></p>
<p class="MsoNormal"> PC Object: 1 MPI processes<u></u><u></u></p>
<p class="MsoNormal"> type: fieldsplit<u></u><u></u></p>
<p class="MsoNormal"> PC has not been set up so information may be incomplete<u></u><u></u></p>
<p class="MsoNormal"> FieldSplit with Schur preconditioner, factorization FULL<u></u><u></u></p>
<p class="MsoNormal"> Preconditioner for the Schur complement formed from S itself<u></u><u></u></p>
<p class="MsoNormal"> Split info:<u></u><u></u></p>
<p class="MsoNormal"> KSP solver for A00 block<u></u><u></u></p>
<p class="MsoNormal"> not yet available<u></u><u></u></p>
<p class="MsoNormal"> KSP solver for S = A11 - A10 inv(A00) A01<u></u><u></u></p>
<p class="MsoNormal"> not yet available<u></u><u></u></p>
<p class="MsoNormal"> linear system matrix = precond matrix:<u></u><u></u></p>
<p class="MsoNormal"> Mat Object: 1 MPI processes<u></u><u></u></p>
<p class="MsoNormal"> type: seqaij<u></u><u></u></p>
<p class="MsoNormal"> rows=659, cols=659<u></u><u></u></p>
<p class="MsoNormal"> total: nonzeros=659, allocated nonzeros=7908<u></u><u></u></p>
<p class="MsoNormal"> total number of mallocs used during MatSetValues calls=0<u></u><u></u></p>
<p class="MsoNormal"> not using I-node routines<u></u><u></u></p>
<p class="MsoNormal"><u></u> <u></u></p>
<p class="MsoNormal">Regards<u></u><u></u></p>
<p class="MsoNormal">Juntao<u></u><u></u></p>
</div>
</div>
</blockquote></div><br clear="all"><div><br></div>-- <br><div dir="ltr" class="gmail_signature"><div dir="ltr"><div><div dir="ltr"><div><div dir="ltr"><div>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>-- Norbert Wiener</div><div><br></div><div><a href="http://www.cse.buffalo.edu/~knepley/" target="_blank">https://www.cse.buffalo.edu/~knepley/</a><br></div></div></div></div></div></div></div></div>