What you're seeing is convergence in the step tolerance, not the relative tolerance applied to the norm of the residual. This means that the norm of the steplength between successive iterations has gone below snes->xtol times the norm of the solution. This appears to be set through -snes_stol.<br>
<br>- Peter<br><br><div class="gmail_quote">On Tue, Mar 20, 2012 at 10:29 AM, Mark F. Adams <span dir="ltr"><<a href="mailto:mark.adams@columbia.edu" target="_blank">mark.adams@columbia.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
We've been trying to drive the nonlinear residual down and SNES seems to return prematurely.<br>
<br>
Appended is Dan's note. We are asking for -snes_rtol 1.e-12 but only getting ~1.e-8.<br>
<br>
Anyone have any ideas?<br>
<br>
Mark<br>
<br>
> Nonlinear solve converged due to CONVERGED_PNORM_RELATIVE iterations 8<br>
<br>
However, here are the initial and final norms:<br>
<br>
> 0 SNES Function norm 1.999411334131e+08<br>
> 8 SNES Function norm 5.999267565004e-01<br>
<br>
2e8/6e-1 = 3.3e8. So, I'm confused. Am I misunderstanding the convergence criterion?<br>
<br>
here's the dump of petsc options at the end:<br>
<br>
> #PETSc Option Table entries:<br>
> -ksp_converged_use_initial_residual_norm<br>
> -ksp_max_it 100<br>
> -ksp_monitor<br>
> -ksp_rtol 1.e-6<br>
> -ksp_type gmres<br>
> -mg_levels_ksp_max_it 4<br>
> -options_left<br>
> -pc_gamg_agg_nsmooths 1<br>
> -pc_gamg_sym_graph<br>
> -pc_gamg_threshold .01<br>
> -pc_gamg_type agg<br>
> -pc_gamg_verbose 2<br>
> -pc_hypre_boomeramg_agg_nl 1<br>
> -pc_hypre_boomeramg_coarsen_type HMIS<br>
> -pc_hypre_boomeramg_interp_type ext+i<br>
> -pc_hypre_boomeramg_no CF<br>
> -pc_hypre_type boomeramg<br>
> -pc_type gamg<br>
> -snes_converged_reason<br>
> -snes_mf_operator<br>
> -snes_monitor<br>
> -snes_rtol 1.e-12<br>
> #End of PETSc Option Table entries<br>
<span><font color="#888888"><br>
<br>
<br>
Mark</font></span></blockquote></div><br>