<div class="gmail_quote">On Wed, Mar 16, 2011 at 14:27, Thomas Witkowski <span dir="ltr"><<a href="mailto:thomas.witkowski@tu-dresden.de" target="_blank">thomas.witkowski@tu-dresden.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div><div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
I'm going to guess that you still have an outer KSP that (in the global norm, rather than the partitioned norm used inside of splits) has a tighter tolerance, therefore it takes a few outer iterations. If you use loose inner tolerances then the preconditioner becomes nonlinear and you'll need to use FGMRES for the outer. When in doubt, run with -ksp_view and show us the results if you don't understand. Additionally, monitoring inner solves separately can be useful, e.g. -fieldsplit_0_ksp_converged_reason -fieldsplit_1_ksp_monitor -ksp_monitor_true_residual.<br>
</blockquote></div></div>
Okay, I run my code with the options "-pc_fieldsplit_type schur -fieldsplit_interior_ksp_type preonly -fieldsplit_interior_pc_type bjacobi -fieldsplit_interior_sub_pc_type lu -fieldsplit_boundary_ksp_monitor -ksp_monitor_true_residual -fieldsplit_interior_ksp_converged_reason". The splits are named "interior" and "boundary". The ksp output is as follows:<br>
<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br></blockquote><div><br></div><div>This is the initial interior solve.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Residual norms for fieldsplit_boundary_ solve.<br>
0 KSP Residual norm 1.790059331071e-04<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
1 KSP Residual norm 1.237356212928e-04<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
2 KSP Residual norm 7.952220245101e-05<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br></blockquote><div><br></div><div>[...]</div><div><br></div><div>Each iteration on the boundary problem requires an interior solve. We see alternating boundary residual and interior solves.</div>
<div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> <br>
[0] Petsc-Iteration 0: 0.574504<br></blockquote><div><br></div><div>You must be printing this somewhere.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
0 KSP preconditioned resid norm 5.745043818120e-01 true resid norm 1.584249437360e-01 ||r(i)||/||b|| 1.000000000000e+00<br></blockquote><div><br></div><div>This is the norm of P^{-1} b, upon which all following norms are computed. Note that this is not actually a "preconditioned residual" because the initial guess was zero, however. In fact, since P is an exact preconditioner for your problem, it is the solution of your problem, but GMRES can't check that until it does one more iteration (the output below). You can use -ksp_type richardson to cheaply confirm that the norm is small, it should show a small residual after one iteration. Once you are happy that the inner solve really is solving your system, you can switch to -ksp_type preonly.</div>
<div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Linear solve converged due to CONVERGED_ITS iterations 1<br>
Residual norms for fieldsplit_boundary_ solve.<br>
0 KSP Residual norm 1.790059331000e-04<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
1 KSP Residual norm 1.237356237744e-04<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
2 KSP Residual norm 7.952220078415e-05<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
3 KSP Residual norm 3.502859285561e-05<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
4 KSP Residual norm 1.601086485810e-05<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
5 KSP Residual norm 8.491169384185e-06<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
6 KSP Residual norm 4.778229157872e-06<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
7 KSP Residual norm 2.571917112191e-06<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
8 KSP Residual norm 1.353977280183e-06<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
9 KSP Residual norm 7.408352317360e-07<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
10 KSP Residual norm 3.810710583336e-07<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
11 KSP Residual norm 1.955898169394e-07<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
12 KSP Residual norm 9.938892788465e-08<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
13 KSP Residual norm 4.893428546840e-08<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
14 KSP Residual norm 2.395120857135e-08<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
15 KSP Residual norm <a href="tel:1.2464367927" target="_blank"></a><a href="tel:1.2464367927" target="_blank"></a><a href="tel:1.2464367927" target="_blank"></a><a href="tel:1.2464367927" target="_blank">1.2464367927</a>84e-08<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
16 KSP Residual norm 6.316409779118e-09<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
17 KSP Residual norm 3.088649355374e-09<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
18 KSP Residual norm 1.593449602302e-09<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br></blockquote><div><br></div><div>The way GMRES computes residuals explains why we needed this Schur solve, but it shouldn't do the one below. Are you using GMRES or some other Krylov solver (perhaps set in an options file or inside your code, -ksp_view would help)?</div>
<div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Linear solve converged due to CONVERGED_ITS iterations 1<br>
Residual norms for fieldsplit_boundary_ solve.<br>
0 KSP Residual norm 1.593449603155e-09<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
1 KSP Residual norm 1.011857676662e-09<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
2 KSP Residual norm 5.660892669788e-10<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
3 KSP Residual norm 3.046799343903e-10<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
4 KSP Residual norm <a href="tel:1.6004713993" target="_blank"></a><a href="tel:1.6004713993" target="_blank"></a><a href="tel:1.6004713993" target="_blank"></a><a href="tel:1.6004713993" target="_blank">1.6004713993</a>29e-10<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
5 KSP Residual norm 8.083216601016e-11<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
6 KSP Residual norm 4.308671648345e-11<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
7 KSP Residual norm 2.381959817139e-11<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
8 KSP Residual norm 1.264839790114e-11<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
9 KSP Residual norm 6.435098853282e-12<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
10 KSP Residual norm 3.184769579797e-12<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
11 KSP Residual norm <a href="tel:1.5022019430" target="_blank"></a><a href="tel:1.5022019430" target="_blank"></a><a href="tel:1.5022019430" target="_blank"></a><a href="tel:1.5022019430" target="_blank">1.5022019430</a>16e-12<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
12 KSP Residual norm 7.099486208425e-13<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
13 KSP Residual norm 3.443222176388e-13<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
14 KSP Residual norm <a href="tel:1.6478434989" target="_blank"></a><a href="tel:1.6478434989" target="_blank"></a><a href="tel:1.6478434989" target="_blank"></a><a href="tel:1.6478434989" target="_blank">1.6478434989</a>12e-13<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
15 KSP Residual norm 8.208508173926e-14<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
16 KSP Residual norm 4.376604355287e-14<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
17 KSP Residual norm 2.596069152755e-14<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
18 KSP Residual norm 1.335744434252e-14<br>
Linear solve converged due to CONVERGED_ITS iterations 1<br>
1 KSP preconditioned resid norm 2.608069232839e-13 true resid norm 1.449869994655e-13 ||r(i)||/||b|| 9.151778504469e-13<br>
<br>
I have no idea how to interpret this output! Could you help me with it?<br><font color="#888888">
<br>
Thomas<br>
</font></blockquote></div><br>