<div class="gmail_quote">On Mon, Dec 26, 2011 at 12:43, Mark F. Adams <span dir="ltr"><<a href="mailto:mark.adams@columbia.edu">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">
I've just started using -ksp_rtol 1.e-10 and retested all of these test and am seeing a very slight improvement going from 1 to 4. I would expect to see a significant drop, say 25%, in the number of iterations going from 1 to 2. The < 5% that I am seeing seems too low but I'm not 100% sure -- maybe there is nothing wrong ...</blockquote>
</div><br><div>Well, here is a different test, just for giggles. This sets up rapid coarsening so that multigrid convergence with weak smoothers isn't already "too good".</div><div><br></div><div><div>$ pwd</div>
<div>/home/jed/petsc/src/snes/examples/tutorials</div></div><div><div>$ for n in {1..4}; do echo Cheby $n; ./ex5 -da_grid_x 5 -da_grid_y 5 -da_refine 3 -da_refine_x 4 -da_refine_y 4 -snes_max_it 1 -ksp_monitor -pc_type mg -pc_mg_levels 4 -mg_levels_pc_type jacobi -mg_levels_ksp_type chebychev -mg_levels_ksp_chebychev_estimate_eigenvalues 0,0.1,0,1.1 -mg_levels_ksp_max_it $n; done</div>
<div>Cheby 1</div><div> 0 KSP Residual norm 1.621385151256e+01 </div><div> 1 KSP Residual norm 1.247824515555e+01 </div><div> 2 KSP Residual norm 1.278794763626e+00 </div><div> 3 KSP Residual norm 1.116456717457e-01 </div>
<div> 4 KSP Residual norm 4.348273343707e-02 </div><div> 5 KSP Residual norm 8.348548286503e-03 </div><div> 6 KSP Residual norm 3.951234403814e-03 </div><div> 7 KSP Residual norm 1.058691144871e-03 </div><div>
8 KSP Residual norm 3.264344380115e-04 </div><div> 9 KSP Residual norm 8.173798438913e-05 </div><div>Cheby 2</div><div> 0 KSP Residual norm 1.891099688325e+01 </div><div> 1 KSP Residual norm 9.980070235568e+00 </div>
<div> 2 KSP Residual norm 8.251956879061e-02 </div><div> 3 KSP Residual norm 8.973973844623e-03 </div><div> 4 KSP Residual norm 6.086189552261e-04 </div><div> 5 KSP Residual norm 8.406518096327e-05 </div><div>
Cheby 3</div><div> 0 KSP Residual norm 1.990124921919e+01 </div><div> 1 KSP Residual norm 1.111981517647e+01 </div><div> 2 KSP Residual norm 4.888111472342e-02 </div><div> 3 KSP Residual norm 2.109153682849e-03 </div>
<div> 4 KSP Residual norm 1.680361259338e-04 </div><div>Cheby 4</div><div> 0 KSP Residual norm 2.051550954711e+01 </div><div> 1 KSP Residual norm 1.049577156306e+01 </div><div> 2 KSP Residual norm 3.445878799134e-02 </div>
<div> 3 KSP Residual norm 9.302079492533e-04 </div><div> 4 KSP Residual norm 3.995254526577e-05 </div><div><br></div></div><div><br></div><div>I don't think the Cheby implementation is buggy and I think we are using it correctly. I agree that it is somewhat disappointing if increasing the order above 1 doesn't make a more noticeable improvement.</div>