<div dir="ltr">What version of PETSc are you using? This has been changing recently.<div><br></div><div>-help will show you the parameters for each level, like -mg_levels_1_eigest_ksp_type <GMRES>. PETSc provides syntactic sugar to do all levels by removing the "_1".</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Jul 8, 2015 at 3:57 PM, Benoit Fabrèges <span dir="ltr"><<a href="mailto:benoit.fabreges@inria.fr" target="_blank">benoit.fabreges@inria.fr</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi,<br>
thank you for the answer. I tried to change the ksp to cg to estimate the eigenvalues like you propose but petsc doesn't find the option:<br>
<br>
mpirun -n 8 ./ex54 -ne 1023 -ksp_rtol 1e-10 -ksp_monitor_true_residual -options_left -mg_levels_esteig_ksp_type cg<br>
0 KSP preconditioned resid norm 2.489979017016e+03 true resid norm 9.779947596961e-04 ||r(i)||/||b|| 1.000000000000e+00<br>
1 KSP preconditioned resid norm 5.415647207494e+01 true resid norm 6.738962907374e-02 ||r(i)||/||b|| 6.890592041073e+01<br>
2 KSP preconditioned resid norm 1.646483778031e+01 true resid norm 4.789220846312e-03 ||r(i)||/||b|| 4.896980069505e+00<br>
3 KSP preconditioned resid norm 1.915371795016e+00 true resid norm 4.570471306345e-03 ||r(i)||/||b|| 4.673308584767e+00<br>
4 KSP preconditioned resid norm 1.234057146730e+00 true resid norm 5.812076545460e-03 ||r(i)||/||b|| 5.942850396526e+00<span class=""><br>
#PETSc Option Table entries:<br>
-ksp_monitor_true_residual<br>
-ksp_rtol 1e-10<br></span>
-mg_levels_esteig_ksp_type cg<span class=""><br>
-ne 1023<br>
-options_left<br>
#End of PETSc Option Table entries<br></span>
There is one unused database option. It is:<br>
Option left: name:-mg_levels_esteig_ksp_type value: cg<br>
<br>
I could not find it either by adding -help to get all the possible options.<br>
By the way, if I reverse the order of the option -options-left and -mg_levels_esteig_ksp_type cg, it doesn't print the left options at the end anymore.<br>
<br>
Is there something else to change apart from the option -mg_levels_esteig_ksp_type cg ?<br>
<br>
Thanks,<br>
Benoit<div><div class="h5"><br>
<br>
<br>
On Wed 08 Jul 2015 03:31:32 PM CEST, Mark Adams wrote:<br>
</div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5">
Sorry for the delay. The problem is that the eigen estimates are bad<br>
for the Chebyshev smoother. Unfortunately this fails catastrophically.<br>
<br>
First, I do not understand why we do not get an error message here:<br>
<br>
13:57 PICell> ~/Codes/petsc/src/ksp/ksp/examples/tutorials$ mpirun -n<br>
8 ./ex54 -ne 1023 -ksp_rtol 1e-10 -ksp_monitor_true_residual<br>
-options_left<br>
0 KSP preconditioned resid norm 2.408779787997e+03 true resid norm<br>
9.779947596963e-04 ||r(i)||/||b|| 1.000000000000e+00<br>
1 KSP preconditioned resid norm 6.605106317291e+01 true resid norm<br>
6.550622855055e-02 ||r(i)||/||b|| 6.698014268593e+01<br>
2 KSP preconditioned resid norm 2.105905914541e+01 true resid norm<br>
5.536477156386e-03 ||r(i)||/||b|| 5.661049920253e+00<br>
3 KSP preconditioned resid norm 2.775390521485e+00 true resid norm<br>
1.822171753713e-03 ||r(i)||/||b|| 1.863171285579e+00<br>
4 KSP preconditioned resid norm 6.367933585867e-01 true resid norm<br>
2.173992967507e-03 ||r(i)||/||b|| 2.222908605545e+00<br>
#PETSc Option Table entries:<br>
-ksp_monitor_true_residual<br>
-ksp_rtol 1e-10<br>
-ne 1023<br>
-options_left<br>
#End of PETSc Option Table entries<br>
There are no unused options.<br>
<br>
It just stops. What is going on here?????<br>
<br>
A problem is that GMRES is the default it is not accuare. CG is much<br>
better. I will add this to all tests (54, 55, 56) that use Cheby.<br>
Here is what we get with CG:<br>
<br>
14:12 PICell> ~/Codes/petsc/src/ksp/ksp/examples/tutorials$ mpirun -n<br>
8 ./ex54 -ne 1023 -ksp_rtol 1e-10 -ksp_monitor_true_residual<br>
-mg_levels_esteig_ksp_type cg<br>
0 KSP preconditioned resid norm 2.397141111771e+03 true resid norm<br>
9.779947596963e-04 ||r(i)||/||b|| 1.000000000000e+00<br>
1 KSP preconditioned resid norm 6.398367771703e+01 true resid norm<br>
5.969698581458e-02 ||r(i)||/||b|| 6.104018986064e+01<br>
2 KSP preconditioned resid norm 2.025487152015e+01 true resid norm<br>
4.795534114089e-03 ||r(i)||/||b|| 4.903435388118e+00<br>
3 KSP preconditioned resid norm 2.693141357521e+00 true resid norm<br>
1.392786125856e-03 ||r(i)||/||b|| 1.424124323824e+00<br>
4 KSP preconditioned resid norm 6.147511674674e-01 true resid norm<br>
2.801964778489e-04 ||r(i)||/||b|| 2.865010012281e-01<br>
5 KSP preconditioned resid norm 1.133515586579e-01 true resid norm<br>
3.760895077288e-05 ||r(i)||/||b|| 3.845516594031e-02<br>
6 KSP preconditioned resid norm 1.169295614319e-02 true resid norm<br>
5.672568123760e-06 ||r(i)||/||b|| 5.800202984239e-03<br>
7 KSP preconditioned resid norm 1.281609546752e-03 true resid norm<br>
8.103035520325e-07 ||r(i)||/||b|| 8.285356787434e-04<br>
8 KSP preconditioned resid norm 1.898415766718e-04 true resid norm<br>
1.250723051310e-07 ||r(i)||/||b|| 1.278864778067e-04<br>
9 KSP preconditioned resid norm 3.325790135426e-05 true resid norm<br>
2.304415060174e-08 ||r(i)||/||b|| 2.356265243067e-05<br>
10 KSP preconditioned resid norm 8.224698701048e-06 true resid norm<br>
4.786148212120e-09 ||r(i)||/||b|| 4.893838299917e-06<br>
11 KSP preconditioned resid norm 1.224414108364e-06 true resid norm<br>
7.170267325093e-10 ||r(i)||/||b|| 7.331600966164e-07<br>
12 KSP preconditioned resid norm 1.760480053719e-07 true resid norm<br>
1.005681798001e-10 ||r(i)||/||b|| 1.028310006807e-07<br>
<br>
So all is well.<br>
<br>
Thanks for pointing this out.<br>
Mark<br>
<br>
On Thu, Jul 2, 2015 at 5:38 PM, Benoit Fabrèges<br></div></div><span class="">
<<a href="mailto:benoit.fabreges@inria.fr" target="_blank">benoit.fabreges@inria.fr</a> <mailto:<a href="mailto:benoit.fabreges@inria.fr" target="_blank">benoit.fabreges@inria.fr</a>>> wrote:<br>
<br>
Dear PETSc developers,<br>
I updated from 3.4 to 3.6 and I have a strange behavior with the<br>
GAMG preconditioner. I am running ex54 with the following options:<br>
-ne 1023 -ksp_rtol 1e-10 -ksp_monitor_true_residual -options_left<br>
<br>
I don't have the same behavior depending on the number of process<br>
used. It is for example not converging with 8 and 12 process with<br>
the reason KSP_DIVERGED_INDEFINITE_PC but it is working fine in<br>
other cases (I didn't try with more than 12 process).<br>
<br>
Do you have the same behavior ? It is working fine with 3.4 version.<br>
<br>
Best,<br>
<br>
Benoit<br>
<br>
<br>
<br>
<br>
</span></blockquote>
<br>
<br>
</blockquote></div><br></div>