<div dir="ltr"><div class="gmail_default" style="font-size:small">Many thanks for the suggestions, Matt.</div><div class="gmail_default" style="font-size:small"><br></div><div class="gmail_default" style="font-size:small">I tried putting the solvers in a loop, like this:</div><div class="gmail_default" style="font-size:small"><br></div><div class="gmail_default" style="font-size:small">do {</div><div class="gmail_default" style="font-size:small"> NewtonLS</div><div class="gmail_default" style="font-size:small"> check convergence</div><div class="gmail_default" style="font-size:small"> if (converged) break</div><div class="gmail_default" style="font-size:small"> NRichardson or NGMRES</div><div class="gmail_default" style="font-size:small">} while (!converged)</div><div class="gmail_default" style="font-size:small"><br></div><div class="gmail_default" style="font-size:small">The results were interesting, to me at least. With NRichardson, there was indeed improvement in the residual norm, followed by improvement with NewtonLS, and so on for a few iterations of this loop. In each case, after a few iterations the NewtonLS appeared to be stuck in the same way as after the first iteration. Eventually neither method was able to reduce the residual norm, which was still significant, so this was not a total success. With NGMRES, the initial behavior was similar, but eventually the NGMRES progress became erratic. The minimal residual norm was a bit better using NGMRES than NRichardson, but neither combination of methods fully converged. For both NRichardson and NGMRES, I simply used the defaults, as I have no knowledge of how to tune the options for my problem.</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Oct 10, 2017 at 4:08 PM, Matthew Knepley <span dir="ltr"><<a href="mailto:knepley@gmail.com" target="_blank">knepley@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote"><span class="">On Tue, Oct 10, 2017 at 12:08 PM, zakaryah . <span dir="ltr"><<a href="mailto:zakaryah@gmail.com" target="_blank">zakaryah@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div style="font-size:small">Thanks for clearing that up.</div><div style="font-size:small"><br></div><div style="font-size:small">I'd appreciate any further help. Here's a summary:</div><div style="font-size:small"><br></div><div style="font-size:small">My ultimate goal is to find a vector field which minimizes an action. The action is a (nonlinear) function of the field and its first spatial derivatives.</div><div style="font-size:small"><br></div><div style="font-size:small">My current approach is to derive the (continuous) Euler-Lagrange equations, which results in a nonlinear PDE that the minimizing field must satisfy. These Euler-Lagrange equations are then discretized, and I'm trying to use an SNES to solve them.</div><div style="font-size:small"><br></div><div style="font-size:small">The problem is that the solver seems to reach a point at which the Jacobian (this corresponds to the second variation of the action, which is like a Hessian of the energy) becomes nearly singular, but where the residual (RHS of PDE) is not close to zero. The residual does not decrease over additional SNES iterations, and the line search results in tiny step sizes. My interpretation is that this point of stagnation is a critical point.</div></div></blockquote><div><br></div></span><div>The normal thing to do here (I think) is to engage solvers which do not depend on that particular point. So using</div><div>NRichardson, or maybe NGMRES, to get past that. I would be interested to see if this is successful.</div><div><br></div><div> Matt</div><div><div class="h5"><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div style="font-size:small">I have checked the hand-coded Jacobian very carefully and I am confident that it is correct.</div><div style="font-size:small"><br></div><div style="font-size:small">I am guessing that such a situation is well-known in the field, but I don't know the lingo or literature. If anyone has suggestions I'd be thrilled. Are there documentation/methodologies within PETSc for this type of situation?</div><div style="font-size:small"><br></div><div style="font-size:small">Is there any advantage to discretizing the action itself and using the optimization routines? With minor modifications I'll have the gradient and Hessian calculations coded. Are the optimization routines likely to stagnate in the same way as the nonlinear solver, or can they take advantage of the structure of the problem to overcome this?</div><div style="font-size:small"><br></div><div style="font-size:small">Thanks a lot in advance for any help.</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Sun, Oct 8, 2017 at 5:57 AM, Barry Smith <span dir="ltr"><<a href="mailto:bsmith@mcs.anl.gov" target="_blank">bsmith@mcs.anl.gov</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
There is apparently confusing in understanding the ordering. Is this all on one process that you get funny results? Are you using MatSetValuesStencil() to provide the matrix (it is generally easier than providing it yourself). In parallel MatView() always maps the rows and columns to the natural ordering before printing, if you use a matrix created from the DMDA. If you create the matrix yourself it has a different MatView in parallel that is in in thePETSc ordering.\<br>
<span class="m_-1393587641323955827m_4104752945513686728HOEnZb"><font color="#888888"><br>
<br>
Barry<br>
</font></span><div class="m_-1393587641323955827m_4104752945513686728HOEnZb"><div class="m_-1393587641323955827m_4104752945513686728h5"><br>
<br>
<br>
> On Oct 8, 2017, at 8:05 AM, zakaryah . <<a href="mailto:zakaryah@gmail.com" target="_blank">zakaryah@gmail.com</a>> wrote:<br>
><br>
> I'm more confused than ever. I don't understand the output of -snes_type test -snes_test_display.<br>
><br>
> For the user-defined state of the vector (where I'd like to test the Jacobian), the finite difference Jacobian at row 0 evaluates as:<br>
><br>
> row 0: (0, 10768.6) (1, 2715.33) (2, -1422.41) (3, -6121.71) (4, 287.797) (5, 744.695) (9, -1454.66) (10, 6.08793) (11, 148.172) (12, 13.1089) (13, -36.5783) (14, -9.99399) (27, -3423.49) (28, -2175.34) (29, 548.662) (30, 145.753) (31, 17.6603) (32, -15.1079) (36, 76.8575) (37, 16.325) (38, 4.83918)<br>
><br>
> But the hand-coded Jacobian at row 0 evaluates as:<br>
><br>
> row 0: (0, 10768.6) (1, 2715.33) (2, -1422.41) (3, -6121.71) (4, 287.797) (5, 744.695) (33, -1454.66) (34, 6.08792) (35, 148.172) (36, 13.1089) (37, -36.5783) (38, -9.99399) (231, -3423.49) (232, -2175.34) (233, 548.662) (234, 145.753) (235, 17.6603) (236, -15.1079) (264, 76.8575) (265, 16.325) (266, 4.83917) (267, 0.) (268, 0.) (269, 0.)<br>
> and the difference between the Jacobians at row 0 evaluates as:<br>
><br>
> row 0: (0, 0.000189908) (1, 7.17315e-05) (2, 9.31778e-05) (3, 0.000514947) (4, 0.000178659) (5, 0.000178217) (9, -2.25457e-05) (10, -6.34278e-06) (11, -5.93241e-07) (12, 9.48544e-06) (13, 4.79709e-06) (14, 2.40016e-06) (27, -0.000335696) (28, -0.000106734) (29, -0.000106653) (30, 2.73119e-06) (31, -7.93382e-07) (32, 1.24048e-07) (36, -4.0302e-06) (37, 3.67494e-06) (38, -2.70115e-06) (39, 0.) (40, 0.) (41, 0.)<br>
><br>
> The difference between the column numbering between the finite difference and the hand-coded Jacobians looks like a serious problem to me, but I'm probably missing something.<br>
><br>
> I am trying to use a 3D DMDA with 3 dof, a box stencil of width 1, and for this test problem the grid dimensions are 11x7x6. For a grid point x,y,z, and dof c, is the index calculated as c + 3*x + 3*11*y + 3*11*7*z? If so, then the column numbers of the hand-coded Jacobian match those of the 27 point stencil I have in mind. However, I am then at a loss to explain the column numbers in the finite difference Jacobian.<br>
><br>
><br>
><br>
><br>
> On Sat, Oct 7, 2017 at 1:49 PM, zakaryah . <<a href="mailto:zakaryah@gmail.com" target="_blank">zakaryah@gmail.com</a>> wrote:<br>
> OK - I ran with -snes_monitor -snes_converged_reason -snes_linesearch_monitor -ksp_monitor -ksp_monitor_true_residual -ksp_converged_reason -pc_type svd -pc_svd_monitor -snes_type newtonls -snes_compare_explicit<br>
><br>
> and here is the full error message, output immediately after<br>
><br>
> Finite difference Jacobian<br>
> Mat Object: 24 MPI processes<br>
> type: mpiaij<br>
><br>
> [0]PETSC ERROR: --------------------- Error Message ------------------------------<wbr>------------------------------<wbr>--<br>
><br>
> [0]PETSC ERROR: Invalid argument<br>
><br>
> [0]PETSC ERROR: Matrix not generated from a DMDA<br>
><br>
> [0]PETSC ERROR: See <a href="http://www.mcs.anl.gov/petsc/documentation/faq.html" rel="noreferrer" target="_blank">http://www.mcs.anl.gov/petsc/d<wbr>ocumentation/faq.html</a> for trouble shooting.<br>
><br>
> [0]PETSC ERROR: Petsc Release Version 3.7.6, Apr, 24, 2017<br>
><br>
> [0]PETSC ERROR: ./CalculateOpticalFlow on a arch-linux2-c-opt named node046.hpc.rockefeller.intern<wbr>al by zfrentz Sat Oct 7 13:44:44 2017<br>
><br>
> [0]PETSC ERROR: Configure options --prefix=/ru-auth/local/home/z<wbr>frentz/PETSc-3.7.6 --download-fblaslapack -with-debugging=0<br>
><br>
> [0]PETSC ERROR: #1 MatView_MPI_DA() line 551 in /rugpfs/fs0/home/zfrentz/PETSc<wbr>/build/petsc-3.7.6/src/dm/impl<wbr>s/da/fdda.c<br>
><br>
> [0]PETSC ERROR: #2 MatView() line 901 in /rugpfs/fs0/home/zfrentz/PETSc<wbr>/build/petsc-3.7.6/src/mat/int<wbr>erface/matrix.c<br>
><br>
> [0]PETSC ERROR: #3 SNESComputeJacobian() line 2371 in /rugpfs/fs0/home/zfrentz/PETSc<wbr>/build/petsc-3.7.6/src/snes/in<wbr>terface/snes.c<br>
><br>
> [0]PETSC ERROR: #4 SNESSolve_NEWTONLS() line 228 in /rugpfs/fs0/home/zfrentz/PETSc<wbr>/build/petsc-3.7.6/src/snes/im<wbr>pls/ls/ls.c<br>
><br>
> [0]PETSC ERROR: #5 SNESSolve() line 4005 in /rugpfs/fs0/home/zfrentz/PETSc<wbr>/build/petsc-3.7.6/src/snes/in<wbr>terface/snes.c<br>
><br>
> [0]PETSC ERROR: #6 solveWarp3D() line 659 in /ru-auth/local/home/zfrentz/Co<wbr>de/OpticalFlow/working/October<wbr>6_2017/mshs.c<br>
><br>
><br>
> On Tue, Oct 3, 2017 at 5:37 PM, Jed Brown <<a href="mailto:jed@jedbrown.org" target="_blank">jed@jedbrown.org</a>> wrote:<br>
> Always always always send the whole error message.<br>
><br>
> "zakaryah ." <<a href="mailto:zakaryah@gmail.com" target="_blank">zakaryah@gmail.com</a>> writes:<br>
><br>
> > I tried -snes_compare_explicit, and got the following error:<br>
> ><br>
> > [0]PETSC ERROR: Invalid argument<br>
> ><br>
> > [0]PETSC ERROR: Matrix not generated from a DMDA<br>
> ><br>
> > What am I doing wrong?<br>
> ><br>
> > On Tue, Oct 3, 2017 at 10:08 AM, Jed Brown <<a href="mailto:jed@jedbrown.org" target="_blank">jed@jedbrown.org</a>> wrote:<br>
> ><br>
> >> Barry Smith <<a href="mailto:bsmith@mcs.anl.gov" target="_blank">bsmith@mcs.anl.gov</a>> writes:<br>
> >><br>
> >> >> On Oct 3, 2017, at 5:54 AM, zakaryah . <<a href="mailto:zakaryah@gmail.com" target="_blank">zakaryah@gmail.com</a>> wrote:<br>
> >> >><br>
> >> >> I'm still working on this. I've made some progress, and it looks like<br>
> >> the issue is with the KSP, at least for now. The Jacobian may be<br>
> >> ill-conditioned. Is it possible to use -snes_test_display during an<br>
> >> intermediate step of the analysis? I would like to inspect the Jacobian<br>
> >> after several solves have already completed,<br>
> >> ><br>
> >> > No, our currently code for testing Jacobians is poor quality and<br>
> >> poorly organized. Needs a major refactoring to do things properly. Sorry<br>
> >><br>
> >> You can use -snes_compare_explicit or -snes_compare_coloring to output<br>
> >> differences on each Newton step.<br>
> >><br>
><br>
><br>
<br>
</div></div></blockquote></div><br></div>
</blockquote></div></div></div><br><br clear="all"><div><br></div>-- <br><div class="m_-1393587641323955827gmail_signature" data-smartmail="gmail_signature"><div dir="ltr"><div><div dir="ltr"><span class=""><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></span><div><a href="http://www.caam.rice.edu/~mk51/" target="_blank">https://www.cse.buffalo.edu/~<wbr>knepley/</a><br></div></div></div></div></div>
</div></div>
</blockquote></div><br></div>