[petsc-dev] Recover from TS failure

Mark Adams mfadams at lbl.gov
Wed Nov 1 12:54:53 CDT 2017


On Wed, Nov 1, 2017 at 1:46 PM, Jed Brown <jed at jedbrown.org> wrote:

> Mark Adams <mfadams at lbl.gov> writes:
>
> > I have added some code in a TS post step method to look at the number of
> > nonlinear iterations and cut the time step if it took too many SNES
> > iterations.  That helped but now I want to go one step further and
> recover
> > from a failed time step (see appended error message).
> >
> > How can/should I recover from a SNES failure inside of TS? I could just
> put
> > TSSolve in a loop and if TS failed because of the solver, then reduce the
> > time step and try to recover the state, and try again.
> >
> > Thoughts?
> >
> > Thanks,
> > Mark
> >
> > Note, this is a direct solver and the linear solver is diverging?  I've
> > seen the same thing with the maximum its in SNES and failed linesearch.
> >
> > 23 TS dt 0.0025 time 0.1025
> >     0 SNES Function norm 9.997473571601e+04
> >     1 SNES Function norm 2.295851692533e-01
> >   Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE
> iterations 1
> > [0]PETSC ERROR: --------------------- Error Message
> > --------------------------------------------------------------
> > [0]PETSC ERROR:
> > [0]PETSC ERROR: TSStep has failed due to DIVERGED_NONLINEAR_SOLVE,
> increase
> > -ts_max_snes_failures or make negative to attempt recovery
>
> Did you increase -ts_max_snes_failures (TSSetMaxSNESFailures)?
>

Ah, read the error message :) with a negative value I get this.

TS can not, or is not going to try, to reconstruct my state and reduce the
time step. So I think I need to do this myself.

I guess just making loop and testing in the TSConvergedReason, is the way
to go. We already do something like this for mesh adpativity.

Thanks,

23 TS dt 0.0025 time 0.1025
    0 SNES Function norm 9.997473571601e+04
    1 SNES Function norm 2.295851692533e-01
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 1
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
    0 SNES Function norm 9.997473571601e+04
  Nonlinear solve did not converge due to DIVERGED_LINEAR_SOLVE iterations 0
[0]PETSC ERROR: --------------------- Error Message
--------------------------------------------------------------
[0]PETSC ERROR:
[0]PETSC ERROR: TSStep has failed due to DIVERGED_STEP_REJECTED
[0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html for
trouble shooting.
[0]PETSC ERROR: Petsc Development GIT revision: v3.8-53-ga2d93b2d43  GIT
Date: 2017-10-09 12:56:49 -0400
[0]PETSC ERROR: ./ex48 on a arch-macosx-gnu-O named MarksMac-5.local by
markadams Wed Nov  1 13:51:24 2017
[0]PETSC ERROR: Configure options --with-cc=clang --with-cc++=clang++
COPTFLAGS="-O2 -g -mavx2" CXXOPTFLAGS="-O2 -g -mavx2" FOPTFLAGS="-O2 -g
-mavx2" --download-mpich=1 --download-parmetis=1 --download-metis=1
--download-hypre=1 --download-ml=1 --download-triangle=1
--download-ctetgen=1 --download-p4est=1 --download-fftw --download-mumps=1
--download-scalapack=1 --with-x=0 --download-superlu_dist
--download-superlu --download-ctetgen --with-debugging=0 --download-hdf5=1
PETSC_ARCH=arch-macosx-gnu-O --download-chaco --with-viewfromoptions=1
[0]PETSC ERROR: #1 TSStep() line 4130 in
/Users/markadams/Codes/petsc/src/ts/interface/ts.c
[0]PETSC ERROR: #2 TSSolve() line 4374 in
/Users/markadams/Codes/petsc/src/ts/interface/ts.c
*********** DIVERGED_STEP_REJECTED ierr=91 reason=-2
L_2 Diff: 1.32241e+07 at time 0.1025




>
> > [0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html
> for
> > trouble shooting.
> > [0]PETSC ERROR: Petsc Development GIT revision: v3.8-53-ga2d93b2d43  GIT
> > Date: 2017-10-09 12:56:49 -0400
> > [0]PETSC ERROR: ./ex48 on a arch-macosx-gnu-O named MarksMac-5.local by
> > markadams Wed Nov  1 13:17:15 2017
> > [0]PETSC ERROR: Configure options --with-cc=clang --with-cc++=clang++
> > COPTFLAGS="-O2 -g -mavx2" CXXOPTFLAGS="-O2 -g -mavx2" FOPTFLAGS="-O2 -g
> > -mavx2" --download-mpich=1 --download-parmetis=1 --download-metis=1
> > --download-hypre=1 --download-ml=1 --download-triangle=1
> > --download-ctetgen=1 --download-p4est=1 --download-fftw
> --download-mumps=1
> > --download-scalapack=1 --with-x=0 --download-superlu_dist
> > --download-superlu --download-ctetgen --with-debugging=0
> --download-hdf5=1
> > PETSC_ARCH=arch-macosx-gnu-O --download-chaco --with-viewfromoptions=1
> > [0]PETSC ERROR: #1 TSStep() line 4129 in
> > /Users/markadams/Codes/petsc/src/ts/interface/ts.c
> > [0]PETSC ERROR: #2 TSSolve() line 4374 in
> > /Users/markadams/Codes/petsc/src/ts/interface/ts.c
> > [0]PETSC ERROR: #3 main() line 1161 in /Users/markadams/Codes/mhd/
> src/ex48.c
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20171101/628729de/attachment.html>


More information about the petsc-dev mailing list