KSP preonly does not support UNPRECONDITIONED

[MGC Nestola]

Dear all,

i am trying to use the last moose version but when i try to use the following options for the executioner i get an error which is reported below

[Executioner]

type = Steady

###  nota che il PJFNK fa un secondo step che fa e piu lento

#    solve_type = 'PJFNK'

solve_type = 'NEWTON'

line_search = 'none'

# petsc_options_iname = ' -pc_type -pc_hypre_type'

# petsc_options_value = ' hypre boomeramg'

petsc_options_iname=' -ksp_type -pc_type -pc_factor_shift_type -pc_factor_mat_solver_package '

petsc_options_value='   preonly   lu       NONZERO               mumps                '

[]

ERROR:

 Nonlinear |R| = 1.176300e-03

[0]PETSC ERROR: --------------------- Error Message --------------------------------------------------------------

[0]PETSC ERROR: No support for this operation for this object type

[0]PETSC ERROR: KSP preonly does not support UNPRECONDITIONED

[0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html for trouble shooting.

[0]PETSC ERROR: Petsc Release Version 3.7.5, unknown

[0]PETSC ERROR: ../../../mech-opt on a arch-darwin-c-opt named marias-mbp.mobile.usilu.net by mariagiuseppina Tue Jun 20 17:33:50 2017

[0]PETSC ERROR: Configure options --prefix=/opt/krause_builds/petsc/openmpi_petsc-3.6.3/gcc-opt --download-hypre=1 --with-ssl=0 --with-debugging=no --with-pic=1 --with-shared-libraries=1 --with-cc=mpicc --with-cxx=mpicxx --with-fc=mpif90 --download-fblaslapack=1 --download-metis=1 --download-parmetis=1 --download-superlu_dist=1 --download-mumps=1 --download-scalapack=1 CC=mpicc CXX=mpicxx FC=mpif90 F77=mpif77 F90=mpif90 CFLAGS="-fPIC -fopenmp" CXXFLAGS="-fPIC -fopenmp" FFLAGS="-fPIC -fopenmp" FCFLAGS="-fPIC -fopenmp" F90FLAGS="-fPIC -fopenmp" F77FLAGS="-fPIC -fopenmp" PETSC_DIR=/opt/krause_builds/petsc

[0]PETSC ERROR: #1 KSPSetUpNorms_Private() line 413 in /opt/krause_builds/petsc/src/ksp/ksp/interface/itcreate.c

[0]PETSC ERROR: #2 KSPSetUp() line 314 in /opt/krause_builds/petsc/src/ksp/ksp/interface/itfunc.c

[0]PETSC ERROR: #3 KSPSolve() line 599 in /opt/krause_builds/petsc/src/ksp/ksp/interface/itfunc.c

[0]PETSC ERROR: #4 SNESSolve_NEWTONLS() line 230 in /opt/krause_builds/petsc/src/snes/impls/ls/ls.c

[0]PETSC ERROR: #5 SNESSolve() line 4005 in /opt/krause_builds/petsc/src/snes/interface/snes.c

--------------------------------------------------------------------------

MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD

with errorcode 1.

Can someone help me to fix the problem?

Maria

[Kong, Fande]

Hi Maria,

Please put "ksp_norm = default" in the preconditioning block to let PETSc decides whatever it wants.  

We recently updated  the some default setting in MOOSE.

Fande,

--
You received this message because you are subscribed to the Google Groups "moose-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to moose-users+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/moose-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/moose-users/5907ca3b-1f0d-4d84-a94a-5c1ba4e5da76%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jed Brown]

It would be -ksp_norm_type NONE.

PETSc: they want -ksp_norm_type UNPRECONDITIONED for everything. Should
PREONLY be made to "support" UNPRECONDITIONED (and PRECONDITIONED)
despite not computing norms?

> --
> You received this message because you are subscribed to the Google Groups "moose-users" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to moose-users...@googlegroups.com.

[MGC Nestola]

Dear all,

i tried with the following options 

petsc_options_iname=' -ksp_type  -pc_type -pc_factor_shift_type -pc_factor_mat_solver_package -ksp_norm '

petsc_options_value='   preonly   lu       NONZERO               mumps          default      '

but it doesn't work

Is there something wrong?

Maria

[Kong, Fande]

On Tue, Jun 20, 2017 at 10:36 AM, MGC Nestola <mgcne...@gmail.com> wrote:

Dear all,

i tried with the following options 

petsc_options_iname=' -ksp_type  -pc_type -pc_factor_shift_type -pc_factor_mat_solver_package -ksp_norm '

petsc_options_value='   preonly   lu       NONZERO               mumps          default      '

Please change default to "none", if you want to use native petsc options.

Fande, 

 

--
You received this message because you are subscribed to the Google Groups "moose-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to moose-users+unsubscribe@googlegroups.com.

Visit this group at https://groups.google.com/group/moose-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/moose-users/e6b31b2b-f8e1-4213-8a0e-b15d91add50c%40googlegroups.com.

[MGC Nestola]

I tried also with none, but still the same.

Maria Nestola

Il giorno martedì 20 giugno 2017 17:37:48 UTC+2, MGC Nestola ha scritto:

[Kong, Fande]

Hi Maria,

In the previous email, I said put "ksp_norm=default" in the preconditiong block. "ksp_norm=defualt" is a moose option. But you now tried to use PETSc options. 

If you like PETSc options, please use:

petsc_options_iname = " -ksp_norm_type"

petsc_options_value = "none"

Fande

--
You received this message because you are subscribed to the Google Groups "moose-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to moose-users+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/moose-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/moose-users/8efa5376-16b2-4044-a441-9c3296e8c43f%40googlegroups.com.

[MGC Nestola]

Hi Fande,

thanks for your help and sorry fro the mistake, 

thus i tried this

petsc_options_iname=' -ksp_type -pc_type -pc_factor_shift_type -pc_factor_mat_solver_package -ksp_norm_type'

petsc_options_value='   preonly   lu       NONZERO               mumps       none         '

but i got the same error.

I was using these options with the previous moose, thus i do not understand why now they do not work anymore

Maria

Il giorno martedì 20 giugno 2017 17:37:48 UTC+2, MGC Nestola ha scritto:

[Kong, Fande]

This one should work 

[Preconditioning]

  [./smp]

    type = SMP

    ksp_norm = default

  [../]

[]

I will take a look at why native options do not work.

Thanks for your report 

Fande,

--
You received this message because you are subscribed to the Google Groups "moose-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to moose-users+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/moose-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/moose-users/9ff561d4-b0af-450b-8490-332103f3aece%40googlegroups.com.

[Kong, Fande]

Hi Maria,

"-ksp_norm_type" does not work in PETSc If you use  KSPOnly. I am not sure it is considered as a bug or not, and I will email to PETSc list.

From now, please use the preconditioning block in your input file.

Thanks,

Fande,