[petsc-users] monitoring the convergence of fieldsplit 0 and 1
Klaij, Christiaan
C.Klaij at marin.nl
Tue Jan 17 01:45:26 CST 2017
Well, that's it, all the rest was hard coded. Here's the relevant part of the code:
CALL PCSetType(pc_system,PCFIELDSPLIT,ierr); CHKERRQ(ierr)
CALL PCFieldSplitSetType(pc_system,PC_COMPOSITE_SCHUR,ierr); CHKERRQ(ierr)
CALL PCFieldSplitSetIS(pc_system,"0",isgs(1),ierr); CHKERRQ(ierr)
CALL PCFieldSplitSetIS(pc_system,"1",isgs(2),ierr); CHKERRQ(ierr)
CALL PCFieldSplitSetSchurFactType(pc_system,PC_FIELDSPLIT_SCHUR_FACT_FULL,ierr);CHKERRQ(ierr)
CALL PCFieldSplitSetSchurPre(pc_system,PC_FIELDSPLIT_SCHUR_PRE_SELFP,PETSC_NULL_OBJECT,ierr);CHKERRQ(ierr)
CALL KSPSetTolerances(ksp_system,tol,PETSC_DEFAULT_REAL,PETSC_DEFAULT_REAL,maxiter,ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_0_ksp_rtol","0.01",ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_1_ksp_rtol","0.01",ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_0_ksp_pc_side","right",ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_1_ksp_pc_side","right",ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_0_ksp_type","gmres",ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_1_upper_ksp_type","preonly",ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_1_upper_pc_type","jacobi",ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_1_inner_ksp_type","preonly",ierr); CHKERRQ(ierr)
CALL PetscOptionsSetValue(PETSC_NULL_OBJECT,"-sys_fieldsplit_1_inner_pc_type","jacobi",ierr); CHKERRQ(ierr)
dr. ir. Christiaan Klaij | CFD Researcher | Research & Development
MARIN | T +31 317 49 33 44 | mailto:C.Klaij at marin.nl | http://www.marin.nl
MARIN news: http://www.marin.nl/web/News/News-items/Verification-and-validation-exercises-for-flow-around-KVLCC2-tanker.htm
________________________________________
From: Barry Smith <bsmith at mcs.anl.gov>
Sent: Monday, January 16, 2017 9:28 PM
To: Klaij, Christiaan
Cc: petsc-users at mcs.anl.gov
Subject: Re: [petsc-users] monitoring the convergence of fieldsplit 0 and 1
Please send all the command line options you use.
> On Jan 16, 2017, at 1:47 AM, Klaij, Christiaan <C.Klaij at marin.nl> wrote:
>
> Barry,
>
> Sure, here's the output with:
>
> -sys_ksp_view -sys_ksp_converged_reason -sys_fieldsplit_0_ksp_converged_reason -sys_fieldsplit_1_ksp_converged_reason
>
> (In my previous email, I rearranged 0 & 1 for easy summing.)
>
> Chris
>
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 1
> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 22
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 1
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 2
> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 6
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 2
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 7
> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 3
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 7
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 7
> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 2
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 7
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 8
> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 2
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 8
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 8
> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 2
> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 8
> Linear sys_ solve converged due to CONVERGED_RTOL iterations 6
> KSP Object:(sys_) 1 MPI processes
> type: fgmres
> GMRES: restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement
> GMRES: happy breakdown tolerance 1e-30
> maximum iterations=300, initial guess is zero
> tolerances: relative=0.01, absolute=1e-50, divergence=10000.
> right preconditioning
> using UNPRECONDITIONED norm type for convergence test
> PC Object:(sys_) 1 MPI processes
> type: fieldsplit
> FieldSplit with Schur preconditioner, factorization FULL
> Preconditioner for the Schur complement formed from Sp, an assembled approximation to S, which uses (lumped, if requested) A00's diagonal's inverse
> Split info:
> Split number 0 Defined by IS
> Split number 1 Defined by IS
> KSP solver for A00 block
> KSP Object: (sys_fieldsplit_0_) 1 MPI processes
> type: gmres
> GMRES: restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement
> GMRES: happy breakdown tolerance 1e-30
> maximum iterations=10000, initial guess is zero
> tolerances: relative=0.01, absolute=1e-50, divergence=10000.
> right preconditioning
> using UNPRECONDITIONED norm type for convergence test
> PC Object: (sys_fieldsplit_0_) 1 MPI processes
> type: ilu
> ILU: out-of-place factorization
> 0 levels of fill
> tolerance for zero pivot 2.22045e-14
> matrix ordering: natural
> factor fill ratio given 1., needed 1.
> Factored matrix follows:
> Mat Object: 1 MPI processes
> type: seqaij
> rows=9600, cols=9600
> package used to perform factorization: petsc
> total: nonzeros=47280, allocated nonzeros=47280
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> linear system matrix = precond matrix:
> Mat Object: (sys_fieldsplit_0_) 1 MPI processes
> type: seqaij
> rows=9600, cols=9600
> total: nonzeros=47280, allocated nonzeros=47280
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> KSP solver for upper A00 in upper triangular factor
> KSP Object: (sys_fieldsplit_1_upper_) 1 MPI processes
> type: preonly
> maximum iterations=10000, initial guess is zero
> tolerances: relative=1e-05, absolute=1e-50, divergence=10000.
> left preconditioning
> using NONE norm type for convergence test
> PC Object: (sys_fieldsplit_1_upper_) 1 MPI processes
> type: jacobi
> linear system matrix = precond matrix:
> Mat Object: (sys_fieldsplit_0_) 1 MPI processes
> type: seqaij
> rows=9600, cols=9600
> total: nonzeros=47280, allocated nonzeros=47280
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> KSP solver for S = A11 - A10 inv(A00) A01
> KSP Object: (sys_fieldsplit_1_) 1 MPI processes
> type: gmres
> GMRES: restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement
> GMRES: happy breakdown tolerance 1e-30
> maximum iterations=10000, initial guess is zero
> tolerances: relative=0.01, absolute=1e-50, divergence=10000.
> right preconditioning
> using UNPRECONDITIONED norm type for convergence test
> PC Object: (sys_fieldsplit_1_) 1 MPI processes
> type: ilu
> ILU: out-of-place factorization
> 0 levels of fill
> tolerance for zero pivot 2.22045e-14
> matrix ordering: natural
> factor fill ratio given 1., needed 1.
> Factored matrix follows:
> Mat Object: 1 MPI processes
> type: seqaij
> rows=3200, cols=3200
> package used to perform factorization: petsc
> total: nonzeros=40404, allocated nonzeros=40404
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> linear system matrix followed by preconditioner matrix:
> Mat Object: (sys_fieldsplit_1_) 1 MPI processes
> type: schurcomplement
> rows=3200, cols=3200
> Schur complement A11 - A10 inv(A00) A01
> A11
> Mat Object: (sys_fieldsplit_1_) 1 MPI processes
> type: seqaij
> rows=3200, cols=3200
> total: nonzeros=40404, allocated nonzeros=40404
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> A10
> Mat Object: 1 MPI processes
> type: seqaij
> rows=3200, cols=9600
> total: nonzeros=47280, allocated nonzeros=47280
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> KSP of A00
> KSP Object: (sys_fieldsplit_1_inner_) 1 MPI processes
> type: preonly
> maximum iterations=10000, initial guess is zero
> tolerances: relative=1e-05, absolute=1e-50, divergence=10000.
> left preconditioning
> using NONE norm type for convergence test
> PC Object: (sys_fieldsplit_1_inner_) 1 MPI processes
> type: jacobi
> linear system matrix = precond matrix:
> Mat Object: (sys_fieldsplit_0_) 1 MPI processes
> type: seqaij
> rows=9600, cols=9600
> total: nonzeros=47280, allocated nonzeros=47280
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> A01
> Mat Object: 1 MPI processes
> type: seqaij
> rows=9600, cols=3200
> total: nonzeros=47280, allocated nonzeros=47280
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> Mat Object: 1 MPI processes
> type: seqaij
> rows=3200, cols=3200
> total: nonzeros=40404, allocated nonzeros=40404
> total number of mallocs used during MatSetValues calls =0
> not using I-node routines
> linear system matrix followed by preconditioner matrix:
> Mat Object: 1 MPI processes
> type: nest
> rows=12800, cols=12800
> Matrix object:
> type=nest, rows=2, cols=2
> MatNest structure:
> (0,0) : prefix="mom_", type=seqaij, rows=9600, cols=9600
> (0,1) : prefix="grad_", type=seqaij, rows=9600, cols=3200
> (1,0) : prefix="div_", type=seqaij, rows=3200, cols=9600
> (1,1) : prefix="stab_", type=seqaij, rows=3200, cols=3200
> Mat Object: 1 MPI processes
> type: nest
> rows=12800, cols=12800
> Matrix object:
> type=nest, rows=2, cols=2
> MatNest structure:
> (0,0) : prefix="sys_fieldsplit_0_", type=seqaij, rows=9600, cols=9600
> (0,1) : type=seqaij, rows=9600, cols=3200
> (1,0) : type=seqaij, rows=3200, cols=9600
> (1,1) : prefix="sys_fieldsplit_1_", type=seqaij, rows=3200, cols=3200
> nusediter_vv 37
> nusediter_pp 37
>
>
>
> dr. ir. Christiaan Klaij | CFD Researcher | Research & Development
> MARIN | T +31 317 49 33 44 | mailto:C.Klaij at marin.nl | http://www.marin.nl
>
> MARIN news: http://www.marin.nl/web/News/News-items/The-Ocean-Cleanup-testing-continues.htm
>
> ________________________________________
> From: Barry Smith <bsmith at mcs.anl.gov>
> Sent: Friday, January 13, 2017 7:51 PM
> To: Klaij, Christiaan
> Cc: petsc-users at mcs.anl.gov
> Subject: Re: [petsc-users] monitoring the convergence of fieldsplit 0 and 1
>
> Yes, I would have expected this to work. Could you send the output from -ksp_view in this case?
>
>
>> On Jan 13, 2017, at 3:46 AM, Klaij, Christiaan <C.Klaij at marin.nl> wrote:
>>
>> Barry,
>>
>> It's been a while but I'm finally using this function in
>> 3.7.4. Is it supposed to work with fieldsplit? Here's why.
>>
>> I'm solving a Navier-Stokes system with fieldsplit (pc has one
>> velocity solve and one pressure solve) and trying to retrieve the
>> totals like this:
>>
>> CALL KSPSolve(ksp_system,rr_system,xx_system,ierr); CHKERRQ(ierr)
>> CALL PCFieldSplitGetSubKSP(pc_system,numsplit,subksp,ierr); CHKERRQ(ierr)
>> CALL KSPGetTotalIterations(subksp(1),nusediter_vv,ierr); CHKERRQ(ierr)
>> CALL KSPGetTotalIterations(subksp(2),nusediter_pp,ierr); CHKERRQ(ierr)
>> print *, 'nusediter_vv', nusediter_vv
>> print *, 'nusediter_pp', nusediter_pp
>>
>> Running the code shows this surprise:
>>
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 1
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 1
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 2
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 2
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 7
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 7
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 7
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 7
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 8
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 8
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 8
>> Linear sys_fieldsplit_0_ solve converged due to CONVERGED_RTOL iterations 8
>>
>> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 22
>> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 6
>> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 3
>> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 2
>> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 2
>> Linear sys_fieldsplit_1_ solve converged due to CONVERGED_RTOL iterations 2
>>
>> nusediter_vv 37
>> nusediter_pp 37
>>
>> So the value of nusediter_pp is indeed 37, but for nusediter_vv
>> it should be 66. Any idea what went wrong?
>>
>> Chris
>>
>>
>>
>> dr. ir. Christiaan Klaij | CFD Researcher | Research & Development
>> MARIN | T +31 317 49 33 44 | mailto:C.Klaij at marin.nl | http://www.marin.nl
>>
>> MARIN news: http://www.marin.nl/web/News/News-items/MARIN-wishes-you-a-challenging-inspiring-2017.htm
>>
>> ________________________________________
>> From: Barry Smith <bsmith at mcs.anl.gov>
>> Sent: Saturday, April 11, 2015 12:27 AM
>> To: Klaij, Christiaan
>> Cc: petsc-users at mcs.anl.gov
>> Subject: Re: [petsc-users] monitoring the convergence of fieldsplit 0 and 1
>>
>> Chris,
>>
>> I have added KSPGetTotalIterations() to the branch barry/add-ksp-total-iterations/master and next. After tests it will go into master
>>
>> Barry
>>
>>> On Apr 10, 2015, at 8:07 AM, Klaij, Christiaan <C.Klaij at marin.nl> wrote:
>>>
>>> Barry,
>>>
>>> Sure, I can call PCFieldSplitGetSubKSP() to get the fieldsplit_0
>>> ksp and then KSPGetIterationNumber, but what does this number
>>> mean?
>>>
>>> It appears to be the number of iterations of the last time that
>>> the subsystem was solved, right? If so, this corresponds to the
>>> last iteration of the coupled system, how about all the previous
>>> iterations?
>>>
>>> Chris
>>> ________________________________________
>>> From: Barry Smith <bsmith at mcs.anl.gov>
>>> Sent: Friday, April 10, 2015 2:48 PM
>>> To: Klaij, Christiaan
>>> Cc: petsc-users at mcs.anl.gov
>>> Subject: Re: [petsc-users] monitoring the convergence of fieldsplit 0 and 1
>>>
>>> Chris,
>>>
>>> It appears you should call PCFieldSplitGetSubKSP() and then get the information you want out of the individual KSPs. If this doesn't work please let us know.
>>>
>>> Barry
>>>
>>>> On Apr 10, 2015, at 6:48 AM, Klaij, Christiaan <C.Klaij at marin.nl> wrote:
>>>>
>>>> A question when using PCFieldSplit: for each linear iteration of
>>>> the system, how many iterations for fielsplit 0 and 1?
>>>>
>>>> One way to find out is to run with -ksp_monitor,
>>>> -fieldsplit_0_ksp_monitor and -fieldsplit_0_ksp_monitor. This
>>>> gives the complete convergence history.
>>>>
>>>> Another way, suggested by Matt, is to use -ksp_monitor,
>>>> -fieldsplit_0_ksp_converged_reason and
>>>> -fieldsplit_1_ksp_converged_reason. This gives only the totals
>>>> for fieldsplit 0 and 1 (but without saying for which one).
>>>>
>>>> Both ways require to somehow process the output, which is a bit
>>>> inconvenient. Could KSPGetResidualHistory perhaps return (some)
>>>> information on the subsystems' convergence for processing inside
>>>> the code?
>>>>
>>>> Chris
>>>>
>>>>
>>>> dr. ir. Christiaan Klaij
>>>> CFD Researcher
>>>> Research & Development
>>>> E mailto:C.Klaij at marin.nl
>>>> T +31 317 49 33 44
>>>>
>>>>
>>>> MARIN
>>>> 2, Haagsteeg, P.O. Box 28, 6700 AA Wageningen, The Netherlands
>>>> T +31 317 49 39 11, F +31 317 49 32 45, I www.marin.nl
>>>>
>>>
>>
>
More information about the petsc-users
mailing list