Re: [petsc-users] subprocess (block) tolerance.
> On Oct 4, 2018, at 6:09 PM, HeeHo Park wrote: > > Barry and Jed, > > Thank you for your answers. I think I need to learn more about domain > decomposition as I am a bit confused. > Is it true that we are using BiCGstab here to solve the system of equations, > using Additive Schwartz as a domain decomposition preconditioner, and that > precondition matrix for each block is preconditioned by ilu[0]? You can say it that way. > > Thanks, > > > On Thu, Oct 4, 2018 at 5:53 PM Smith, Barry F. wrote: > > Since > > KSP Object: (flow_sub_) 1 MPI processes > type: preonly > > this means only a single iteration of the inner solver is used so the > numbers in red are not used. > > You could do something like -flow_ksp_type fgmres -flow_sub_ksp_type gmres > -flow_sub_ksp_rtol 1.e-2 but it wouldn't help maters. Likely the current > values are the best. > >Barry > > > > On Oct 4, 2018, at 5:44 PM, HeeHo Park wrote: > > > > Hi, I'm running PFLOTRAN and in PFLOTRAN, we have flow_ and flow_sub_ > > processes. I was wondering what the red underlined values meant (each block > > tolerance?) and how to change them (would it affect convergence?). Blue > > marked bold values are changed from the default values for linear solvers. > > > > FLOW Linear Solver > >solver: bcgs > >preconditioner: asm > > atol: 1.00E-10 > > rtol: 1.00E-10 > > dtol: 1.00E+04 > > maximum iteration: 1 > > KSP Object: (flow_) 8 MPI processes > > type: bcgs > > maximum iterations=1, initial guess is zero > > tolerances: relative=1e-10, absolute=1e-10, divergence=1. > > left preconditioning > > using PRECONDITIONED norm type for convergence test > > PC Object: (flow_) 8 MPI processes > > type: asm > > Additive Schwarz: total subdomain blocks = 8, amount of overlap = 1 > > Additive Schwarz: restriction/interpolation type - RESTRICT > > [0] number of local blocks = 1 > > [1] number of local blocks = 1 > > [2] number of local blocks = 1 > > [3] number of local blocks = 1 > > [4] number of local blocks = 1 > > [5] number of local blocks = 1 > > [6] number of local blocks = 1 > > [7] number of local blocks = 1 > > Local solve info for each block is in the following KSP and PC objects: > > - - - - - - - - - - - - - - - - - - > > [0] local block number 0, size = 1389 > > KSP Object: (flow_sub_) 1 MPI processes > > type: preonly > > maximum iterations=1, initial guess is zero > > >>> tolerances: relative=1e-05, absolute=1e-50, divergence=1. > > left preconditioning > > using DEFAULT norm type for convergence test > > PC Object: (flow_sub_) 1 MPI processes > > type: ilu > > PC has not been set up so information may be incomplete > > out-of-place factorization > > 0 levels of fill > > tolerance for zero pivot 2.22045e-14 > > using diagonal shift on blocks to prevent zero pivot [INBLOCKS] > > matrix ordering: natural > > linear system matrix = precond matrix: > > Mat Object: (flow_) 1 MPI processes > > type: seqbaij > > rows=1389, cols=1389, bs=3 > > total: nonzeros=20025, allocated nonzeros=20025 > > total number of mallocs used during MatSetValues calls =0 > > block size is 3 > > - - - - - - - - - - - - - - - - - - > > > > -- > > HeeHo Daniel Park > > > > -- > HeeHo Daniel Park
Re: [petsc-users] subprocess (block) tolerance.
Barry and Jed, Thank you for your answers. I think I need to learn more about domain decomposition as I am a bit confused. Is it true that we are using BiCGstab here to solve the system of equations, using Additive Schwartz as a domain decomposition preconditioner, and that precondition matrix for each block is preconditioned by ilu[0]? Thanks, On Thu, Oct 4, 2018 at 5:53 PM Smith, Barry F. wrote: > > Since > > KSP Object: (flow_sub_) 1 MPI processes > type: preonly > > this means only a single iteration of the inner solver is used so the > numbers in red are not used. > > You could do something like -flow_ksp_type fgmres -flow_sub_ksp_type gmres > -flow_sub_ksp_rtol 1.e-2 but it wouldn't help maters. Likely the current > values are the best. > >Barry > > > > On Oct 4, 2018, at 5:44 PM, HeeHo Park wrote: > > > > Hi, I'm running PFLOTRAN and in PFLOTRAN, we have flow_ and flow_sub_ > processes. I was wondering what the red underlined values meant (each block > tolerance?) and how to change them (would it affect convergence?). Blue > marked bold values are changed from the default values for linear solvers. > > > > FLOW Linear Solver > >solver: bcgs > >preconditioner: asm > > atol: 1.00E-10 > > rtol: 1.00E-10 > > dtol: 1.00E+04 > > maximum iteration: 1 > > KSP Object: (flow_) 8 MPI processes > > type: bcgs > > maximum iterations=1, initial guess is zero > > tolerances: relative=1e-10, absolute=1e-10, divergence=1. > > left preconditioning > > using PRECONDITIONED norm type for convergence test > > PC Object: (flow_) 8 MPI processes > > type: asm > > Additive Schwarz: total subdomain blocks = 8, amount of overlap = 1 > > Additive Schwarz: restriction/interpolation type - RESTRICT > > [0] number of local blocks = 1 > > [1] number of local blocks = 1 > > [2] number of local blocks = 1 > > [3] number of local blocks = 1 > > [4] number of local blocks = 1 > > [5] number of local blocks = 1 > > [6] number of local blocks = 1 > > [7] number of local blocks = 1 > > Local solve info for each block is in the following KSP and PC > objects: > > - - - - - - - - - - - - - - - - - - > > [0] local block number 0, size = 1389 > > KSP Object: (flow_sub_) 1 MPI processes > > type: preonly > > maximum iterations=1, initial guess is zero > > >>> tolerances: relative=1e-05, absolute=1e-50, divergence=1. > > left preconditioning > > using DEFAULT norm type for convergence test > > PC Object: (flow_sub_) 1 MPI processes > > type: ilu > > PC has not been set up so information may be incomplete > > out-of-place factorization > > 0 levels of fill > > tolerance for zero pivot 2.22045e-14 > > using diagonal shift on blocks to prevent zero pivot [INBLOCKS] > > matrix ordering: natural > > linear system matrix = precond matrix: > > Mat Object: (flow_) 1 MPI processes > > type: seqbaij > > rows=1389, cols=1389, bs=3 > > total: nonzeros=20025, allocated nonzeros=20025 > > total number of mallocs used during MatSetValues calls =0 > > block size is 3 > > - - - - - - - - - - - - - - - - - - > > > > -- > > HeeHo Daniel Park > > -- HeeHo Daniel Park
Re: [petsc-users] subprocess (block) tolerance.
Since KSP Object: (flow_sub_) 1 MPI processes type: preonly this means only a single iteration of the inner solver is used so the numbers in red are not used. You could do something like -flow_ksp_type fgmres -flow_sub_ksp_type gmres -flow_sub_ksp_rtol 1.e-2 but it wouldn't help maters. Likely the current values are the best. Barry > On Oct 4, 2018, at 5:44 PM, HeeHo Park wrote: > > Hi, I'm running PFLOTRAN and in PFLOTRAN, we have flow_ and flow_sub_ > processes. I was wondering what the red underlined values meant (each block > tolerance?) and how to change them (would it affect convergence?). Blue > marked bold values are changed from the default values for linear solvers. > > FLOW Linear Solver >solver: bcgs >preconditioner: asm > atol: 1.00E-10 > rtol: 1.00E-10 > dtol: 1.00E+04 > maximum iteration: 1 > KSP Object: (flow_) 8 MPI processes > type: bcgs > maximum iterations=1, initial guess is zero > tolerances: relative=1e-10, absolute=1e-10, divergence=1. > left preconditioning > using PRECONDITIONED norm type for convergence test > PC Object: (flow_) 8 MPI processes > type: asm > Additive Schwarz: total subdomain blocks = 8, amount of overlap = 1 > Additive Schwarz: restriction/interpolation type - RESTRICT > [0] number of local blocks = 1 > [1] number of local blocks = 1 > [2] number of local blocks = 1 > [3] number of local blocks = 1 > [4] number of local blocks = 1 > [5] number of local blocks = 1 > [6] number of local blocks = 1 > [7] number of local blocks = 1 > Local solve info for each block is in the following KSP and PC objects: > - - - - - - - - - - - - - - - - - - > [0] local block number 0, size = 1389 > KSP Object: (flow_sub_) 1 MPI processes > type: preonly > maximum iterations=1, initial guess is zero > >>> tolerances: relative=1e-05, absolute=1e-50, divergence=1. > left preconditioning > using DEFAULT norm type for convergence test > PC Object: (flow_sub_) 1 MPI processes > type: ilu > PC has not been set up so information may be incomplete > out-of-place factorization > 0 levels of fill > tolerance for zero pivot 2.22045e-14 > using diagonal shift on blocks to prevent zero pivot [INBLOCKS] > matrix ordering: natural > linear system matrix = precond matrix: > Mat Object: (flow_) 1 MPI processes > type: seqbaij > rows=1389, cols=1389, bs=3 > total: nonzeros=20025, allocated nonzeros=20025 > total number of mallocs used during MatSetValues calls =0 > block size is 3 > - - - - - - - - - - - - - - - - - - > > -- > HeeHo Daniel Park
Re: [petsc-users] subprocess (block) tolerance.
The subdomain KSP (flow_sub_) has type "preonly" so it always does exactly one iteration. If you were to use an iterative subdomain solver (e.g., -flow_sub_ksp_type gmres) then those tolerances would be used. HeeHo Park writes: > Hi, I'm running PFLOTRAN and in PFLOTRAN, we have flow_ and flow_sub_ > processes. I was wondering what the red underlined values meant (each block > tolerance?) and how to change them (would it affect convergence?). Blue > marked bold values are changed from the default values for linear solvers. > > FLOW Linear Solver >solver: bcgs >preconditioner: asm > *atol: 1.00E-10* > * rtol: 1.00E-10* > dtol: 1.00E+04 > maximum iteration: 1 > KSP Object: (flow_) 8 MPI processes > type: bcgs > maximum iterations=1, initial guess is zero > tolerances: * relative=1e-10, absolute=1e-10*, divergence=1. > left preconditioning > using PRECONDITIONED norm type for convergence test > PC Object: (flow_) 8 MPI processes > type: asm > Additive Schwarz: total subdomain blocks = 8, amount of overlap = 1 > Additive Schwarz: restriction/interpolation type - RESTRICT > [0] number of local blocks = 1 > [1] number of local blocks = 1 > [2] number of local blocks = 1 > [3] number of local blocks = 1 > [4] number of local blocks = 1 > [5] number of local blocks = 1 > [6] number of local blocks = 1 > [7] number of local blocks = 1 > Local solve info for each block is in the following KSP and PC objects: > - - - - - - - - - - - - - - - - - - > [0] local block number 0, size = 1389 > KSP Object: (flow_sub_) 1 MPI processes > type: preonly > maximum iterations=1, initial guess is zero tolerances: *relative=1e-05, absolute=1e-50*, divergence=1. > left preconditioning > using DEFAULT norm type for convergence test > PC Object: (flow_sub_) 1 MPI processes > type: ilu > PC has not been set up so information may be incomplete > out-of-place factorization > 0 levels of fill > tolerance for zero pivot 2.22045e-14 > using diagonal shift on blocks to prevent zero pivot [INBLOCKS] > matrix ordering: natural > linear system matrix = precond matrix: > Mat Object: (flow_) 1 MPI processes > type: seqbaij > rows=1389, cols=1389, bs=3 > total: nonzeros=20025, allocated nonzeros=20025 > total number of mallocs used during MatSetValues calls =0 > block size is 3 > - - - - - - - - - - - - - - - - - - > > -- > HeeHo Daniel Park
