xref: /petsc/src/snes/tutorials/output/ex12_tri_parmetis_bddc.out (revision c87ba875e4007ad659b117ea274f03d5f4cd5ea7)
1  0 SNES Function norm 5.83179
2    0 KSP Residual norm 5.0581
3    1 KSP Residual norm 0.0106943
4    2 KSP Residual norm 7.46624e-07
5    3 KSP Residual norm < 1.e-11
6  Linear solve converged due to CONVERGED_RTOL iterations 3
7  1 SNES Function norm < 1.e-11
8L_2 Error: 0.00527027
9Nonlinear solve converged due to CONVERGED_FNORM_RELATIVE iterations 1
10SNES Object: 4 MPI processes
11  type: newtonls
12  maximum iterations=50, maximum function evaluations=10000
13  tolerances: relative=1e-08, absolute=1e-50, solution=1e-08
14  total number of linear solver iterations=3
15  total number of function evaluations=2
16  norm schedule ALWAYS
17  SNESLineSearch Object: 4 MPI processes
18    type: bt
19      interpolation: cubic
20      alpha=1.000000e-04
21    maxstep=1.000000e+08, minlambda=1.000000e-12
22    tolerances: relative=1.000000e-08, absolute=1.000000e-15, lambda=1.000000e-08
23    maximum iterations=40
24  KSP Object: 4 MPI processes
25    type: gmres
26      restart=100, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement
27      happy breakdown tolerance 1e-30
28    maximum iterations=10000, initial guess is zero
29    tolerances:  relative=1e-09, absolute=1e-50, divergence=10000.
30    left preconditioning
31    using PRECONDITIONED norm type for convergence test
32  PC Object: 4 MPI processes
33    type: bddc
34      Use verbose output: 0
35      Use user-defined CSR: 0
36      Use local mat graph: 1
37      Connectivity graph topological dimension: 2
38      Use vertices: 1 (vertex size 1)
39      Use edges: 1
40      Use faces: 0
41      Use true near null space: 0
42      Use QR for single constraints on cc: 0
43      Use change of basis on local edge nodes: 0
44      Use change of basis on local face nodes: 0
45      User defined change of basis matrix: 0
46      Has change of basis matrix: 0
47      Eliminate dirichlet boundary dofs: 0
48      Switch on static condensation ops around the interface preconditioner: 0
49      Use exact dirichlet trick: 0
50      Interface extension: DIRICHLET
51      Multilevel max levels: 0
52      Multilevel coarsening ratio: 8
53      Use estimated eigs for coarse problem: 0
54      Use deluxe scaling: 0
55      Use deluxe zerorows: 0
56      Use deluxe singlemat: 0
57      Rebuild interface graph for Schur principal minors: 0
58      Number of dofs' layers for the computation of principal minors: -1
59      Use user CSR graph to compute successive layers: 0
60      Adaptive constraint selection threshold (active 0, userdefined 0): 0.
61      Min constraints / connected component: 0
62      Max constraints / connected component: 0
63      Invert exact Schur complement for adaptive selection: 0
64      Symmetric computation of primal basis functions: 1
65      Num. Procs. to map coarse adjacency list: 0
66      Coarse eqs per proc (significant at the coarsest level): 1
67      Detect disconnected: 0 (filter 0)
68      Benign subspace trick: 0 (change explicit 0)
69      Benign subspace trick is active: 0
70      Algebraic computation of no-net-flux: 0
71    ********************************** STATISTICS AT LEVEL 0 **********************************
72      Global dofs sizes: all 49 interface 13 coarse 5
73      Coarsening ratios: all/coarse 9 interface/coarse 2
74      Active processes : 4
75      Total subdomains : 4
76      Dofs type        :	MIN	MAX	MEAN
77      Interior  dofs   :	9	9	9
78      Interface dofs   :	7	7	7
79      Primal    dofs   :	3	3	3
80      Local     dofs   :	16	16	16
81      Local     subs   :	1	1
82    --- Interior solver (rank 0)
83      KSP Object: (pc_bddc_dirichlet_) 1 MPI processes
84        type: preonly
85        maximum iterations=10000, initial guess is zero
86        tolerances:  relative=1e-05, absolute=1e-50, divergence=10000.
87        left preconditioning
88        using NONE norm type for convergence test
89      PC Object: (pc_bddc_dirichlet_) 1 MPI processes
90        type: lu
91          out-of-place factorization
92          tolerance for zero pivot 2.22045e-14
93          matrix ordering: nd
94          factor fill ratio given 5., needed 1.43902
95            Factored matrix follows:
96              Mat Object: 1 MPI processes
97                type: seqaij
98                rows=9, cols=9
99                package used to perform factorization: petsc
100                total: nonzeros=59, allocated nonzeros=59
101                total number of mallocs used during MatSetValues calls=0
102                  using I-node routines: found 7 nodes, limit used is 5
103        linear system matrix = precond matrix:
104        Mat Object: (pc_bddc_dirichlet_) 1 MPI processes
105          type: seqaij
106          rows=9, cols=9
107          total: nonzeros=41, allocated nonzeros=41
108          total number of mallocs used during MatSetValues calls=0
109            not using I-node routines
110    --- Correction solver (rank 0)
111      KSP Object: (pc_bddc_neumann_) 1 MPI processes
112        type: preonly
113        maximum iterations=10000, initial guess is zero
114        tolerances:  relative=1e-05, absolute=1e-50, divergence=10000.
115        left preconditioning
116        using NONE norm type for convergence test
117      PC Object: (pc_bddc_neumann_) 1 MPI processes
118        type: lu
119          out-of-place factorization
120          tolerance for zero pivot 2.22045e-14
121          matrix ordering: nd
122          factor fill ratio given 5., needed 1.29333
123            Factored matrix follows:
124              Mat Object: 1 MPI processes
125                type: seqaij
126                rows=15, cols=15
127                package used to perform factorization: petsc
128                total: nonzeros=97, allocated nonzeros=97
129                total number of mallocs used during MatSetValues calls=0
130                  not using I-node routines
131        linear system matrix = precond matrix:
132        Mat Object: (pc_bddc_neumann_) 1 MPI processes
133          type: seqaij
134          rows=15, cols=15
135          total: nonzeros=75, allocated nonzeros=75
136          total number of mallocs used during MatSetValues calls=0
137            not using I-node routines
138    --- Coarse solver
139      KSP Object: (pc_bddc_coarse_) 4 MPI processes
140        type: preonly
141        maximum iterations=1, initial guess is zero
142        tolerances:  relative=1e-05, absolute=1e-50, divergence=10000.
143        left preconditioning
144        using NONE norm type for convergence test
145      PC Object: (pc_bddc_coarse_) 4 MPI processes
146        type: redundant
147          First (color=0) of 4 PCs follows
148        linear system matrix = precond matrix:
149        Mat Object: (pc_bddc_coarse_) 4 MPI processes
150          type: mpiaij
151          rows=5, cols=5
152          total: nonzeros=21, allocated nonzeros=25
153          total number of mallocs used during MatSetValues calls=0
154            using I-node (on process 0) routines: found 1 nodes, limit used is 5
155              KSP Object:     (pc_bddc_coarse_redundant_)     1 MPI processes
156                type: preonly
157                maximum iterations=10000, initial guess is zero
158                tolerances:  relative=1e-05, absolute=1e-50, divergence=10000.
159                left preconditioning
160                using NONE norm type for convergence test
161              PC Object:     (pc_bddc_coarse_redundant_)     1 MPI processes
162                type: lu
163                  out-of-place factorization
164                  tolerance for zero pivot 2.22045e-14
165                  matrix ordering: nd
166                  factor fill ratio given 5., needed 1.09524
167                    Factored matrix follows:
168                      Mat Object:     1 MPI processes
169                        type: seqaij
170                        rows=5, cols=5
171                        package used to perform factorization: petsc
172                        total: nonzeros=23, allocated nonzeros=23
173                        total number of mallocs used during MatSetValues calls=0
174                          using I-node routines: found 3 nodes, limit used is 5
175                linear system matrix = precond matrix:
176                Mat Object:     1 MPI processes
177                  type: seqaij
178                  rows=5, cols=5
179                  total: nonzeros=21, allocated nonzeros=21
180                  total number of mallocs used during MatSetValues calls=0
181                    not using I-node routines
182    linear system matrix = precond matrix:
183    Mat Object: 4 MPI processes
184      type: is
185      rows=49, cols=49
186      total: nonzeros=328, allocated nonzeros=386
187      total number of mallocs used during MatSetValues calls=0
188