xref: /petsc/src/snes/tutorials/output/ex69_q2p1fetidp.out (revision d5c9c0c4eebc2f2a01a1bd0c86fca87e2acd2a03)
1DM Object: Distributed Mesh 5 MPI processes
2  type: plex
3Distributed Mesh in 2 dimensions:
4  0-cells: 15 18 15 15 9
5  1-cells: 22 24 22 22 12
6  2-cells: 8 8 8 8 4
7Labels:
8  celltype: 3 strata with value/size (0 (15), 1 (22), 4 (8))
9  depth: 3 strata with value/size (0 (15), 1 (22), 2 (8))
10  marker: 2 strata with value/size (1 (9), 4 (5))
11  Face Sets: 2 strata with value/size (1 (6), 4 (3))
12  markerBottom: 1 strata with value/size (1 (9))
13  markerRight: 0 strata with value/size ()
14  markerTop: 0 strata with value/size ()
15  markerLeft: 1 strata with value/size (1 (5))
16Integral of pressure: -0.0387881
17SNES Object: 5 MPI processes
18  type: newtonls
19  tolerances: relative=1e-08, absolute=1e-50, solution=1e-08
20  norm schedule ALWAYS
21  SNESLineSearch Object: 5 MPI processes
22    type: bt
23      interpolation: cubic
24      alpha=1.000000e-04
25    maxstep=1.000000e+08, minlambda=1.000000e-12
26    tolerances: relative=1.000000e-08, absolute=1.000000e-15, lambda=1.000000e-08
27    maximum iterations=40
28  KSP Object: 5 MPI processes
29    type: fetidp
30      fully redundant: 0
31      saddle point:    1
32    Inner KSP solver details
33      KSP Object: (fetidp_) 5 MPI processes
34        type: cg
35        maximum iterations=10000, initial guess is zero
36        tolerances:  relative=1e-12, absolute=1e-10, divergence=10000.
37        left preconditioning
38        using NATURAL norm type for convergence test
39      PC Object: (fetidp_) 5 MPI processes
40        type: shell
41          FETI-DP multipliers
42            Dirichlet preconditioner (just from rank 0)
43            Mat Object: 1 MPI processes
44              type: schurcomplement
45              rows=24, cols=24
46                Schur complement A11 - A10 inv(A00) A01
47                A11
48                  Mat Object: 1 MPI processes
49                    type: seqaij
50                    rows=24, cols=24
51                    total: nonzeros=200, allocated nonzeros=200
52                    total number of mallocs used during MatSetValues calls=0
53                      using I-node routines: found 16 nodes, limit used is 5
54                A10
55                  Mat Object: 1 MPI processes
56                    type: seqaij
57                    rows=24, cols=76
58                    total: nonzeros=470, allocated nonzeros=470
59                    total number of mallocs used during MatSetValues calls=0
60                      not using I-node routines
61                KSP of A00
62                  KSP Object: (fetidp_bddc_pc_bddc_dirichlet_) 1 MPI processes
63                    type: preonly
64                    maximum iterations=10000, initial guess is zero
65                    tolerances:  relative=1e-05, absolute=1e-50, divergence=10000.
66                    left preconditioning
67                    using NONE norm type for convergence test
68                  PC Object: (fetidp_bddc_pc_bddc_dirichlet_) 1 MPI processes
69                    type: svd
70                      All singular values smaller than 1e-12 treated as zero
71                      Provided essential rank of the matrix 0 (all other eigenvalues are zeroed)
72                    linear system matrix = precond matrix:
73                    Mat Object: (fetidp_bddc_pc_bddc_dirichlet_) 1 MPI processes
74                      type: seqaij
75                      rows=76, cols=76
76                      total: nonzeros=1576, allocated nonzeros=1576
77                      total number of mallocs used during MatSetValues calls=0
78                        not using I-node routines
79                A01
80                  Mat Object: 1 MPI processes
81                    type: seqaij
82                    rows=76, cols=24
83                    total: nonzeros=470, allocated nonzeros=470
84                    total number of mallocs used during MatSetValues calls=0
85                      using I-node routines: found 48 nodes, limit used is 5
86        linear system matrix = precond matrix:
87        Mat Object: F (fetidp_) 5 MPI processes
88          type: shell
89          rows=60, cols=60
90    Inner BDDC solver details
91      PC Object: (fetidp_bddc_) 5 MPI processes
92        type: bddc
93          Use verbose output: 0
94          Use user-defined CSR: 0
95          Use local mat graph: 1
96          Connectivity graph topological dimension: 2
97          Graph max count: 2
98          Use vertices: 1 (vertex size 3)
99          Use edges: 1
100          Use faces: 1
101          Use true near null space: 0
102          Use QR for single constraints on cc: 0
103          Use change of basis on local edge nodes: 0
104          Use change of basis on local face nodes: 0
105          User defined change of basis matrix: 0
106          Has change of basis matrix: 1
107          Eliminate dirichlet boundary dofs: 1
108          Switch on static condensation ops around the interface preconditioner: 0
109          Use exact dirichlet trick: 1
110          Interface extension: DIRICHLET
111          Multilevel max levels: 0
112          Multilevel coarsening ratio: 8
113          Use estimated eigs for coarse problem: 0
114          Use deluxe scaling: 0
115          Use deluxe zerorows: 0
116          Use deluxe singlemat: 0
117          Rebuild interface graph for Schur principal minors: 0
118          Number of dofs' layers for the computation of principal minors: -1
119          Use user CSR graph to compute successive layers: 0
120          Adaptive constraint selection threshold (active 0, userdefined 0): 0.
121          Min constraints / connected component: 0
122          Max constraints / connected component: 0
123          Invert exact Schur complement for adaptive selection: 0
124          Symmetric computation of primal basis functions: 1
125          Num. Procs. to map coarse adjacency list: 0
126          Coarse eqs per proc (significant at the coarsest level): 1
127          Detect disconnected: 1 (filter 0)
128          Benign subspace trick: 1 (change explicit 1)
129          Benign subspace trick is active: 1
130          Algebraic computation of no-net-flux: 1
131        ********************************** STATISTICS AT LEVEL 0 **********************************
132          Global dofs sizes: all 394 interface 68 coarse 23
133          Coarsening ratios: all/coarse 17 interface/coarse 2
134          Active processes : 5
135          Total subdomains : 6
136          Benign subs      : 6
137          Dofs type        :	MIN	MAX	MEAN
138          Interior  dofs   :	36	76	65
139          Interface dofs   :	16	40	28
140          Primal    dofs   :	5	14	9
141          Local     dofs   :	52	109	94
142          Local     subs   :	1	2
143        --- Interior solver (rank 0)
144          KSP Object: (fetidp_bddc_pc_bddc_dirichlet_) 1 MPI processes
145            type: preonly
146            maximum iterations=10000, initial guess is zero
147            tolerances:  relative=1e-05, absolute=1e-50, divergence=10000.
148            left preconditioning
149            using NONE norm type for convergence test
150          PC Object: (fetidp_bddc_pc_bddc_dirichlet_) 1 MPI processes
151            type: svd
152              All singular values smaller than 1e-12 treated as zero
153              Provided essential rank of the matrix 0 (all other eigenvalues are zeroed)
154            linear system matrix = precond matrix:
155            Mat Object: (fetidp_bddc_pc_bddc_dirichlet_) 1 MPI processes
156              type: seqaij
157              rows=76, cols=76
158              total: nonzeros=1576, allocated nonzeros=1576
159              total number of mallocs used during MatSetValues calls=0
160                not using I-node routines
161        --- Correction solver (rank 0)
162          KSP Object: (fetidp_bddc_pc_bddc_neumann_) 1 MPI processes
163            type: preonly
164            maximum iterations=1000, initial guess is zero
165            tolerances:  relative=1e-10, absolute=1e-10, divergence=10000.
166            left preconditioning
167            using NONE norm type for convergence test
168          PC Object: (fetidp_bddc_pc_bddc_neumann_) 1 MPI processes
169            type: svd
170              All singular values smaller than 1e-12 treated as zero
171              Provided essential rank of the matrix 0 (all other eigenvalues are zeroed)
172            linear system matrix = precond matrix:
173            Mat Object: (fetidp_bddc_pc_bddc_neumann_) 1 MPI processes
174              type: seqaij
175              rows=96, cols=96
176              total: nonzeros=2424, allocated nonzeros=2424
177              total number of mallocs used during MatSetValues calls=0
178                not using I-node routines
179        --- Coarse solver
180          KSP Object: (fetidp_bddc_pc_bddc_coarse_) 5 MPI processes
181            type: preonly
182            maximum iterations=1000, initial guess is zero
183            tolerances:  relative=1e-10, absolute=1e-10, divergence=10000.
184            left preconditioning
185            using NONE norm type for convergence test
186          PC Object: (fetidp_bddc_pc_bddc_coarse_) 5 MPI processes
187            type: redundant
188              First (color=0) of 5 PCs follows
189            linear system matrix = precond matrix:
190            Mat Object: (fetidp_bddc_pc_bddc_coarse_) 5 MPI processes
191              type: mpiaij
192              rows=23, cols=23
193              total: nonzeros=365, allocated nonzeros=414
194              total number of mallocs used during MatSetValues calls=0
195                has attached null space
196                using I-node (on process 0) routines: found 3 nodes, limit used is 5
197                      KSP Object:         (fetidp_bddc_pc_bddc_coarse_redundant_)         1 MPI processes
198                        type: preonly
199                        maximum iterations=10000, initial guess is zero
200                        tolerances:  relative=1e-05, absolute=1e-50, divergence=10000.
201                        left preconditioning
202                        using NONE norm type for convergence test
203                      PC Object:         (fetidp_bddc_pc_bddc_coarse_redundant_)         1 MPI processes
204                        type: svd
205                          All singular values smaller than 1e-12 treated as zero
206                          Provided essential rank of the matrix 0 (all other eigenvalues are zeroed)
207                        linear system matrix = precond matrix:
208                        Mat Object:         1 MPI processes
209                          type: seqaij
210                          rows=23, cols=23
211                          total: nonzeros=365, allocated nonzeros=365
212                          total number of mallocs used during MatSetValues calls=0
213                            using I-node routines: found 16 nodes, limit used is 5
214        linear system matrix = precond matrix:
215        Mat Object: 5 MPI processes
216          type: is
217          rows=394, cols=394
218          total: nonzeros=11830, allocated nonzeros=11830
219          total number of mallocs used during MatSetValues calls=0
220            has attached near null space
221    maximum iterations=10000, initial guess is zero
222    tolerances:  relative=1e-12, absolute=1e-10, divergence=10000.
223    left preconditioning
224    using NONE norm type for convergence test
225  PC Object: 5 MPI processes
226    type: none
227    linear system matrix followed by preconditioner matrix:
228    Mat Object: 5 MPI processes
229      type: is
230      rows=394, cols=394
231      total: nonzeros=11830, allocated nonzeros=14730
232      total number of mallocs used during MatSetValues calls=0
233        has attached null space
234    Mat Object: 5 MPI processes
235      type: is
236      rows=394, cols=394
237      total: nonzeros=11830, allocated nonzeros=14730
238      total number of mallocs used during MatSetValues calls=0
239L_2 Error: 0.00418 [1.2e-05, 0.00418]
240Integral of pressure: 0.
241