1KSP Object: 8 MPI processes 2 type: cg 3 maximum iterations=10000, initial guess is zero 4 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 5 left preconditioning 6 using PRECONDITIONED norm type for convergence test 7PC Object: 8 MPI processes 8 type: bddc 9 Use verbose output: 0 10 Use user-defined CSR: 0 11 Use local mat graph: 1 12 Connectivity graph topological dimension: 3 13 Corner selection: 1 (selected 1) 14 Use vertices: 1 (vertex size 3) 15 Use edges: 1 16 Use faces: 1 17 Use true near null space: 0 18 Use QR for single constraints on cc: 0 19 Use change of basis on local edge nodes: 0 20 Use change of basis on local face nodes: 0 21 User defined change of basis matrix: 0 22 Has change of basis matrix: 0 23 Eliminate dirichlet boundary dofs: 0 24 Switch on static condensation ops around the interface preconditioner: 0 25 Use exact dirichlet trick: 1 26 Interface extension: DIRICHLET 27 Multilevel max levels: 2 28 Multilevel coarsening ratio: 2 29 Use estimated eigs for coarse problem: 0 30 Use deluxe scaling: 0 31 Use deluxe zerorows: 0 32 Use deluxe singlemat: 0 33 Rebuild interface graph for Schur principal minors: 0 34 Number of dofs' layers for the computation of principal minors: -1 35 Use user CSR graph to compute successive layers: 0 36 Adaptive constraint selection threshold (active 0, userdefined 0): 0. 37 Min constraints / connected component: 0 38 Max constraints / connected component: 0 39 Invert exact Schur complement for adaptive selection: 0 40 Symmetric computation of primal basis functions: 1 41 Num. Procs. to map coarse adjacency list: 0 42 Coarse eqs per proc (significant at the coarsest level): 1 43 Detect disconnected: 0 (filter 0) 44 Benign subspace trick: 0 (change explicit 0) 45 Benign subspace trick is active: 0 46 Algebraic computation of no-net-flux: 0 47 ********************************** STATISTICS AT LEVEL 0 ********************************** 48 Global dofs sizes: all 2160 interface 648 coarse 159 49 Coarsening ratios: all/coarse 13 interface/coarse 4 50 Active processes : 8 51 Total subdomains : 8 52 Dofs type : MIN MAX MEAN 53 Interior dofs : 144 240 189 54 Interface dofs : 156 210 182 55 Primal dofs : 54 54 54 56 Local dofs : 300 450 371 57 Local subs : 1 1 58 --- Interior solver (rank 0) 59 KSP Object: (pc_bddc_dirichlet_) 1 MPI process 60 type: preonly 61 maximum iterations=10000, initial guess is zero 62 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 63 left preconditioning 64 not checking for convergence 65 PC Object: (pc_bddc_dirichlet_) 1 MPI process 66 type: lu 67 out-of-place factorization 68 tolerance for zero pivot 2.22045e-14 69 matrix ordering: nd 70 factor fill ratio given 5., needed 1.55143 71 Factored matrix: 72 Mat Object: (pc_bddc_dirichlet_) 1 MPI process 73 type: seqaij 74 rows=144, cols=144, bs=3 75 package used to perform factorization: petsc 76 total: nonzeros=9774, allocated nonzeros=9774 77 using I-node routines: found 45 nodes, limit used is 5 78 linear system matrix, which is also used to construct the preconditioner: 79 Mat Object: (pc_bddc_dirichlet_) 1 MPI process 80 type: seqaij 81 rows=144, cols=144, bs=3 82 total: nonzeros=6300, allocated nonzeros=6300 83 total number of mallocs used during MatSetValues calls=0 84 using I-node routines: found 48 nodes, limit used is 5 85 --- Correction solver (rank 0) 86 KSP Object: (pc_bddc_neumann_) 1 MPI process 87 type: preonly 88 maximum iterations=10000, initial guess is zero 89 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 90 left preconditioning 91 not checking for convergence 92 PC Object: (pc_bddc_neumann_) 1 MPI process 93 type: lu 94 out-of-place factorization 95 tolerance for zero pivot 2.22045e-14 96 matrix ordering: nd 97 factor fill ratio given 5., needed 2.16719 98 Factored matrix: 99 Mat Object: (pc_bddc_neumann_) 1 MPI process 100 type: seqaij 101 rows=279, cols=279, bs=3 102 package used to perform factorization: petsc 103 total: nonzeros=30915, allocated nonzeros=30915 104 using I-node routines: found 85 nodes, limit used is 5 105 linear system matrix, which is also used to construct the preconditioner: 106 Mat Object: (pc_bddc_neumann_) 1 MPI process 107 type: seqaij 108 rows=279, cols=279, bs=3 109 total: nonzeros=14265, allocated nonzeros=14265 110 total number of mallocs used during MatSetValues calls=0 111 using I-node routines: found 93 nodes, limit used is 5 112 --- Coarse solver 113 KSP Object: (pc_bddc_coarse_) 8 MPI processes 114 type: richardson 115 damping factor=1. 116 maximum iterations=1, initial guess is zero 117 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 118 left preconditioning 119 not checking for convergence 120 PC Object: (pc_bddc_coarse_) 8 MPI processes 121 type: bddc 122 Use verbose output: 0 123 Use user-defined CSR: 0 124 Use local mat graph: 1 125 Connectivity graph topological dimension: 3 126 Corner selection: 1 (selected 1) 127 Use vertices: 1 (vertex size 1) 128 Use edges: 1 129 Use faces: 0 130 Use true near null space: 0 131 Use QR for single constraints on cc: 0 132 Use change of basis on local edge nodes: 0 133 Use change of basis on local face nodes: 0 134 User defined change of basis matrix: 0 135 Has change of basis matrix: 0 136 Eliminate dirichlet boundary dofs: 0 137 Switch on static condensation ops around the interface preconditioner: 0 138 Use exact dirichlet trick: 0 139 Interface extension: DIRICHLET 140 Multilevel max levels: 2 141 Multilevel coarsening ratio: 2 142 Use estimated eigs for coarse problem: 0 143 Use deluxe scaling: 0 144 Use deluxe zerorows: 0 145 Use deluxe singlemat: 0 146 Rebuild interface graph for Schur principal minors: 0 147 Number of dofs' layers for the computation of principal minors: -1 148 Use user CSR graph to compute successive layers: 0 149 Adaptive constraint selection threshold (active 0, userdefined 0): 0. 150 Min constraints / connected component: 0 151 Max constraints / connected component: 0 152 Invert exact Schur complement for adaptive selection: 0 153 Symmetric computation of primal basis functions: 1 154 Num. Procs. to map coarse adjacency list: 0 155 Coarse eqs per proc (significant at the coarsest level): 1 156 Detect disconnected: 1 (filter 0) 157 Benign subspace trick: 0 (change explicit 0) 158 Benign subspace trick is active: 0 159 Algebraic computation of no-net-flux: 0 160 ********************************** STATISTICS AT LEVEL 1 ********************************** 161 Global dofs sizes: all 159 interface 123 coarse 31 162 Coarsening ratios: all/coarse 5 interface/coarse 3 163 Active processes : 4 164 Total subdomains : 4 165 Dofs type : MIN MAX MEAN 166 Interior dofs : 9 9 9 167 Interface dofs : 71 71 71 168 Primal dofs : 19 19 19 169 Local dofs : 80 80 80 170 Local subs : 1 1 171 --- Interior solver (rank 0) 172 KSP Object: (pc_bddc_dirichlet_l1_) 1 MPI process 173 type: preonly 174 maximum iterations=10000, initial guess is zero 175 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 176 left preconditioning 177 not checking for convergence 178 PC Object: (pc_bddc_dirichlet_l1_) 1 MPI process 179 type: lu 180 out-of-place factorization 181 tolerance for zero pivot 2.22045e-14 182 matrix ordering: nd 183 factor fill ratio given 5., needed 1. 184 Factored matrix: 185 Mat Object: (pc_bddc_dirichlet_l1_) 1 MPI process 186 type: seqaij 187 rows=9, cols=9 188 package used to perform factorization: petsc 189 total: nonzeros=81, allocated nonzeros=81 190 using I-node routines: found 2 nodes, limit used is 5 191 linear system matrix, which is also used to construct the preconditioner: 192 Mat Object: (pc_bddc_dirichlet_l1_) 1 MPI process 193 type: seqaij 194 rows=9, cols=9 195 total: nonzeros=81, allocated nonzeros=81 196 total number of mallocs used during MatSetValues calls=0 197 using I-node routines: found 2 nodes, limit used is 5 198 --- Correction solver (rank 0) 199 KSP Object: (pc_bddc_neumann_l1_) 1 MPI process 200 type: preonly 201 maximum iterations=10000, initial guess is zero 202 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 203 left preconditioning 204 not checking for convergence 205 PC Object: (pc_bddc_neumann_l1_) 1 MPI process 206 type: lu 207 out-of-place factorization 208 tolerance for zero pivot 2.22045e-14 209 matrix ordering: nd 210 factor fill ratio given 5., needed 1. 211 Factored matrix: 212 Mat Object: (pc_bddc_neumann_l1_) 1 MPI process 213 type: seqaij 214 rows=62, cols=62 215 package used to perform factorization: petsc 216 total: nonzeros=3266, allocated nonzeros=3266 217 using I-node routines: found 14 nodes, limit used is 5 218 linear system matrix, which is also used to construct the preconditioner: 219 Mat Object: (pc_bddc_neumann_l1_) 1 MPI process 220 type: seqaij 221 rows=62, cols=62 222 total: nonzeros=3266, allocated nonzeros=3266 223 total number of mallocs used during MatSetValues calls=0 224 using I-node routines: found 16 nodes, limit used is 5 225 linear system matrix, which is also used to construct the preconditioner: 226 Mat Object: (pc_bddc_coarse_) 8 MPI processes 227 type: is 228 rows=159, cols=159 229 total: nonzeros=20192, allocated nonzeros=20192 230 total number of mallocs used during MatSetValues calls=0 231 --- Coarse solver 232 KSP Object: (pc_bddc_coarse_l1_) 4 MPI processes 233 type: richardson 234 damping factor=1. 235 maximum iterations=1, initial guess is zero 236 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 237 left preconditioning 238 not checking for convergence 239 PC Object: (pc_bddc_coarse_l1_) 4 MPI processes 240 type: bddc 241 Use verbose output: 0 242 Use user-defined CSR: 0 243 Use local mat graph: 1 244 Connectivity graph topological dimension: 2 245 Corner selection: 1 (selected 1) 246 Use vertices: 1 (vertex size 1) 247 Use edges: 1 248 Use faces: 0 249 Use true near null space: 0 250 Use QR for single constraints on cc: 0 251 Use change of basis on local edge nodes: 0 252 Use change of basis on local face nodes: 0 253 User defined change of basis matrix: 0 254 Has change of basis matrix: 0 255 Eliminate dirichlet boundary dofs: 0 256 Switch on static condensation ops around the interface preconditioner: 0 257 Use exact dirichlet trick: 0 258 Interface extension: DIRICHLET 259 Multilevel max levels: 2 260 Multilevel coarsening ratio: 2 261 Use estimated eigs for coarse problem: 0 262 Use deluxe scaling: 0 263 Use deluxe zerorows: 0 264 Use deluxe singlemat: 0 265 Rebuild interface graph for Schur principal minors: 0 266 Number of dofs' layers for the computation of principal minors: -1 267 Use user CSR graph to compute successive layers: 0 268 Adaptive constraint selection threshold (active 0, userdefined 0): 0. 269 Min constraints / connected component: 0 270 Max constraints / connected component: 0 271 Invert exact Schur complement for adaptive selection: 0 272 Symmetric computation of primal basis functions: 1 273 Num. Procs. to map coarse adjacency list: 0 274 Coarse eqs per proc (significant at the coarsest level): 1 275 Detect disconnected: 1 (filter 0) 276 Benign subspace trick: 0 (change explicit 0) 277 Benign subspace trick is active: 0 278 Algebraic computation of no-net-flux: 0 279 ********************************** STATISTICS AT LEVEL 2 ********************************** 280 Global dofs sizes: all 31 interface 19 coarse 13 281 Coarsening ratios: all/coarse 2 interface/coarse 1 282 Active processes : 2 283 Total subdomains : 2 284 Dofs type : MIN MAX MEAN 285 Interior dofs : 6 6 6 286 Interface dofs : 19 19 19 287 Primal dofs : 13 13 13 288 Local dofs : 25 25 25 289 Local subs : 1 1 290 --- Interior solver (rank 0) 291 KSP Object: (pc_bddc_dirichlet_l2_) 1 MPI process 292 type: preonly 293 maximum iterations=10000, initial guess is zero 294 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 295 left preconditioning 296 not checking for convergence 297 PC Object: (pc_bddc_dirichlet_l2_) 1 MPI process 298 type: lu 299 out-of-place factorization 300 tolerance for zero pivot 2.22045e-14 301 matrix ordering: nd 302 factor fill ratio given 5., needed 1. 303 Factored matrix: 304 Mat Object: (pc_bddc_dirichlet_l2_) 1 MPI process 305 type: seqaij 306 rows=6, cols=6 307 package used to perform factorization: petsc 308 total: nonzeros=36, allocated nonzeros=36 309 using I-node routines: found 2 nodes, limit used is 5 310 linear system matrix, which is also used to construct the preconditioner: 311 Mat Object: (pc_bddc_dirichlet_l2_) 1 MPI process 312 type: seqaij 313 rows=6, cols=6 314 total: nonzeros=36, allocated nonzeros=36 315 total number of mallocs used during MatSetValues calls=0 316 using I-node routines: found 2 nodes, limit used is 5 317 --- Correction solver (rank 0) 318 KSP Object: (pc_bddc_neumann_l2_) 1 MPI process 319 type: preonly 320 maximum iterations=10000, initial guess is zero 321 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 322 left preconditioning 323 not checking for convergence 324 PC Object: (pc_bddc_neumann_l2_) 1 MPI process 325 type: lu 326 out-of-place factorization 327 tolerance for zero pivot 2.22045e-14 328 matrix ordering: nd 329 factor fill ratio given 5., needed 1. 330 Factored matrix: 331 Mat Object: (pc_bddc_neumann_l2_) 1 MPI process 332 type: seqaij 333 rows=13, cols=13 334 package used to perform factorization: petsc 335 total: nonzeros=169, allocated nonzeros=169 336 using I-node routines: found 3 nodes, limit used is 5 337 linear system matrix, which is also used to construct the preconditioner: 338 Mat Object: (pc_bddc_neumann_l2_) 1 MPI process 339 type: seqaij 340 rows=13, cols=13 341 total: nonzeros=169, allocated nonzeros=169 342 total number of mallocs used during MatSetValues calls=0 343 using I-node routines: found 3 nodes, limit used is 5 344 linear system matrix, which is also used to construct the preconditioner: 345 Mat Object: (pc_bddc_coarse_l1_) 4 MPI processes 346 type: is 347 rows=31, cols=31 348 total: nonzeros=1106, allocated nonzeros=1106 349 total number of mallocs used during MatSetValues calls=0 350 --- Coarse solver 351 KSP Object: (pc_bddc_coarse_l2_) 2 MPI processes 352 type: preonly 353 maximum iterations=1, initial guess is zero 354 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 355 left preconditioning 356 not checking for convergence 357 PC Object: (pc_bddc_coarse_l2_) 2 MPI processes 358 type: redundant 359 First (color=0) of 2 PCs follows 360 KSP Object: (pc_bddc_coarse_l2_redundant_) 1 MPI process 361 type: preonly 362 maximum iterations=10000, initial guess is zero 363 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 364 left preconditioning 365 not checking for convergence 366 PC Object: (pc_bddc_coarse_l2_redundant_) 1 MPI process 367 type: lu 368 out-of-place factorization 369 tolerance for zero pivot 2.22045e-14 370 matrix ordering: nd 371 factor fill ratio given 5., needed 1. 372 Factored matrix: 373 Mat Object: (pc_bddc_coarse_l2_redundant_) 1 MPI process 374 type: seqaij 375 rows=13, cols=13 376 package used to perform factorization: petsc 377 total: nonzeros=169, allocated nonzeros=169 378 using I-node routines: found 3 nodes, limit used is 5 379 linear system matrix, which is also used to construct the preconditioner: 380 Mat Object: 1 MPI process 381 type: seqaij 382 rows=13, cols=13 383 total: nonzeros=169, allocated nonzeros=169 384 total number of mallocs used during MatSetValues calls=0 385 using I-node routines: found 3 nodes, limit used is 5 386 linear system matrix, which is also used to construct the preconditioner: 387 Mat Object: (pc_bddc_coarse_l2_) 2 MPI processes 388 type: mpiaij 389 rows=13, cols=13 390 total: nonzeros=169, allocated nonzeros=169 391 total number of mallocs used during MatSetValues calls=0 392 using I-node (on process 0) routines: found 2 nodes, limit used is 5 393 linear system matrix, which is also used to construct the preconditioner: 394 Mat Object: 8 MPI processes 395 type: is 396 rows=2160, cols=2160, bs=3 397 total: nonzeros=156078, allocated nonzeros=156078 398 total number of mallocs used during MatSetValues calls=0 399 has attached near null space 400