10acf423eSBarry SmithKSP Object: 1 MPI process 20acf423eSBarry Smith type: gmres 3f971d498SPierre Jolivet restart=30, using classical (unmodified) Gram-Schmidt orthogonalization with no iterative refinement 4*143f2514SPierre Jolivet happy breakdown tolerance=1e-30 50acf423eSBarry Smith maximum iterations=10000, initial guess is zero 60acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 70acf423eSBarry Smith left preconditioning 80acf423eSBarry Smith using DEFAULT norm type for convergence test 90acf423eSBarry SmithPC Object: 1 MPI process 100acf423eSBarry Smith type: gamg 110acf423eSBarry Smith PC has not been set up so information may be incomplete 120acf423eSBarry Smith type is MULTIPLICATIVE, levels=0 cycles=unknown 130acf423eSBarry Smith Cycles per PCApply=0 140acf423eSBarry Smith Using externally compute Galerkin coarse grid matrices 150acf423eSBarry Smith GAMG specific options 160acf423eSBarry Smith Threshold for dropping small values in graph on each level = 170acf423eSBarry Smith Threshold scaling factor for each level not specified = 1. 180acf423eSBarry Smith AGG specific options 190acf423eSBarry Smith Number of levels of aggressive coarsening 1 200acf423eSBarry Smith Square graph aggressive coarsening 210acf423eSBarry Smith Coarsening algorithm not yet selected 220acf423eSBarry Smith Number smoothing steps to construct prolongation 1 230acf423eSBarry Smith Complexity: grid = 0. operator = 0. 246c34c54dSStefano Zampini Per-level complexity: op = operator, int = interpolation 256c34c54dSStefano Zampini #equations | #active PEs | avg nnz/row op | avg nnz/row int 26ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 270acf423eSBarry Smith Mat Object: 1 MPI process 280acf423eSBarry Smith type: seqaij 290acf423eSBarry Smith rows=16, cols=16 300acf423eSBarry Smith total: nonzeros=64, allocated nonzeros=64 310acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 320acf423eSBarry Smith not using I-node routines 330acf423eSBarry SmithKSP Object: 1 MPI process 340acf423eSBarry Smith type: gmres 35f971d498SPierre Jolivet restart=30, using classical (unmodified) Gram-Schmidt orthogonalization with no iterative refinement 36*143f2514SPierre Jolivet happy breakdown tolerance=1e-30 370acf423eSBarry Smith maximum iterations=10000, initial guess is zero 380acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 390acf423eSBarry Smith left preconditioning 400acf423eSBarry Smith using PRECONDITIONED norm type for convergence test 410acf423eSBarry SmithPC Object: 1 MPI process 420acf423eSBarry Smith type: gamg 430acf423eSBarry Smith type is MULTIPLICATIVE, levels=2 cycles=v 440acf423eSBarry Smith Cycles per PCApply=1 450acf423eSBarry Smith Using externally compute Galerkin coarse grid matrices 460acf423eSBarry Smith GAMG specific options 470acf423eSBarry Smith Threshold for dropping small values in graph on each level = -1. -1. 480acf423eSBarry Smith Threshold scaling factor for each level not specified = 1. 490acf423eSBarry Smith AGG specific options 500acf423eSBarry Smith Number of levels of aggressive coarsening 1 510acf423eSBarry Smith Square graph aggressive coarsening 520acf423eSBarry Smith MatCoarsen Object: (pc_gamg_) 1 MPI process 530acf423eSBarry Smith type: mis 540acf423eSBarry Smith Number smoothing steps to construct prolongation 1 550acf423eSBarry Smith Complexity: grid = 1.1875 operator = 1.14062 566c34c54dSStefano Zampini Per-level complexity: op = operator, int = interpolation 576c34c54dSStefano Zampini #equations | #active PEs | avg nnz/row op | avg nnz/row int 586c34c54dSStefano Zampini 3 1 3 0 596c34c54dSStefano Zampini 16 1 4 2 600acf423eSBarry Smith Coarse grid solver -- level 0 ------------------------------- 610acf423eSBarry Smith KSP Object: (mg_coarse_) 1 MPI process 620acf423eSBarry Smith type: preonly 630acf423eSBarry Smith maximum iterations=10000, initial guess is zero 640acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 650acf423eSBarry Smith left preconditioning 668405740aSBarry Smith not checking for convergence 670acf423eSBarry Smith PC Object: (mg_coarse_) 1 MPI process 680acf423eSBarry Smith type: bjacobi 690acf423eSBarry Smith number of blocks = 1 700acf423eSBarry Smith Local solver information for first block is in the following KSP and PC objects on rank 0: 710acf423eSBarry Smith Use -mg_coarse_ksp_view ::ascii_info_detail to display information for all blocks 720acf423eSBarry Smith KSP Object: (mg_coarse_sub_) 1 MPI process 730acf423eSBarry Smith type: preonly 740acf423eSBarry Smith maximum iterations=1, initial guess is zero 750acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 760acf423eSBarry Smith left preconditioning 778405740aSBarry Smith not checking for convergence 780acf423eSBarry Smith PC Object: (mg_coarse_sub_) 1 MPI process 790acf423eSBarry Smith type: lu 800acf423eSBarry Smith out-of-place factorization 810acf423eSBarry Smith tolerance for zero pivot 2.22045e-14 820acf423eSBarry Smith using diagonal shift on blocks to prevent zero pivot [INBLOCKS] 830acf423eSBarry Smith matrix ordering: nd 840acf423eSBarry Smith factor fill ratio given 5., needed 1. 85ecf3d421SBarry Smith Factored matrix: 860acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 870acf423eSBarry Smith type: seqaij 880acf423eSBarry Smith rows=3, cols=3 890acf423eSBarry Smith package used to perform factorization: petsc 900acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 910acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 92ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 930acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 940acf423eSBarry Smith type: seqaij 950acf423eSBarry Smith rows=3, cols=3 960acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 970acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 980acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 99ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 1000acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 1010acf423eSBarry Smith type: seqaij 1020acf423eSBarry Smith rows=3, cols=3 1030acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 1040acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 1050acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 1060acf423eSBarry Smith Down solver (pre-smoother) on level 1 ------------------------------- 1070acf423eSBarry Smith KSP Object: (mg_levels_1_) 1 MPI process 1080acf423eSBarry Smith type: chebyshev 1090acf423eSBarry Smith Chebyshev polynomial of first kind 1106c34c54dSStefano Zampini eigenvalue targets used: min 0.514268, max 5.65695 1116c34c54dSStefano Zampini eigenvalues provided (min 0.299461, max 5.14268) with transform: [0. 0.1; 0. 1.1] 1120acf423eSBarry Smith maximum iterations=2, nonzero initial guess 1130acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 1140acf423eSBarry Smith left preconditioning 1158405740aSBarry Smith not checking for convergence 1160acf423eSBarry Smith PC Object: (mg_levels_1_) 1 MPI process 1170acf423eSBarry Smith type: jacobi 1180acf423eSBarry Smith type DIAGONAL 119ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 1200acf423eSBarry Smith Mat Object: 1 MPI process 1210acf423eSBarry Smith type: seqaij 1220acf423eSBarry Smith rows=16, cols=16 1230acf423eSBarry Smith total: nonzeros=64, allocated nonzeros=64 1240acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 1250acf423eSBarry Smith not using I-node routines 1260acf423eSBarry Smith Up solver (post-smoother) same as down solver (pre-smoother) 127ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 1280acf423eSBarry Smith Mat Object: 1 MPI process 1290acf423eSBarry Smith type: seqaij 1300acf423eSBarry Smith rows=16, cols=16 1310acf423eSBarry Smith total: nonzeros=64, allocated nonzeros=64 1320acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 1330acf423eSBarry Smith not using I-node routines 1340acf423eSBarry SmithKSP Object: 1 MPI process 1350acf423eSBarry Smith type: gmres 136f971d498SPierre Jolivet restart=30, using classical (unmodified) Gram-Schmidt orthogonalization with no iterative refinement 137*143f2514SPierre Jolivet happy breakdown tolerance=1e-30 1380acf423eSBarry Smith maximum iterations=10000, initial guess is zero 1390acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 1400acf423eSBarry Smith left preconditioning 1410acf423eSBarry Smith using PRECONDITIONED norm type for convergence test 1420acf423eSBarry SmithPC Object: 1 MPI process 1430acf423eSBarry Smith type: gamg 1440acf423eSBarry Smith type is MULTIPLICATIVE, levels=2 cycles=v 1450acf423eSBarry Smith Cycles per PCApply=1 1460acf423eSBarry Smith Using externally compute Galerkin coarse grid matrices 1470acf423eSBarry Smith GAMG specific options 1480acf423eSBarry Smith Threshold for dropping small values in graph on each level = -1. -1. 1490acf423eSBarry Smith Threshold scaling factor for each level not specified = 1. 1500acf423eSBarry Smith AGG specific options 1510acf423eSBarry Smith Number of levels of aggressive coarsening 1 1520acf423eSBarry Smith Square graph aggressive coarsening 1530acf423eSBarry Smith MatCoarsen Object: (pc_gamg_) 1 MPI process 1540acf423eSBarry Smith type: mis 1550acf423eSBarry Smith Number smoothing steps to construct prolongation 1 1560acf423eSBarry Smith Complexity: grid = 1.1875 operator = 1.14062 1576c34c54dSStefano Zampini Per-level complexity: op = operator, int = interpolation 1586c34c54dSStefano Zampini #equations | #active PEs | avg nnz/row op | avg nnz/row int 1596c34c54dSStefano Zampini 3 1 3 0 1606c34c54dSStefano Zampini 16 1 4 2 1610acf423eSBarry Smith Coarse grid solver -- level 0 ------------------------------- 1620acf423eSBarry Smith KSP Object: (mg_coarse_) 1 MPI process 1630acf423eSBarry Smith type: preonly 1640acf423eSBarry Smith maximum iterations=10000, initial guess is zero 1650acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 1660acf423eSBarry Smith left preconditioning 1678405740aSBarry Smith not checking for convergence 1680acf423eSBarry Smith PC Object: (mg_coarse_) 1 MPI process 1690acf423eSBarry Smith type: bjacobi 1700acf423eSBarry Smith number of blocks = 1 1710acf423eSBarry Smith Local solver information for first block is in the following KSP and PC objects on rank 0: 1720acf423eSBarry Smith Use -mg_coarse_ksp_view ::ascii_info_detail to display information for all blocks 1730acf423eSBarry Smith KSP Object: (mg_coarse_sub_) 1 MPI process 1740acf423eSBarry Smith type: preonly 1750acf423eSBarry Smith maximum iterations=1, initial guess is zero 1760acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 1770acf423eSBarry Smith left preconditioning 1788405740aSBarry Smith not checking for convergence 1790acf423eSBarry Smith PC Object: (mg_coarse_sub_) 1 MPI process 1800acf423eSBarry Smith type: lu 1810acf423eSBarry Smith out-of-place factorization 1820acf423eSBarry Smith tolerance for zero pivot 2.22045e-14 1830acf423eSBarry Smith using diagonal shift on blocks to prevent zero pivot [INBLOCKS] 1840acf423eSBarry Smith matrix ordering: nd 1850acf423eSBarry Smith factor fill ratio given 5., needed 1. 186ecf3d421SBarry Smith Factored matrix: 1870acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 1880acf423eSBarry Smith type: seqaij 1890acf423eSBarry Smith rows=3, cols=3 1900acf423eSBarry Smith package used to perform factorization: petsc 1910acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 1920acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 193ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 1940acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 1950acf423eSBarry Smith type: seqaij 1960acf423eSBarry Smith rows=3, cols=3 1970acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 1980acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 1990acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 200ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 2010acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 2020acf423eSBarry Smith type: seqaij 2030acf423eSBarry Smith rows=3, cols=3 2040acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 2050acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 2060acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 2070acf423eSBarry Smith Down solver (pre-smoother) on level 1 ------------------------------- 2080acf423eSBarry Smith KSP Object: (mg_levels_1_) 1 MPI process 2090acf423eSBarry Smith type: chebyshev 2100acf423eSBarry Smith Chebyshev polynomial of first kind 2110acf423eSBarry Smith eigenvalue targets used: min 0.159372, max 1.75309 2120acf423eSBarry Smith eigenvalues estimated via gmres: min 0.406283, max 1.59372 2130acf423eSBarry Smith eigenvalues estimated using gmres with transform: [0. 0.1; 0. 1.1] 2140acf423eSBarry Smith KSP Object: (mg_levels_1_esteig_) 1 MPI process 2150acf423eSBarry Smith type: gmres 216f971d498SPierre Jolivet restart=30, using classical (unmodified) Gram-Schmidt orthogonalization with no iterative refinement 217*143f2514SPierre Jolivet happy breakdown tolerance=1e-30 2180acf423eSBarry Smith maximum iterations=10, initial guess is zero 2190acf423eSBarry Smith tolerances: relative=1e-12, absolute=1e-50, divergence=10000. 2200acf423eSBarry Smith left preconditioning 2210acf423eSBarry Smith using PRECONDITIONED norm type for convergence test 2220acf423eSBarry Smith estimating eigenvalues using a noisy random number generated right-hand side 2230acf423eSBarry Smith maximum iterations=2, nonzero initial guess 2240acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 2250acf423eSBarry Smith left preconditioning 2268405740aSBarry Smith not checking for convergence 2270acf423eSBarry Smith PC Object: (mg_levels_1_) 1 MPI process 2280acf423eSBarry Smith type: jacobi 2290acf423eSBarry Smith type DIAGONAL 230ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 2310acf423eSBarry Smith Mat Object: 1 MPI process 2320acf423eSBarry Smith type: seqaij 2330acf423eSBarry Smith rows=16, cols=16 2340acf423eSBarry Smith total: nonzeros=64, allocated nonzeros=64 2350acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 2360acf423eSBarry Smith not using I-node routines 2370acf423eSBarry Smith Up solver (post-smoother) same as down solver (pre-smoother) 238ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 2390acf423eSBarry Smith Mat Object: 1 MPI process 2400acf423eSBarry Smith type: seqaij 2410acf423eSBarry Smith rows=16, cols=16 2420acf423eSBarry Smith total: nonzeros=64, allocated nonzeros=64 2430acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 2440acf423eSBarry Smith not using I-node routines 2450acf423eSBarry SmithKSP Object: 1 MPI process 2460acf423eSBarry Smith type: gmres 247f971d498SPierre Jolivet restart=30, using classical (unmodified) Gram-Schmidt orthogonalization with no iterative refinement 248*143f2514SPierre Jolivet happy breakdown tolerance=1e-30 2490acf423eSBarry Smith maximum iterations=10000, initial guess is zero 2500acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 2510acf423eSBarry Smith left preconditioning 2520acf423eSBarry Smith using PRECONDITIONED norm type for convergence test 2530acf423eSBarry SmithPC Object: 1 MPI process 2540acf423eSBarry Smith type: gamg 2550acf423eSBarry Smith type is MULTIPLICATIVE, levels=2 cycles=v 2560acf423eSBarry Smith Cycles per PCApply=1 2570acf423eSBarry Smith Using externally compute Galerkin coarse grid matrices 2580acf423eSBarry Smith GAMG specific options 2590acf423eSBarry Smith Threshold for dropping small values in graph on each level = -1. -1. 2600acf423eSBarry Smith Threshold scaling factor for each level not specified = 1. 2610acf423eSBarry Smith AGG specific options 2620acf423eSBarry Smith Number of levels of aggressive coarsening 1 2630acf423eSBarry Smith Square graph aggressive coarsening 2640acf423eSBarry Smith MatCoarsen Object: (pc_gamg_) 1 MPI process 2650acf423eSBarry Smith type: mis 2660acf423eSBarry Smith Number smoothing steps to construct prolongation 1 2670acf423eSBarry Smith Complexity: grid = 1.1875 operator = 1.14062 2686c34c54dSStefano Zampini Per-level complexity: op = operator, int = interpolation 2696c34c54dSStefano Zampini #equations | #active PEs | avg nnz/row op | avg nnz/row int 2706c34c54dSStefano Zampini 3 1 3 0 2716c34c54dSStefano Zampini 16 1 4 2 2720acf423eSBarry Smith Coarse grid solver -- level 0 ------------------------------- 2730acf423eSBarry Smith KSP Object: (mg_coarse_) 1 MPI process 2740acf423eSBarry Smith type: preonly 2750acf423eSBarry Smith maximum iterations=10000, initial guess is zero 2760acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 2770acf423eSBarry Smith left preconditioning 2788405740aSBarry Smith not checking for convergence 2790acf423eSBarry Smith PC Object: (mg_coarse_) 1 MPI process 2800acf423eSBarry Smith type: bjacobi 2810acf423eSBarry Smith number of blocks = 1 2820acf423eSBarry Smith Local solver information for first block is in the following KSP and PC objects on rank 0: 2830acf423eSBarry Smith Use -mg_coarse_ksp_view ::ascii_info_detail to display information for all blocks 2840acf423eSBarry Smith KSP Object: (mg_coarse_sub_) 1 MPI process 2850acf423eSBarry Smith type: preonly 2860acf423eSBarry Smith maximum iterations=1, initial guess is zero 2870acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 2880acf423eSBarry Smith left preconditioning 2898405740aSBarry Smith not checking for convergence 2900acf423eSBarry Smith PC Object: (mg_coarse_sub_) 1 MPI process 2910acf423eSBarry Smith type: lu 2920acf423eSBarry Smith out-of-place factorization 2930acf423eSBarry Smith tolerance for zero pivot 2.22045e-14 2940acf423eSBarry Smith using diagonal shift on blocks to prevent zero pivot [INBLOCKS] 2950acf423eSBarry Smith matrix ordering: nd 2960acf423eSBarry Smith factor fill ratio given 5., needed 1. 297ecf3d421SBarry Smith Factored matrix: 2980acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 2990acf423eSBarry Smith type: seqaij 3000acf423eSBarry Smith rows=3, cols=3 3010acf423eSBarry Smith package used to perform factorization: petsc 3020acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 3030acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 304ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 3050acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 3060acf423eSBarry Smith type: seqaij 3070acf423eSBarry Smith rows=3, cols=3 3080acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 3090acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 3100acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 311ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 3120acf423eSBarry Smith Mat Object: (mg_coarse_sub_) 1 MPI process 3130acf423eSBarry Smith type: seqaij 3140acf423eSBarry Smith rows=3, cols=3 3150acf423eSBarry Smith total: nonzeros=9, allocated nonzeros=9 3160acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 3170acf423eSBarry Smith using I-node routines: found 1 nodes, limit used is 5 3180acf423eSBarry Smith Down solver (pre-smoother) on level 1 ------------------------------- 3190acf423eSBarry Smith KSP Object: (mg_levels_1_) 1 MPI process 3200acf423eSBarry Smith type: chebyshev 3210acf423eSBarry Smith Chebyshev polynomial of first kind 3220acf423eSBarry Smith eigenvalue targets used: min 0.160581, max 1.76639 3230acf423eSBarry Smith eigenvalues estimated via gmres: min 0.394193, max 1.60581 3240acf423eSBarry Smith eigenvalues estimated using gmres with transform: [0. 0.1; 0. 1.1] 3250acf423eSBarry Smith KSP Object: (mg_levels_1_esteig_) 1 MPI process 3260acf423eSBarry Smith type: gmres 327f971d498SPierre Jolivet restart=30, using classical (unmodified) Gram-Schmidt orthogonalization with no iterative refinement 328*143f2514SPierre Jolivet happy breakdown tolerance=1e-30 3290acf423eSBarry Smith maximum iterations=10, initial guess is zero 3300acf423eSBarry Smith tolerances: relative=1e-12, absolute=1e-50, divergence=10000. 3310acf423eSBarry Smith left preconditioning 3320acf423eSBarry Smith using PRECONDITIONED norm type for convergence test 3330acf423eSBarry Smith estimating eigenvalues using a noisy random number generated right-hand side 3340acf423eSBarry Smith maximum iterations=2, nonzero initial guess 3350acf423eSBarry Smith tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 3360acf423eSBarry Smith left preconditioning 3378405740aSBarry Smith not checking for convergence 3380acf423eSBarry Smith PC Object: (mg_levels_1_) 1 MPI process 3390acf423eSBarry Smith type: jacobi 3400acf423eSBarry Smith type DIAGONAL 341ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 3420acf423eSBarry Smith Mat Object: 1 MPI process 3430acf423eSBarry Smith type: seqaij 3440acf423eSBarry Smith rows=16, cols=16 3450acf423eSBarry Smith total: nonzeros=64, allocated nonzeros=64 3460acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 3470acf423eSBarry Smith not using I-node routines 3480acf423eSBarry Smith Up solver (post-smoother) same as down solver (pre-smoother) 349ecf3d421SBarry Smith linear system matrix, which is also used to construct the preconditioner: 3500acf423eSBarry Smith Mat Object: 1 MPI process 3510acf423eSBarry Smith type: seqaij 3520acf423eSBarry Smith rows=16, cols=16 3530acf423eSBarry Smith total: nonzeros=64, allocated nonzeros=64 3540acf423eSBarry Smith total number of mallocs used during MatSetValues calls=0 3550acf423eSBarry Smith not using I-node routines 356