1atol=1e-50, rtol=1e-08, stol=1e-08, maxit=50, maxf=10000 2 0 SNES Function norm 5.41468 3 ---------- Testing Jacobian ------------- 4 Testing hand-coded Jacobian, if (for double precision runs) ||J - Jfd||_F/||J||_F is 5 O(1.e-8), the hand-coded Jacobian is probably correct. 6 ||J - Jfd||_F/||J||_F = 0.481271, ||J - Jfd||_F = 32. 7 Hand-coded Jacobian ---------- 8Mat Object: 1 MPI process 9 type: seqaij 10row 0: (0, 1.) (1, 16.) 11row 1: (1, -31.) (2, 16.) 12row 2: (1, 16.) (2, -31.) (3, 16.) 13row 3: (2, 16.) (3, -31.) 14row 4: (3, 16.) (4, 1.) 15 Finite difference Jacobian ---------- 16Mat Object: 1 MPI process 17 type: seqaij 18row 0: (0, 1.) 19row 1: (0, 16.) (1, -31.) (2, 16.) 20row 2: (1, 16.) (2, -31.) (3, 16.) 21row 3: (2, 16.) (3, -31.) (4, 16.) 22row 4: (4, 1.) 23 Hand-coded minus finite-difference Jacobian with tolerance 1e-05 ---------- 24Mat Object: 1 MPI process 25 type: seqaij 26row 0: (1, 16.) 27row 1: (0, -16.) 28row 2: 29row 3: (4, -16.) 30row 4: (3, 16.) 31 0 KSP Residual norm 6.945929944146e+00 32 1 KSP Residual norm 2.792520156365e-15 33Vec Object: 1 MPI process 34 type: seq 350.5 36-1.25024 37-2.76563 38-4.42798 39-0.5 40Vec Object: 1 MPI process 41 type: seq 42-3.39308 430.243317 440.393287 450.345825 46-6.0332 47Mat Object: 1 MPI process 48 type: seqaij 49row 0: (0, 1.) (1, 16.) 50row 1: (1, -31.) (2, 16.) 51row 2: (1, 16.) (2, -31.) (3, 16.) 52row 3: (2, 16.) (3, -31.) 53row 4: (3, 16.) (4, 1.) 54Mat Object: 1 MPI process 55 type: seqaij 56row 0: (0, 1.) 57row 1: (1, 1.) 58row 2: (2, 1.) 59row 3: (3, 1.) 60row 4: (4, 1.) 61 1 SNES Function norm 111.09 62SNES Object: 1 MPI process 63 type: ksptransposeonly 64 maximum iterations=50, maximum function evaluations=10000 65 tolerances: relative=1e-08, absolute=1e-50, solution=1e-08 66 total number of linear solver iterations=1 67 total number of function evaluations=4 68 norm schedule ALWAYS 69 KSP Object: 1 MPI process 70 type: gmres 71 restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement 72 happy breakdown tolerance 1e-30 73 maximum iterations=10000, initial guess is zero 74 tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 75 left preconditioning 76 using PRECONDITIONED norm type for convergence test 77 PC Object: 1 MPI process 78 type: ilu 79 out-of-place factorization 80 0 levels of fill 81 tolerance for zero pivot 2.22045e-14 82 matrix ordering: natural 83 factor fill ratio given 1., needed 1. 84 Factored matrix follows: 85 Mat Object: 1 MPI process 86 type: seqaij 87 rows=5, cols=5, bs=5 88 package used to perform factorization: petsc 89 total: nonzeros=11, allocated nonzeros=11 90 not using I-node routines 91 linear system matrix = precond matrix: 92 Mat Object: 1 MPI process 93 type: seqaij 94 rows=5, cols=5, bs=5 variable blocks set 95 total: nonzeros=11, allocated nonzeros=15 96 total number of mallocs used during MatSetValues calls=0 97 not using I-node routines 98number of SNES iterations = 1 99 100Norm of error 6.77514, Iterations 1 101