1*c4762a1bSJed Brownatol=1e-50, rtol=1e-08, stol=1e-08, maxit=50, maxf=10000 2*c4762a1bSJed Brown 0 SNES Function norm 5.41468 3*c4762a1bSJed Brown ---------- Testing Jacobian ------------- 4*c4762a1bSJed Brown Testing hand-coded Jacobian, if (for double precision runs) ||J - Jfd||_F/||J||_F is 5*c4762a1bSJed Brown O(1.e-8), the hand-coded Jacobian is probably correct. 6*c4762a1bSJed Brown ||J - Jfd||_F/||J||_F = 0.481271, ||J - Jfd||_F = 32. 7*c4762a1bSJed Brown Hand-coded Jacobian ---------- 8*c4762a1bSJed BrownMat Object: 1 MPI processes 9*c4762a1bSJed Brown type: seqaij 10*c4762a1bSJed Brownrow 0: (0, 1.) (1, 16.) 11*c4762a1bSJed Brownrow 1: (1, -31.) (2, 16.) 12*c4762a1bSJed Brownrow 2: (1, 16.) (2, -31.) (3, 16.) 13*c4762a1bSJed Brownrow 3: (2, 16.) (3, -31.) 14*c4762a1bSJed Brownrow 4: (3, 16.) (4, 1.) 15*c4762a1bSJed Brown Finite difference Jacobian ---------- 16*c4762a1bSJed BrownMat Object: 1 MPI processes 17*c4762a1bSJed Brown type: seqaij 18*c4762a1bSJed Brownrow 0: (0, 1.) 19*c4762a1bSJed Brownrow 1: (0, 16.) (1, -31.) (2, 16.) 20*c4762a1bSJed Brownrow 2: (1, 16.) (2, -31.) (3, 16.) 21*c4762a1bSJed Brownrow 3: (2, 16.) (3, -31.) (4, 16.) 22*c4762a1bSJed Brownrow 4: (4, 1.) 23*c4762a1bSJed Brown Hand-coded minus finite-difference Jacobian with tolerance 1e-05 ---------- 24*c4762a1bSJed BrownMat Object: 1 MPI processes 25*c4762a1bSJed Brown type: seqaij 26*c4762a1bSJed Brownrow 0: (1, 16.) 27*c4762a1bSJed Brownrow 1: (0, -16.) 28*c4762a1bSJed Brownrow 2: 29*c4762a1bSJed Brownrow 3: (4, -16.) 30*c4762a1bSJed Brownrow 4: (3, 16.) 31*c4762a1bSJed Brown 0 KSP Residual norm 6.945929944146e+00 32*c4762a1bSJed Brown 1 KSP Residual norm 2.792520156365e-15 33*c4762a1bSJed BrownVec Object: 1 MPI processes 34*c4762a1bSJed Brown type: seq 35*c4762a1bSJed Brown0.5 36*c4762a1bSJed Brown-1.25024 37*c4762a1bSJed Brown-2.76563 38*c4762a1bSJed Brown-4.42798 39*c4762a1bSJed Brown-0.5 40*c4762a1bSJed BrownVec Object: 1 MPI processes 41*c4762a1bSJed Brown type: seq 42*c4762a1bSJed Brown-3.39308 43*c4762a1bSJed Brown0.243317 44*c4762a1bSJed Brown0.393287 45*c4762a1bSJed Brown0.345825 46*c4762a1bSJed Brown-6.0332 47*c4762a1bSJed BrownMat Object: 1 MPI processes 48*c4762a1bSJed Brown type: seqaij 49*c4762a1bSJed Brownrow 0: (0, 1.) (1, 16.) 50*c4762a1bSJed Brownrow 1: (1, -31.) (2, 16.) 51*c4762a1bSJed Brownrow 2: (1, 16.) (2, -31.) (3, 16.) 52*c4762a1bSJed Brownrow 3: (2, 16.) (3, -31.) 53*c4762a1bSJed Brownrow 4: (3, 16.) (4, 1.) 54*c4762a1bSJed BrownMat Object: 1 MPI processes 55*c4762a1bSJed Brown type: seqaij 56*c4762a1bSJed Brownrow 0: (0, 1.) 57*c4762a1bSJed Brownrow 1: (1, 1.) 58*c4762a1bSJed Brownrow 2: (2, 1.) 59*c4762a1bSJed Brownrow 3: (3, 1.) 60*c4762a1bSJed Brownrow 4: (4, 1.) 61*c4762a1bSJed Brown 1 SNES Function norm 111.09 62*c4762a1bSJed BrownSNES Object: 1 MPI processes 63*c4762a1bSJed Brown type: ksptransposeonly 64*c4762a1bSJed Brown maximum iterations=50, maximum function evaluations=10000 65*c4762a1bSJed Brown tolerances: relative=1e-08, absolute=1e-50, solution=1e-08 66*c4762a1bSJed Brown total number of linear solver iterations=1 67*c4762a1bSJed Brown total number of function evaluations=9 68*c4762a1bSJed Brown norm schedule ALWAYS 69*c4762a1bSJed Brown KSP Object: 1 MPI processes 70*c4762a1bSJed Brown type: gmres 71*c4762a1bSJed Brown restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement 72*c4762a1bSJed Brown happy breakdown tolerance 1e-30 73*c4762a1bSJed Brown maximum iterations=10000, initial guess is zero 74*c4762a1bSJed Brown tolerances: relative=1e-05, absolute=1e-50, divergence=10000. 75*c4762a1bSJed Brown left preconditioning 76*c4762a1bSJed Brown using PRECONDITIONED norm type for convergence test 77*c4762a1bSJed Brown PC Object: 1 MPI processes 78*c4762a1bSJed Brown type: ilu 79*c4762a1bSJed Brown out-of-place factorization 80*c4762a1bSJed Brown 0 levels of fill 81*c4762a1bSJed Brown tolerance for zero pivot 2.22045e-14 82*c4762a1bSJed Brown matrix ordering: natural 83*c4762a1bSJed Brown factor fill ratio given 1., needed 1. 84*c4762a1bSJed Brown Factored matrix follows: 85*c4762a1bSJed Brown Mat Object: 1 MPI processes 86*c4762a1bSJed Brown type: seqaij 87*c4762a1bSJed Brown rows=5, cols=5 88*c4762a1bSJed Brown package used to perform factorization: petsc 89*c4762a1bSJed Brown total: nonzeros=11, allocated nonzeros=11 90*c4762a1bSJed Brown total number of mallocs used during MatSetValues calls=0 91*c4762a1bSJed Brown not using I-node routines 92*c4762a1bSJed Brown linear system matrix = precond matrix: 93*c4762a1bSJed Brown Mat Object: 1 MPI processes 94*c4762a1bSJed Brown type: seqaij 95*c4762a1bSJed Brown rows=5, cols=5 96*c4762a1bSJed Brown total: nonzeros=11, allocated nonzeros=15 97*c4762a1bSJed Brown total number of mallocs used during MatSetValues calls=0 98*c4762a1bSJed Brown not using I-node routines 99*c4762a1bSJed Brownnumber of SNES iterations = 1 100*c4762a1bSJed Brown 101*c4762a1bSJed BrownNorm of error 6.77514, Iterations 1 102