Albany_IOSS: Loading STKMesh from Exodus file antarctica_2d.exo IOSS-STK: number of node sets = 1 IOSS-STK: number of side sets = 1 Restart Index not set. Not reading solution from exodus (-1) [GenericSTKMeshStruct] Processing field requirements... - Reading Node Scalar field 'ice_thickness' from file 'thickness.ascii' ... done! - Reading Node Scalar field 'surface_height' from file 'surface_height.ascii' ... done! - Reading Node Layered Scalar field 'temperature' from file 'temperature.ascii' ... done! - Reading Node Scalar field 'basal_friction' from file 'basal_friction_reg.ascii' ... done! [ExtrudedSTKMesh] Adding nodes... done! [ExtrudedSTKMesh] Adding elements... done! [ExtrudedSTKMesh] Adding basalside sides... done! [ExtrudedSTKMesh] Adding upperside sides... done! [ExtrudedSTKMesh] Adding lateral sides... done! [ExtrudedSTKMesh] Extruding basal fields... - Extruding Scalar Node field ice_thickness...done! - Extruding Scalar Node field surface_height...done! - Extruding Scalar Node field basal_friction...done! [ExtrudedSTKMesh] Interpolating layered basal fields... - Interpolating Scalar field temperature...done! [GenericSTKMeshStruct] Processing field requirements... - Skipping field 'temperature' since it's listed as present in the mesh. - Skipping field 'ice_thickness' since it's listed as present in the mesh. - Skipping field 'surface_height' since it's listed as present in the mesh. - Skipping field 'basal_friction' since it's listed as present in the mesh. [GenericSTKMeshStruct] Processing field requirements... - Reading Node Vector field 'surface_velocity' from file 'surface_velocity.ascii' ... done! STKDisc: 261040 elements on Proc 0 STKDisc: nodeset bottom has size 13455 on Proc 0. STKDisc: nodeset extruded_node has size 282555 on Proc 0. STKDisc: nodeset lateral has size 84 on Proc 0. STKDisc: nodeset top has size 13455 on Proc 0. STKDisc: sideset basalside has size 13052 on Proc 0. STKDisc: sideset extruded_lateralside has size 60 on Proc 0. STKDisc: sideset extruded_surface_quad4_edge2_1 has size 60 on Proc 0. STKDisc: sideset lateralside has size 60 on Proc 0. STKDisc: sideset upperside has size 13052 on Proc 0. STKDisc: 13052 elements on Proc 0 STKDisc: nodeset node has size 13455 on Proc 0. STKDisc: sideset lateralside has size 3 on Proc 0. STKDisc: 13052 elements on Proc 0 STKDisc: nodeset all_nodes has size 13455 on Proc 0. StateManager: initializing state: temperature StateManager: initializing state: ice_thickness StateManager: initializing state: surface_height StateManager: initializing state: basal_friction StateManager: initializing state: ice_thickness StateManager: initializing state: surface_height StateManager: initializing state: temperature StateManager: initializing state: basal_friction StateManager: initializing state: surface_velocity xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Sacado ParameterLibrary has been initialized: Library of all registered parameters: Bed Roughness: Supports AD = 1, Supports_Analytic = 0 Coulomb Friction Coefficient: Supports AD = 1, Supports_Analytic = 0 Glen's Law Homotopy Parameter: Supports AD = 1, Supports_Analytic = 0 Hydraulic-Over-Hydrostatic Potential Ratio: Supports AD = 1, Supports_Analytic = 0 Power Exponent: Supports AD = 1, Supports_Analytic = 0 Power Law Coefficient: Supports AD = 1, Supports_Analytic = 0 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Number of parameters vectors = 1 Number of parameters in parameter vector 0 = 1 Number of distributed parameters vectors = 0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Beginning Continuation Run Stepper Method: Natural Initial Parameter Value = 0.000e+00 Maximum Parameter Value = 1.000e+00 Minimum Parameter Value = 0.000e+00 Maximum Number of Continuation Steps = 15 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Solver Type = LOCA Jacobian Operator = Have Jacobian [default] Lean Matrix Free = 0 [default] LOCASolver: Create Second Matrix = 0 [default] Write Only Converged Solution = 1 [default] LOCA -> Bifurcation -> Type = None [default] Constraints -> [empty list] Predictor -> Method = Constant First Step Predictor -> [empty list] Last Step Predictor -> [empty list] Stepper -> Initial Value = 0 Continuation Parameter = Glen's Law Homotopy Parameter Continuation Method = Natural Max Steps = 15 Max Value = 1 Min Value = 0 Compute Eigenvalues = 0 [default] Max Nonlinear Iterations = 15 [default] Enable Tangent Factor Step Size Scaling = 0 [default] Min Tangent Factor = 0.1 [default] Tangent Factor Exponent = 1 [default] Compute Eigenvalues On Target Step = 0 [default] Return Failed on Reaching Max Steps = 1 [default] Eigensolver -> Save Eigen Data Method = User-Defined User-Defined Save Eigen Data Name = Piro Strategy Piro Strategy = Teuchos::RCP{ptr=0x42d18820,node=0x42d18910,strong_count=6,weak_count=0} Method = Default [default] Step Size -> Initial Step Size = 0.1 Method = Adaptive [default] Max Step Size = 1e+12 [default] Min Step Size = 1e-12 [default] Failed Step Reduction Factor = 0.5 [default] Successful Step Increase Factor = 1.26 [default] Aggressiveness = 0.5 [default] NOX -> Nonlinear Solver = Line Search Based Status Tests -> Test Type = Combo Combo Type = OR Number of Tests = 2 Test 0 -> Test Type = Combo Combo Type = AND Number of Tests = 2 Tolerance = 1e-08 [unused] Norm Type = Two Norm [unused] Scale Type = Unscaled [unused] Test 0 -> Test Type = NormF Norm Type = Two Norm Scale Type = Scaled Tolerance = 1e-05 Test 1 -> Test Type = NormWRMS Absolute Tolerance = 0.01 Relative Tolerance = 1e-07 BDF Multiplier = 1 [default] Tolerance = 1 [default] Alpha = 1 [default] Beta = 0.5 [default] Disable Implicit Weighting = 1 [default] Test 1 -> Test Type = MaxIters Maximum Iterations = 40 Direction -> Method = Newton Newton -> Forcing Term Method = Constant Rescue Bad Newton Solve = 1 Linear Solver -> Write Linear System = 0 [unused] Aztec Solver = GMRES [unused] Max Iterations = 43 [unused] Tolerance = 0.0001 Output Frequency = 20 [unused] Preconditioner = Ifpack [unused] Stratimikos Linear Solver -> NOX Stratimikos Options -> Use Linear Solve Tolerance From NOX = 0 Zero Initial Guess = 0 [default] Compute Scaling Manually = 1 [default] Output Solver Details = 1 [default] Throw Error on Prec Failure = 1 [default] Preconditioner Reuse Policy = Rebuild [default] Max Age Of Prec = 1 [default] Stratimikos -> Linear Solver Type = AztecOO Preconditioner Type = ML Enable Delayed Solver Construction = 0 [default] Linear Solver Types -> AztecOO -> Output Every RHS = 0 [default] Forward Solve -> Max Iterations = 500 Tolerance = 1e-06 AztecOO Settings -> Aztec Solver = GMRES [unused] Convergence Test = r0 [unused] Size of Krylov Subspace = 200 [unused] Output Frequency = 1 [unused] Adjoint Solve -> Max Iterations = 500 Tolerance = 1e-06 AztecOO Settings -> Aztec Solver = GMRES [unused] Convergence Test = r0 [unused] Size of Krylov Subspace = 200 [unused] Output Frequency = 1 [unused] VerboseObject -> Output File = none [default] Verbosity Level = default [default] Preconditioner Types -> Ifpack -> Overlap = 0 [unused] Prec Type = ILU [unused] Ifpack Settings -> fact: level-of-fill = 0 [unused] ML -> Base Method Defaults = none Reuse Fine Level Smoother = 0 [unused] ML Settings -> default values = SA [unused] ML output = 10 [unused] eigen-analysis: iterations = 49 [unused] repartition: enable = 1 [unused] repartition: max min ratio = 1.327 [unused] repartition: min per proc = 600 [unused] repartition: Zoltan dimensions = 2 [unused] repartition: start level = 4 [unused] semicoarsen: number of levels = 2 [unused] semicoarsen: coarsen rate = 14 [unused] smoother: sweeps = 4 [unused] smoother: type = Gauss-Seidel [unused] smoother: Chebyshev eig boost = 1.2 [unused] smoother: sweeps (level 0) = 1 [unused] smoother: type (level 0) = line Gauss-Seidel [unused] smoother: line GS Type = standard [unused] smoother: line orientation = vertical [unused] smoother: line direction nodes = 21 [unused] smoother: damping factor = 1 [unused] smoother: pre or post = both [unused] coarse: type = Gauss-Seidel [unused] coarse: sweeps = 4 [unused] coarse: max size = 2000 [unused] coarse: pre or post = pre [unused] max levels = 7 [unused] PDE equations = 2 [unused] x-coordinates = 0x2aaafe7ec0c0 [unused] y-coordinates = 0x2aaafea13e98 [unused] z-coordinates = 0x2aaafec3bc70 [unused] null space: type = pre-computed [unused] null space: dimension = 3 [unused] null space: vectors = 0x2aaafee64010 [unused] null space: add default vectors = 0 [unused] Line Search -> Method = Backtrack Full Step -> Full Step = 1 [unused] Backtrack -> Minimum Step = 1e-12 [default] Default Step = 1 [default] Recovery Step = 1 [default] Max Iters = 100 [default] Reduction Factor = 0.5 [default] Printing -> Output Precision = 3 Output Processor = 0 MyPID = 0 Output Information -> Error = 1 Warning = 1 Outer Iteration = 1 Parameters = 0 Details = 0 Linear Solver Details = 0 Stepper Iteration = 1 Stepper Details = 1 Stepper Parameters = 1 Debug = 0 [default] Inner Iteration = 0 [default] Outer Iteration StatusTest = 0 [default] Test Details = 0 [default] Solver Options -> Status Test Check Type = Minimal Observer = Teuchos::RCP{ptr=0,node=0,strong_count=0,weak_count=0} [default] User Defined Pre/Post Operator = Teuchos::RCP{ptr=0,node=0,strong_count=0,weak_count=0} [default] User Defined Merit Function = Teuchos::RCP{ptr=0,node=0,strong_count=0,weak_count=0} [default] Fixed Point Iteration Type = Seidel [default] NOX Observer -> [empty list] Save Eigen Data Strategy -> [empty list] Status Test -> [empty list] eval pVector LOCA::ParameterVector (size = 1) 0 Glen's Law Homotopy Parameter = 0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Start of Continuation Step 0 : Attempting to converge initial guess at initial parameter values. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ************************************************************************ -- Nonlinear Solver Step 0 -- ||F|| = 1.612e+05 step = 0.000e+00 dx = 0.000e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.5279e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.7785e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.5279e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.5747e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 5.1774e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.5747e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2032e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 6.0502e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2032e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 123456789 (123456789) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 8.315116e-02 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8679e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.1365e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8679e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 2.069758263 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1272057788 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.0230372129 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.006825279444 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001117754728 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001041349024 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.1487477506 (s) - for hierarchy setup = 2.072722641 (s) - for smoothers setup = 0.1572841816 (s) - for coarse setup = 2.474244684e-05 (s) - for final setup = 0.001429252326 (s) Total for this setup = 2.380589838 (s) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 7.044410e-01 iter: 2 residual = 9.549404e-02 iter: 3 residual = 1.813600e-02 iter: 4 residual = 4.536494e-03 iter: 5 residual = 1.711376e-03 iter: 6 residual = 8.414135e-04 iter: 7 residual = 4.202608e-04 iter: 8 residual = 2.038524e-04 iter: 9 residual = 9.028437e-05 iter: 10 residual = 4.318520e-05 iter: 11 residual = 2.418024e-05 iter: 12 residual = 1.301893e-05 iter: 13 residual = 6.832270e-06 iter: 14 residual = 3.846479e-06 iter: 15 residual = 2.304918e-06 iter: 16 residual = 1.476209e-06 iter: 17 residual = 1.004261e-06 iter: 18 residual = 7.093759e-07 Solution time: 9.330586 (sec.) total iterations: 18 Total solve time = 9.33288 sec ************************************************************************ -- Nonlinear Solver Step 1 -- ||F|| = 1.337e+05 step = 2.500e-01 dx = 8.364e+05 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.6135e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.6889e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.6135e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.6059e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 5.1153e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.6059e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2037e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 6.0488e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2037e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 460185016 (460185016) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.201567e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8990e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0194e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8990e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.740943816 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1224247282 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.01770214178 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001443335786 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001023104414 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001029679552 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.0578924939 (s) - for hierarchy setup = 1.740977285 (s) - for smoothers setup = 0.140476482 (s) - for coarse setup = 2.030935138e-05 (s) - for final setup = 0.0002892166376 (s) Total for this setup = 1.939901443 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.231840794 2.231840794 2- Preconditioner apply = 5.812154622 a- first application(s) only = 0.3048382774 0.3048382774 b- remaining applications = 5.507316344 0.289858755 3- Total time required by ML so far is 8.043995416 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 6.655449e-01 iter: 2 residual = 1.002981e-01 iter: 3 residual = 2.079207e-02 iter: 4 residual = 5.865402e-03 iter: 5 residual = 2.393320e-03 iter: 6 residual = 1.131605e-03 iter: 7 residual = 5.792189e-04 iter: 8 residual = 3.207693e-04 iter: 9 residual = 1.600072e-04 iter: 10 residual = 7.415652e-05 iter: 11 residual = 3.768208e-05 iter: 12 residual = 2.042202e-05 iter: 13 residual = 1.164285e-05 iter: 14 residual = 7.092648e-06 iter: 15 residual = 4.346706e-06 iter: 16 residual = 2.745283e-06 iter: 17 residual = 1.850944e-06 iter: 18 residual = 1.295318e-06 iter: 19 residual = 9.132367e-07 Solution time: 10.111701 (sec.) total iterations: 19 Total solve time = 10.1128 sec ************************************************************************ -- Nonlinear Solver Step 2 -- ||F|| = 1.265e+05 step = 2.500e-01 dx = 8.160e+05 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.6107e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.6918e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.6107e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.6574e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 5.0162e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.6574e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.1865e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 6.0965e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.1865e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 120144603 (120144603) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.495032e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9062e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9929e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9062e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 2.068657593 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1196047813 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05386164226 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001848917454 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001081386581 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.000101714395 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07308472507 (s) - for hierarchy setup = 2.068690523 (s) - for smoothers setup = 0.1738611683 (s) - for coarse setup = 2.878997475e-05 (s) - for final setup = 0.001416124403 (s) Total for this setup = 2.317336868 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.882013014 1.882013014 2- Preconditioner apply = 6.098045744 a- first application(s) only = 0.3002596544 0.3002596544 b- remaining applications = 5.79778609 0.2898893045 3- Total time required by ML so far is 7.980058758 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 5.924325e-01 iter: 2 residual = 1.007603e-01 iter: 3 residual = 2.141449e-02 iter: 4 residual = 6.058096e-03 iter: 5 residual = 2.522660e-03 iter: 6 residual = 1.293977e-03 iter: 7 residual = 7.132985e-04 iter: 8 residual = 4.075790e-04 iter: 9 residual = 2.167431e-04 iter: 10 residual = 1.049353e-04 iter: 11 residual = 5.335055e-05 iter: 12 residual = 2.877550e-05 iter: 13 residual = 1.559840e-05 iter: 14 residual = 9.063377e-06 iter: 15 residual = 5.748215e-06 iter: 16 residual = 3.882518e-06 iter: 17 residual = 2.697265e-06 iter: 18 residual = 1.835493e-06 iter: 19 residual = 1.264400e-06 iter: 20 residual = 8.968265e-07 Solution time: 10.692418 (sec.) total iterations: 20 Total solve time = 10.6931 sec ************************************************************************ -- Nonlinear Solver Step 3 -- ||F|| = 1.183e+05 step = 2.500e-01 dx = 8.672e+05 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.6337e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.6685e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.6337e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.7097e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.9193e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.7097e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.1744e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 6.1303e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.1744e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1220988223 (1220988223) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.291942e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9133e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9670e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9133e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.735846289 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1275399756 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.009435362183 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001483317465 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001023281366 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001037726179 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05637811311 (s) - for hierarchy setup = 1.735878841 (s) - for smoothers setup = 0.1373297703 (s) - for coarse setup = 2.071354538e-05 (s) - for final setup = 0.0002684788778 (s) Total for this setup = 1.930106839 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.244260147 2.244260147 2- Preconditioner apply = 6.441137746 a- first application(s) only = 0.3097679252 0.3097679252 b- remaining applications = 6.131369821 0.2919699915 3- Total time required by ML so far is 8.685397893 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 5.356798e-01 iter: 2 residual = 1.007187e-01 iter: 3 residual = 2.175870e-02 iter: 4 residual = 6.124371e-03 iter: 5 residual = 2.533747e-03 iter: 6 residual = 1.349345e-03 iter: 7 residual = 7.810624e-04 iter: 8 residual = 4.664975e-04 iter: 9 residual = 2.614912e-04 iter: 10 residual = 1.319896e-04 iter: 11 residual = 6.772759e-05 iter: 12 residual = 3.624091e-05 iter: 13 residual = 1.968152e-05 iter: 14 residual = 1.115031e-05 iter: 15 residual = 6.663471e-06 iter: 16 residual = 4.349663e-06 iter: 17 residual = 3.075604e-06 iter: 18 residual = 2.186935e-06 iter: 19 residual = 1.529942e-06 iter: 20 residual = 1.065039e-06 iter: 21 residual = 7.270533e-07 Solution time: 11.140791 (sec.) total iterations: 21 Total solve time = 11.1416 sec ************************************************************************ -- Nonlinear Solver Step 4 -- ||F|| = 1.103e+05 step = 5.000e-01 dx = 9.013e+05 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.6996e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.6031e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.6996e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.7699e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.8125e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.7699e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.1891e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 6.0892e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.1891e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 398113993 (398113993) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.581461e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8711e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.1240e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8711e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.781000352 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1226642542 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.07544380985 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001553343609 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001030787826 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001031495631 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.076672622 (s) - for hierarchy setup = 1.781033392 (s) - for smoothers setup = 0.1984696267 (s) - for coarse setup = 2.100039274e-05 (s) - for final setup = 0.0002579651773 (s) Total for this setup = 2.056697997 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.873739076 1.873739076 2- Preconditioner apply = 6.720721637 a- first application(s) only = 0.3016659142 0.3016659142 b- remaining applications = 6.419055723 0.2917752601 3- Total time required by ML so far is 8.594460713 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 4.059848e-01 iter: 2 residual = 8.399566e-02 iter: 3 residual = 1.894057e-02 iter: 4 residual = 5.771113e-03 iter: 5 residual = 2.604522e-03 iter: 6 residual = 1.368465e-03 iter: 7 residual = 7.289323e-04 iter: 8 residual = 4.418829e-04 iter: 9 residual = 2.911243e-04 iter: 10 residual = 1.768737e-04 iter: 11 residual = 9.917711e-05 iter: 12 residual = 5.091407e-05 iter: 13 residual = 2.551514e-05 iter: 14 residual = 1.434790e-05 iter: 15 residual = 8.428386e-06 iter: 16 residual = 4.992766e-06 iter: 17 residual = 3.209243e-06 iter: 18 residual = 2.144619e-06 iter: 19 residual = 1.462271e-06 iter: 20 residual = 1.065331e-06 iter: 21 residual = 7.972539e-07 Solution time: 11.671552 (sec.) total iterations: 21 Total solve time = 11.6724 sec ************************************************************************ -- Nonlinear Solver Step 5 -- ||F|| = 7.507e+04 step = 1.000e+00 dx = 8.452e+05 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.7759e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.5303e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.7759e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8128e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.7390e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8128e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2201e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 6.0043e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2201e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 602001354 (602001354) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.097704e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8955e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0324e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8955e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.707453495 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1228168169 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.04967997503 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001496477053 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001006945968 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001034531742 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05467423983 (s) - for hierarchy setup = 1.707486022 (s) - for smoothers setup = 0.1728505874 (s) - for coarse setup = 2.238899469e-05 (s) - for final setup = 0.0003047175705 (s) Total for this setup = 1.935586032 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.980031889 1.980031889 2- Preconditioner apply = 6.983529361 a- first application(s) only = 0.3621901413 0.3621901413 b- remaining applications = 6.62133922 0.3009699645 3- Total time required by ML so far is 8.96356125 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 2.305749e-01 iter: 2 residual = 5.340415e-02 iter: 3 residual = 1.311691e-02 iter: 4 residual = 4.937258e-03 iter: 5 residual = 2.744052e-03 iter: 6 residual = 1.662124e-03 iter: 7 residual = 9.388043e-04 iter: 8 residual = 5.057995e-04 iter: 9 residual = 2.855660e-04 iter: 10 residual = 1.661274e-04 iter: 11 residual = 9.929675e-05 iter: 12 residual = 6.620186e-05 iter: 13 residual = 4.193169e-05 iter: 14 residual = 2.304128e-05 iter: 15 residual = 1.360630e-05 iter: 16 residual = 8.422656e-06 iter: 17 residual = 5.005668e-06 iter: 18 residual = 3.128964e-06 iter: 19 residual = 2.033664e-06 iter: 20 residual = 1.285109e-06 iter: 21 residual = 7.864988e-07 Solution time: 11.926418 (sec.) total iterations: 21 Total solve time = 11.9272 sec ************************************************************************ -- Nonlinear Solver Step 6 -- ||F|| = 1.168e+04 step = 1.000e+00 dx = 3.715e+05 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.7856e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.5212e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.7856e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8238e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.7205e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8238e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2258e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.9889e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2258e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 691371600 (691371600) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 5.767250e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8841e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0752e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8841e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.826277699 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1225621011 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.06196413189 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001356378198 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001034885645 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001031355932 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07169538923 (s) - for hierarchy setup = 1.826309795 (s) - for smoothers setup = 0.184868495 (s) - for coarse setup = 2.119783312e-05 (s) - for final setup = 0.0002681342885 (s) Total for this setup = 2.08341424 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.880915793 1.880915793 2- Preconditioner apply = 7.159069621 a- first application(s) only = 0.304100574 0.304100574 b- remaining applications = 6.854969047 0.3115895021 3- Total time required by ML so far is 9.039985414 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 2.357714e-01 iter: 2 residual = 5.913673e-02 iter: 3 residual = 1.595471e-02 iter: 4 residual = 7.414499e-03 iter: 5 residual = 4.780850e-03 iter: 6 residual = 3.513965e-03 iter: 7 residual = 2.656120e-03 iter: 8 residual = 2.021359e-03 iter: 9 residual = 1.629451e-03 iter: 10 residual = 1.303847e-03 iter: 11 residual = 9.951388e-04 iter: 12 residual = 7.415314e-04 iter: 13 residual = 5.004084e-04 iter: 14 residual = 3.341993e-04 iter: 15 residual = 2.528998e-04 iter: 16 residual = 1.899350e-04 iter: 17 residual = 1.364277e-04 iter: 18 residual = 1.016627e-04 iter: 19 residual = 7.555713e-05 iter: 20 residual = 5.540045e-05 iter: 21 residual = 4.148561e-05 iter: 22 residual = 3.075964e-05 iter: 23 residual = 2.272542e-05 iter: 24 residual = 1.685838e-05 iter: 25 residual = 1.230795e-05 iter: 26 residual = 8.993878e-06 iter: 27 residual = 6.614321e-06 iter: 28 residual = 4.815634e-06 iter: 29 residual = 3.519356e-06 iter: 30 residual = 2.555498e-06 iter: 31 residual = 1.825331e-06 iter: 32 residual = 1.299808e-06 iter: 33 residual = 9.148179e-07 Solution time: 20.086424 (sec.) total iterations: 33 Total solve time = 20.0873 sec ************************************************************************ -- Nonlinear Solver Step 7 -- ||F|| = 4.874e+02 step = 1.000e+00 dx = 6.087e+04 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.7861e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.5208e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.7861e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8254e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.7179e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8254e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2308e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.9755e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2308e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 2099584695 (2099584695) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 5.194127e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8891e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0563e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8891e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.79887993 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1229092516 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.02945852559 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.000155328773 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 9.916163981e-05 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001099128276 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05660712253 (s) - for hierarchy setup = 1.798915742 (s) - for smoothers setup = 0.1527321804 (s) - for coarse setup = 2.195965499e-05 (s) - for final setup = 0.0002648495138 (s) Total for this setup = 2.008781109 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.011726635 2.011726635 2- Preconditioner apply = 11.01088872 a- first application(s) only = 0.3027710225 0.3027710225 b- remaining applications = 10.70811769 0.314944638 3- Total time required by ML so far is 13.02261535 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 2.223553e-01 iter: 2 residual = 5.707662e-02 iter: 3 residual = 1.676087e-02 iter: 4 residual = 8.685271e-03 iter: 5 residual = 6.194534e-03 iter: 6 residual = 5.160828e-03 iter: 7 residual = 4.317877e-03 iter: 8 residual = 3.371157e-03 iter: 9 residual = 2.648395e-03 iter: 10 residual = 2.030119e-03 iter: 11 residual = 1.481762e-03 iter: 12 residual = 1.092153e-03 iter: 13 residual = 7.787794e-04 iter: 14 residual = 5.603071e-04 iter: 15 residual = 4.346111e-04 iter: 16 residual = 3.326909e-04 iter: 17 residual = 2.410702e-04 iter: 18 residual = 1.799145e-04 iter: 19 residual = 1.371172e-04 iter: 20 residual = 1.021393e-04 iter: 21 residual = 7.700164e-05 iter: 22 residual = 5.813327e-05 iter: 23 residual = 4.323994e-05 iter: 24 residual = 3.234692e-05 iter: 25 residual = 2.402900e-05 iter: 26 residual = 1.769883e-05 iter: 27 residual = 1.316633e-05 iter: 28 residual = 9.731553e-06 iter: 29 residual = 7.109477e-06 iter: 30 residual = 5.188300e-06 iter: 31 residual = 3.742211e-06 iter: 32 residual = 2.672608e-06 iter: 33 residual = 1.889805e-06 iter: 34 residual = 1.324865e-06 iter: 35 residual = 9.249704e-07 Solution time: 22.292115 (sec.) total iterations: 35 Total solve time = 22.293 sec ************************************************************************ -- Nonlinear Solver Step 8 -- ||F|| = 2.682e+01 step = 1.000e+00 dx = 2.490e+03 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.7861e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.5208e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.7861e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8255e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.7178e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8255e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2280e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.9829e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2280e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 817280597 (817280597) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.936155e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8838e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0761e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8838e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.792855466 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1233920818 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05541256256 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001428313553 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001005111262 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.000109359622 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.06866157893 (s) - for hierarchy setup = 1.79289209 (s) - for smoothers setup = 0.1791573465 (s) - for coarse setup = 2.13868916e-05 (s) - for final setup = 0.0002576624975 (s) Total for this setup = 2.041230067 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.952180901 1.952180901 2- Preconditioner apply = 11.89226252 a- first application(s) only = 0.30876939 0.30876939 b- remaining applications = 11.58349313 0.321763698 3- Total time required by ML so far is 13.84444342 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 7.160541e-02 iter: 2 residual = 1.937017e-02 iter: 3 residual = 7.360665e-03 iter: 4 residual = 3.413513e-03 iter: 5 residual = 1.653656e-03 iter: 6 residual = 9.360932e-04 iter: 7 residual = 7.446814e-04 iter: 8 residual = 6.729044e-04 iter: 9 residual = 4.994202e-04 iter: 10 residual = 2.810684e-04 iter: 11 residual = 1.883637e-04 iter: 12 residual = 1.473063e-04 iter: 13 residual = 1.119689e-04 iter: 14 residual = 8.062927e-05 iter: 15 residual = 5.998784e-05 iter: 16 residual = 4.587332e-05 iter: 17 residual = 3.428118e-05 iter: 18 residual = 2.480866e-05 iter: 19 residual = 1.789891e-05 iter: 20 residual = 1.301296e-05 iter: 21 residual = 9.661968e-06 iter: 22 residual = 7.138130e-06 iter: 23 residual = 5.120036e-06 iter: 24 residual = 3.710413e-06 iter: 25 residual = 2.752054e-06 iter: 26 residual = 2.019835e-06 iter: 27 residual = 1.436398e-06 iter: 28 residual = 1.002906e-06 iter: 29 residual = 7.289408e-07 Solution time: 17.792090 (sec.) total iterations: 29 Total solve time = 17.793 sec ************************************************************************ -- Nonlinear Solver Step 9 -- ||F|| = 1.751e+00 step = 1.000e+00 dx = 7.980e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.7861e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.5208e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.7861e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8255e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.7178e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8255e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2335e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.9682e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2335e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 676628952 (676628952) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.201976e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8889e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0570e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8889e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.946528755 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.123247074 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.01285629906 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001541916281 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001030983403 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001023849472 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05755105801 (s) - for hierarchy setup = 1.946565944 (s) - for smoothers setup = 0.1364630479 (s) - for coarse setup = 2.112146467e-05 (s) - for final setup = 0.000282545574 (s) Total for this setup = 2.141130249 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.972573607 1.972573607 2- Preconditioner apply = 9.952421581 a- first application(s) only = 0.3038937198 0.3038937198 b- remaining applications = 9.648527862 0.3216175954 3- Total time required by ML so far is 11.92499519 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 4.263260e-02 iter: 2 residual = 1.324492e-02 iter: 3 residual = 5.303678e-03 iter: 4 residual = 2.109163e-03 iter: 5 residual = 8.821211e-04 iter: 6 residual = 3.462532e-04 iter: 7 residual = 1.253264e-04 iter: 8 residual = 3.879381e-05 iter: 9 residual = 1.306482e-05 iter: 10 residual = 7.936562e-06 iter: 11 residual = 6.726564e-06 iter: 12 residual = 5.643216e-06 iter: 13 residual = 4.130168e-06 iter: 14 residual = 2.610688e-06 iter: 15 residual = 1.542853e-06 iter: 16 residual = 1.100944e-06 iter: 17 residual = 9.018117e-07 Solution time: 9.197760 (sec.) total iterations: 17 Total solve time = 9.19862 sec ************************************************************************ -- Nonlinear Solver Step 10 -- ||F|| = 4.201e-03 step = 1.000e+00 dx = 2.104e-01 (Converged!) ************************************************************************ ************************************************************************ -- Final Status Test Results -- Converged....OR Combination -> Converged....AND Combination -> Converged....F-Norm = 3.532e-07 < 1.000e-05 (Length-Scaled Two-Norm, Absolute Tolerance) Converged....WRMS-Norm = 1.763e-03 < 1 (Min Step Size: 1.000e+00 >= 1) ??...........Number of Iterations = -1 < 40 ************************************************************************ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ End of Continuation Step 0 : Parameter: Glen's Law Homotopy Parameter = 0.000e+00 --> Step Converged in 10 Nonlinear Solver Iterations! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Calling Predictor with method: Constant ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Start of Continuation Step 1 : Parameter: Glen's Law Homotopy Parameter = 1.000e-01 from 0.000e+00 Continuation Method: Natural Current step size = 1.000e-01 Previous step size = 0.000e+00 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ************************************************************************ -- Nonlinear Solver Step 0 -- ||F|| = 6.677e+04 step = 0.000e+00 dx = 0.000e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.9738e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3545e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.9738e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8370e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6986e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8370e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2830e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.8387e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2830e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 920584619 (920584619) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.222380e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8960e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0304e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8960e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.784264698 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1233044416 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.0198436575 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001415293664 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001011779532 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001018168405 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07187443972 (s) - for hierarchy setup = 1.784299612 (s) - for smoothers setup = 0.1434926232 (s) - for coarse setup = 2.688262612e-05 (s) - for final setup = 0.0003247205168 (s) Total for this setup = 2.000252309 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.083587914 2.083587914 2- Preconditioner apply = 5.685316719 a- first application(s) only = 0.3423115434 0.3423115434 b- remaining applications = 5.343005176 0.2968336209 3- Total time required by ML so far is 7.768904633 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.509139e-01 iter: 2 residual = 2.250726e-02 iter: 3 residual = 4.151251e-03 iter: 4 residual = 9.314207e-04 iter: 5 residual = 3.148325e-04 iter: 6 residual = 1.908841e-04 iter: 7 residual = 1.306880e-04 iter: 8 residual = 7.794926e-05 iter: 9 residual = 4.787958e-05 iter: 10 residual = 2.880788e-05 iter: 11 residual = 1.728697e-05 iter: 12 residual = 1.121808e-05 iter: 13 residual = 7.247847e-06 iter: 14 residual = 4.233341e-06 iter: 15 residual = 2.324426e-06 iter: 16 residual = 1.317400e-06 iter: 17 residual = 8.108907e-07 Solution time: 8.900928 (sec.) total iterations: 17 Total solve time = 8.90174 sec ************************************************************************ -- Nonlinear Solver Step 1 -- ||F|| = 2.584e+04 step = 1.000e+00 dx = 1.189e+05 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.9640e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3627e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.9640e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8398e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6940e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8398e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.3102e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.7702e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.3102e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 84907399 (84907399) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.549838e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8825e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0809e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8825e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.797136608 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1209205035 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.0558557054 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001947814599 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001066587865 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.00010341499 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05688947625 (s) - for hierarchy setup = 1.797168702 (s) - for smoothers setup = 0.1771810642 (s) - for coarse setup = 2.128258348e-05 (s) - for final setup = 0.0002522869036 (s) Total for this setup = 2.031777093 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.928384188 1.928384188 2- Preconditioner apply = 5.548475154 a- first application(s) only = 0.317479345 0.317479345 b- remaining applications = 5.230995809 0.2906108783 3- Total time required by ML so far is 7.476859341 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 9.712860e-02 iter: 2 residual = 1.166011e-02 iter: 3 residual = 1.966141e-03 iter: 4 residual = 4.349986e-04 iter: 5 residual = 1.198973e-04 iter: 6 residual = 5.518526e-05 iter: 7 residual = 3.527879e-05 iter: 8 residual = 2.125602e-05 iter: 9 residual = 1.296914e-05 iter: 10 residual = 7.673631e-06 iter: 11 residual = 4.078990e-06 iter: 12 residual = 2.384486e-06 iter: 13 residual = 1.595478e-06 iter: 14 residual = 1.111599e-06 iter: 15 residual = 7.437625e-07 Solution time: 8.019859 (sec.) total iterations: 15 Total solve time = 8.02079 sec ************************************************************************ -- Nonlinear Solver Step 2 -- ||F|| = 2.789e+03 step = 1.000e+00 dx = 4.426e+03 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.9676e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3598e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.9676e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8401e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6934e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8401e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.3118e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.7661e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.3118e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 654191261 (654191261) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.112663e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8235e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.3099e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8235e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.780356069 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1225268263 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.06410335936 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001396303996 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001008147374 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001013027504 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07067028154 (s) - for hierarchy setup = 1.780395244 (s) - for smoothers setup = 0.1869719336 (s) - for coarse setup = 2.099573612e-05 (s) - for final setup = 0.0002784477547 (s) Total for this setup = 2.038586616 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.974894616 1.974894616 2- Preconditioner apply = 5.013622643 a- first application(s) only = 0.3355892263 0.3355892263 b- remaining applications = 4.678033417 0.2923770886 3- Total time required by ML so far is 6.988517259 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 9.818433e-02 iter: 2 residual = 1.352458e-02 iter: 3 residual = 2.610029e-03 iter: 4 residual = 5.785815e-04 iter: 5 residual = 1.352940e-04 iter: 6 residual = 3.620248e-05 iter: 7 residual = 1.550704e-05 iter: 8 residual = 9.865626e-06 iter: 9 residual = 6.136662e-06 iter: 10 residual = 3.870032e-06 iter: 11 residual = 2.548304e-06 iter: 12 residual = 1.559315e-06 iter: 13 residual = 1.064764e-06 iter: 14 residual = 8.359992e-07 Solution time: 8.237384 (sec.) total iterations: 14 Total solve time = 8.23826 sec ************************************************************************ -- Nonlinear Solver Step 3 -- ||F|| = 5.746e+01 step = 1.000e+00 dx = 1.381e+02 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 3.9676e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3597e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 3.9676e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8402e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6934e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8402e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.2971e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.8031e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.2971e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1389581740 (1389581740) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.048555e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8904e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0514e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8904e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.855485885 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1256598476 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.02681899443 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0004557669163 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001033907756 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001039542258 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05859547481 (s) - for hierarchy setup = 1.855521744 (s) - for smoothers setup = 0.153141954 (s) - for coarse setup = 2.15517357e-05 (s) - for final setup = 0.0002581384033 (s) Total for this setup = 2.067777006 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.967922441 1.967922441 2- Preconditioner apply = 5.039860121 a- first application(s) only = 0.34465631 0.34465631 b- remaining applications = 4.695203811 0.3130135874 3- Total time required by ML so far is 7.007782562 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.050102e-01 iter: 2 residual = 1.770420e-02 iter: 3 residual = 3.727843e-03 iter: 4 residual = 8.625316e-04 iter: 5 residual = 2.056868e-04 iter: 6 residual = 5.526161e-05 iter: 7 residual = 2.731940e-05 iter: 8 residual = 2.321535e-05 iter: 9 residual = 2.028872e-05 iter: 10 residual = 1.312431e-05 iter: 11 residual = 7.641337e-06 iter: 12 residual = 5.664241e-06 iter: 13 residual = 4.266404e-06 iter: 14 residual = 2.959672e-06 iter: 15 residual = 2.021535e-06 iter: 16 residual = 1.298124e-06 iter: 17 residual = 8.328908e-07 Solution time: 9.593891 (sec.) total iterations: 17 Total solve time = 9.59444 sec ************************************************************************ -- Nonlinear Solver Step 4 -- ||F|| = 5.696e-02 step = 1.000e+00 dx = 6.762e+00 (Converged!) ************************************************************************ ************************************************************************ -- Final Status Test Results -- Converged....OR Combination -> Converged....AND Combination -> Converged....F-Norm = 4.789e-06 < 1.000e-05 (Length-Scaled Two-Norm, Absolute Tolerance) Converged....WRMS-Norm = 5.623e-02 < 1 (Min Step Size: 1.000e+00 >= 1) ??...........Number of Iterations = -1 < 40 ************************************************************************ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ End of Continuation Step 1 : Parameter: Glen's Law Homotopy Parameter = 1.000e-01 from 0.000e+00 --> Step Converged in 4 Nonlinear Solver Iterations! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Calling Predictor with method: Constant ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Start of Continuation Step 2 : Parameter: Glen's Law Homotopy Parameter = 2.269e-01 from 1.000e-01 Continuation Method: Natural Current step size = 1.269e-01 Previous step size = 1.000e-01 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ************************************************************************ -- Nonlinear Solver Step 0 -- ||F|| = 5.396e+04 step = 0.000e+00 dx = 0.000e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0298e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3078e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0298e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8437e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6875e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8437e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.4473e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.4467e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.4473e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 25979533 (25979533) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.049055e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8965e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0288e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8965e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.7782989 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1260562018 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.04420465603 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001623062417 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001050690189 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.000104052946 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07202984951 (s) - for hierarchy setup = 1.778333483 (s) - for smoothers setup = 0.170632286 (s) - for coarse setup = 2.092961222e-05 (s) - for final setup = 0.000271323137 (s) Total for this setup = 2.021526522 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.009188998 2.009188998 2- Preconditioner apply = 6.032352105 a- first application(s) only = 0.3637622818 0.3637622818 b- remaining applications = 5.668589823 0.3149216568 3- Total time required by ML so far is 8.041541102 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.391904e-01 iter: 2 residual = 2.214245e-02 iter: 3 residual = 4.373758e-03 iter: 4 residual = 1.076007e-03 iter: 5 residual = 3.183095e-04 iter: 6 residual = 1.171524e-04 iter: 7 residual = 5.800927e-05 iter: 8 residual = 3.815200e-05 iter: 9 residual = 2.499538e-05 iter: 10 residual = 1.438047e-05 iter: 11 residual = 8.462813e-06 iter: 12 residual = 5.228220e-06 iter: 13 residual = 3.212755e-06 iter: 14 residual = 2.072344e-06 iter: 15 residual = 1.430998e-06 iter: 16 residual = 9.495798e-07 Solution time: 8.454967 (sec.) total iterations: 16 Total solve time = 8.45539 sec ************************************************************************ -- Nonlinear Solver Step 1 -- ||F|| = 3.171e+04 step = 1.000e+00 dx = 1.635e+04 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0355e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3032e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0355e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9215e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5627e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9215e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.5943e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.1383e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.5943e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1432667460 (1432667460) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.143100e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8998e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0166e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8998e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.761544126 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.12594839 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05401988514 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001656571403 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001017106697 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001047262922 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05866029393 (s) - for hierarchy setup = 1.761576791 (s) - for smoothers setup = 0.1803403692 (s) - for coarse setup = 2.083182335e-05 (s) - for final setup = 0.0002899216488 (s) Total for this setup = 2.001120427 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.949504433 1.949504433 2- Preconditioner apply = 5.342825414 a- first application(s) only = 0.3117262339 0.3117262339 b- remaining applications = 5.03109918 0.2959470106 3- Total time required by ML so far is 7.292329847 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 7.234888e-02 iter: 2 residual = 8.803542e-03 iter: 3 residual = 1.681207e-03 iter: 4 residual = 4.105585e-04 iter: 5 residual = 1.181020e-04 iter: 6 residual = 3.876541e-05 iter: 7 residual = 1.386024e-05 iter: 8 residual = 6.090857e-06 iter: 9 residual = 3.610871e-06 iter: 10 residual = 2.369746e-06 iter: 11 residual = 1.445351e-06 iter: 12 residual = 8.500693e-07 Solution time: 6.402820 (sec.) total iterations: 12 Total solve time = 6.40328 sec ************************************************************************ -- Nonlinear Solver Step 2 -- ||F|| = 4.806e+03 step = 1.000e+00 dx = 4.238e+02 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0353e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3034e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0353e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8942e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6058e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8942e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.5621e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.2027e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.5621e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 2120550377 (2120550377) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.078461e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9001e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0154e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9001e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.788772672 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1279153023 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05301466491 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.000199043192 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001060720533 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001080511138 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.0739309378 (s) - for hierarchy setup = 1.788804323 (s) - for smoothers setup = 0.1813431336 (s) - for coarse setup = 2.214405686e-05 (s) - for final setup = 0.0002589896321 (s) Total for this setup = 2.04460809 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.942468893 1.942468893 2- Preconditioner apply = 4.117532282 a- first application(s) only = 0.3047466036 0.3047466036 b- remaining applications = 3.812785679 0.293291206 3- Total time required by ML so far is 6.060001175 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 8.579394e-02 iter: 2 residual = 1.191717e-02 iter: 3 residual = 2.335095e-03 iter: 4 residual = 5.348403e-04 iter: 5 residual = 1.356457e-04 iter: 6 residual = 3.954483e-05 iter: 7 residual = 1.338662e-05 iter: 8 residual = 5.046781e-06 iter: 9 residual = 2.412356e-06 iter: 10 residual = 1.579887e-06 iter: 11 residual = 1.121067e-06 iter: 12 residual = 7.003427e-07 Solution time: 7.134409 (sec.) total iterations: 12 Total solve time = 7.13511 sec ************************************************************************ -- Nonlinear Solver Step 3 -- ||F|| = 2.637e+02 step = 1.000e+00 dx = 4.389e+01 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0353e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3034e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0353e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8907e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6114e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8907e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.5594e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.2082e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.5594e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1463556122 (1463556122) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.288259e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.7844e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.4702e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.7844e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.791861813 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1265496314 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.04698245227 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001562386751 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001007998362 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001066606492 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05968500581 (s) - for hierarchy setup = 1.791894284 (s) - for smoothers setup = 0.1738957828 (s) - for coarse setup = 2.167746425e-05 (s) - for final setup = 0.0002606827766 (s) Total for this setup = 2.025992636 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.970686134 1.970686134 2- Preconditioner apply = 4.562447651 a- first application(s) only = 0.3780242261 0.3780242261 b- remaining applications = 4.184423425 0.321878725 3- Total time required by ML so far is 6.533133785 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.020602e-01 iter: 2 residual = 1.717649e-02 iter: 3 residual = 3.522484e-03 iter: 4 residual = 7.976228e-04 iter: 5 residual = 1.902489e-04 iter: 6 residual = 4.813161e-05 iter: 7 residual = 1.396978e-05 iter: 8 residual = 5.810082e-06 iter: 9 residual = 3.930936e-06 iter: 10 residual = 3.186242e-06 iter: 11 residual = 2.377614e-06 iter: 12 residual = 1.559464e-06 iter: 13 residual = 1.152787e-06 iter: 14 residual = 9.805549e-07 Solution time: 7.659418 (sec.) total iterations: 14 Total solve time = 7.65994 sec ************************************************************************ -- Nonlinear Solver Step 4 -- ||F|| = 1.665e+00 step = 1.000e+00 dx = 5.216e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0352e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3034e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0352e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8905e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6116e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8905e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.5594e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.2083e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.5594e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 2068857944 (2068857944) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.160572e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8981e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0229e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8981e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 2.273370731 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1260156464 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05153094698 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001578424126 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001023188233 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001022499055 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07403703965 (s) - for hierarchy setup = 2.273404445 (s) - for smoothers setup = 0.1779090045 (s) - for coarse setup = 2.121180296e-05 (s) - for final setup = 0.0002651605755 (s) Total for this setup = 2.525870762 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.96631813 1.96631813 2- Preconditioner apply = 4.926583112 a- first application(s) only = 0.3306755563 0.3306755563 b- remaining applications = 4.595907556 0.3063938371 3- Total time required by ML so far is 6.892901242 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.198168e-01 iter: 2 residual = 2.243633e-02 iter: 3 residual = 4.839011e-03 iter: 4 residual = 1.146316e-03 iter: 5 residual = 3.097478e-04 iter: 6 residual = 1.329103e-04 iter: 7 residual = 1.036944e-04 iter: 8 residual = 9.347586e-05 iter: 9 residual = 7.236262e-05 iter: 10 residual = 4.081642e-05 iter: 11 residual = 2.672831e-05 iter: 12 residual = 2.102975e-05 iter: 13 residual = 1.419997e-05 iter: 14 residual = 6.932830e-06 iter: 15 residual = 3.958620e-06 iter: 16 residual = 3.190100e-06 iter: 17 residual = 2.733192e-06 iter: 18 residual = 2.016797e-06 iter: 19 residual = 1.327341e-06 iter: 20 residual = 9.164210e-07 Solution time: 11.214538 (sec.) total iterations: 20 Total solve time = 11.2151 sec ************************************************************************ -- Nonlinear Solver Step 5 -- ||F|| = 7.454e-04 step = 1.000e+00 dx = 2.525e-01 (Converged!) ************************************************************************ ************************************************************************ -- Final Status Test Results -- Converged....OR Combination -> Converged....AND Combination -> Converged....F-Norm = 6.267e-08 < 1.000e-05 (Length-Scaled Two-Norm, Absolute Tolerance) Converged....WRMS-Norm = 2.108e-03 < 1 (Min Step Size: 1.000e+00 >= 1) ??...........Number of Iterations = -1 < 40 ************************************************************************ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ End of Continuation Step 2 : Parameter: Glen's Law Homotopy Parameter = 2.269e-01 from 1.000e-01 --> Step Converged in 5 Nonlinear Solver Iterations! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Calling Predictor with method: Constant ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Start of Continuation Step 3 : Parameter: Glen's Law Homotopy Parameter = 3.820e-01 from 2.269e-01 Continuation Method: Natural Current step size = 1.551e-01 Previous step size = 1.269e-01 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ************************************************************************ -- Nonlinear Solver Step 0 -- ||F|| = 2.883e+04 step = 0.000e+00 dx = 0.000e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0401e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2994e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0401e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.8840e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.6221e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.8840e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.5970e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 5.1329e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.5970e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 329956843 (329956843) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.145755e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8848e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0724e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8848e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.839301012 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1296353163 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.02933955286 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001537287608 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001073749736 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001060860232 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07165346015 (s) - for hierarchy setup = 1.839333929 (s) - for smoothers setup = 0.1593420589 (s) - for coarse setup = 2.140924335e-05 (s) - for final setup = 0.0002877525985 (s) Total for this setup = 2.070871699 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.451841886 2.451841886 2- Preconditioner apply = 6.772880689 a- first application(s) only = 0.3381091878 0.3381091878 b- remaining applications = 6.434771501 0.3064176905 3- Total time required by ML so far is 9.224722575 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.311519e-01 iter: 2 residual = 2.300372e-02 iter: 3 residual = 5.294280e-03 iter: 4 residual = 1.461529e-03 iter: 5 residual = 4.317840e-04 iter: 6 residual = 1.372970e-04 iter: 7 residual = 4.651249e-05 iter: 8 residual = 1.771078e-05 iter: 9 residual = 8.141889e-06 iter: 10 residual = 4.898952e-06 iter: 11 residual = 3.389900e-06 iter: 12 residual = 2.215470e-06 iter: 13 residual = 1.260453e-06 iter: 14 residual = 7.031760e-07 Solution time: 7.689289 (sec.) total iterations: 14 Total solve time = 7.6898 sec ************************************************************************ -- Nonlinear Solver Step 1 -- ||F|| = 2.132e+04 step = 1.000e+00 dx = 1.268e+03 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0406e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2990e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0406e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9272e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5538e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9272e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7371e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8701e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7371e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1742511399 (1742511399) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.161414e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8960e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0306e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8960e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.892438041 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1828493215 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.0001228256151 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0006111804396 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001019528136 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.000112737529 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07419436239 (s) - for hierarchy setup = 1.892473374 (s) - for smoothers setup = 0.1837980179 (s) - for coarse setup = 2.223439515e-05 (s) - for final setup = 0.0002592615783 (s) Total for this setup = 2.150982032 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.999227561 1.999227561 2- Preconditioner apply = 4.892603434 a- first application(s) only = 0.3331096638 0.3331096638 b- remaining applications = 4.55949377 0.3039662513 3- Total time required by ML so far is 6.891830995 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 6.041552e-02 iter: 2 residual = 8.273816e-03 iter: 3 residual = 1.976138e-03 iter: 4 residual = 5.515431e-04 iter: 5 residual = 1.707578e-04 iter: 6 residual = 5.545487e-05 iter: 7 residual = 1.813980e-05 iter: 8 residual = 6.493428e-06 iter: 9 residual = 2.378263e-06 iter: 10 residual = 8.805289e-07 Solution time: 5.817180 (sec.) total iterations: 10 Total solve time = 5.81777 sec ************************************************************************ -- Nonlinear Solver Step 2 -- ||F|| = 3.755e+03 step = 1.000e+00 dx = 9.825e+01 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0406e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2990e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0406e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9234e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5597e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9234e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7091e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.9204e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7091e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1627896877 (1627896877) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.948407e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9059e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9942e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9059e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.88035378 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1237033755 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05577942915 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.000179531984 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001069065183 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001008845866 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05943224486 (s) - for hierarchy setup = 1.880385698 (s) - for smoothers setup = 0.1798701277 (s) - for coarse setup = 2.725422382e-05 (s) - for final setup = 0.0002615023404 (s) Total for this setup = 2.120222498 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.076796189 2.076796189 2- Preconditioner apply = 3.730081534 a- first application(s) only = 0.3489543442 0.3489543442 b- remaining applications = 3.38112719 0.3073751991 3- Total time required by ML so far is 5.806877723 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 7.775227e-02 iter: 2 residual = 1.214668e-02 iter: 3 residual = 2.750264e-03 iter: 4 residual = 7.755095e-04 iter: 5 residual = 2.352712e-04 iter: 6 residual = 7.631456e-05 iter: 7 residual = 2.542175e-05 iter: 8 residual = 8.604333e-06 iter: 9 residual = 3.107087e-06 iter: 10 residual = 1.188467e-06 iter: 11 residual = 5.024096e-07 Solution time: 6.403108 (sec.) total iterations: 11 Total solve time = 6.40368 sec ************************************************************************ -- Nonlinear Solver Step 3 -- ||F|| = 3.926e+02 step = 1.000e+00 dx = 2.084e+01 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0405e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2991e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0405e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9233e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5599e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9233e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7076e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.9232e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7076e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1335441364 (1335441364) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.073014e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9071e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9897e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9071e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.861027014 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1254428439 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05401917547 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001187622547 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001086443663 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.000105814077 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.06840681843 (s) - for hierarchy setup = 1.861059436 (s) - for smoothers setup = 0.1797952401 (s) - for coarse setup = 2.441927791e-05 (s) - for final setup = 0.0002546776086 (s) Total for this setup = 2.109773268 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.060798974 2.060798974 2- Preconditioner apply = 4.120096179 a- first application(s) only = 0.3505441546 0.3505441546 b- remaining applications = 3.769552024 0.3141293353 3- Total time required by ML so far is 6.180895153 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.291930e-01 iter: 2 residual = 2.829096e-02 iter: 3 residual = 7.357812e-03 iter: 4 residual = 1.905954e-03 iter: 5 residual = 4.911277e-04 iter: 6 residual = 1.272517e-04 iter: 7 residual = 3.459800e-05 iter: 8 residual = 1.061162e-05 iter: 9 residual = 3.969745e-06 iter: 10 residual = 1.773349e-06 iter: 11 residual = 9.925827e-07 Solution time: 6.209726 (sec.) total iterations: 11 Total solve time = 6.21023 sec ************************************************************************ -- Nonlinear Solver Step 4 -- ||F|| = 2.212e+01 step = 1.000e+00 dx = 5.550e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0406e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2990e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0406e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9234e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5598e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9234e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7076e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.9232e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7076e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1670684817 (1670684817) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.054737e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8968e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0277e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8968e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.746495215 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1299385773 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.0477498211 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001498078927 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001071086153 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001057377085 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05968348496 (s) - for hierarchy setup = 1.746530572 (s) - for smoothers setup = 0.1780510526 (s) - for coarse setup = 2.188701183e-05 (s) - for final setup = 0.0002603642642 (s) Total for this setup = 1.984799294 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.041374233 2.041374233 2- Preconditioner apply = 4.025134125 a- first application(s) only = 0.3429045072 0.3429045072 b- remaining applications = 3.682229618 0.3068524681 3- Total time required by ML so far is 6.066508357 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.337001e-01 iter: 2 residual = 3.070996e-02 iter: 3 residual = 9.113150e-03 iter: 4 residual = 2.479135e-03 iter: 5 residual = 6.369839e-04 iter: 6 residual = 1.599666e-04 iter: 7 residual = 4.070521e-05 iter: 8 residual = 1.229576e-05 iter: 9 residual = 7.071990e-06 iter: 10 residual = 5.952462e-06 iter: 11 residual = 4.395355e-06 iter: 12 residual = 2.005980e-06 iter: 13 residual = 1.075776e-06 iter: 14 residual = 8.727745e-07 Solution time: 8.004019 (sec.) total iterations: 14 Total solve time = 8.00459 sec ************************************************************************ -- Nonlinear Solver Step 5 -- ||F|| = 8.430e-01 step = 1.000e+00 dx = 2.988e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0405e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2991e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0405e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9234e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5597e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9234e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7076e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.9232e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7076e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 412020830 (412020830) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.811909e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8725e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.1189e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8725e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 2.371517743 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1274312502 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.04978829157 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001626266167 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001019202173 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.000111383386 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07668140903 (s) - for hierarchy setup = 2.371550431 (s) - for smoothers setup = 0.177595472 (s) - for coarse setup = 2.175569534e-05 (s) - for final setup = 0.0002743313089 (s) Total for this setup = 2.626370852 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.925126449 1.925126449 2- Preconditioner apply = 5.030931413 a- first application(s) only = 0.3221142795 0.3221142795 b- remaining applications = 4.708817134 0.3139211422 3- Total time required by ML so far is 6.956057862 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.290875e-01 iter: 2 residual = 3.295507e-02 iter: 3 residual = 1.014389e-02 iter: 4 residual = 2.739751e-03 iter: 5 residual = 7.110000e-04 iter: 6 residual = 1.769441e-04 iter: 7 residual = 4.620608e-05 iter: 8 residual = 1.814561e-05 iter: 9 residual = 1.406590e-05 iter: 10 residual = 1.278017e-05 iter: 11 residual = 9.161048e-06 iter: 12 residual = 3.852289e-06 iter: 13 residual = 2.206428e-06 iter: 14 residual = 1.832341e-06 iter: 15 residual = 1.369539e-06 iter: 16 residual = 8.203897e-07 Solution time: 9.062006 (sec.) total iterations: 16 Total solve time = 9.06259 sec ************************************************************************ -- Nonlinear Solver Step 6 -- ||F|| = 2.196e-03 step = 1.000e+00 dx = 4.157e-02 (Converged!) ************************************************************************ ************************************************************************ -- Final Status Test Results -- Converged....OR Combination -> Converged....AND Combination -> Converged....F-Norm = 1.846e-07 < 1.000e-05 (Length-Scaled Two-Norm, Absolute Tolerance) Converged....WRMS-Norm = 3.465e-04 < 1 (Min Step Size: 1.000e+00 >= 1) ??...........Number of Iterations = -1 < 40 ************************************************************************ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ End of Continuation Step 3 : Parameter: Glen's Law Homotopy Parameter = 3.820e-01 from 2.269e-01 --> Step Converged in 6 Nonlinear Solver Iterations! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Calling Predictor with method: Constant ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Start of Continuation Step 4 : Parameter: Glen's Law Homotopy Parameter = 5.650e-01 from 3.820e-01 Continuation Method: Natural Current step size = 1.830e-01 Previous step size = 1.551e-01 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ************************************************************************ -- Nonlinear Solver Step 0 -- ||F|| = 7.383e+03 step = 0.000e+00 dx = 0.000e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9224e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5613e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9224e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7071e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.9241e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7071e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1951550994 (1951550994) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.200248e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9036e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0026e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9036e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.998482938 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1254000869 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05339440145 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.004142715596 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001065153629 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001033497974 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05946479272 (s) - for hierarchy setup = 1.998516414 (s) - for smoothers setup = 0.1831470691 (s) - for coarse setup = 2.142507583e-05 (s) - for final setup = 0.0002642218024 (s) Total for this setup = 2.241669415 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.549700283 2.549700283 2- Preconditioner apply = 5.617107807 a- first application(s) only = 0.3432590561 0.3432590561 b- remaining applications = 5.273848751 0.3102263971 3- Total time required by ML so far is 8.16680809 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.151984e-01 iter: 2 residual = 2.465328e-02 iter: 3 residual = 6.594981e-03 iter: 4 residual = 2.071237e-03 iter: 5 residual = 6.729089e-04 iter: 6 residual = 2.255327e-04 iter: 7 residual = 7.708069e-05 iter: 8 residual = 2.751004e-05 iter: 9 residual = 1.010374e-05 iter: 10 residual = 3.758446e-06 iter: 11 residual = 1.452895e-06 iter: 12 residual = 6.414079e-07 Solution time: 6.581127 (sec.) total iterations: 12 Total solve time = 6.58175 sec ************************************************************************ -- Nonlinear Solver Step 1 -- ||F|| = 4.706e+03 step = 1.000e+00 dx = 1.090e+02 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7340e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8756e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7340e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1203571492 (1203571492) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.009249e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8993e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0184e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8993e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.85036161 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1237793081 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05238816701 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0002168072388 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001051463187 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001044729725 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07078851666 (s) - for hierarchy setup = 1.850398105 (s) - for smoothers setup = 0.1765939016 (s) - for coarse setup = 2.208352089e-05 (s) - for final setup = 0.0002522123978 (s) Total for this setup = 2.098296378 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.182215245 2.182215245 2- Preconditioner apply = 4.328117775 a- first application(s) only = 0.3154544616 0.3154544616 b- remaining applications = 4.012663313 0.3086664087 3- Total time required by ML so far is 6.51033302 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 5.460595e-02 iter: 2 residual = 7.943674e-03 iter: 3 residual = 2.157319e-03 iter: 4 residual = 6.853582e-04 iter: 5 residual = 2.344264e-04 iter: 6 residual = 8.525344e-05 iter: 7 residual = 3.345737e-05 iter: 8 residual = 1.389339e-05 iter: 9 residual = 5.829900e-06 iter: 10 residual = 2.301448e-06 iter: 11 residual = 8.761918e-07 Solution time: 6.333568 (sec.) total iterations: 11 Total solve time = 6.33417 sec ************************************************************************ -- Nonlinear Solver Step 2 -- ||F|| = 9.122e+02 step = 1.000e+00 dx = 7.837e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7312e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8807e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7312e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 620654105 (620654105) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.053711e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9047e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9984e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9047e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.949736614 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1282772245 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05129309744 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001511201262 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001049535349 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001083267853 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.06004251912 (s) - for hierarchy setup = 1.949773381 (s) - for smoothers setup = 0.1799347224 (s) - for coarse setup = 2.221390605e-05 (s) - for final setup = 0.0002567470074 (s) Total for this setup = 2.190276323 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.027516393 2.027516393 2- Preconditioner apply = 4.050841521 a- first application(s) only = 0.3425862622 0.3425862622 b- remaining applications = 3.708255259 0.3090212716 3- Total time required by ML so far is 6.078357914 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 7.104222e-02 iter: 2 residual = 1.074694e-02 iter: 3 residual = 2.713192e-03 iter: 4 residual = 8.904259e-04 iter: 5 residual = 3.085506e-04 iter: 6 residual = 1.095286e-04 iter: 7 residual = 4.085284e-05 iter: 8 residual = 1.579427e-05 iter: 9 residual = 6.293765e-06 iter: 10 residual = 2.651304e-06 iter: 11 residual = 1.122771e-06 iter: 12 residual = 4.576638e-07 Solution time: 7.070630 (sec.) total iterations: 12 Total solve time = 7.07128 sec ************************************************************************ -- Nonlinear Solver Step 3 -- ||F|| = 1.182e+02 step = 1.000e+00 dx = 1.651e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0405e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2991e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0405e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7310e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8810e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7310e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 517259346 (517259346) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 6.651076e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8806e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0883e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8806e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.874283757 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1385902651 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.03618159704 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.000149436295 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001314738765 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001071458682 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07325576991 (s) - for hierarchy setup = 1.874316057 (s) - for smoothers setup = 0.1751599181 (s) - for coarse setup = 3.216788173e-05 (s) - for final setup = 0.0003472305834 (s) Total for this setup = 2.12335436 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.130242415 2.130242415 2- Preconditioner apply = 4.460790899 a- first application(s) only = 0.3664595438 0.3664595438 b- remaining applications = 4.094331355 0.3149485658 3- Total time required by ML so far is 6.591033313 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 8.315969e-02 iter: 2 residual = 1.190661e-02 iter: 3 residual = 2.487689e-03 iter: 4 residual = 6.823218e-04 iter: 5 residual = 2.265443e-04 iter: 6 residual = 8.682742e-05 iter: 7 residual = 3.394128e-05 iter: 8 residual = 1.265459e-05 iter: 9 residual = 4.757834e-06 iter: 10 residual = 1.918703e-06 iter: 11 residual = 8.157621e-07 Solution time: 6.617700 (sec.) total iterations: 11 Total solve time = 6.61833 sec ************************************************************************ -- Nonlinear Solver Step 4 -- ||F|| = 6.240e+00 step = 1.000e+00 dx = 3.493e-01 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0404e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2992e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0404e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7310e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8809e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7310e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 753906116 (753906116) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.219788e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8872e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0635e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8872e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.796009197 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1253842385 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05480952188 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001449482515 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001050392166 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001027416438 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.0596293658 (s) - for hierarchy setup = 1.796045232 (s) - for smoothers setup = 0.1805464895 (s) - for coarse setup = 2.139247954e-05 (s) - for final setup = 0.0002684481442 (s) Total for this setup = 2.036762018 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.050108647 2.050108647 2- Preconditioner apply = 4.234933971 a- first application(s) only = 0.368636149 0.368636149 b- remaining applications = 3.866297822 0.3221914852 3- Total time required by ML so far is 6.285042618 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 8.896548e-02 iter: 2 residual = 1.237153e-02 iter: 3 residual = 2.435976e-03 iter: 4 residual = 5.909596e-04 iter: 5 residual = 1.585008e-04 iter: 6 residual = 4.614393e-05 iter: 7 residual = 1.526607e-05 iter: 8 residual = 6.112549e-06 iter: 9 residual = 2.892183e-06 iter: 10 residual = 1.667516e-06 iter: 11 residual = 1.251493e-06 iter: 12 residual = 9.999978e-07 Solution time: 6.827169 (sec.) total iterations: 12 Total solve time = 6.8279 sec ************************************************************************ -- Nonlinear Solver Step 5 -- ||F|| = 1.330e-01 step = 1.000e+00 dx = 1.240e-01 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7310e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8810e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7310e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 818514217 (818514217) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.187396e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9018e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0091e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9018e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.836785209 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1260878369 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05208005756 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001526568085 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001045400277 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001042932272 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.06888455059 (s) - for hierarchy setup = 1.836820429 (s) - for smoothers setup = 0.1785293845 (s) - for coarse setup = 2.113729715e-05 (s) - for final setup = 0.0003506150097 (s) Total for this setup = 2.08484611 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.977142915 1.977142915 2- Preconditioner apply = 4.311474845 a- first application(s) only = 0.3374475939 0.3374475939 b- remaining applications = 3.974027251 0.3056944039 3- Total time required by ML so far is 6.28861776 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 9.629732e-02 iter: 2 residual = 1.585119e-02 iter: 3 residual = 3.446811e-03 iter: 4 residual = 8.539294e-04 iter: 5 residual = 2.237242e-04 iter: 6 residual = 6.356326e-05 iter: 7 residual = 2.746881e-05 iter: 8 residual = 2.097445e-05 iter: 9 residual = 1.727501e-05 iter: 10 residual = 9.668437e-06 iter: 11 residual = 4.128854e-06 iter: 12 residual = 2.828393e-06 iter: 13 residual = 2.509740e-06 iter: 14 residual = 2.077448e-06 iter: 15 residual = 1.285701e-06 iter: 16 residual = 7.747418e-07 Solution time: 9.327104 (sec.) total iterations: 16 Total solve time = 9.32784 sec ************************************************************************ -- Nonlinear Solver Step 6 -- ||F|| = 1.754e-04 step = 1.000e+00 dx = 4.372e-02 (Converged!) ************************************************************************ ************************************************************************ -- Final Status Test Results -- Converged....OR Combination -> Converged....AND Combination -> Converged....F-Norm = 1.475e-08 < 1.000e-05 (Length-Scaled Two-Norm, Absolute Tolerance) Converged....WRMS-Norm = 3.636e-04 < 1 (Min Step Size: 1.000e+00 >= 1) ??...........Number of Iterations = -1 < 40 ************************************************************************ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ End of Continuation Step 4 : Parameter: Glen's Law Homotopy Parameter = 5.650e-01 from 3.820e-01 --> Step Converged in 6 Nonlinear Solver Iterations! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Calling Predictor with method: Constant ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Start of Continuation Step 5 : Parameter: Glen's Law Homotopy Parameter = 7.809e-01 from 5.650e-01 Continuation Method: Natural Current step size = 2.159e-01 Previous step size = 1.830e-01 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ************************************************************************ -- Nonlinear Solver Step 0 -- ||F|| = 9.358e+02 step = 0.000e+00 dx = 0.000e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7326e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8781e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7326e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 875110458 (875110458) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.202014e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8992e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0189e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8992e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.799254078 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1266156044 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05087492149 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001961141825 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001073190942 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001042103395 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05872923322 (s) - for hierarchy setup = 1.799288813 (s) - for smoothers setup = 0.1778981695 (s) - for coarse setup = 2.392102033e-05 (s) - for final setup = 0.0002871546894 (s) Total for this setup = 2.036468451 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.015972883 2.015972883 2- Preconditioner apply = 5.821463775 a- first application(s) only = 0.3671266548 0.3671266548 b- remaining applications = 5.45433712 0.32084336 3- Total time required by ML so far is 7.837436657 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 9.952232e-02 iter: 2 residual = 2.335761e-02 iter: 3 residual = 6.634569e-03 iter: 4 residual = 2.172020e-03 iter: 5 residual = 7.525405e-04 iter: 6 residual = 2.754761e-04 iter: 7 residual = 9.842517e-05 iter: 8 residual = 3.603422e-05 iter: 9 residual = 1.358247e-05 iter: 10 residual = 5.080279e-06 iter: 11 residual = 1.967384e-06 iter: 12 residual = 8.256005e-07 Solution time: 6.483789 (sec.) total iterations: 12 Total solve time = 6.48433 sec ************************************************************************ -- Nonlinear Solver Step 1 -- ||F|| = 4.321e+02 step = 1.000e+00 dx = 4.123e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0403e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2993e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0403e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5575e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7356e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8728e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7356e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 897800488 (897800488) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 5.299564e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9250e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9246e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9250e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.8076499 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1267525554 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.0491929939 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001438781619 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001042876393 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001071346924 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07388265152 (s) - for hierarchy setup = 1.807681795 (s) - for smoothers setup = 0.1763008498 (s) - for coarse setup = 2.799089998e-05 (s) - for final setup = 0.0002623191103 (s) Total for this setup = 2.058398393 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.977749299 1.977749299 2- Preconditioner apply = 4.195266554 a- first application(s) only = 0.3333973046 0.3333973046 b- remaining applications = 3.861869249 0.2970668653 3- Total time required by ML so far is 6.173015853 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 5.183906e-02 iter: 2 residual = 1.155028e-02 iter: 3 residual = 8.290108e-03 iter: 4 residual = 6.262156e-03 iter: 5 residual = 3.284050e-03 iter: 6 residual = 1.263000e-03 iter: 7 residual = 4.597154e-04 iter: 8 residual = 1.737133e-04 iter: 9 residual = 6.813871e-05 iter: 10 residual = 2.654345e-05 iter: 11 residual = 1.009778e-05 iter: 12 residual = 3.732809e-06 iter: 13 residual = 1.329801e-06 iter: 14 residual = 4.728512e-07 Solution time: 7.948353 (sec.) total iterations: 14 Total solve time = 7.949 sec ************************************************************************ -- Nonlinear Solver Step 2 -- ||F|| = 1.048e+02 step = 1.000e+00 dx = 1.837e-01 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0363e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.3025e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0363e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9250e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5572e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9250e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7356e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8728e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7356e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 2135176499 (2135176499) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.015988e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9054e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9958e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9054e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.786298066 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1239651265 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05344059225 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.000210037455 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001060860232 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001062257215 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05954343732 (s) - for hierarchy setup = 1.786331123 (s) - for smoothers setup = 0.1778280679 (s) - for coarse setup = 2.460181713e-05 (s) - for final setup = 0.0002552848309 (s) Total for this setup = 2.024232842 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.984526256 1.984526256 2- Preconditioner apply = 4.980336213 a- first application(s) only = 0.3327474669 0.3327474669 b- remaining applications = 4.647588747 0.3098392498 3- Total time required by ML so far is 6.96486247 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 6.457826e-02 iter: 2 residual = 7.821342e-03 iter: 3 residual = 2.110069e-03 iter: 4 residual = 9.093015e-04 iter: 5 residual = 3.371330e-04 iter: 6 residual = 1.101309e-04 iter: 7 residual = 3.554736e-05 iter: 8 residual = 1.172159e-05 iter: 9 residual = 4.024627e-06 iter: 10 residual = 1.408288e-06 iter: 11 residual = 4.800163e-07 Solution time: 6.393690 (sec.) total iterations: 11 Total solve time = 6.3943 sec ************************************************************************ -- Nonlinear Solver Step 3 -- ||F|| = 5.265e+01 step = 1.000e+00 dx = 1.636e-02 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7353e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8733e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7353e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 151616635 (151616635) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.166232e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9023e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0072e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9023e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.929595393 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1284302492 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05209047999 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001463415101 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001089069992 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001118145883 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07153728977 (s) - for hierarchy setup = 1.9296274 (s) - for smoothers setup = 0.1808877923 (s) - for coarse setup = 2.491101623e-05 (s) - for final setup = 0.000261801295 (s) Total for this setup = 2.182581238 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.964697891 1.964697891 2- Preconditioner apply = 4.008261213 a- first application(s) only = 0.3357735155 0.3357735155 b- remaining applications = 3.672487698 0.3060406415 3- Total time required by ML so far is 5.972959104 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 6.690076e-02 iter: 2 residual = 7.979465e-03 iter: 3 residual = 1.242946e-03 iter: 4 residual = 2.354031e-04 iter: 5 residual = 5.786183e-05 iter: 6 residual = 2.037926e-05 iter: 7 residual = 8.959140e-06 iter: 8 residual = 3.430978e-06 iter: 9 residual = 1.148759e-06 iter: 10 residual = 4.035716e-07 Solution time: 5.699007 (sec.) total iterations: 10 Total solve time = 5.69969 sec ************************************************************************ -- Nonlinear Solver Step 4 -- ||F|| = 2.361e+00 step = 1.000e+00 dx = 1.201e-03 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9247e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5578e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9247e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7353e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8733e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7353e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1354965967 (1354965967) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.296651e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8903e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0519e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8903e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.810060092 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1315035131 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.04799638968 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001569148153 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001051314175 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001032985747 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.06059014704 (s) - for hierarchy setup = 1.810095973 (s) - for smoothers setup = 0.1798652476 (s) - for coarse setup = 2.547539771e-05 (s) - for final setup = 0.0002764118835 (s) Total for this setup = 2.051120718 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.111051974 2.111051974 2- Preconditioner apply = 3.638771482 a- first application(s) only = 0.3078455115 0.3078455115 b- remaining applications = 3.33092597 0.3028114518 3- Total time required by ML so far is 5.749823456 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 1.053834e-01 iter: 2 residual = 1.757590e-02 iter: 3 residual = 3.336412e-03 iter: 4 residual = 6.550028e-04 iter: 5 residual = 1.376546e-04 iter: 6 residual = 3.993409e-05 iter: 7 residual = 1.567258e-05 iter: 8 residual = 6.093725e-06 iter: 9 residual = 2.103906e-06 iter: 10 residual = 7.645896e-07 Solution time: 5.499018 (sec.) total iterations: 10 Total solve time = 5.49977 sec ************************************************************************ -- Nonlinear Solver Step 5 -- ||F|| = 3.743e-02 step = 1.000e+00 dx = 1.854e-04 (Converged!) ************************************************************************ ************************************************************************ -- Final Status Test Results -- Converged....OR Combination -> Converged....AND Combination -> Converged....F-Norm = 3.147e-06 < 1.000e-05 (Length-Scaled Two-Norm, Absolute Tolerance) Converged....WRMS-Norm = 1.553e-06 < 1 (Min Step Size: 1.000e+00 >= 1) ??...........Number of Iterations = -1 < 40 ************************************************************************ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ End of Continuation Step 5 : Parameter: Glen's Law Homotopy Parameter = 7.809e-01 from 5.650e-01 --> Step Converged in 5 Nonlinear Solver Iterations! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Calling Predictor with method: Constant ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Start of Continuation Step 6 : Parameter: Glen's Law Homotopy Parameter = 1.000e+00 from 7.809e-01 Continuation Method: Natural Current step size = 2.191e-01 Previous step size = 2.159e-01 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ************************************************************************ -- Nonlinear Solver Step 0 -- ||F|| = 4.483e+01 step = 0.000e+00 dx = 0.000e+00 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7353e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8733e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7353e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1256824689 (1256824689) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.110464e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9261e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9207e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9261e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.750183145 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1292273123 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.04521856364 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0007318006828 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.000108201988 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001020133495 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07533970568 (s) - for hierarchy setup = 1.750221847 (s) - for smoothers setup = 0.1753878919 (s) - for coarse setup = 2.185348421e-05 (s) - for final setup = 0.0002573067322 (s) Total for this setup = 2.001472936 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.990539062 1.990539062 2- Preconditioner apply = 3.691760406 a- first application(s) only = 0.3543220116 0.3543220116 b- remaining applications = 3.337438394 0.3034034904 3- Total time required by ML so far is 5.682299468 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 7.906521e-02 iter: 2 residual = 1.652649e-02 iter: 3 residual = 4.500975e-03 iter: 4 residual = 1.565606e-03 iter: 5 residual = 5.344594e-04 iter: 6 residual = 1.846938e-04 iter: 7 residual = 6.896019e-05 iter: 8 residual = 2.734101e-05 iter: 9 residual = 1.056601e-05 iter: 10 residual = 3.868140e-06 iter: 11 residual = 1.344509e-06 iter: 12 residual = 4.793410e-07 Solution time: 6.583227 (sec.) total iterations: 12 Total solve time = 6.58388 sec ************************************************************************ -- Nonlinear Solver Step 1 -- ||F|| = 5.206e+00 step = 1.000e+00 dx = 4.690e-02 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0407e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2989e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0407e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9219e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5622e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9219e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7357e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8726e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7357e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1491994894 (1491994894) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.486518e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.8958e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 7.0313e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.8958e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 1.740900062 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1277559996 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.05289341509 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.000147420913 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001048212871 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.0001055141911 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.05719226319 (s) - for hierarchy setup = 1.740934821 (s) - for smoothers setup = 0.181007171 (s) - for coarse setup = 2.461671829e-05 (s) - for final setup = 0.0002787392586 (s) Total for this setup = 1.979679598 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.9261422 1.9261422 2- Preconditioner apply = 4.322863805 a- first application(s) only = 0.3598683039 0.3598683039 b- remaining applications = 3.962995501 0.3048458078 3- Total time required by ML so far is 6.249006005 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 5.079342e-02 iter: 2 residual = 7.412457e-03 iter: 3 residual = 1.590533e-03 iter: 4 residual = 5.289462e-04 iter: 5 residual = 2.058244e-04 iter: 6 residual = 7.067264e-05 iter: 7 residual = 2.255660e-05 iter: 8 residual = 7.091326e-06 iter: 9 residual = 2.364610e-06 iter: 10 residual = 8.624940e-07 Solution time: 5.976243 (sec.) total iterations: 10 Total solve time = 5.97687 sec ************************************************************************ -- Nonlinear Solver Step 2 -- ||F|| = 1.896e-01 step = 1.000e+00 dx = 9.496e-04 ************************************************************************ ------------------------------------------------------------------------------ Using `power-method' for eigen-computations *** *** ML_Epetra::MultiLevelPreconditioner *** Matrix has 141466332 rows and 7358049360 nonzeros, distributed over 256 process(es) The linear system matrix is an Epetra_CrsMatrix ** Leaving column map of Main linear system matrix unchanged Default values for `SA' Maximum number of levels = 7 Using increasing levels. Finest level = 0, coarsest level = 6 Number of applications of the ML cycle = 1 Number of PDE equations = 2 Aggregation threshold = 0 Max coarse size = 2000 R and P smoothing : P = (I-\omega A) P_t, R = P^T R and P smoothing : \omega = 1.333/lambda_max Null space type = user-supplied Null space dimension = 3 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 0) : Gen Restriction and Prolongator ----------------------------------------------------------------------- SemiCoarsening: Coarsening from 141466332 to 6736492 Warning:No communication information given with Pmat's getrow on level 0 (finest = 0).!!!! ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 0 to level 1 ML_Gen_MultilevelHierarchy (level 1): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 1) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 1) begins ML_Aggregate_CoarsenUncoupled : current level = 1 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 120623760 (Nrows=6736492) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 3101601 (3368246) Aggregation(UC) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 3101601 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 0 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 357272 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 12 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 1) : Max eigenvalue = 4.0405e+00 Prolongator/Restriction smoother (level 1) : damping factor #1 = 3.2991e-01 Prolongator/Restriction smoother (level 1) : ( = 1.3330e+00 / 4.0405e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 1 to level 2 ML_Gen_MultilevelHierarchy (level 2): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 2) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 2) begins ML_Aggregate_CoarsenUncoupled : current level = 2 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 40167540 (Nrows=1071816) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 287696 (357272) Aggregation(UC) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 287696 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 27598 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 32 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 27630 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 124 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 2) : Max eigenvalue = 2.9248e+00 Prolongator/Restriction smoother (level 2) : damping factor #1 = 4.5576e-01 Prolongator/Restriction smoother (level 2) : ( = 1.3330e+00 / 2.9248e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 2 to level 3 ML_Gen_MultilevelHierarchy (level 3): repartitioning suppressed until level 4 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 3) : Gen Restriction and Prolongator ----------------------------------------------------------------------- ML_Aggregate_Coarsen (level 3) begins ML_Aggregate_CoarsenUncoupled : current level = 3 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 4551084 (Nrows=82890) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 21257 (27630) Aggregation(UC) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 21257 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 1957 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 42 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 1999 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 43 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using power method Gen_Prolongator (level 3) : Max eigenvalue = 2.7354e+00 Prolongator/Restriction smoother (level 3) : damping factor #1 = 4.8732e-01 Prolongator/Restriction smoother (level 3) : ( = 1.3330e+00 / 2.7354e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 3 to level 4 Repartitioning (level 4): min rows per proc = 600 Repartitioning (level 4): largest max/min ratio = 1.327e+00 Repartitioning (level 4): max #rows (global) that fits on one proc = 5000 Repartitioning (level 4): #proc to use in repartitioning = 9 Repartitioning using Zoltan 3.83 ZOLTAN Load balancing method = 3 (RCB) Build configuration: ZOLTAN_ID_TYPE: unsigned int (4 bytes) ZOLTAN_GNO_TYPE: ssize_t, (8 bytes) MPI_Datatype for ZOLTAN_ID_TYPE: MPI_UNSIGNED MPI_Datatype for ZOLTAN_GNO_TYPE: MPI_LONG ZOLTAN Parameter IMBALANCE_TOL[0] = 1.100000 ZOLTAN Parameter AUTO_MIGRATE = FALSE ZOLTAN Parameter MIGRATE_ONLY_PROC_CHANGES = 1 ZOLTAN Parameter OBJ_WEIGHT_DIM = 1 ZOLTAN Parameter EDGE_WEIGHT_DIM = 0 ZOLTAN Parameter DEBUG_LEVEL = 1 ZOLTAN Parameter DEBUG_PROCESSOR = 0 ZOLTAN Parameter DETERMINISTIC = TRUE ZOLTAN Parameter TIMER = 1 (wall) ZOLTAN Parameter NUM_GID_ENTRIES = 1 ZOLTAN Parameter NUM_LID_ENTRIES = 0 ZOLTAN Parameter RETURN_LISTS = IMPORT AND EXPORT ZOLTAN Parameter NUM_GLOBAL_PARTS = 9 ZOLTAN Parameter NUM_LOCAL_PARTS = -1 ZOLTAN Parameter REMAP = 1 ZOLTAN Parameter SEED = 1027571338 (1027571338) ZOLTAN Parameter LB_APPROACH = repartition ZOLTAN Parameter RCB_OVERALLOC = 1.200000 ZOLTAN Parameter RCB_REUSE = 0 ZOLTAN Parameter CHECK_GEOM = 1 ZOLTAN Parameter RCB_OUTPUT_LEVEL = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter RCB_LOCK_DIRECTIONS = 0 ZOLTAN Parameter RCB_SET_DIRECTIONS = 0 ZOLTAN Parameter RCB_RECTILINEAR_BLOCKS = 0 ZOLTAN Parameter OBJ_WEIGHTS_COMPARABLE = 0 ZOLTAN Parameter RCB_MULTICRITERIA_NORM = 1 ZOLTAN Parameter RCB_MAX_ASPECT_RATIO = 10.000000 ZOLTAN Parameter AVERAGE_CUTS = 0 ZOLTAN Parameter RANDOM_PIVOTS = 0 ZOLTAN Parameter RCB_RECOMPUTE_BOX = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 0.000000 ZOLTAN Parameter FINAL_OUTPUT = 0 ZOLTAN Parameter KEEP_CUTS = 0 ZOLTAN Parameter REDUCE_DIMENSIONS = 0 ZOLTAN Parameter DEGENERATE_RATIO = 10.000000 Zoltan (level 4) : time required = 4.105886e-03 ----------------------------------------------------------------------- ML_Gen_MultiLevelHierarchy (level 4) : Gen Restriction and Prolongator ----------------------------------------------------------------------- It appears that repartitioning has been performed and so further semicoarsening is aborted. Any further line smoothing is going to numerically act as point smoothing . ML_Aggregate_Coarsen (level 4) begins ML_Aggregate_CoarsenUncoupled : current level = 4 ML_Aggregate_CoarsenUncoupled : current eps = 0.000000e+00 Aggregation(UVB) : Total nonzeros = 398457 (Nrows=5997) Aggregation(UVB) : Amalgamated matrix done Aggregation(UC) : Phase 0 - no. of bdry pts = 0 Aggregation(UC) : Phase 1 - nodes aggregated = 1320 (1999) Aggregation(UC) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 1 - nodes aggregated = 1320 Aggregation(UC_Phase2_3) : Phase 1 - total aggregates = 89 Aggregation(UC_Phase2_3) : Phase 2a- additional aggregates = 12 Aggregation(UC_Phase2_3) : Phase 2 - total aggregates = 101 Aggregation(UC_Phase2_3) : Phase 2 - boundary nodes = 7 Aggregation(UC_Phase2_3) : Phase 3 - leftovers = 0 and singletons = 0 Aggregation(UC) : QR factorization - too small aggregates = 0 Calculating eigenvalue estimate using CG method Gen_Prolongator (level 4) : Max eigenvalue = 1.9062e+00 Prolongator/Restriction smoother (level 4) : damping factor #1 = 6.9928e-01 Prolongator/Restriction smoother (level 4) : ( = 1.3330e+00 / 1.9062e+00) ML_Gen_MultilevelHierarchy: Projecting node coordinates from level 4 to level 5 Repartitioning (level 5): min rows per proc = 600 Repartitioning (level 5): largest max/min ratio = 1.327e+00 Repartitioning (level 5): max #rows (global) that fits on one proc = 5000 Repartitioning (level 5): #proc to use in repartitioning = 1 Smoothed Aggregation : operator complexity = 1.374156e+00. Time to build the hierarchy = 2.361874164 (s) Number of actual levels : 6 Smoother (level 0) : # global rows = 141466332, # estim. global nnz = 7358049360, # nnz per row = 52.01 Smoother (level 0) : line Gauss-Seidel(sweeps=1,omega=1,both) Smoother (level 0) : Setup time : 0.1285520531 (s) Smoother (level 1) : # global rows = 6736492, # estim. global nnz = 120623760, # nnz per row = 17.91 Smoother (level 1) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 1) : Setup time : 0.04746858496 (s) Smoother (level 2) : # global rows = 1071816, # estim. global nnz = 40167540, # nnz per row = 37.48 Smoother (level 2) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 2) : Setup time : 0.0001437636092 (s) Smoother (level 3) : # global rows = 82890, # estim. global nnz = 4551084, # nnz per row = 54.91 Smoother (level 3) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 3) : Setup time : 0.0001031402498 (s) Smoother (level 4) : # global rows = 5997, # estim. global nnz = 398457, # nnz per row = 66.44 Smoother (level 4) : ML Gauss-Seidel (sweeps=4,omega=1,both) Smoother (level 4) : Setup time : 0.000104974024 (s) Coarse solve (level 5) : ML Gauss-Seidel (sweeps=4,omega=1,pre) Cumulative timing for construction so far: - for initial setup = 0.07468972541 (s) - for hierarchy setup = 2.361910118 (s) - for smoothers setup = 0.176372516 (s) - for coarse setup = 2.479553223e-05 (s) - for final setup = 0.0002793818712 (s) Total for this setup = 2.613527822 (s) ------------------------------------------------------------------------------ ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 1.922495901 1.922495901 2- Preconditioner apply = 3.829536621 a- first application(s) only = 0.3549414957 0.3549414957 b- remaining applications = 3.474595126 0.3158722842 3- Total time required by ML so far is 5.752032522 seconds (constr + all applications) ------------------------------------------------------------------------------ Solving block system using AztecOO ... ******************************************************* ***** Problem: Epetra::CrsMatrix ***** Preconditioned GMRES solution ***** ML (L=6, LineGS_pre0/LineGS_post0, GS_pre5/~) ***** No scaling ******************************************************* iter: 0 residual = 1.000000e+00 iter: 1 residual = 4.905404e-02 iter: 2 residual = 6.470179e-03 iter: 3 residual = 1.100554e-03 iter: 4 residual = 2.272147e-04 iter: 5 residual = 6.594523e-05 iter: 6 residual = 2.178495e-05 iter: 7 residual = 7.723012e-06 iter: 8 residual = 3.061560e-06 iter: 9 residual = 1.236582e-06 iter: 10 residual = 4.730648e-07 Solution time: 5.713766 (sec.) total iterations: 10 Total solve time = 5.71433 sec ************************************************************************ -- Nonlinear Solver Step 3 -- ||F|| = 8.241e-04 step = 1.000e+00 dx = 1.659e-04 (Converged!) ************************************************************************ ************************************************************************ -- Final Status Test Results -- Converged....OR Combination -> Converged....AND Combination -> Converged....F-Norm = 6.929e-08 < 1.000e-05 (Length-Scaled Two-Norm, Absolute Tolerance) Converged....WRMS-Norm = 1.380e-06 < 1 (Min Step Size: 1.000e+00 >= 1) ??...........Number of Iterations = -1 < 40 ************************************************************************ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ End of Continuation Step 6 : Parameter: Glen's Law Homotopy Parameter = 1.000e+00 from 7.809e-01 --> Step Converged in 3 Nonlinear Solver Iterations! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Calling Predictor with method: Constant Continuation run stopping: parameter reached bound of 1.000e+00 Continuation Stepper Finished Convergence Stats: for step #1 : NumLinSolves, Krylov, Kr/Solve; LastKrylov, LastTol: 39 610 15.64102564 10 -1 Finished eval of first model: Params, Responses Parameter vector 0: MyPID GID Value 0 0 0 Main_Solve: MeanValue of final solution -2.49044163148 Number of Failed Comparisons: 0 ------------------------------------------------------------------------------ ML time information (seconds) total avg 1- Construction = 2.538846727 2.538846727 2- Preconditioner apply = 3.760958121 a- first application(s) only = 0.3175189951 0.3175189951 b- remaining applications = 3.443439126 0.3130399205 3- Total time required by ML so far is 6.299804848 seconds (constr + all applications) ------------------------------------------------------------------------------ *** Teuchos::StackedTimer::report() - Remainder for a level will be *** *** incorrect if a timer in the level does not exist on every rank *** *** of the MPI Communicator. *** Albany Stacked Timer: 1772.52 [1] {min=1772.43, max=1772.56, std dev=0.024081} | Albany: Total Time: 1772.52 - 99.9999% [1] {min=1772.43, max=1772.56, std dev=0.0244474} | | Albany: Setup Time: 47.2282 - 2.66447% [1] {min=47.2228, max=47.2378, std dev=0.00390634} | | Thyra::EpetraModelEvaluator::evalModel(...): 1724.24 - 97.2765% [1] {min=1724.24, max=1724.25, std dev=0.00216204} | | | Albany: **Total Fill Time**: 1274.62 - 73.9232% [93] {min=1271.98, max=1282.15, std dev=1.39023} | | | | > Albany Fill: Residual: 326.017 - 25.5777% [53] {min=324.229, max=329.01, std dev=0.983782} | | | | | Phalanx::SortAndOrderEvaluators: 1.54807e-05 - 4.74842e-06% [3] {min=1.2544e-05, max=2.0982e-05, std dev=2.46803e-06} | | | | | Phalanx: Evaluator 74: [] Gather Coordinate Vector: 3.20499 - 0.983074% [345904] {min=3.04858, max=3.31415, std dev=0.0444038} | | | | | Phalanx: Evaluator 44: [] DOFCellToSide (Coord Vec -> Coord Vec lateralside): 0.571175 - 0.175198% [345904] {min=0.546321, max=0.613248, std dev=0.012247} | | | | | Phalanx: Evaluator 41: [] ComputeBasisFunctionsSide: 0.874456 - 0.268224% [345904] {min=0.501878, max=4.63528, std dev=0.745362} | | | | | Phalanx: Evaluator 2: [] Load State Field: 1.79561 - 0.550772% [345904] {min=1.69094, max=1.88562, std dev=0.0371906} | | | | | Phalanx: Evaluator 32: [] DOFCellToSide (surface_height -> surface_height_lateralside): 0.496308 - 0.152233% [345904] {min=0.477416, max=0.519784, std dev=0.0072052} | | | | | Phalanx: Evaluator 31: [] DOFInterpolationSide: 0.502932 - 0.154265% [345904] {min=0.482081, max=0.547119, std dev=0.0101469} | | | | | Phalanx: Evaluator 52: [] Shared Parameter Glen's Law Homotopy Parameter: 0.524418 - 0.160856% [345904] {min=0.500921, max=0.551629, std dev=0.0080565} | | | | | Phalanx: Evaluator 48: [] ComputeBasisFunctions: 244.484 - 74.9911% [345904] {min=243.299, max=245.707, std dev=0.356333} | | | | | Phalanx: Evaluator 72: [] Gather Solution: 2.75116 - 0.84387% [345904] {min=2.68067, max=2.88156, std dev=0.0242126} | | | | | Phalanx: Evaluator 10: [] DOFVecGradInterpolationBase: 10.9772 - 3.36707% [345904] {min=10.8912, max=11.1843, std dev=0.0386164} | | | | | Phalanx: Evaluator 0: [] Load State Field: 1.55604 - 0.477287% [345904] {min=1.51487, max=1.62713, std dev=0.0186202} | | | | | Phalanx: Evaluator 14: [] NodesToCellInterpolation: 2.41828 - 0.741764% [345904] {min=2.38091, max=2.53205, std dev=0.0247526} | | | | | Phalanx: Evaluator 55: [] ViscosityFO: 7.67297 - 2.35355% [345904] {min=7.56623, max=7.89202, std dev=0.0340393} | | | | | Phalanx: Evaluator 13: [] DOFGradInterpolationBase: 4.45351 - 1.36603% [345904] {min=4.37143, max=4.55177, std dev=0.0344483} | | | | | Phalanx: Evaluator 56: [] StokesFOBodyForce: 2.43416 - 0.746635% [345904] {min=2.38905, max=2.5462, std dev=0.0200683} | | | | | Phalanx: Evaluator 9: [] DOFVecInterpolationBase: 2.95404 - 0.906099% [345904] {min=2.9206, max=2.99595, std dev=0.0142324} | | | | | Phalanx: Evaluator 51: [] StokesFOResid: 9.91843 - 3.0423% [345904] {min=9.83691, max=10.2515, std dev=0.0450607} | | | | | Phalanx: Evaluator 1: [] Load State Field: 1.55239 - 0.476168% [345904] {min=1.50593, max=1.60219, std dev=0.0195085} | | | | | Phalanx: Evaluator 30: [] DOFCellToSide (ice_thickness -> ice_thickness_lateralside): 0.568515 - 0.174382% [345904] {min=0.546173, max=0.599837, std dev=0.00836657} | | | | | Phalanx: Evaluator 29: [] DOFInterpolationSide: 0.523771 - 0.160657% [345904] {min=0.500563, max=0.556208, std dev=0.00921812} | | | | | Phalanx: Evaluator 57: [] StokesFOLateralResid: 0.552089 - 0.169343% [345904] {min=0.524748, max=0.650259, std dev=0.0176524} | | | | | Phalanx: Evaluator 40: [] DOFCellToSide (Coord Vec -> Coord Vec basalside): 0.595892 - 0.182779% [345904] {min=0.569955, max=0.623781, std dev=0.00900944} | | | | | Phalanx: Evaluator 37: [] ComputeBasisFunctionsSide: 1.30241 - 0.399492% [345904] {min=1.25627, max=1.36529, std dev=0.0212811} | | | | | Phalanx: Evaluator 7: [] Load Side Set Field basal_friction_basalside from Side Set State basal_friction: 2.79548 - 0.857462% [345904] {min=2.71, max=2.88248, std dev=0.0369219} | | | | | Phalanx: Evaluator 19: [] DOFInterpolationSide: 0.576372 - 0.176792% [345904] {min=0.550915, max=0.606853, std dev=0.00918402} | | | | | Phalanx: Evaluator 70: [] BasalFrictionCoefficient: 0.622574 - 0.190963% [345904] {min=0.600162, max=0.645184, std dev=0.00868511} | | | | | Phalanx: Evaluator 17: [] DOFCellToSide (Velocity -> Velocity_basalside): 0.548584 - 0.168268% [345904] {min=0.529337, max=0.573454, std dev=0.00842815} | | | | | Phalanx: Evaluator 15: [] DOFVecInterpolationSide: 0.607509 - 0.186342% [345904] {min=0.584341, max=0.647295, std dev=0.0106097} | | | | | Phalanx: Evaluator 58: [] StokesFOBasalResid: 0.67775 - 0.207888% [345904] {min=0.653864, max=0.706095, std dev=0.00844489} | | | | | Phalanx: Evaluator 73: [] Scatter Velocity Residual: 2.28873 - 0.702026% [345904] {min=2.24849, max=2.35137, std dev=0.0165745} | | | | | Phalanx: Evaluator 312: [] Neumann Aggregator: 0.525029 - 0.161043% [345904] {min=0.499555, max=0.549316, std dev=0.00772368} | | | | | Phalanx: Evaluator 300: [] Dirichlet Aggregator: 0.000239388 - 7.3428e-05% [53] {min=0.000164691, max=0.000453286, std dev=3.78491e-05} | | | | | Remainder: 14.6902 - 4.50595% | | | | > Albany Fill: Jacobian: 761.038 - 59.7072% [39] {min=761.012, max=761.063, std dev=0.014136} | | | | | Phalanx::SortAndOrderEvaluators: 4.1538e-06 - 5.45807e-07% [3] {min=3.03e-06, max=1.1517e-05, std dev=7.97178e-07} | | | | | Phalanx: Evaluator 149: [] Gather Coordinate Vector: 2.32852 - 0.305967% [254533] {min=2.23635, max=2.45921, std dev=0.0403702} | | | | | Phalanx: Evaluator 119: [] DOFCellToSide (Coord Vec -> Coord Vec lateralside): 0.447253 - 0.0587688% [254533] {min=0.428571, max=0.493127, std dev=0.0115981} | | | | | Phalanx: Evaluator 116: [] ComputeBasisFunctionsSide: 0.684314 - 0.0899185% [254533] {min=0.394947, max=3.6688, std dev=0.589543} | | | | | Phalanx: Evaluator 77: [] Load State Field: 1.36569 - 0.179451% [254533] {min=1.28398, max=1.45838, std dev=0.0355455} | | | | | Phalanx: Evaluator 107: [] DOFCellToSide (surface_height -> surface_height_lateralside): 0.371988 - 0.048879% [254533] {min=0.356752, max=0.399702, std dev=0.00574465} | | | | | Phalanx: Evaluator 106: [] DOFInterpolationSide: 0.382705 - 0.0502872% [254533] {min=0.362842, max=0.414464, std dev=0.00849951} | | | | | Phalanx: Evaluator 127: [] Shared Parameter Glen's Law Homotopy Parameter: 0.446173 - 0.0586268% [254533] {min=0.430025, max=0.470764, std dev=0.00627801} | | | | | Phalanx: Evaluator 123: [] ComputeBasisFunctions: 187.967 - 24.6987% [254533] {min=180.576, max=190.95, std dev=1.83947} | | | | | Phalanx: Evaluator 147: [] Gather Solution: 15.0484 - 1.97735% [254533] {min=14.2686, max=16.3924, std dev=0.358613} | | | | | Phalanx: Evaluator 85: [] DOFVecGradInterpolationBase: 83.4741 - 10.9685% [254533] {min=83.2446, max=83.9419, std dev=0.107627} | | | | | Phalanx: Evaluator 75: [] Load State Field: 1.20085 - 0.157791% [254533] {min=1.16628, max=1.23728, std dev=0.012864} | | | | | Phalanx: Evaluator 89: [] NodesToCellInterpolation: 1.79647 - 0.236055% [254533] {min=1.76862, max=1.87072, std dev=0.0181443} | | | | | Phalanx: Evaluator 130: [] ViscosityFO: 84.2241 - 11.067% [254533] {min=83.4009, max=85.269, std dev=0.426861} | | | | | Phalanx: Evaluator 88: [] DOFGradInterpolationBase: 3.53061 - 0.463921% [254533] {min=3.45856, max=3.59756, std dev=0.023688} | | | | | Phalanx: Evaluator 131: [] StokesFOBodyForce: 5.3637 - 0.704787% [254533] {min=5.31051, max=5.44678, std dev=0.0233843} | | | | | Phalanx: Evaluator 84: [] DOFVecInterpolationBase: 25.8419 - 3.39561% [254533] {min=25.2669, max=26.7246, std dev=0.184179} | | | | | Phalanx: Evaluator 126: [] StokesFOResid: 142.598 - 18.7373% [254533] {min=138.913, max=146.202, std dev=1.63238} | | | | | Phalanx: Evaluator 76: [] Load State Field: 1.19818 - 0.15744% [254533] {min=1.1655, max=1.23967, std dev=0.012578} | | | | | Phalanx: Evaluator 105: [] DOFCellToSide (ice_thickness -> ice_thickness_lateralside): 0.442849 - 0.0581901% [254533] {min=0.425478, max=0.468551, std dev=0.00625207} | | | | | Phalanx: Evaluator 104: [] DOFInterpolationSide: 0.395558 - 0.0519761% [254533] {min=0.378963, max=0.422302, std dev=0.00740828} | | | | | Phalanx: Evaluator 132: [] StokesFOLateralResid: 0.41915 - 0.055076% [254533] {min=0.39052, max=0.518881, std dev=0.0210682} | | | | | Phalanx: Evaluator 115: [] DOFCellToSide (Coord Vec -> Coord Vec basalside): 0.447958 - 0.0588614% [254533] {min=0.432905, max=0.482243, std dev=0.00829768} | | | | | Phalanx: Evaluator 112: [] ComputeBasisFunctionsSide: 0.987037 - 0.129696% [254533] {min=0.948858, max=1.02981, std dev=0.0157261} | | | | | Phalanx: Evaluator 82: [] Load Side Set Field basal_friction_basalside from Side Set State basal_friction: 2.13536 - 0.280585% [254533] {min=2.05613, max=2.22378, std dev=0.0304898} | | | | | Phalanx: Evaluator 94: [] DOFInterpolationSide: 0.427068 - 0.0561165% [254533] {min=0.410921, max=0.44473, std dev=0.00606829} | | | | | Phalanx: Evaluator 145: [] BasalFrictionCoefficient: 0.572537 - 0.0752311% [254533] {min=0.55172, max=0.599733, std dev=0.00688209} | | | | | Phalanx: Evaluator 92: [] DOFCellToSide (Velocity -> Velocity_basalside): 0.588145 - 0.0772819% [254533] {min=0.569972, max=0.611715, std dev=0.00829409} | | | | | Phalanx: Evaluator 90: [] DOFVecInterpolationSide: 0.836 - 0.10985% [254533] {min=0.816723, max=0.864362, std dev=0.0077824} | | | | | Phalanx: Evaluator 133: [] StokesFOBasalResid: 1.68986 - 0.222047% [254533] {min=1.67446, max=1.72508, std dev=0.00778146} | | | | | Phalanx: Evaluator 148: [] Scatter Velocity Residual: 43.3483 - 5.69594% [254533] {min=42.2678, max=43.9299, std dev=0.209977} | | | | | Phalanx: Evaluator 313: [] Neumann Aggregator: 0.382341 - 0.0502394% [254533] {min=0.367129, max=0.407659, std dev=0.00604762} | | | | | > Albany Fill: Jacobian Export: 118.228 - 15.5351% [39] {min=110.002, max=130.102, std dev=3.58217} | | | | | Phalanx: Evaluator 301: [] Dirichlet Aggregator: 0.000114545 - 1.50511e-05% [39] {min=9.5201e-05, max=0.000161987, std dev=1.36574e-05} | | | | | Remainder: 31.868 - 4.18744% | | | | Remainder: 187.562 - 14.7152% | | | NOX: Total Preconditioner Generation Time: 89.6269 - 5.19804% [39] {min=84.513, max=91.2337, std dev=0.723778} | | | NOX: Total Linear Solve Time: 349.215 - 20.2532% [39] {min=349.189, max=349.242, std dev=0.0124997} | | | | Stratimikos: AztecOOLOWS: 346.135 - 99.1178% [39] {min=346.113, max=346.146, std dev=0.00769602} | | | | | Stratimikos: AztecOOLOWS:SingleSolve: 346.13 - 99.9986% [39] {min=346.109, max=346.139, std dev=0.00751209} | | | | | | Epetra_CrsMatrix::Multiply(TransA,X,Y): 63.0099 - 18.2041% [688] {min=59.7655, max=70.6268, std dev=1.46865} | | | | | | Epetra_CrsMatrix::Multiply(TransA,x,y): 61.8899 - 17.8805% [649] {min=60.4985, max=65.3184, std dev=0.700647} | | | | | | Remainder: 221.23 - 63.9153% | | | | | Remainder: 0.00484778 - 0.00140055% | | | | Remainder: 3.08077 - 0.882197% | | | Phalanx::SortAndOrderEvaluators: 6.0344e-06 - 3.49974e-07% [4] {min=3.653e-06, max=1.2075e-05, std dev=2.09478e-06} | | | Albany: Output to File: 0.691871 - 0.040126% [7] {min=0.648788, max=0.749924, std dev=0.0161686} | | | Remainder: 10.0928 - 0.585349% | | Remainder: 1.04546 - 0.0589813% | Remainder: 0.00144623 - 8.15915e-05%