chrono::ChSolverMKL Class Reference
Description
Interface to the Intel MKL Pardiso parallel sparse direct solver.
Sparse linear direct solver. Cannot handle VI and complementarity problems, so it cannot be used with NSC formulations.
The solver is equipped with two main features:
- sparsity pattern lock
- sparsity pattern learning
See ChDirectSolverLS for more details.
If appropriate and warranted by the problem setup, it is highly recommended to enable the sparsity pattern lock. This can significantly improve performance for more complex problems (larger size and/or problems which include constraints).
Minimal usage example, to be put anywhere in the code, before starting the main simulation loop:
auto mkl_solver = chrono_types::make_shared<ChSolverMKL>();
system.SetSolver(mkl_solver);
See ChSystemDescriptor for more information about the problem formulation and the data structures passed to the solver.
#include <ChSolverMKL.h>
Inheritance diagram for chrono::ChSolverMKL:
Collaboration diagram for chrono::ChSolverMKL:
Public Member Functions | |
| ChSolverMKL (int num_threads=0) | |
| Construct an MKL Pardiso sparse direct solver object and specify the number of OpenMP threads. More... | |
| virtual Type | GetType () const override |
| Return type of the solver. | |
| Eigen::PardisoLU< ChSparseMatrix > & | GetMklEngine () |
| Get a handle to the underlying MKL engine. | |
 Public Member Functions inherited from chrono::ChDirectSolverLS | |
| void | LockSparsityPattern (bool val) |
| Enable/disable locking the sparsity pattern (default: false). More... | |
| void | UseSparsityPatternLearner (bool val) |
| Enable/disable use of the sparsity pattern learner (default: enabled). More... | |
| void | ForceSparsityPatternUpdate () |
| Force a call to the sparsity pattern learner to update sparsity pattern on the underlying matrix. More... | |
| void | SetSparsityEstimate (double sparsity) |
| Set estimate for matrix sparsity, a value in [0,1], with 0 indicating a fully dense matrix (default: 0.9). More... | |
| virtual void | SetMatrixSymmetryType (MatrixSymmetryType symmetry) |
| Set the matrix symmetry type (default: GENERAL). | |
| void | UsePermutationVector (bool val) |
| Enable/disable use of permutation vector (default: false). More... | |
| void | LeverageRhsSparsity (bool val) |
| Enable/disable leveraging sparsity in right-hand side vector (default: false). More... | |
| virtual void | EnableNullPivotDetection (bool val, double threshold=0) |
| Enable detection of null pivots. More... | |
| void | ResetTimers () |
| Reset timers for internal phases in Solve and Setup. | |
| double | GetTimeSolve_Assembly () const |
| Get cumulative time for assembly operations in Solve phase. | |
| double | GetTimeSolve_SolverCall () const |
| Get cumulative time for Pardiso calls in Solve phase. | |
| double | GetTimeSetup_Assembly () const |
| Get cumulative time for assembly operations in Setup phase. | |
| double | GetTimeSetup_SolverCall () const |
| Get cumulative time for Pardiso calls in Setup phase. | |
| int | GetNumSetupCalls () const |
| Return the number of calls to the solver's Setup function. | |
| int | GetNumSolveCalls () const |
| Return the number of calls to the solver's Setup function. | |
| ChSparseMatrix & | GetMatrix () |
| Get a handle to the underlying matrix. | |
| virtual bool | Setup (ChSystemDescriptor &sysd) override |
| Perform the solver setup operations. More... | |
| virtual double | Solve (ChSystemDescriptor &sysd) override |
| Solve linear system. More... | |
| virtual void | ArchiveOUT (ChArchiveOut &marchive) override |
| Method to allow serialization of transient data to archives. | |
| virtual void | ArchiveIN (ChArchiveIn &marchive) override |
| Method to allow de serialization of transient data from archives. | |
| Public Member Functions inherited from chrono::ChSolver | |
| void | SetVerbose (bool mv) |
| Set verbose output from solver. | |
Additional Inherited Members | |
| Public Types inherited from chrono::ChDirectSolverLS | |
| enum | MatrixSymmetryType { MatrixSymmetryType::GENERAL, MatrixSymmetryType::SYMMETRIC_POSDEF, MatrixSymmetryType::SYMMETRIC_INDEF, MatrixSymmetryType::STRUCTURAL_SYMMETRIC } |
| Public Types inherited from chrono::ChSolver | |
| enum | Type { Type::PSOR = 0, Type::PSSOR, Type::PJACOBI, Type::PMINRES, Type::BARZILAIBORWEIN, Type::APGD, Type::SPARSE_LU, Type::SPARSE_QR, Type::PARDISO, Type::MUMPS, Type::GMRES, Type::MINRES, Type::BICGSTAB, CUSTOM } |
| Available types of solvers. More... | |
| Protected Member Functions inherited from chrono::ChDirectSolverLS | |
| virtual bool | SolveRequiresMatrix () const override |
| Indicate whether or not the Solve() phase requires an up-to-date problem matrix. More... | |
| Protected Attributes inherited from chrono::ChDirectSolverLS | |
| ChSparseMatrix | m_mat |
| problem matrix | |
| int | m_dim |
| problem size | |
| MatrixSymmetryType | m_symmetry |
| symmetry of problem matrix | |
| double | m_sparsity |
| user-supplied estimate of matrix sparsity | |
| ChVectorDynamic< double > | m_rhs |
| right-hand side vector | |
| ChVectorDynamic< double > | m_sol |
| solution vector | |
| int | m_solve_call |
| counter for calls to Solve | |
| int | m_setup_call |
| counter for calls to Setup | |
| bool | m_lock |
| is the matrix sparsity pattern locked? | |
| bool | m_use_learner |
| use the sparsity pattern learner? | |
| bool | m_force_update |
| force a call to the sparsity pattern learner? | |
| bool | m_use_perm |
| use of the permutation vector? | |
| bool | m_use_rhs_sparsity |
| leverage right-hand side sparsity? | |
| bool | m_null_pivot_detection |
| enable detection of zero pivots? | |
| ChTimer | m_timer_setup_assembly |
| timer for matrix assembly | |
| ChTimer | m_timer_setup_solvercall |
| timer for factorization | |
| ChTimer | m_timer_solve_assembly |
| timer for RHS assembly | |
| ChTimer | m_timer_solve_solvercall |
| timer for solution | |
| Protected Attributes inherited from chrono::ChSolver | |
| bool | verbose |
Constructor & Destructor Documentation
◆ ChSolverMKL()
| chrono::ChSolverMKL::ChSolverMKL | ( | int | num_threads = 0 | ) |
Construct an MKL Pardiso sparse direct solver object and specify the number of OpenMP threads.
Passing the default value num_threads=0 results in using a number of threads equal to the number of available processors (as returned by the function omp_get_num_procs)
The documentation for this class was generated from the following files:
- /builds/uwsbel/chrono/src/chrono_mkl/ChSolverMKL.h
- /builds/uwsbel/chrono/src/chrono_mkl/ChSolverMKL.cpp
