Inheritance diagram for Seldon::MatrixPardiso< T >:
Public Member Functions | |
| MatrixPardiso () | |
| default constructor | |
| ~MatrixPardiso () | |
| destructor | |
| bool | UseInteger8 () const |
| void | Clear () |
| LU factorization is cleared. | |
| void | SelectOrdering (int type) |
| sets ordering algorithm to use | |
| void | SetPermutation (const IVect &permut) |
| Sets ordering. | |
| void | HideMessages () |
| Disables output messages. | |
| void | ShowMessages () |
| Enables output messages. | |
| int | GetInfoFactorization () const |
| returns the error code obtained during numerical factorization | |
| size_t | GetMemorySize () const |
| returns the size of memory used by numerical factorization in bytes | |
| template<class T0 , class Prop , class Storage , class Allocator > | |
| void | FactorizeMatrix (Matrix< T0, Prop, Storage, Allocator > &mat, bool keep_matrix=false) |
| performs analysis and factorization of matrix mat | |
| void | FactorizeCSR (Vector< pardiso_int_t > &Ptr, Vector< pardiso_int_t > &IndCol, Vector< T > &Values, bool sym) |
| Performs analysis and factorisation of a matrix given in CSR form. | |
| void | FactorizeCSR (bool sym) |
| template<class Allocator2 > | |
| void | Solve (Vector< T, VectFull, Allocator2 > &x) |
| solves A x = b, x contains the source b on input, the solution x on output | |
| template<class Allocator2 > | |
| void | Solve (const SeldonTranspose &TransA, Vector< T, VectFull, Allocator2 > &x) |
| solves A x = b or A^T x = b x contains the source b on input, the solution x on output | |
| void | Solve (const SeldonTranspose &TransA, T *x_ptr, int nrhs_) |
| solves A x = b or A^T x = b x contains the source b on input, the solution x on output | |
| template<class Allocator2 , class Prop > | |
| void | Solve (const SeldonTranspose &TransA, Matrix< T, Prop, ColMajor, Allocator2 > &x) |
| solves A x = b or A^T x = b (x, b are matrices) x contains the source b on input, the solution x on output | |
| virtual void | SetPivotThreshold (double) |
| Sets the threshold for pivot. | |
| virtual void | RefineSolution () |
| Tells to the direct solver that refinement is required. | |
| virtual void | DoNotRefineSolution () |
| Tells to the direct solver that no refinement is required. | |
| virtual void | SetCoefficientEstimationNeededMemory (double coef) |
| Method overloaded in Mumps solver. | |
| virtual void | SetMaximumCoefficientEstimationNeededMemory (double coef) |
| Method overloaded in Mumps solver. | |
| virtual void | SetIncreaseCoefficientEstimationNeededMemory (double coef) |
| Method overloaded in Mumps solver. | |
| virtual void | SelectParallelOrdering (int) |
| selects ordering to use in parallel for the interfaced solver | |
| virtual void | SetPermutation (const Vector< int > &) |
| gives the ordering array to the interface solver | |
| virtual void | SetNumberOfThreadPerNode (int n) |
| Sets the number of threads per mpi process. | |
Protected Attributes | |
| void * | pt [64] |
| pointer to Pardiso object | |
| pardiso_int_t | mtype |
| matrix type | |
| pardiso_int_t | iparm [64] |
| integer parameters | |
| double | dparm [64] |
| pardiso_int_t | size_matrix |
| pardiso_int_t | maxfct |
| maximum number of factors with identical sparsity structure | |
| pardiso_int_t | mnum |
| factor number, 1 <= mnum <= maxfct | |
| pardiso_int_t | msglvl |
| verbosity level | |
| int | info_facto |
| Vector< pardiso_int_t > | ptrA |
| Vector< pardiso_int_t > | indA |
| matrix in CSR form | |
| Vector< T > | valA |
| values of non-zero entries | |
| Vector< pardiso_int_t > | perm |
| permutation array | |
| pardiso_int_t | type_ordering |
Detailed Description
template<class T>
class Seldon::MatrixPardiso< T >
Definition at line 58 of file Pardiso.hxx.
The documentation for this class was generated from the following files:
- computation/interfaces/direct/Pardiso.hxx
- computation/interfaces/direct/Pardiso.cxx
- lib/Compil/Seldon/Pardiso.cpp