10 #ifndef EIGEN_PASTIXSUPPORT_H 11 #define EIGEN_PASTIXSUPPORT_H 23 template<
typename _MatrixType,
bool IsStrSym = false>
class PastixLU;
24 template<
typename _MatrixType,
int Options>
class PastixLLT;
25 template<
typename _MatrixType,
int Options>
class PastixLDLT;
32 template<
typename _MatrixType>
36 typedef typename _MatrixType::Scalar
Scalar;
38 typedef typename _MatrixType::Index
Index;
41 template<
typename _MatrixType,
int Options>
45 typedef typename _MatrixType::Scalar
Scalar;
47 typedef typename _MatrixType::Index
Index;
50 template<
typename _MatrixType,
int Options>
54 typedef typename _MatrixType::Scalar
Scalar;
56 typedef typename _MatrixType::Index
Index;
59 void eigen_pastix(pastix_data_t **pastix_data,
int pastix_comm,
int n,
int *ptr,
int *idx,
float *vals,
int *perm,
int * invp,
float *x,
int nbrhs,
int *iparm,
double *dparm)
61 if (n == 0) { ptr = NULL; idx = NULL; vals = NULL; }
62 if (nbrhs == 0) {x = NULL; nbrhs=1;}
63 s_pastix(pastix_data, pastix_comm, n, ptr, idx, vals, perm, invp, x, nbrhs, iparm, dparm);
66 void eigen_pastix(pastix_data_t **pastix_data,
int pastix_comm,
int n,
int *ptr,
int *idx,
double *vals,
int *perm,
int * invp,
double *x,
int nbrhs,
int *iparm,
double *dparm)
68 if (n == 0) { ptr = NULL; idx = NULL; vals = NULL; }
69 if (nbrhs == 0) {x = NULL; nbrhs=1;}
70 d_pastix(pastix_data, pastix_comm, n, ptr, idx, vals, perm, invp, x, nbrhs, iparm, dparm);
73 void eigen_pastix(pastix_data_t **pastix_data,
int pastix_comm,
int n,
int *ptr,
int *idx, std::complex<float> *vals,
int *perm,
int * invp, std::complex<float> *x,
int nbrhs,
int *iparm,
double *dparm)
75 if (n == 0) { ptr = NULL; idx = NULL; vals = NULL; }
76 if (nbrhs == 0) {x = NULL; nbrhs=1;}
77 c_pastix(pastix_data, pastix_comm, n, ptr, idx, reinterpret_cast<COMPLEX*>(vals), perm, invp, reinterpret_cast<COMPLEX*>(x), nbrhs, iparm, dparm);
80 void eigen_pastix(pastix_data_t **pastix_data,
int pastix_comm,
int n,
int *ptr,
int *idx, std::complex<double> *vals,
int *perm,
int * invp, std::complex<double> *x,
int nbrhs,
int *iparm,
double *dparm)
82 if (n == 0) { ptr = NULL; idx = NULL; vals = NULL; }
83 if (nbrhs == 0) {x = NULL; nbrhs=1;}
84 z_pastix(pastix_data, pastix_comm, n, ptr, idx, reinterpret_cast<DCOMPLEX*>(vals), perm, invp, reinterpret_cast<DCOMPLEX*>(x), nbrhs, iparm, dparm);
88 template <
typename MatrixType>
91 if ( !(mat.outerIndexPtr()[0]) )
94 for(i = 0; i <= mat.rows(); ++i)
95 ++mat.outerIndexPtr()[i];
96 for(i = 0; i < mat.nonZeros(); ++i)
97 ++mat.innerIndexPtr()[i];
102 template <
typename MatrixType>
106 if ( mat.outerIndexPtr()[0] == 1 )
109 for(i = 0; i <= mat.rows(); ++i)
110 --mat.outerIndexPtr()[i];
111 for(i = 0; i < mat.nonZeros(); ++i)
112 --mat.innerIndexPtr()[i];
119 template <
class Derived>
125 typedef typename MatrixType::Scalar
Scalar;
127 typedef typename MatrixType::Index
Index;
133 PastixBase() : m_initisOk(false), m_analysisIsOk(false), m_factorizationIsOk(false), m_isInitialized(false), m_pastixdata(0), m_size(0)
147 template<
typename Rhs>
151 eigen_assert(m_isInitialized &&
"Pastix solver is not initialized.");
153 &&
"PastixBase::solve(): invalid number of rows of the right hand side matrix b");
157 template<
typename Rhs,
typename Dest>
162 return *
static_cast<Derived*
>(
this);
166 return *
static_cast<const Derived*
>(
this);
185 return m_iparm(idxparam);
203 return m_dparm(idxparam);
206 inline Index
cols()
const {
return m_size; }
207 inline Index
rows()
const {
return m_size; }
219 eigen_assert(m_isInitialized &&
"Decomposition is not initialized.");
227 template<
typename Rhs>
231 eigen_assert(m_isInitialized &&
"Pastix LU, LLT or LDLT is not initialized.");
233 &&
"PastixBase::solve(): invalid number of rows of the right hand side matrix b");
243 void analyzePattern(ColSpMatrix& mat);
246 void factorize(ColSpMatrix& mat);
251 eigen_assert(m_initisOk &&
"The Pastix structure should be allocated first");
252 m_iparm(IPARM_START_TASK) = API_TASK_CLEAN;
253 m_iparm(IPARM_END_TASK) = API_TASK_CLEAN;
255 m_perm.data(), m_invp.data(), 0, 0, m_iparm.data(), m_dparm.data());
258 void compute(ColSpMatrix& mat);
278 template <
class Derived>
282 m_iparm.setZero(IPARM_SIZE);
283 m_dparm.setZero(DPARM_SIZE);
285 m_iparm(IPARM_MODIFY_PARAMETER) = API_NO;
286 pastix(&m_pastixdata, MPI_COMM_WORLD,
288 0, 0, 0, 1, m_iparm.data(), m_dparm.data());
290 m_iparm[IPARM_MATRIX_VERIFICATION] = API_NO;
291 m_iparm[IPARM_VERBOSE] = 2;
292 m_iparm[IPARM_ORDERING] = API_ORDER_SCOTCH;
293 m_iparm[IPARM_INCOMPLETE] = API_NO;
294 m_iparm[IPARM_OOC_LIMIT] = 2000;
295 m_iparm[IPARM_RHS_MAKING] = API_RHS_B;
296 m_iparm(IPARM_MATRIX_VERIFICATION) = API_NO;
298 m_iparm(IPARM_START_TASK) = API_TASK_INIT;
299 m_iparm(IPARM_END_TASK) = API_TASK_INIT;
301 0, 0, 0, 0, m_iparm.data(), m_dparm.data());
304 if(m_iparm(IPARM_ERROR_NUMBER)) {
314 template <
class Derived>
322 m_iparm(IPARM_MATRIX_VERIFICATION) = API_NO;
323 m_isInitialized = m_factorizationIsOk;
327 template <
class Derived>
330 eigen_assert(m_initisOk &&
"The initialization of PaSTiX failed");
337 m_perm.resize(m_size);
338 m_invp.resize(m_size);
340 m_iparm(IPARM_START_TASK) = API_TASK_ORDERING;
341 m_iparm(IPARM_END_TASK) = API_TASK_ANALYSE;
343 mat.
valuePtr(), m_perm.data(), m_invp.data(), 0, 0, m_iparm.data(), m_dparm.data());
346 if(m_iparm(IPARM_ERROR_NUMBER))
349 m_analysisIsOk =
false;
354 m_analysisIsOk =
true;
358 template <
class Derived>
362 eigen_assert(m_analysisIsOk &&
"The analysis phase should be called before the factorization phase");
363 m_iparm(IPARM_START_TASK) = API_TASK_NUMFACT;
364 m_iparm(IPARM_END_TASK) = API_TASK_NUMFACT;
368 mat.
valuePtr(), m_perm.data(), m_invp.data(), 0, 0, m_iparm.data(), m_dparm.data());
371 if(m_iparm(IPARM_ERROR_NUMBER))
374 m_factorizationIsOk =
false;
375 m_isInitialized =
false;
380 m_factorizationIsOk =
true;
381 m_isInitialized =
true;
386 template<
typename Base>
387 template<
typename Rhs,
typename Dest>
390 eigen_assert(m_isInitialized &&
"The matrix should be factorized first");
392 THIS_METHOD_IS_ONLY_FOR_COLUMN_MAJOR_MATRICES);
397 for (
int i = 0; i < b.cols(); i++){
398 m_iparm[IPARM_START_TASK] = API_TASK_SOLVE;
399 m_iparm[IPARM_END_TASK] = API_TASK_REFINE;
402 m_perm.data(), m_invp.data(), &x(0, i),
rhs, m_iparm.data(), m_dparm.data());
408 return m_iparm(IPARM_ERROR_NUMBER)==0;
430 template<
typename _MatrixType,
bool IsStrSym>
437 typedef typename MatrixType::Index
Index;
457 m_structureIsUptodate =
false;
459 grabMatrix(matrix, temp);
469 m_structureIsUptodate =
false;
471 grabMatrix(matrix, temp);
472 Base::analyzePattern(temp);
483 grabMatrix(matrix, temp);
484 Base::factorize(temp);
490 m_structureIsUptodate =
false;
491 m_iparm(IPARM_SYM) = API_SYM_NO;
492 m_iparm(IPARM_FACTORIZATION) = API_FACT_LU;
501 if(!m_structureIsUptodate)
504 m_transposedStructure = matrix.
transpose();
507 for (Index j=0; j<m_transposedStructure.outerSize(); ++j)
508 for(
typename ColSpMatrix::InnerIterator it(m_transposedStructure, j); it; ++it)
511 m_structureIsUptodate =
true;
514 out = m_transposedStructure + matrix;
540 template<
typename _MatrixType,
int _UpLo>
549 enum { UpLo = _UpLo };
567 grabMatrix(matrix, temp);
578 grabMatrix(matrix, temp);
579 Base::analyzePattern(temp);
587 grabMatrix(matrix, temp);
588 Base::factorize(temp);
595 m_iparm(IPARM_SYM) = API_SYM_YES;
596 m_iparm(IPARM_FACTORIZATION) = API_FACT_LLT;
602 out.template selfadjointView<Lower>() = matrix.template selfadjointView<UpLo>();
621 template<
typename _MatrixType,
int _UpLo>
630 enum { UpLo = _UpLo };
648 grabMatrix(matrix, temp);
659 grabMatrix(matrix, temp);
660 Base::analyzePattern(temp);
668 grabMatrix(matrix, temp);
669 Base::factorize(temp);
677 m_iparm(IPARM_SYM) = API_SYM_YES;
678 m_iparm(IPARM_FACTORIZATION) = API_FACT_LDLT;
684 out.template selfadjointView<Lower>() = matrix.template selfadjointView<UpLo>();
691 template<
typename _MatrixType,
typename Rhs>
698 template<typename Dest>
void evalTo(Dest& dst)
const 700 dec()._solve(
rhs(),dst);
704 template<
typename _MatrixType,
typename Rhs>
711 template<typename Dest>
void evalTo(Dest& dst)
const 713 this->defaultEvalTo(dst);
SparseMatrix< Scalar, ColMajor > ColSpMatrix
void grabMatrix(const MatrixType &matrix, ColSpMatrix &out)
Matrix< Index, Dynamic, 1 > m_perm
_MatrixType::RealScalar RealScalar
void analyzePattern(ColSpMatrix &mat)
PastixLU(const MatrixType &matrix)
void compute(const MatrixType &matrix)
Base::ColSpMatrix ColSpMatrix
PastixLDLT(const MatrixType &matrix)
Matrix< int, IPARM_SIZE, 1 > m_iparm
PastixBase< _MatrixType > Dec
void factorize(const MatrixType &matrix)
iterative scaling algorithm to equilibrate rows and column norms in matrices
PastixLLT(const MatrixType &matrix)
void factorize(const MatrixType &matrix)
MatrixType::RealScalar RealScalar
#define EIGEN_STATIC_ASSERT(CONDITION, MSG)
void compute(const MatrixType &matrix)
pastix_data_t * m_pastixdata
const unsigned int RowMajorBit
void compute(const MatrixType &matrix)
const Index * outerIndexPtr() const
Array< RealScalar, IPARM_SIZE, 1 > & dparm()
const Derived & derived() const
double & dparm(int idxparam)
void fortran_to_c_numbering(MatrixType &mat)
bool _solve(const MatrixBase< Rhs > &b, MatrixBase< Dest > &x) const
void compute(ColSpMatrix &mat)
_MatrixType::Scalar Scalar
void analyzePattern(const MatrixType &matrix)
Base class of any sparse matrices or sparse expressions.
Matrix< Scalar, Dynamic, 1 > Vector
_MatrixType::RealScalar RealScalar
A sparse direct supernodal Cholesky (LLT) factorization and solver based on the PaStiX library...
PastixBase< _MatrixType > Dec
PastixBase< PastixLLT< MatrixType, _UpLo > > Base
bool m_structureIsUptodate
Provides a generic way to set and pass user-specified options.
_MatrixType::RealScalar RealScalar
PastixBase< PastixLU< MatrixType > > Base
int & iparm(int idxparam)
const internal::solve_retval< PastixBase, Rhs > solve(const MatrixBase< Rhs > &b) const
Transpose< Derived > transpose()
_MatrixType::Scalar Scalar
_MatrixType::Scalar Scalar
#define EIGEN_MAKE_SPARSE_SOLVE_HELPERS(DecompositionType, Rhs)
void c_to_fortran_numbering(MatrixType &mat)
void analyzePattern(const MatrixType &matrix)
void out(const real_t *x, real_t *f)
void rhs(const real_t *x, real_t *f)
void eigen_pastix(pastix_data_t **pastix_data, int pastix_comm, int n, int *ptr, int *idx, float *vals, int *perm, int *invp, float *x, int nbrhs, int *iparm, double *dparm)
void grabMatrix(const MatrixType &matrix, ColSpMatrix &out)
Base::ColSpMatrix ColSpMatrix
Array< Index, IPARM_SIZE, 1 > & iparm()
const Derived & derived() const
Base::ColSpMatrix ColSpMatrix
void factorize(ColSpMatrix &mat)
void factorize(const MatrixType &matrix)
const internal::sparse_solve_retval< PastixBase, Rhs > solve(const SparseMatrixBase< Rhs > &b) const
General-purpose arrays with easy API for coefficient-wise operations.
const Index * innerIndexPtr() const
Interface to the PaStix solver.
Matrix< Index, Dynamic, 1 > m_invp
#define EIGEN_MAKE_SOLVE_HELPERS(DecompositionType, Rhs)
A sparse direct supernodal Cholesky (LLT) factorization and solver based on the PaStiX library...
ComputationInfo info() const
Reports whether previous computation was successful.
ColSpMatrix m_transposedStructure
void grabMatrix(const MatrixType &matrix, ColSpMatrix &out)
internal::pastix_traits< Derived >::MatrixType _MatrixType
const Scalar * valuePtr() const
void analyzePattern(const MatrixType &matrix)
Matrix< double, DPARM_SIZE, 1 > m_dparm
Base class for all dense matrices, vectors, and expressions.
void init(int nV, int nC, SymmetricMatrix *H, real_t *g, Matrix *A, const real_t *const lb, const real_t *const ub, const real_t *const lbA, const real_t *const ubA, int nWSR, const real_t *const x0, Options *options, int nOutputs, mxArray *plhs[])
PastixBase< PastixLDLT< MatrixType, _UpLo > > Base
MatrixType::Scalar Scalar