11 #ifndef EIGEN_INCOMPLETE_LUT_H 12 #define EIGEN_INCOMPLETE_LUT_H 28 template <
typename VectorV,
typename VectorI>
31 typedef typename VectorV::RealScalar RealScalar;
41 if (ncut < first || ncut > last )
return 0;
45 RealScalar abskey =
abs(
row(mid));
46 for (
Index j = first + 1; j <= last; j++) {
47 if (
abs(
row(j)) > abskey) {
50 swap(ind(mid), ind(j));
55 swap(ind(mid), ind(first));
57 if (mid > ncut) last = mid - 1;
58 else if (mid < ncut ) first = mid + 1;
59 }
while (mid != ncut );
98 template <
typename _Scalar,
typename _StorageIndex =
int>
103 using Base::m_isInitialized;
120 : m_droptol(
NumTraits<Scalar>::dummy_precision()), m_fillfactor(10),
121 m_analysisIsOk(false), m_factorizationIsOk(false)
124 template<
typename MatrixType>
126 : m_droptol(droptol),m_fillfactor(fillfactor),
127 m_analysisIsOk(false),m_factorizationIsOk(false)
144 eigen_assert(m_isInitialized &&
"IncompleteLUT is not initialized.");
148 template<
typename MatrixType>
149 void analyzePattern(
const MatrixType& amat);
151 template<
typename MatrixType>
152 void factorize(
const MatrixType& amat);
159 template<
typename MatrixType>
162 analyzePattern(amat);
167 void setDroptol(
const RealScalar& droptol);
168 void setFillfactor(
int fillfactor);
170 template<
typename Rhs,
typename Dest>
174 x = m_lu.template triangularView<UnitLower>().solve(x);
175 x = m_lu.template triangularView<Upper>().solve(x);
205 template<
typename Scalar,
typename StorageIndex>
208 this->m_droptol = droptol;
215 template<
typename Scalar,
typename StorageIndex>
218 this->m_fillfactor = fillfactor;
221 template <
typename Scalar,
typename StorageIndex>
222 template<
typename _MatrixType>
228 #ifndef EIGEN_MPL2_ONLY 237 m_Pinv = m_P.inverse();
242 ordering(mat1,m_Pinv);
243 m_P = m_Pinv.inverse();
246 m_analysisIsOk =
true;
247 m_factorizationIsOk =
false;
248 m_isInitialized =
true;
251 template <
typename Scalar,
typename StorageIndex>
252 template<
typename _MatrixType>
260 eigen_assert((amat.rows() == amat.cols()) &&
"The factorization should be done on a square matrix");
261 Index n = amat.cols();
269 eigen_assert(m_analysisIsOk &&
"You must first call analyzePattern()");
279 Index fill_in = (amat.nonZeros()*m_fillfactor)/n + 1;
280 if (fill_in > n) fill_in = n;
283 Index nnzL = fill_in/2;
285 m_lu.reserve(n * (nnzL + nnzU + 1));
288 for (
Index ii = 0; ii < n; ii++)
294 ju(ii) = convert_index<StorageIndex>(ii);
296 jr(ii) = convert_index<StorageIndex>(ii);
297 RealScalar rownorm = 0;
302 Index k = j_it.index();
306 ju(sizel) = convert_index<StorageIndex>(k);
307 u(sizel) = j_it.value();
308 jr(k) = convert_index<StorageIndex>(sizel);
313 u(ii) = j_it.value();
318 Index jpos = ii + sizeu;
319 ju(jpos) = convert_index<StorageIndex>(k);
320 u(jpos) = j_it.value();
321 jr(k) = convert_index<StorageIndex>(jpos);
334 rownorm =
sqrt(rownorm);
344 Index minrow = ju.segment(jj,sizel-jj).minCoeff(&k);
346 if (minrow != ju(jj))
351 jr(minrow) = convert_index<StorageIndex>(jj);
352 jr(j) = convert_index<StorageIndex>(k);
360 while (ki_it && ki_it.index() < minrow) ++ki_it;
362 Scalar fact = u(jj) / ki_it.value();
365 if(
abs(fact) <= m_droptol)
373 for (; ki_it; ++ki_it)
375 Scalar prod = fact * ki_it.value();
376 Index j = ki_it.index();
393 ju(newpos) = convert_index<StorageIndex>(j);
395 jr(j) = convert_index<StorageIndex>(newpos);
402 ju(len) = convert_index<StorageIndex>(minrow);
409 for(
Index k = 0; k <sizeu; k++) jr(ju(ii+k)) = -1;
415 len = (std::min)(sizel, nnzL);
422 for(
Index k = 0; k < len; k++)
423 m_lu.insertBackByOuterInnerUnordered(ii,ju(k)) = u(k);
427 if (u(ii) == Scalar(0))
428 u(ii) =
sqrt(m_droptol) * rownorm;
429 m_lu.insertBackByOuterInnerUnordered(ii, ii) = u(ii);
434 for(
Index k = 1; k < sizeu; k++)
436 if(
abs(u(ii+k)) > m_droptol * rownorm )
439 u(ii + len) = u(ii + k);
440 ju(ii + len) = ju(ii + k);
444 len = (std::min)(sizeu, nnzU);
450 for(
Index k = ii + 1; k < ii + len; k++)
451 m_lu.insertBackByOuterInnerUnordered(ii,ju(k)) = u(k);
454 m_lu.makeCompressed();
456 m_factorizationIsOk =
true;
462 #endif // EIGEN_INCOMPLETE_LUT_H void setDroptol(const RealScalar &droptol)
IncompleteLUT(const MatrixType &mat, const RealScalar &droptol=NumTraits< Scalar >::dummy_precision(), int fillfactor=10)
VectorBlock< Derived > SegmentReturnType
void factorize(const MatrixType &amat)
SparseSolverBase< IncompleteLUT > Base
A base class for sparse solvers.
ComputationInfo info() const
Reports whether previous computation was successful.
EIGEN_DEVICE_FUNC const SqrtReturnType sqrt() const
Holds information about the various numeric (i.e. scalar) types allowed by Eigen. ...
EIGEN_DEVICE_FUNC IndexDest convert_index(const IndexSrc &idx)
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE const AbsReturnType abs() const
PermutationMatrix< Dynamic, Dynamic, StorageIndex > m_Pinv
Matrix< StorageIndex, Dynamic, 1 > VectorI
void _solve_impl(const Rhs &b, Dest &x) const
EIGEN_DEVICE_FUNC ColXpr col(Index i)
This is the const version of col().
PermutationMatrix< Dynamic, Dynamic, StorageIndex > m_P
Index QuickSplit(VectorV &row, VectorI &ind, Index ncut)
TransposeReturnType transpose()
EIGEN_DEFAULT_DENSE_INDEX_TYPE Index
The Index type as used for the API.
EIGEN_DEVICE_FUNC RowXpr row(Index i)
This is the const version of row(). */.
Base::InnerIterator InnerIterator
Incomplete LU factorization with dual-threshold strategy.
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE const Abs2ReturnType abs2() const
SparseSymmetricPermutationProduct< Derived, Upper|Lower > twistedBy(const PermutationMatrix< Dynamic, Dynamic, StorageIndex > &perm) const
Matrix< Scalar, Dynamic, 1 > Vector
_StorageIndex StorageIndex
SparseMatrix< Scalar, RowMajor, StorageIndex > FactorType
void analyzePattern(const MatrixType &amat)
NumTraits< Scalar >::Real RealScalar
IncompleteLUT & compute(const MatrixType &amat)
#define eigen_internal_assert(x)
EIGEN_DEVICE_FUNC const Scalar & b
void swap(mpfr::mpreal &x, mpfr::mpreal &y)
void swap(scoped_array< T > &a, scoped_array< T > &b)
void setFillfactor(int fillfactor)