1 // This file is part of Eigen, a lightweight C++ template library
2 // for linear algebra.
3 //
4 // Copyright (C) 2010 Benoit Jacob <jacob.benoit.1@gmail.com>
5 // Copyright (C) 2013-2014 Gael Guennebaud <gael.guennebaud@inria.fr>
6 //
7 // This Source Code Form is subject to the terms of the Mozilla
8 // Public License v. 2.0. If a copy of the MPL was not distributed
9 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
10 
11 #ifndef EIGEN_BIDIAGONALIZATION_H
12 #define EIGEN_BIDIAGONALIZATION_H
13 
14 namespace Eigen {
15 
16 namespace internal {
17 // UpperBidiagonalization will probably be replaced by a Bidiagonalization class, don't want to make it stable API.
18 // At the same time, it's useful to keep for now as it's about the only thing that is testing the BandMatrix class.
19 
20 template<typename _MatrixType> class UpperBidiagonalization
21 {
22   public:
23 
24     typedef _MatrixType MatrixType;
25     enum {
26       RowsAtCompileTime = MatrixType::RowsAtCompileTime,
27       ColsAtCompileTime = MatrixType::ColsAtCompileTime,
28       ColsAtCompileTimeMinusOne = internal::decrement_size<ColsAtCompileTime>::ret
29     };
30     typedef typename MatrixType::Scalar Scalar;
31     typedef typename MatrixType::RealScalar RealScalar;
32     typedef Eigen::Index Index; ///< \deprecated since Eigen 3.3
33     typedef Matrix<Scalar, 1, ColsAtCompileTime> RowVectorType;
34     typedef Matrix<Scalar, RowsAtCompileTime, 1> ColVectorType;
35     typedef BandMatrix<RealScalar, ColsAtCompileTime, ColsAtCompileTime, 1, 0, RowMajor> BidiagonalType;
36     typedef Matrix<Scalar, ColsAtCompileTime, 1> DiagVectorType;
37     typedef Matrix<Scalar, ColsAtCompileTimeMinusOne, 1> SuperDiagVectorType;
38     typedef HouseholderSequence<
39               const MatrixType,
40               const typename internal::remove_all<typename Diagonal<const MatrixType,0>::ConjugateReturnType>::type
41             > HouseholderUSequenceType;
42     typedef HouseholderSequence<
43               const typename internal::remove_all<typename MatrixType::ConjugateReturnType>::type,
44               Diagonal<const MatrixType,1>,
45               OnTheRight
46             > HouseholderVSequenceType;
47 
48     /**
49     * \brief Default Constructor.
50     *
51     * The default constructor is useful in cases in which the user intends to
52     * perform decompositions via Bidiagonalization::compute(const MatrixType&).
53     */
UpperBidiagonalization()54     UpperBidiagonalization() : m_householder(), m_bidiagonal(), m_isInitialized(false) {}
55 
UpperBidiagonalization(const MatrixType & matrix)56     explicit UpperBidiagonalization(const MatrixType& matrix)
57       : m_householder(matrix.rows(), matrix.cols()),
58         m_bidiagonal(matrix.cols(), matrix.cols()),
59         m_isInitialized(false)
60     {
61       compute(matrix);
62     }
63 
64     UpperBidiagonalization& compute(const MatrixType& matrix);
65     UpperBidiagonalization& computeUnblocked(const MatrixType& matrix);
66 
householder()67     const MatrixType& householder() const { return m_householder; }
bidiagonal()68     const BidiagonalType& bidiagonal() const { return m_bidiagonal; }
69 
householderU()70     const HouseholderUSequenceType householderU() const
71     {
72       eigen_assert(m_isInitialized && "UpperBidiagonalization is not initialized.");
73       return HouseholderUSequenceType(m_householder, m_householder.diagonal().conjugate());
74     }
75 
householderV()76     const HouseholderVSequenceType householderV() // const here gives nasty errors and i'm lazy
77     {
78       eigen_assert(m_isInitialized && "UpperBidiagonalization is not initialized.");
79       return HouseholderVSequenceType(m_householder.conjugate(), m_householder.const_derived().template diagonal<1>())
80              .setLength(m_householder.cols()-1)
81              .setShift(1);
82     }
83 
84   protected:
85     MatrixType m_householder;
86     BidiagonalType m_bidiagonal;
87     bool m_isInitialized;
88 };
89 
90 // Standard upper bidiagonalization without fancy optimizations
91 // This version should be faster for small matrix size
92 template<typename MatrixType>
93 void upperbidiagonalization_inplace_unblocked(MatrixType& mat,
94                                               typename MatrixType::RealScalar *diagonal,
95                                               typename MatrixType::RealScalar *upper_diagonal,
96                                               typename MatrixType::Scalar* tempData = 0)
97 {
98   typedef typename MatrixType::Scalar Scalar;
99 
100   Index rows = mat.rows();
101   Index cols = mat.cols();
102 
103   typedef Matrix<Scalar,Dynamic,1,ColMajor,MatrixType::MaxRowsAtCompileTime,1> TempType;
104   TempType tempVector;
105   if(tempData==0)
106   {
107     tempVector.resize(rows);
108     tempData = tempVector.data();
109   }
110 
111   for (Index k = 0; /* breaks at k==cols-1 below */ ; ++k)
112   {
113     Index remainingRows = rows - k;
114     Index remainingCols = cols - k - 1;
115 
116     // construct left householder transform in-place in A
117     mat.col(k).tail(remainingRows)
118        .makeHouseholderInPlace(mat.coeffRef(k,k), diagonal[k]);
119     // apply householder transform to remaining part of A on the left
120     mat.bottomRightCorner(remainingRows, remainingCols)
121        .applyHouseholderOnTheLeft(mat.col(k).tail(remainingRows-1), mat.coeff(k,k), tempData);
122 
123     if(k == cols-1) break;
124 
125     // construct right householder transform in-place in mat
126     mat.row(k).tail(remainingCols)
127        .makeHouseholderInPlace(mat.coeffRef(k,k+1), upper_diagonal[k]);
128     // apply householder transform to remaining part of mat on the left
129     mat.bottomRightCorner(remainingRows-1, remainingCols)
130        .applyHouseholderOnTheRight(mat.row(k).tail(remainingCols-1).transpose(), mat.coeff(k,k+1), tempData);
131   }
132 }
133 
134 /** \internal
135   * Helper routine for the block reduction to upper bidiagonal form.
136   *
137   * Let's partition the matrix A:
138   *
139   *      | A00 A01 |
140   *  A = |         |
141   *      | A10 A11 |
142   *
143   * This function reduces to bidiagonal form the left \c rows x \a blockSize vertical panel [A00/A10]
144   * and the \a blockSize x \c cols horizontal panel [A00 A01] of the matrix \a A. The bottom-right block A11
145   * is updated using matrix-matrix products:
146   *   A22 -= V * Y^T - X * U^T
147   * where V and U contains the left and right Householder vectors. U and V are stored in A10, and A01
148   * respectively, and the update matrices X and Y are computed during the reduction.
149   *
150   */
151 template<typename MatrixType>
upperbidiagonalization_blocked_helper(MatrixType & A,typename MatrixType::RealScalar * diagonal,typename MatrixType::RealScalar * upper_diagonal,Index bs,Ref<Matrix<typename MatrixType::Scalar,Dynamic,Dynamic,traits<MatrixType>::Flags & RowMajorBit>> X,Ref<Matrix<typename MatrixType::Scalar,Dynamic,Dynamic,traits<MatrixType>::Flags & RowMajorBit>> Y)152 void upperbidiagonalization_blocked_helper(MatrixType& A,
153                                            typename MatrixType::RealScalar *diagonal,
154                                            typename MatrixType::RealScalar *upper_diagonal,
155                                            Index bs,
156                                            Ref<Matrix<typename MatrixType::Scalar, Dynamic, Dynamic,
157                                                       traits<MatrixType>::Flags & RowMajorBit> > X,
158                                            Ref<Matrix<typename MatrixType::Scalar, Dynamic, Dynamic,
159                                                       traits<MatrixType>::Flags & RowMajorBit> > Y)
160 {
161   typedef typename MatrixType::Scalar Scalar;
162   typedef typename MatrixType::RealScalar RealScalar;
163   typedef typename NumTraits<RealScalar>::Literal Literal;
164   enum { StorageOrder = traits<MatrixType>::Flags & RowMajorBit };
165   typedef InnerStride<int(StorageOrder) == int(ColMajor) ? 1 : Dynamic> ColInnerStride;
166   typedef InnerStride<int(StorageOrder) == int(ColMajor) ? Dynamic : 1> RowInnerStride;
167   typedef Ref<Matrix<Scalar, Dynamic, 1>, 0, ColInnerStride>    SubColumnType;
168   typedef Ref<Matrix<Scalar, 1, Dynamic>, 0, RowInnerStride>    SubRowType;
169   typedef Ref<Matrix<Scalar, Dynamic, Dynamic, StorageOrder > > SubMatType;
170 
171   Index brows = A.rows();
172   Index bcols = A.cols();
173 
174   Scalar tau_u, tau_u_prev(0), tau_v;
175 
176   for(Index k = 0; k < bs; ++k)
177   {
178     Index remainingRows = brows - k;
179     Index remainingCols = bcols - k - 1;
180 
181     SubMatType X_k1( X.block(k,0, remainingRows,k) );
182     SubMatType V_k1( A.block(k,0, remainingRows,k) );
183 
184     // 1 - update the k-th column of A
185     SubColumnType v_k = A.col(k).tail(remainingRows);
186           v_k -= V_k1 * Y.row(k).head(k).adjoint();
187     if(k) v_k -= X_k1 * A.col(k).head(k);
188 
189     // 2 - construct left Householder transform in-place
190     v_k.makeHouseholderInPlace(tau_v, diagonal[k]);
191 
192     if(k+1<bcols)
193     {
194       SubMatType Y_k  ( Y.block(k+1,0, remainingCols, k+1) );
195       SubMatType U_k1 ( A.block(0,k+1, k,remainingCols) );
196 
197       // this eases the application of Householder transforAions
198       // A(k,k) will store tau_v later
199       A(k,k) = Scalar(1);
200 
201       // 3 - Compute y_k^T = tau_v * ( A^T*v_k - Y_k-1*V_k-1^T*v_k - U_k-1*X_k-1^T*v_k )
202       {
203         SubColumnType y_k( Y.col(k).tail(remainingCols) );
204 
205         // let's use the begining of column k of Y as a temporary vector
206         SubColumnType tmp( Y.col(k).head(k) );
207         y_k.noalias()  = A.block(k,k+1, remainingRows,remainingCols).adjoint() * v_k; // bottleneck
208         tmp.noalias()  = V_k1.adjoint()  * v_k;
209         y_k.noalias() -= Y_k.leftCols(k) * tmp;
210         tmp.noalias()  = X_k1.adjoint()  * v_k;
211         y_k.noalias() -= U_k1.adjoint()  * tmp;
212         y_k *= numext::conj(tau_v);
213       }
214 
215       // 4 - update k-th row of A (it will become u_k)
216       SubRowType u_k( A.row(k).tail(remainingCols) );
217       u_k = u_k.conjugate();
218       {
219         u_k -= Y_k * A.row(k).head(k+1).adjoint();
220         if(k) u_k -= U_k1.adjoint() * X.row(k).head(k).adjoint();
221       }
222 
223       // 5 - construct right Householder transform in-place
224       u_k.makeHouseholderInPlace(tau_u, upper_diagonal[k]);
225 
226       // this eases the application of Householder transformations
227       // A(k,k+1) will store tau_u later
228       A(k,k+1) = Scalar(1);
229 
230       // 6 - Compute x_k = tau_u * ( A*u_k - X_k-1*U_k-1^T*u_k - V_k*Y_k^T*u_k )
231       {
232         SubColumnType x_k ( X.col(k).tail(remainingRows-1) );
233 
234         // let's use the begining of column k of X as a temporary vectors
235         // note that tmp0 and tmp1 overlaps
236         SubColumnType tmp0 ( X.col(k).head(k) ),
237                       tmp1 ( X.col(k).head(k+1) );
238 
239         x_k.noalias()   = A.block(k+1,k+1, remainingRows-1,remainingCols) * u_k.transpose(); // bottleneck
240         tmp0.noalias()  = U_k1 * u_k.transpose();
241         x_k.noalias()  -= X_k1.bottomRows(remainingRows-1) * tmp0;
242         tmp1.noalias()  = Y_k.adjoint() * u_k.transpose();
243         x_k.noalias()  -= A.block(k+1,0, remainingRows-1,k+1) * tmp1;
244         x_k *= numext::conj(tau_u);
245         tau_u = numext::conj(tau_u);
246         u_k = u_k.conjugate();
247       }
248 
249       if(k>0) A.coeffRef(k-1,k) = tau_u_prev;
250       tau_u_prev = tau_u;
251     }
252     else
253       A.coeffRef(k-1,k) = tau_u_prev;
254 
255     A.coeffRef(k,k) = tau_v;
256   }
257 
258   if(bs<bcols)
259     A.coeffRef(bs-1,bs) = tau_u_prev;
260 
261   // update A22
262   if(bcols>bs && brows>bs)
263   {
264     SubMatType A11( A.bottomRightCorner(brows-bs,bcols-bs) );
265     SubMatType A10( A.block(bs,0, brows-bs,bs) );
266     SubMatType A01( A.block(0,bs, bs,bcols-bs) );
267     Scalar tmp = A01(bs-1,0);
268     A01(bs-1,0) = Literal(1);
269     A11.noalias() -= A10 * Y.topLeftCorner(bcols,bs).bottomRows(bcols-bs).adjoint();
270     A11.noalias() -= X.topLeftCorner(brows,bs).bottomRows(brows-bs) * A01;
271     A01(bs-1,0) = tmp;
272   }
273 }
274 
275 /** \internal
276   *
277   * Implementation of a block-bidiagonal reduction.
278   * It is based on the following paper:
279   *   The Design of a Parallel Dense Linear Algebra Software Library: Reduction to Hessenberg, Tridiagonal, and Bidiagonal Form.
280   *   by Jaeyoung Choi, Jack J. Dongarra, David W. Walker. (1995)
281   *   section 3.3
282   */
283 template<typename MatrixType, typename BidiagType>
284 void upperbidiagonalization_inplace_blocked(MatrixType& A, BidiagType& bidiagonal,
285                                             Index maxBlockSize=32,
286                                             typename MatrixType::Scalar* /*tempData*/ = 0)
287 {
288   typedef typename MatrixType::Scalar Scalar;
289   typedef Block<MatrixType,Dynamic,Dynamic> BlockType;
290 
291   Index rows = A.rows();
292   Index cols = A.cols();
293   Index size = (std::min)(rows, cols);
294 
295   // X and Y are work space
296   enum { StorageOrder = traits<MatrixType>::Flags & RowMajorBit };
297   Matrix<Scalar,
298          MatrixType::RowsAtCompileTime,
299          Dynamic,
300          StorageOrder,
301          MatrixType::MaxRowsAtCompileTime> X(rows,maxBlockSize);
302   Matrix<Scalar,
303          MatrixType::ColsAtCompileTime,
304          Dynamic,
305          StorageOrder,
306          MatrixType::MaxColsAtCompileTime> Y(cols,maxBlockSize);
307   Index blockSize = (std::min)(maxBlockSize,size);
308 
309   Index k = 0;
310   for(k = 0; k < size; k += blockSize)
311   {
312     Index bs = (std::min)(size-k,blockSize);  // actual size of the block
313     Index brows = rows - k;                   // rows of the block
314     Index bcols = cols - k;                   // columns of the block
315 
316     // partition the matrix A:
317     //
318     //      | A00 A01 A02 |
319     //      |             |
320     // A  = | A10 A11 A12 |
321     //      |             |
322     //      | A20 A21 A22 |
323     //
324     // where A11 is a bs x bs diagonal block,
325     // and let:
326     //      | A11 A12 |
327     //  B = |         |
328     //      | A21 A22 |
329 
330     BlockType B = A.block(k,k,brows,bcols);
331 
332     // This stage performs the bidiagonalization of A11, A21, A12, and updating of A22.
333     // Finally, the algorithm continue on the updated A22.
334     //
335     // However, if B is too small, or A22 empty, then let's use an unblocked strategy
336     if(k+bs==cols || bcols<48) // somewhat arbitrary threshold
337     {
338       upperbidiagonalization_inplace_unblocked(B,
339                                                &(bidiagonal.template diagonal<0>().coeffRef(k)),
340                                                &(bidiagonal.template diagonal<1>().coeffRef(k)),
341                                                X.data()
342                                               );
343       break; // We're done
344     }
345     else
346     {
347       upperbidiagonalization_blocked_helper<BlockType>( B,
348                                                         &(bidiagonal.template diagonal<0>().coeffRef(k)),
349                                                         &(bidiagonal.template diagonal<1>().coeffRef(k)),
350                                                         bs,
351                                                         X.topLeftCorner(brows,bs),
352                                                         Y.topLeftCorner(bcols,bs)
353                                                       );
354     }
355   }
356 }
357 
358 template<typename _MatrixType>
computeUnblocked(const _MatrixType & matrix)359 UpperBidiagonalization<_MatrixType>& UpperBidiagonalization<_MatrixType>::computeUnblocked(const _MatrixType& matrix)
360 {
361   Index rows = matrix.rows();
362   Index cols = matrix.cols();
363   EIGEN_ONLY_USED_FOR_DEBUG(cols);
364 
365   eigen_assert(rows >= cols && "UpperBidiagonalization is only for Arices satisfying rows>=cols.");
366 
367   m_householder = matrix;
368 
369   ColVectorType temp(rows);
370 
371   upperbidiagonalization_inplace_unblocked(m_householder,
372                                            &(m_bidiagonal.template diagonal<0>().coeffRef(0)),
373                                            &(m_bidiagonal.template diagonal<1>().coeffRef(0)),
374                                            temp.data());
375 
376   m_isInitialized = true;
377   return *this;
378 }
379 
380 template<typename _MatrixType>
compute(const _MatrixType & matrix)381 UpperBidiagonalization<_MatrixType>& UpperBidiagonalization<_MatrixType>::compute(const _MatrixType& matrix)
382 {
383   Index rows = matrix.rows();
384   Index cols = matrix.cols();
385   EIGEN_ONLY_USED_FOR_DEBUG(rows);
386   EIGEN_ONLY_USED_FOR_DEBUG(cols);
387 
388   eigen_assert(rows >= cols && "UpperBidiagonalization is only for Arices satisfying rows>=cols.");
389 
390   m_householder = matrix;
391   upperbidiagonalization_inplace_blocked(m_householder, m_bidiagonal);
392 
393   m_isInitialized = true;
394   return *this;
395 }
396 
397 #if 0
398 /** \return the Householder QR decomposition of \c *this.
399   *
400   * \sa class Bidiagonalization
401   */
402 template<typename Derived>
403 const UpperBidiagonalization<typename MatrixBase<Derived>::PlainObject>
404 MatrixBase<Derived>::bidiagonalization() const
405 {
406   return UpperBidiagonalization<PlainObject>(eval());
407 }
408 #endif
409 
410 } // end namespace internal
411 
412 } // end namespace Eigen
413 
414 #endif // EIGEN_BIDIAGONALIZATION_H
415