1// This file is part of Eigen, a lightweight C++ template library
2// for linear algebra.
3//
4// Copyright (C) 2008-2010 Gael Guennebaud <gael.guennebaud@inria.fr>
5// Copyright (C) 2006-2008 Benoit Jacob <jacob.benoit.1@gmail.com>
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_PARTIAL_REDUX_H
12#define EIGEN_PARTIAL_REDUX_H
13
14namespace Eigen {
15
16/** \class PartialReduxExpr
17 * \ingroup Core_Module
18 *
19 * \brief Generic expression of a partially reduxed matrix
20 *
21 * \tparam MatrixType the type of the matrix we are applying the redux operation
22 * \tparam MemberOp type of the member functor
23 * \tparam Direction indicates the direction of the redux (#Vertical or #Horizontal)
24 *
25 * This class represents an expression of a partial redux operator of a matrix.
26 * It is the return type of some VectorwiseOp functions,
27 * and most of the time this is the only way it is used.
28 *
29 * \sa class VectorwiseOp
30 */
31
32template< typename MatrixType, typename MemberOp, int Direction>
33class PartialReduxExpr;
34
35namespace internal {
36template<typename MatrixType, typename MemberOp, int Direction>
37struct traits<PartialReduxExpr<MatrixType, MemberOp, Direction> >
38 : traits<MatrixType>
39{
40 typedef typename MemberOp::result_type Scalar;
41 typedef typename traits<MatrixType>::StorageKind StorageKind;
42 typedef typename traits<MatrixType>::XprKind XprKind;
43 typedef typename MatrixType::Scalar InputScalar;
44 enum {
45 RowsAtCompileTime = Direction==Vertical ? 1 : MatrixType::RowsAtCompileTime,
46 ColsAtCompileTime = Direction==Horizontal ? 1 : MatrixType::ColsAtCompileTime,
47 MaxRowsAtCompileTime = Direction==Vertical ? 1 : MatrixType::MaxRowsAtCompileTime,
48 MaxColsAtCompileTime = Direction==Horizontal ? 1 : MatrixType::MaxColsAtCompileTime,
49 Flags = RowsAtCompileTime == 1 ? RowMajorBit : 0,
50 TraversalSize = Direction==Vertical ? MatrixType::RowsAtCompileTime : MatrixType::ColsAtCompileTime
51 };
52};
53}
54
55template< typename MatrixType, typename MemberOp, int Direction>
56class PartialReduxExpr : public internal::dense_xpr_base< PartialReduxExpr<MatrixType, MemberOp, Direction> >::type,
57 internal::no_assignment_operator
58{
59 public:
60
61 typedef typename internal::dense_xpr_base<PartialReduxExpr>::type Base;
62 EIGEN_DENSE_PUBLIC_INTERFACE(PartialReduxExpr)
63
64 EIGEN_DEVICE_FUNC
65 explicit PartialReduxExpr(const MatrixType& mat, const MemberOp& func = MemberOp())
66 : m_matrix(mat), m_functor(func) {}
67
68 EIGEN_DEVICE_FUNC
69 Index rows() const { return (Direction==Vertical ? 1 : m_matrix.rows()); }
70 EIGEN_DEVICE_FUNC
71 Index cols() const { return (Direction==Horizontal ? 1 : m_matrix.cols()); }
72
73 EIGEN_DEVICE_FUNC
74 typename MatrixType::Nested nestedExpression() const { return m_matrix; }
75
76 EIGEN_DEVICE_FUNC
77 const MemberOp& functor() const { return m_functor; }
78
79 protected:
80 typename MatrixType::Nested m_matrix;
81 const MemberOp m_functor;
82};
83
84#define EIGEN_MEMBER_FUNCTOR(MEMBER,COST) \
85 template <typename ResultType> \
86 struct member_##MEMBER { \
87 EIGEN_EMPTY_STRUCT_CTOR(member_##MEMBER) \
88 typedef ResultType result_type; \
89 template<typename Scalar, int Size> struct Cost \
90 { enum { value = COST }; }; \
91 template<typename XprType> \
92 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE \
93 ResultType operator()(const XprType& mat) const \
94 { return mat.MEMBER(); } \
95 }
96
97namespace internal {
98
99EIGEN_MEMBER_FUNCTOR(squaredNorm, Size * NumTraits<Scalar>::MulCost + (Size-1)*NumTraits<Scalar>::AddCost);
100EIGEN_MEMBER_FUNCTOR(norm, (Size+5) * NumTraits<Scalar>::MulCost + (Size-1)*NumTraits<Scalar>::AddCost);
101EIGEN_MEMBER_FUNCTOR(stableNorm, (Size+5) * NumTraits<Scalar>::MulCost + (Size-1)*NumTraits<Scalar>::AddCost);
102EIGEN_MEMBER_FUNCTOR(blueNorm, (Size+5) * NumTraits<Scalar>::MulCost + (Size-1)*NumTraits<Scalar>::AddCost);
103EIGEN_MEMBER_FUNCTOR(hypotNorm, (Size-1) * functor_traits<scalar_hypot_op<Scalar> >::Cost );
104EIGEN_MEMBER_FUNCTOR(sum, (Size-1)*NumTraits<Scalar>::AddCost);
105EIGEN_MEMBER_FUNCTOR(mean, (Size-1)*NumTraits<Scalar>::AddCost + NumTraits<Scalar>::MulCost);
106EIGEN_MEMBER_FUNCTOR(minCoeff, (Size-1)*NumTraits<Scalar>::AddCost);
107EIGEN_MEMBER_FUNCTOR(maxCoeff, (Size-1)*NumTraits<Scalar>::AddCost);
108EIGEN_MEMBER_FUNCTOR(all, (Size-1)*NumTraits<Scalar>::AddCost);
109EIGEN_MEMBER_FUNCTOR(any, (Size-1)*NumTraits<Scalar>::AddCost);
110EIGEN_MEMBER_FUNCTOR(count, (Size-1)*NumTraits<Scalar>::AddCost);
111EIGEN_MEMBER_FUNCTOR(prod, (Size-1)*NumTraits<Scalar>::MulCost);
112
113template <int p, typename ResultType>
114struct member_lpnorm {
115 typedef ResultType result_type;
116 template<typename Scalar, int Size> struct Cost
117 { enum { value = (Size+5) * NumTraits<Scalar>::MulCost + (Size-1)*NumTraits<Scalar>::AddCost }; };
118 EIGEN_DEVICE_FUNC member_lpnorm() {}
119 template<typename XprType>
120 EIGEN_DEVICE_FUNC inline ResultType operator()(const XprType& mat) const
121 { return mat.template lpNorm<p>(); }
122};
123
124template <typename BinaryOp, typename Scalar>
125struct member_redux {
126 typedef typename result_of<
127 BinaryOp(const Scalar&,const Scalar&)
128 >::type result_type;
129 template<typename _Scalar, int Size> struct Cost
130 { enum { value = (Size-1) * functor_traits<BinaryOp>::Cost }; };
131 EIGEN_DEVICE_FUNC explicit member_redux(const BinaryOp func) : m_functor(func) {}
132 template<typename Derived>
133 EIGEN_DEVICE_FUNC inline result_type operator()(const DenseBase<Derived>& mat) const
134 { return mat.redux(m_functor); }
135 const BinaryOp m_functor;
136};
137}
138
139/** \class VectorwiseOp
140 * \ingroup Core_Module
141 *
142 * \brief Pseudo expression providing partial reduction operations
143 *
144 * \tparam ExpressionType the type of the object on which to do partial reductions
145 * \tparam Direction indicates the direction of the redux (#Vertical or #Horizontal)
146 *
147 * This class represents a pseudo expression with partial reduction features.
148 * It is the return type of DenseBase::colwise() and DenseBase::rowwise()
149 * and most of the time this is the only way it is used.
150 *
151 * Example: \include MatrixBase_colwise.cpp
152 * Output: \verbinclude MatrixBase_colwise.out
153 *
154 * \sa DenseBase::colwise(), DenseBase::rowwise(), class PartialReduxExpr
155 */
156template<typename ExpressionType, int Direction> class VectorwiseOp
157{
158 public:
159
160 typedef typename ExpressionType::Scalar Scalar;
161 typedef typename ExpressionType::RealScalar RealScalar;
162 typedef Eigen::Index Index; ///< \deprecated since Eigen 3.3
163 typedef typename internal::ref_selector<ExpressionType>::non_const_type ExpressionTypeNested;
164 typedef typename internal::remove_all<ExpressionTypeNested>::type ExpressionTypeNestedCleaned;
165
166 template<template<typename _Scalar> class Functor,
167 typename Scalar_=Scalar> struct ReturnType
168 {
169 typedef PartialReduxExpr<ExpressionType,
170 Functor<Scalar_>,
171 Direction
172 > Type;
173 };
174
175 template<typename BinaryOp> struct ReduxReturnType
176 {
177 typedef PartialReduxExpr<ExpressionType,
178 internal::member_redux<BinaryOp,Scalar>,
179 Direction
180 > Type;
181 };
182
183 enum {
184 isVertical = (Direction==Vertical) ? 1 : 0,
185 isHorizontal = (Direction==Horizontal) ? 1 : 0
186 };
187
188 protected:
189
190 typedef typename internal::conditional<isVertical,
191 typename ExpressionType::ColXpr,
192 typename ExpressionType::RowXpr>::type SubVector;
193 /** \internal
194 * \returns the i-th subvector according to the \c Direction */
195 EIGEN_DEVICE_FUNC
196 SubVector subVector(Index i)
197 {
198 return SubVector(m_matrix.derived(),i);
199 }
200
201 /** \internal
202 * \returns the number of subvectors in the direction \c Direction */
203 EIGEN_DEVICE_FUNC
204 Index subVectors() const
205 { return isVertical?m_matrix.cols():m_matrix.rows(); }
206
207 template<typename OtherDerived> struct ExtendedType {
208 typedef Replicate<OtherDerived,
209 isVertical ? 1 : ExpressionType::RowsAtCompileTime,
210 isHorizontal ? 1 : ExpressionType::ColsAtCompileTime> Type;
211 };
212
213 /** \internal
214 * Replicates a vector to match the size of \c *this */
215 template<typename OtherDerived>
216 EIGEN_DEVICE_FUNC
217 typename ExtendedType<OtherDerived>::Type
218 extendedTo(const DenseBase<OtherDerived>& other) const
219 {
220 EIGEN_STATIC_ASSERT(EIGEN_IMPLIES(isVertical, OtherDerived::MaxColsAtCompileTime==1),
221 YOU_PASSED_A_ROW_VECTOR_BUT_A_COLUMN_VECTOR_WAS_EXPECTED)
222 EIGEN_STATIC_ASSERT(EIGEN_IMPLIES(isHorizontal, OtherDerived::MaxRowsAtCompileTime==1),
223 YOU_PASSED_A_COLUMN_VECTOR_BUT_A_ROW_VECTOR_WAS_EXPECTED)
224 return typename ExtendedType<OtherDerived>::Type
225 (other.derived(),
226 isVertical ? 1 : m_matrix.rows(),
227 isHorizontal ? 1 : m_matrix.cols());
228 }
229
230 template<typename OtherDerived> struct OppositeExtendedType {
231 typedef Replicate<OtherDerived,
232 isHorizontal ? 1 : ExpressionType::RowsAtCompileTime,
233 isVertical ? 1 : ExpressionType::ColsAtCompileTime> Type;
234 };
235
236 /** \internal
237 * Replicates a vector in the opposite direction to match the size of \c *this */
238 template<typename OtherDerived>
239 EIGEN_DEVICE_FUNC
240 typename OppositeExtendedType<OtherDerived>::Type
241 extendedToOpposite(const DenseBase<OtherDerived>& other) const
242 {
243 EIGEN_STATIC_ASSERT(EIGEN_IMPLIES(isHorizontal, OtherDerived::MaxColsAtCompileTime==1),
244 YOU_PASSED_A_ROW_VECTOR_BUT_A_COLUMN_VECTOR_WAS_EXPECTED)
245 EIGEN_STATIC_ASSERT(EIGEN_IMPLIES(isVertical, OtherDerived::MaxRowsAtCompileTime==1),
246 YOU_PASSED_A_COLUMN_VECTOR_BUT_A_ROW_VECTOR_WAS_EXPECTED)
247 return typename OppositeExtendedType<OtherDerived>::Type
248 (other.derived(),
249 isHorizontal ? 1 : m_matrix.rows(),
250 isVertical ? 1 : m_matrix.cols());
251 }
252
253 public:
254 EIGEN_DEVICE_FUNC
255 explicit inline VectorwiseOp(ExpressionType& matrix) : m_matrix(matrix) {}
256
257 /** \internal */
258 EIGEN_DEVICE_FUNC
259 inline const ExpressionType& _expression() const { return m_matrix; }
260
261 /** \returns a row or column vector expression of \c *this reduxed by \a func
262 *
263 * The template parameter \a BinaryOp is the type of the functor
264 * of the custom redux operator. Note that func must be an associative operator.
265 *
266 * \sa class VectorwiseOp, DenseBase::colwise(), DenseBase::rowwise()
267 */
268 template<typename BinaryOp>
269 EIGEN_DEVICE_FUNC
270 const typename ReduxReturnType<BinaryOp>::Type
271 redux(const BinaryOp& func = BinaryOp()) const
272 { return typename ReduxReturnType<BinaryOp>::Type(_expression(), internal::member_redux<BinaryOp,Scalar>(func)); }
273
274 typedef typename ReturnType<internal::member_minCoeff>::Type MinCoeffReturnType;
275 typedef typename ReturnType<internal::member_maxCoeff>::Type MaxCoeffReturnType;
276 typedef typename ReturnType<internal::member_squaredNorm,RealScalar>::Type SquaredNormReturnType;
277 typedef typename ReturnType<internal::member_norm,RealScalar>::Type NormReturnType;
278 typedef typename ReturnType<internal::member_blueNorm,RealScalar>::Type BlueNormReturnType;
279 typedef typename ReturnType<internal::member_stableNorm,RealScalar>::Type StableNormReturnType;
280 typedef typename ReturnType<internal::member_hypotNorm,RealScalar>::Type HypotNormReturnType;
281 typedef typename ReturnType<internal::member_sum>::Type SumReturnType;
282 typedef typename ReturnType<internal::member_mean>::Type MeanReturnType;
283 typedef typename ReturnType<internal::member_all>::Type AllReturnType;
284 typedef typename ReturnType<internal::member_any>::Type AnyReturnType;
285 typedef PartialReduxExpr<ExpressionType, internal::member_count<Index>, Direction> CountReturnType;
286 typedef typename ReturnType<internal::member_prod>::Type ProdReturnType;
287 typedef Reverse<const ExpressionType, Direction> ConstReverseReturnType;
288 typedef Reverse<ExpressionType, Direction> ReverseReturnType;
289
290 template<int p> struct LpNormReturnType {
291 typedef PartialReduxExpr<ExpressionType, internal::member_lpnorm<p,RealScalar>,Direction> Type;
292 };
293
294 /** \returns a row (or column) vector expression of the smallest coefficient
295 * of each column (or row) of the referenced expression.
296 *
297 * \warning the result is undefined if \c *this contains NaN.
298 *
299 * Example: \include PartialRedux_minCoeff.cpp
300 * Output: \verbinclude PartialRedux_minCoeff.out
301 *
302 * \sa DenseBase::minCoeff() */
303 EIGEN_DEVICE_FUNC
304 const MinCoeffReturnType minCoeff() const
305 { return MinCoeffReturnType(_expression()); }
306
307 /** \returns a row (or column) vector expression of the largest coefficient
308 * of each column (or row) of the referenced expression.
309 *
310 * \warning the result is undefined if \c *this contains NaN.
311 *
312 * Example: \include PartialRedux_maxCoeff.cpp
313 * Output: \verbinclude PartialRedux_maxCoeff.out
314 *
315 * \sa DenseBase::maxCoeff() */
316 EIGEN_DEVICE_FUNC
317 const MaxCoeffReturnType maxCoeff() const
318 { return MaxCoeffReturnType(_expression()); }
319
320 /** \returns a row (or column) vector expression of the squared norm
321 * of each column (or row) of the referenced expression.
322 * This is a vector with real entries, even if the original matrix has complex entries.
323 *
324 * Example: \include PartialRedux_squaredNorm.cpp
325 * Output: \verbinclude PartialRedux_squaredNorm.out
326 *
327 * \sa DenseBase::squaredNorm() */
328 EIGEN_DEVICE_FUNC
329 const SquaredNormReturnType squaredNorm() const
330 { return SquaredNormReturnType(_expression()); }
331
332 /** \returns a row (or column) vector expression of the norm
333 * of each column (or row) of the referenced expression.
334 * This is a vector with real entries, even if the original matrix has complex entries.
335 *
336 * Example: \include PartialRedux_norm.cpp
337 * Output: \verbinclude PartialRedux_norm.out
338 *
339 * \sa DenseBase::norm() */
340 EIGEN_DEVICE_FUNC
341 const NormReturnType norm() const
342 { return NormReturnType(_expression()); }
343
344 /** \returns a row (or column) vector expression of the norm
345 * of each column (or row) of the referenced expression.
346 * This is a vector with real entries, even if the original matrix has complex entries.
347 *
348 * Example: \include PartialRedux_norm.cpp
349 * Output: \verbinclude PartialRedux_norm.out
350 *
351 * \sa DenseBase::norm() */
352 template<int p>
353 EIGEN_DEVICE_FUNC
354 const typename LpNormReturnType<p>::Type lpNorm() const
355 { return typename LpNormReturnType<p>::Type(_expression()); }
356
357
358 /** \returns a row (or column) vector expression of the norm
359 * of each column (or row) of the referenced expression, using
360 * Blue's algorithm.
361 * This is a vector with real entries, even if the original matrix has complex entries.
362 *
363 * \sa DenseBase::blueNorm() */
364 EIGEN_DEVICE_FUNC
365 const BlueNormReturnType blueNorm() const
366 { return BlueNormReturnType(_expression()); }
367
368
369 /** \returns a row (or column) vector expression of the norm
370 * of each column (or row) of the referenced expression, avoiding
371 * underflow and overflow.
372 * This is a vector with real entries, even if the original matrix has complex entries.
373 *
374 * \sa DenseBase::stableNorm() */
375 EIGEN_DEVICE_FUNC
376 const StableNormReturnType stableNorm() const
377 { return StableNormReturnType(_expression()); }
378
379
380 /** \returns a row (or column) vector expression of the norm
381 * of each column (or row) of the referenced expression, avoiding
382 * underflow and overflow using a concatenation of hypot() calls.
383 * This is a vector with real entries, even if the original matrix has complex entries.
384 *
385 * \sa DenseBase::hypotNorm() */
386 EIGEN_DEVICE_FUNC
387 const HypotNormReturnType hypotNorm() const
388 { return HypotNormReturnType(_expression()); }
389
390 /** \returns a row (or column) vector expression of the sum
391 * of each column (or row) of the referenced expression.
392 *
393 * Example: \include PartialRedux_sum.cpp
394 * Output: \verbinclude PartialRedux_sum.out
395 *
396 * \sa DenseBase::sum() */
397 EIGEN_DEVICE_FUNC
398 const SumReturnType sum() const
399 { return SumReturnType(_expression()); }
400
401 /** \returns a row (or column) vector expression of the mean
402 * of each column (or row) of the referenced expression.
403 *
404 * \sa DenseBase::mean() */
405 EIGEN_DEVICE_FUNC
406 const MeanReturnType mean() const
407 { return MeanReturnType(_expression()); }
408
409 /** \returns a row (or column) vector expression representing
410 * whether \b all coefficients of each respective column (or row) are \c true.
411 * This expression can be assigned to a vector with entries of type \c bool.
412 *
413 * \sa DenseBase::all() */
414 EIGEN_DEVICE_FUNC
415 const AllReturnType all() const
416 { return AllReturnType(_expression()); }
417
418 /** \returns a row (or column) vector expression representing
419 * whether \b at \b least one coefficient of each respective column (or row) is \c true.
420 * This expression can be assigned to a vector with entries of type \c bool.
421 *
422 * \sa DenseBase::any() */
423 EIGEN_DEVICE_FUNC
424 const AnyReturnType any() const
425 { return AnyReturnType(_expression()); }
426
427 /** \returns a row (or column) vector expression representing
428 * the number of \c true coefficients of each respective column (or row).
429 * This expression can be assigned to a vector whose entries have the same type as is used to
430 * index entries of the original matrix; for dense matrices, this is \c std::ptrdiff_t .
431 *
432 * Example: \include PartialRedux_count.cpp
433 * Output: \verbinclude PartialRedux_count.out
434 *
435 * \sa DenseBase::count() */
436 EIGEN_DEVICE_FUNC
437 const CountReturnType count() const
438 { return CountReturnType(_expression()); }
439
440 /** \returns a row (or column) vector expression of the product
441 * of each column (or row) of the referenced expression.
442 *
443 * Example: \include PartialRedux_prod.cpp
444 * Output: \verbinclude PartialRedux_prod.out
445 *
446 * \sa DenseBase::prod() */
447 EIGEN_DEVICE_FUNC
448 const ProdReturnType prod() const
449 { return ProdReturnType(_expression()); }
450
451
452 /** \returns a matrix expression
453 * where each column (or row) are reversed.
454 *
455 * Example: \include Vectorwise_reverse.cpp
456 * Output: \verbinclude Vectorwise_reverse.out
457 *
458 * \sa DenseBase::reverse() */
459 EIGEN_DEVICE_FUNC
460 const ConstReverseReturnType reverse() const
461 { return ConstReverseReturnType( _expression() ); }
462
463 /** \returns a writable matrix expression
464 * where each column (or row) are reversed.
465 *
466 * \sa reverse() const */
467 EIGEN_DEVICE_FUNC
468 ReverseReturnType reverse()
469 { return ReverseReturnType( _expression() ); }
470
471 typedef Replicate<ExpressionType,(isVertical?Dynamic:1),(isHorizontal?Dynamic:1)> ReplicateReturnType;
472 EIGEN_DEVICE_FUNC
473 const ReplicateReturnType replicate(Index factor) const;
474
475 /**
476 * \return an expression of the replication of each column (or row) of \c *this
477 *
478 * Example: \include DirectionWise_replicate.cpp
479 * Output: \verbinclude DirectionWise_replicate.out
480 *
481 * \sa VectorwiseOp::replicate(Index), DenseBase::replicate(), class Replicate
482 */
483 // NOTE implemented here because of sunstudio's compilation errors
484 // isVertical*Factor+isHorizontal instead of (isVertical?Factor:1) to handle CUDA bug with ternary operator
485 template<int Factor> const Replicate<ExpressionType,isVertical*Factor+isHorizontal,isHorizontal*Factor+isVertical>
486 EIGEN_DEVICE_FUNC
487 replicate(Index factor = Factor) const
488 {
489 return Replicate<ExpressionType,(isVertical?Factor:1),(isHorizontal?Factor:1)>
490 (_expression(),isVertical?factor:1,isHorizontal?factor:1);
491 }
492
493/////////// Artithmetic operators ///////////
494
495 /** Copies the vector \a other to each subvector of \c *this */
496 template<typename OtherDerived>
497 EIGEN_DEVICE_FUNC
498 ExpressionType& operator=(const DenseBase<OtherDerived>& other)
499 {
500 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
501 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
502 //eigen_assert((m_matrix.isNull()) == (other.isNull())); FIXME
503 return const_cast<ExpressionType&>(m_matrix = extendedTo(other.derived()));
504 }
505
506 /** Adds the vector \a other to each subvector of \c *this */
507 template<typename OtherDerived>
508 EIGEN_DEVICE_FUNC
509 ExpressionType& operator+=(const DenseBase<OtherDerived>& other)
510 {
511 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
512 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
513 return const_cast<ExpressionType&>(m_matrix += extendedTo(other.derived()));
514 }
515
516 /** Substracts the vector \a other to each subvector of \c *this */
517 template<typename OtherDerived>
518 EIGEN_DEVICE_FUNC
519 ExpressionType& operator-=(const DenseBase<OtherDerived>& other)
520 {
521 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
522 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
523 return const_cast<ExpressionType&>(m_matrix -= extendedTo(other.derived()));
524 }
525
526 /** Multiples each subvector of \c *this by the vector \a other */
527 template<typename OtherDerived>
528 EIGEN_DEVICE_FUNC
529 ExpressionType& operator*=(const DenseBase<OtherDerived>& other)
530 {
531 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
532 EIGEN_STATIC_ASSERT_ARRAYXPR(ExpressionType)
533 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
534 m_matrix *= extendedTo(other.derived());
535 return const_cast<ExpressionType&>(m_matrix);
536 }
537
538 /** Divides each subvector of \c *this by the vector \a other */
539 template<typename OtherDerived>
540 EIGEN_DEVICE_FUNC
541 ExpressionType& operator/=(const DenseBase<OtherDerived>& other)
542 {
543 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
544 EIGEN_STATIC_ASSERT_ARRAYXPR(ExpressionType)
545 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
546 m_matrix /= extendedTo(other.derived());
547 return const_cast<ExpressionType&>(m_matrix);
548 }
549
550 /** Returns the expression of the sum of the vector \a other to each subvector of \c *this */
551 template<typename OtherDerived> EIGEN_STRONG_INLINE EIGEN_DEVICE_FUNC
552 CwiseBinaryOp<internal::scalar_sum_op<Scalar,typename OtherDerived::Scalar>,
553 const ExpressionTypeNestedCleaned,
554 const typename ExtendedType<OtherDerived>::Type>
555 operator+(const DenseBase<OtherDerived>& other) const
556 {
557 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
558 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
559 return m_matrix + extendedTo(other.derived());
560 }
561
562 /** Returns the expression of the difference between each subvector of \c *this and the vector \a other */
563 template<typename OtherDerived>
564 EIGEN_DEVICE_FUNC
565 CwiseBinaryOp<internal::scalar_difference_op<Scalar,typename OtherDerived::Scalar>,
566 const ExpressionTypeNestedCleaned,
567 const typename ExtendedType<OtherDerived>::Type>
568 operator-(const DenseBase<OtherDerived>& other) const
569 {
570 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
571 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
572 return m_matrix - extendedTo(other.derived());
573 }
574
575 /** Returns the expression where each subvector is the product of the vector \a other
576 * by the corresponding subvector of \c *this */
577 template<typename OtherDerived> EIGEN_STRONG_INLINE EIGEN_DEVICE_FUNC
578 CwiseBinaryOp<internal::scalar_product_op<Scalar>,
579 const ExpressionTypeNestedCleaned,
580 const typename ExtendedType<OtherDerived>::Type>
581 EIGEN_DEVICE_FUNC
582 operator*(const DenseBase<OtherDerived>& other) const
583 {
584 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
585 EIGEN_STATIC_ASSERT_ARRAYXPR(ExpressionType)
586 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
587 return m_matrix * extendedTo(other.derived());
588 }
589
590 /** Returns the expression where each subvector is the quotient of the corresponding
591 * subvector of \c *this by the vector \a other */
592 template<typename OtherDerived>
593 EIGEN_DEVICE_FUNC
594 CwiseBinaryOp<internal::scalar_quotient_op<Scalar>,
595 const ExpressionTypeNestedCleaned,
596 const typename ExtendedType<OtherDerived>::Type>
597 operator/(const DenseBase<OtherDerived>& other) const
598 {
599 EIGEN_STATIC_ASSERT_VECTOR_ONLY(OtherDerived)
600 EIGEN_STATIC_ASSERT_ARRAYXPR(ExpressionType)
601 EIGEN_STATIC_ASSERT_SAME_XPR_KIND(ExpressionType, OtherDerived)
602 return m_matrix / extendedTo(other.derived());
603 }
604
605 /** \returns an expression where each column (or row) of the referenced matrix are normalized.
606 * The referenced matrix is \b not modified.
607 * \sa MatrixBase::normalized(), normalize()
608 */
609 EIGEN_DEVICE_FUNC
610 CwiseBinaryOp<internal::scalar_quotient_op<Scalar>,
611 const ExpressionTypeNestedCleaned,
612 const typename OppositeExtendedType<typename ReturnType<internal::member_norm,RealScalar>::Type>::Type>
613 normalized() const { return m_matrix.cwiseQuotient(extendedToOpposite(this->norm())); }
614
615
616 /** Normalize in-place each row or columns of the referenced matrix.
617 * \sa MatrixBase::normalize(), normalized()
618 */
619 EIGEN_DEVICE_FUNC void normalize() {
620 m_matrix = this->normalized();
621 }
622
623 EIGEN_DEVICE_FUNC inline void reverseInPlace();
624
625/////////// Geometry module ///////////
626
627 typedef Homogeneous<ExpressionType,Direction> HomogeneousReturnType;
628 EIGEN_DEVICE_FUNC
629 HomogeneousReturnType homogeneous() const;
630
631 typedef typename ExpressionType::PlainObject CrossReturnType;
632 template<typename OtherDerived>
633 EIGEN_DEVICE_FUNC
634 const CrossReturnType cross(const MatrixBase<OtherDerived>& other) const;
635
636 enum {
637 HNormalized_Size = Direction==Vertical ? internal::traits<ExpressionType>::RowsAtCompileTime
638 : internal::traits<ExpressionType>::ColsAtCompileTime,
639 HNormalized_SizeMinusOne = HNormalized_Size==Dynamic ? Dynamic : HNormalized_Size-1
640 };
641 typedef Block<const ExpressionType,
642 Direction==Vertical ? int(HNormalized_SizeMinusOne)
643 : int(internal::traits<ExpressionType>::RowsAtCompileTime),
644 Direction==Horizontal ? int(HNormalized_SizeMinusOne)
645 : int(internal::traits<ExpressionType>::ColsAtCompileTime)>
646 HNormalized_Block;
647 typedef Block<const ExpressionType,
648 Direction==Vertical ? 1 : int(internal::traits<ExpressionType>::RowsAtCompileTime),
649 Direction==Horizontal ? 1 : int(internal::traits<ExpressionType>::ColsAtCompileTime)>
650 HNormalized_Factors;
651 typedef CwiseBinaryOp<internal::scalar_quotient_op<typename internal::traits<ExpressionType>::Scalar>,
652 const HNormalized_Block,
653 const Replicate<HNormalized_Factors,
654 Direction==Vertical ? HNormalized_SizeMinusOne : 1,
655 Direction==Horizontal ? HNormalized_SizeMinusOne : 1> >
656 HNormalizedReturnType;
657
658 EIGEN_DEVICE_FUNC
659 const HNormalizedReturnType hnormalized() const;
660
661 protected:
662 ExpressionTypeNested m_matrix;
663};
664
665//const colwise moved to DenseBase.h due to CUDA compiler bug
666
667
668/** \returns a writable VectorwiseOp wrapper of *this providing additional partial reduction operations
669 *
670 * \sa rowwise(), class VectorwiseOp, \ref TutorialReductionsVisitorsBroadcasting
671 */
672template<typename Derived>
673inline typename DenseBase<Derived>::ColwiseReturnType
674DenseBase<Derived>::colwise()
675{
676 return ColwiseReturnType(derived());
677}
678
679//const rowwise moved to DenseBase.h due to CUDA compiler bug
680
681
682/** \returns a writable VectorwiseOp wrapper of *this providing additional partial reduction operations
683 *
684 * \sa colwise(), class VectorwiseOp, \ref TutorialReductionsVisitorsBroadcasting
685 */
686template<typename Derived>
687inline typename DenseBase<Derived>::RowwiseReturnType
688DenseBase<Derived>::rowwise()
689{
690 return RowwiseReturnType(derived());
691}
692
693} // end namespace Eigen
694
695#endif // EIGEN_PARTIAL_REDUX_H
696