Home | History | Annotate | Download | only in Core
      1 // This file is part of Eigen, a lightweight C++ template library
      2 // for linear algebra.
      3 //
      4 // Copyright (C) 2006-2010 Benoit Jacob <jacob.benoit.1 (at) gmail.com>
      5 // Copyright (C) 2008-2009 Gael Guennebaud <gael.guennebaud (at) 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_MATRIX_H
     12 #define EIGEN_MATRIX_H
     13 
     14 namespace Eigen {
     15 
     16 namespace internal {
     17 template<typename _Scalar, int _Rows, int _Cols, int _Options, int _MaxRows, int _MaxCols>
     18 struct traits<Matrix<_Scalar, _Rows, _Cols, _Options, _MaxRows, _MaxCols> >
     19 {
     20 private:
     21   enum { size = internal::size_at_compile_time<_Rows,_Cols>::ret };
     22   typedef typename find_best_packet<_Scalar,size>::type PacketScalar;
     23   enum {
     24       row_major_bit = _Options&RowMajor ? RowMajorBit : 0,
     25       is_dynamic_size_storage = _MaxRows==Dynamic || _MaxCols==Dynamic,
     26       max_size = is_dynamic_size_storage ? Dynamic : _MaxRows*_MaxCols,
     27       default_alignment = compute_default_alignment<_Scalar,max_size>::value,
     28       actual_alignment = ((_Options&DontAlign)==0) ? default_alignment : 0,
     29       required_alignment = unpacket_traits<PacketScalar>::alignment,
     30       packet_access_bit = (packet_traits<_Scalar>::Vectorizable && (EIGEN_UNALIGNED_VECTORIZE || (actual_alignment>=required_alignment))) ? PacketAccessBit : 0
     31     };
     32 
     33 public:
     34   typedef _Scalar Scalar;
     35   typedef Dense StorageKind;
     36   typedef Eigen::Index StorageIndex;
     37   typedef MatrixXpr XprKind;
     38   enum {
     39     RowsAtCompileTime = _Rows,
     40     ColsAtCompileTime = _Cols,
     41     MaxRowsAtCompileTime = _MaxRows,
     42     MaxColsAtCompileTime = _MaxCols,
     43     Flags = compute_matrix_flags<_Scalar, _Rows, _Cols, _Options, _MaxRows, _MaxCols>::ret,
     44     Options = _Options,
     45     InnerStrideAtCompileTime = 1,
     46     OuterStrideAtCompileTime = (Options&RowMajor) ? ColsAtCompileTime : RowsAtCompileTime,
     47 
     48     // FIXME, the following flag in only used to define NeedsToAlign in PlainObjectBase
     49     EvaluatorFlags = LinearAccessBit | DirectAccessBit | packet_access_bit | row_major_bit,
     50     Alignment = actual_alignment
     51   };
     52 };
     53 }
     54 
     55 /** \class Matrix
     56   * \ingroup Core_Module
     57   *
     58   * \brief The matrix class, also used for vectors and row-vectors
     59   *
     60   * The %Matrix class is the work-horse for all \em dense (\ref dense "note") matrices and vectors within Eigen.
     61   * Vectors are matrices with one column, and row-vectors are matrices with one row.
     62   *
     63   * The %Matrix class encompasses \em both fixed-size and dynamic-size objects (\ref fixedsize "note").
     64   *
     65   * The first three template parameters are required:
     66   * \tparam _Scalar Numeric type, e.g. float, double, int or std::complex<float>.
     67   *                 User defined scalar types are supported as well (see \ref user_defined_scalars "here").
     68   * \tparam _Rows Number of rows, or \b Dynamic
     69   * \tparam _Cols Number of columns, or \b Dynamic
     70   *
     71   * The remaining template parameters are optional -- in most cases you don't have to worry about them.
     72   * \tparam _Options A combination of either \b #RowMajor or \b #ColMajor, and of either
     73   *                 \b #AutoAlign or \b #DontAlign.
     74   *                 The former controls \ref TopicStorageOrders "storage order", and defaults to column-major. The latter controls alignment, which is required
     75   *                 for vectorization. It defaults to aligning matrices except for fixed sizes that aren't a multiple of the packet size.
     76   * \tparam _MaxRows Maximum number of rows. Defaults to \a _Rows (\ref maxrows "note").
     77   * \tparam _MaxCols Maximum number of columns. Defaults to \a _Cols (\ref maxrows "note").
     78   *
     79   * Eigen provides a number of typedefs covering the usual cases. Here are some examples:
     80   *
     81   * \li \c Matrix2d is a 2x2 square matrix of doubles (\c Matrix<double, 2, 2>)
     82   * \li \c Vector4f is a vector of 4 floats (\c Matrix<float, 4, 1>)
     83   * \li \c RowVector3i is a row-vector of 3 ints (\c Matrix<int, 1, 3>)
     84   *
     85   * \li \c MatrixXf is a dynamic-size matrix of floats (\c Matrix<float, Dynamic, Dynamic>)
     86   * \li \c VectorXf is a dynamic-size vector of floats (\c Matrix<float, Dynamic, 1>)
     87   *
     88   * \li \c Matrix2Xf is a partially fixed-size (dynamic-size) matrix of floats (\c Matrix<float, 2, Dynamic>)
     89   * \li \c MatrixX3d is a partially dynamic-size (fixed-size) matrix of double (\c Matrix<double, Dynamic, 3>)
     90   *
     91   * See \link matrixtypedefs this page \endlink for a complete list of predefined \em %Matrix and \em Vector typedefs.
     92   *
     93   * You can access elements of vectors and matrices using normal subscripting:
     94   *
     95   * \code
     96   * Eigen::VectorXd v(10);
     97   * v[0] = 0.1;
     98   * v[1] = 0.2;
     99   * v(0) = 0.3;
    100   * v(1) = 0.4;
    101   *
    102   * Eigen::MatrixXi m(10, 10);
    103   * m(0, 1) = 1;
    104   * m(0, 2) = 2;
    105   * m(0, 3) = 3;
    106   * \endcode
    107   *
    108   * This class can be extended with the help of the plugin mechanism described on the page
    109   * \ref TopicCustomizing_Plugins by defining the preprocessor symbol \c EIGEN_MATRIX_PLUGIN.
    110   *
    111   * <i><b>Some notes:</b></i>
    112   *
    113   * <dl>
    114   * <dt><b>\anchor dense Dense versus sparse:</b></dt>
    115   * <dd>This %Matrix class handles dense, not sparse matrices and vectors. For sparse matrices and vectors, see the Sparse module.
    116   *
    117   * Dense matrices and vectors are plain usual arrays of coefficients. All the coefficients are stored, in an ordinary contiguous array.
    118   * This is unlike Sparse matrices and vectors where the coefficients are stored as a list of nonzero coefficients.</dd>
    119   *
    120   * <dt><b>\anchor fixedsize Fixed-size versus dynamic-size:</b></dt>
    121   * <dd>Fixed-size means that the numbers of rows and columns are known are compile-time. In this case, Eigen allocates the array
    122   * of coefficients as a fixed-size array, as a class member. This makes sense for very small matrices, typically up to 4x4, sometimes up
    123   * to 16x16. Larger matrices should be declared as dynamic-size even if one happens to know their size at compile-time.
    124   *
    125   * Dynamic-size means that the numbers of rows or columns are not necessarily known at compile-time. In this case they are runtime
    126   * variables, and the array of coefficients is allocated dynamically on the heap.
    127   *
    128   * Note that \em dense matrices, be they Fixed-size or Dynamic-size, <em>do not</em> expand dynamically in the sense of a std::map.
    129   * If you want this behavior, see the Sparse module.</dd>
    130   *
    131   * <dt><b>\anchor maxrows _MaxRows and _MaxCols:</b></dt>
    132   * <dd>In most cases, one just leaves these parameters to the default values.
    133   * These parameters mean the maximum size of rows and columns that the matrix may have. They are useful in cases
    134   * when the exact numbers of rows and columns are not known are compile-time, but it is known at compile-time that they cannot
    135   * exceed a certain value. This happens when taking dynamic-size blocks inside fixed-size matrices: in this case _MaxRows and _MaxCols
    136   * are the dimensions of the original matrix, while _Rows and _Cols are Dynamic.</dd>
    137   * </dl>
    138   *
    139   * <i><b>ABI and storage layout</b></i>
    140   *
    141   * The table below summarizes the ABI of some possible Matrix instances which is fixed thorough the lifetime of Eigen 3.
    142   * <table  class="manual">
    143   * <tr><th>Matrix type</th><th>Equivalent C structure</th></tr>
    144   * <tr><td>\code Matrix<T,Dynamic,Dynamic> \endcode</td><td>\code
    145   * struct {
    146   *   T *data;                  // with (size_t(data)%EIGEN_MAX_ALIGN_BYTES)==0
    147   *   Eigen::Index rows, cols;
    148   *  };
    149   * \endcode</td></tr>
    150   * <tr class="alt"><td>\code
    151   * Matrix<T,Dynamic,1>
    152   * Matrix<T,1,Dynamic> \endcode</td><td>\code
    153   * struct {
    154   *   T *data;                  // with (size_t(data)%EIGEN_MAX_ALIGN_BYTES)==0
    155   *   Eigen::Index size;
    156   *  };
    157   * \endcode</td></tr>
    158   * <tr><td>\code Matrix<T,Rows,Cols> \endcode</td><td>\code
    159   * struct {
    160   *   T data[Rows*Cols];        // with (size_t(data)%A(Rows*Cols*sizeof(T)))==0
    161   *  };
    162   * \endcode</td></tr>
    163   * <tr class="alt"><td>\code Matrix<T,Dynamic,Dynamic,0,MaxRows,MaxCols> \endcode</td><td>\code
    164   * struct {
    165   *   T data[MaxRows*MaxCols];  // with (size_t(data)%A(MaxRows*MaxCols*sizeof(T)))==0
    166   *   Eigen::Index rows, cols;
    167   *  };
    168   * \endcode</td></tr>
    169   * </table>
    170   * Note that in this table Rows, Cols, MaxRows and MaxCols are all positive integers. A(S) is defined to the largest possible power-of-two
    171   * smaller to EIGEN_MAX_STATIC_ALIGN_BYTES.
    172   *
    173   * \see MatrixBase for the majority of the API methods for matrices, \ref TopicClassHierarchy,
    174   * \ref TopicStorageOrders
    175   */
    176 
    177 template<typename _Scalar, int _Rows, int _Cols, int _Options, int _MaxRows, int _MaxCols>
    178 class Matrix
    179   : public PlainObjectBase<Matrix<_Scalar, _Rows, _Cols, _Options, _MaxRows, _MaxCols> >
    180 {
    181   public:
    182 
    183     /** \brief Base class typedef.
    184       * \sa PlainObjectBase
    185       */
    186     typedef PlainObjectBase<Matrix> Base;
    187 
    188     enum { Options = _Options };
    189 
    190     EIGEN_DENSE_PUBLIC_INTERFACE(Matrix)
    191 
    192     typedef typename Base::PlainObject PlainObject;
    193 
    194     using Base::base;
    195     using Base::coeffRef;
    196 
    197     /**
    198       * \brief Assigns matrices to each other.
    199       *
    200       * \note This is a special case of the templated operator=. Its purpose is
    201       * to prevent a default operator= from hiding the templated operator=.
    202       *
    203       * \callgraph
    204       */
    205     EIGEN_DEVICE_FUNC
    206     EIGEN_STRONG_INLINE Matrix& operator=(const Matrix& other)
    207     {
    208       return Base::_set(other);
    209     }
    210 
    211     /** \internal
    212       * \brief Copies the value of the expression \a other into \c *this with automatic resizing.
    213       *
    214       * *this might be resized to match the dimensions of \a other. If *this was a null matrix (not already initialized),
    215       * it will be initialized.
    216       *
    217       * Note that copying a row-vector into a vector (and conversely) is allowed.
    218       * The resizing, if any, is then done in the appropriate way so that row-vectors
    219       * remain row-vectors and vectors remain vectors.
    220       */
    221     template<typename OtherDerived>
    222     EIGEN_DEVICE_FUNC
    223     EIGEN_STRONG_INLINE Matrix& operator=(const DenseBase<OtherDerived>& other)
    224     {
    225       return Base::_set(other);
    226     }
    227 
    228     /* Here, doxygen failed to copy the brief information when using \copydoc */
    229 
    230     /**
    231       * \brief Copies the generic expression \a other into *this.
    232       * \copydetails DenseBase::operator=(const EigenBase<OtherDerived> &other)
    233       */
    234     template<typename OtherDerived>
    235     EIGEN_DEVICE_FUNC
    236     EIGEN_STRONG_INLINE Matrix& operator=(const EigenBase<OtherDerived> &other)
    237     {
    238       return Base::operator=(other);
    239     }
    240 
    241     template<typename OtherDerived>
    242     EIGEN_DEVICE_FUNC
    243     EIGEN_STRONG_INLINE Matrix& operator=(const ReturnByValue<OtherDerived>& func)
    244     {
    245       return Base::operator=(func);
    246     }
    247 
    248     /** \brief Default constructor.
    249       *
    250       * For fixed-size matrices, does nothing.
    251       *
    252       * For dynamic-size matrices, creates an empty matrix of size 0. Does not allocate any array. Such a matrix
    253       * is called a null matrix. This constructor is the unique way to create null matrices: resizing
    254       * a matrix to 0 is not supported.
    255       *
    256       * \sa resize(Index,Index)
    257       */
    258     EIGEN_DEVICE_FUNC
    259     EIGEN_STRONG_INLINE Matrix() : Base()
    260     {
    261       Base::_check_template_params();
    262       EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED
    263     }
    264 
    265     // FIXME is it still needed
    266     EIGEN_DEVICE_FUNC
    267     explicit Matrix(internal::constructor_without_unaligned_array_assert)
    268       : Base(internal::constructor_without_unaligned_array_assert())
    269     { Base::_check_template_params(); EIGEN_INITIALIZE_COEFFS_IF_THAT_OPTION_IS_ENABLED }
    270 
    271 #if EIGEN_HAS_RVALUE_REFERENCES
    272     EIGEN_DEVICE_FUNC
    273     Matrix(Matrix&& other) EIGEN_NOEXCEPT_IF(std::is_nothrow_move_constructible<Scalar>::value)
    274       : Base(std::move(other))
    275     {
    276       Base::_check_template_params();
    277       if (RowsAtCompileTime!=Dynamic && ColsAtCompileTime!=Dynamic)
    278         Base::_set_noalias(other);
    279     }
    280     EIGEN_DEVICE_FUNC
    281     Matrix& operator=(Matrix&& other) EIGEN_NOEXCEPT_IF(std::is_nothrow_move_assignable<Scalar>::value)
    282     {
    283       other.swap(*this);
    284       return *this;
    285     }
    286 #endif
    287 
    288     #ifndef EIGEN_PARSED_BY_DOXYGEN
    289 
    290     // This constructor is for both 1x1 matrices and dynamic vectors
    291     template<typename T>
    292     EIGEN_DEVICE_FUNC
    293     EIGEN_STRONG_INLINE explicit Matrix(const T& x)
    294     {
    295       Base::_check_template_params();
    296       Base::template _init1<T>(x);
    297     }
    298 
    299     template<typename T0, typename T1>
    300     EIGEN_DEVICE_FUNC
    301     EIGEN_STRONG_INLINE Matrix(const T0& x, const T1& y)
    302     {
    303       Base::_check_template_params();
    304       Base::template _init2<T0,T1>(x, y);
    305     }
    306     #else
    307     /** \brief Constructs a fixed-sized matrix initialized with coefficients starting at \a data */
    308     EIGEN_DEVICE_FUNC
    309     explicit Matrix(const Scalar *data);
    310 
    311     /** \brief Constructs a vector or row-vector with given dimension. \only_for_vectors
    312       *
    313       * This is useful for dynamic-size vectors. For fixed-size vectors,
    314       * it is redundant to pass these parameters, so one should use the default constructor
    315       * Matrix() instead.
    316       *
    317       * \warning This constructor is disabled for fixed-size \c 1x1 matrices. For instance,
    318       * calling Matrix<double,1,1>(1) will call the initialization constructor: Matrix(const Scalar&).
    319       * For fixed-size \c 1x1 matrices it is therefore recommended to use the default
    320       * constructor Matrix() instead, especially when using one of the non standard
    321       * \c EIGEN_INITIALIZE_MATRICES_BY_{ZERO,\c NAN} macros (see \ref TopicPreprocessorDirectives).
    322       */
    323     EIGEN_STRONG_INLINE explicit Matrix(Index dim);
    324     /** \brief Constructs an initialized 1x1 matrix with the given coefficient */
    325     Matrix(const Scalar& x);
    326     /** \brief Constructs an uninitialized matrix with \a rows rows and \a cols columns.
    327       *
    328       * This is useful for dynamic-size matrices. For fixed-size matrices,
    329       * it is redundant to pass these parameters, so one should use the default constructor
    330       * Matrix() instead.
    331       *
    332       * \warning This constructor is disabled for fixed-size \c 1x2 and \c 2x1 vectors. For instance,
    333       * calling Matrix2f(2,1) will call the initialization constructor: Matrix(const Scalar& x, const Scalar& y).
    334       * For fixed-size \c 1x2 or \c 2x1 vectors it is therefore recommended to use the default
    335       * constructor Matrix() instead, especially when using one of the non standard
    336       * \c EIGEN_INITIALIZE_MATRICES_BY_{ZERO,\c NAN} macros (see \ref TopicPreprocessorDirectives).
    337       */
    338     EIGEN_DEVICE_FUNC
    339     Matrix(Index rows, Index cols);
    340 
    341     /** \brief Constructs an initialized 2D vector with given coefficients */
    342     Matrix(const Scalar& x, const Scalar& y);
    343     #endif
    344 
    345     /** \brief Constructs an initialized 3D vector with given coefficients */
    346     EIGEN_DEVICE_FUNC
    347     EIGEN_STRONG_INLINE Matrix(const Scalar& x, const Scalar& y, const Scalar& z)
    348     {
    349       Base::_check_template_params();
    350       EIGEN_STATIC_ASSERT_VECTOR_SPECIFIC_SIZE(Matrix, 3)
    351       m_storage.data()[0] = x;
    352       m_storage.data()[1] = y;
    353       m_storage.data()[2] = z;
    354     }
    355     /** \brief Constructs an initialized 4D vector with given coefficients */
    356     EIGEN_DEVICE_FUNC
    357     EIGEN_STRONG_INLINE Matrix(const Scalar& x, const Scalar& y, const Scalar& z, const Scalar& w)
    358     {
    359       Base::_check_template_params();
    360       EIGEN_STATIC_ASSERT_VECTOR_SPECIFIC_SIZE(Matrix, 4)
    361       m_storage.data()[0] = x;
    362       m_storage.data()[1] = y;
    363       m_storage.data()[2] = z;
    364       m_storage.data()[3] = w;
    365     }
    366 
    367 
    368     /** \brief Copy constructor */
    369     EIGEN_DEVICE_FUNC
    370     EIGEN_STRONG_INLINE Matrix(const Matrix& other) : Base(other)
    371     { }
    372 
    373     /** \brief Copy constructor for generic expressions.
    374       * \sa MatrixBase::operator=(const EigenBase<OtherDerived>&)
    375       */
    376     template<typename OtherDerived>
    377     EIGEN_DEVICE_FUNC
    378     EIGEN_STRONG_INLINE Matrix(const EigenBase<OtherDerived> &other)
    379       : Base(other.derived())
    380     { }
    381 
    382     EIGEN_DEVICE_FUNC inline Index innerStride() const { return 1; }
    383     EIGEN_DEVICE_FUNC inline Index outerStride() const { return this->innerSize(); }
    384 
    385     /////////// Geometry module ///////////
    386 
    387     template<typename OtherDerived>
    388     EIGEN_DEVICE_FUNC
    389     explicit Matrix(const RotationBase<OtherDerived,ColsAtCompileTime>& r);
    390     template<typename OtherDerived>
    391     EIGEN_DEVICE_FUNC
    392     Matrix& operator=(const RotationBase<OtherDerived,ColsAtCompileTime>& r);
    393 
    394     // allow to extend Matrix outside Eigen
    395     #ifdef EIGEN_MATRIX_PLUGIN
    396     #include EIGEN_MATRIX_PLUGIN
    397     #endif
    398 
    399   protected:
    400     template <typename Derived, typename OtherDerived, bool IsVector>
    401     friend struct internal::conservative_resize_like_impl;
    402 
    403     using Base::m_storage;
    404 };
    405 
    406 /** \defgroup matrixtypedefs Global matrix typedefs
    407   *
    408   * \ingroup Core_Module
    409   *
    410   * Eigen defines several typedef shortcuts for most common matrix and vector types.
    411   *
    412   * The general patterns are the following:
    413   *
    414   * \c MatrixSizeType where \c Size can be \c 2,\c 3,\c 4 for fixed size square matrices or \c X for dynamic size,
    415   * and where \c Type can be \c i for integer, \c f for float, \c d for double, \c cf for complex float, \c cd
    416   * for complex double.
    417   *
    418   * For example, \c Matrix3d is a fixed-size 3x3 matrix type of doubles, and \c MatrixXf is a dynamic-size matrix of floats.
    419   *
    420   * There are also \c VectorSizeType and \c RowVectorSizeType which are self-explanatory. For example, \c Vector4cf is
    421   * a fixed-size vector of 4 complex floats.
    422   *
    423   * \sa class Matrix
    424   */
    425 
    426 #define EIGEN_MAKE_TYPEDEFS(Type, TypeSuffix, Size, SizeSuffix)   \
    427 /** \ingroup matrixtypedefs */                                    \
    428 typedef Matrix<Type, Size, Size> Matrix##SizeSuffix##TypeSuffix;  \
    429 /** \ingroup matrixtypedefs */                                    \
    430 typedef Matrix<Type, Size, 1>    Vector##SizeSuffix##TypeSuffix;  \
    431 /** \ingroup matrixtypedefs */                                    \
    432 typedef Matrix<Type, 1, Size>    RowVector##SizeSuffix##TypeSuffix;
    433 
    434 #define EIGEN_MAKE_FIXED_TYPEDEFS(Type, TypeSuffix, Size)         \
    435 /** \ingroup matrixtypedefs */                                    \
    436 typedef Matrix<Type, Size, Dynamic> Matrix##Size##X##TypeSuffix;  \
    437 /** \ingroup matrixtypedefs */                                    \
    438 typedef Matrix<Type, Dynamic, Size> Matrix##X##Size##TypeSuffix;
    439 
    440 #define EIGEN_MAKE_TYPEDEFS_ALL_SIZES(Type, TypeSuffix) \
    441 EIGEN_MAKE_TYPEDEFS(Type, TypeSuffix, 2, 2) \
    442 EIGEN_MAKE_TYPEDEFS(Type, TypeSuffix, 3, 3) \
    443 EIGEN_MAKE_TYPEDEFS(Type, TypeSuffix, 4, 4) \
    444 EIGEN_MAKE_TYPEDEFS(Type, TypeSuffix, Dynamic, X) \
    445 EIGEN_MAKE_FIXED_TYPEDEFS(Type, TypeSuffix, 2) \
    446 EIGEN_MAKE_FIXED_TYPEDEFS(Type, TypeSuffix, 3) \
    447 EIGEN_MAKE_FIXED_TYPEDEFS(Type, TypeSuffix, 4)
    448 
    449 EIGEN_MAKE_TYPEDEFS_ALL_SIZES(int,                  i)
    450 EIGEN_MAKE_TYPEDEFS_ALL_SIZES(float,                f)
    451 EIGEN_MAKE_TYPEDEFS_ALL_SIZES(double,               d)
    452 EIGEN_MAKE_TYPEDEFS_ALL_SIZES(std::complex<float>,  cf)
    453 EIGEN_MAKE_TYPEDEFS_ALL_SIZES(std::complex<double>, cd)
    454 
    455 #undef EIGEN_MAKE_TYPEDEFS_ALL_SIZES
    456 #undef EIGEN_MAKE_TYPEDEFS
    457 #undef EIGEN_MAKE_FIXED_TYPEDEFS
    458 
    459 } // end namespace Eigen
    460 
    461 #endif // EIGEN_MATRIX_H
    462