|
|
|
|
#ifndef Magnum_Math_RectangularMatrix_h
|
|
|
|
|
#define Magnum_Math_RectangularMatrix_h
|
|
|
|
|
/*
|
|
|
|
|
This file is part of Magnum.
|
|
|
|
|
|
|
|
|
|
Copyright © 2010, 2011, 2012, 2013, 2014
|
|
|
|
|
Vladimír Vondruš <mosra@centrum.cz>
|
|
|
|
|
|
|
|
|
|
Permission is hereby granted, free of charge, to any person obtaining a
|
|
|
|
|
copy of this software and associated documentation files (the "Software"),
|
|
|
|
|
to deal in the Software without restriction, including without limitation
|
|
|
|
|
the rights to use, copy, modify, merge, publish, distribute, sublicense,
|
|
|
|
|
and/or sell copies of the Software, and to permit persons to whom the
|
|
|
|
|
Software is furnished to do so, subject to the following conditions:
|
|
|
|
|
|
|
|
|
|
The above copyright notice and this permission notice shall be included
|
|
|
|
|
in all copies or substantial portions of the Software.
|
|
|
|
|
|
|
|
|
|
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
|
|
|
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
|
|
|
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
|
|
|
|
|
THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
|
|
|
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
|
|
|
|
|
FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
|
|
|
|
|
DEALINGS IN THE SOFTWARE.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
/** @file
|
|
|
|
|
* @brief Class @ref Magnum::Math::RectangularMatrix, typedef @ref Magnum::Math::Matrix2x3, @ref Magnum::Math::Matrix3x2, @ref Magnum::Math::Matrix2x4, @ref Magnum::Math::Matrix4x2, @ref Magnum::Math::Matrix3x4, @ref Magnum::Math::Matrix4x3
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include "Magnum/Math/Vector.h"
|
|
|
|
|
|
|
|
|
|
namespace Magnum { namespace Math {
|
|
|
|
|
|
|
|
|
|
namespace Implementation {
|
|
|
|
|
template<std::size_t, std::size_t, class, class> struct RectangularMatrixConverter;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
@brief Rectangular matrix
|
|
|
|
|
@tparam cols Column count
|
|
|
|
|
@tparam rows Row count
|
|
|
|
|
@tparam T Underlying data type
|
|
|
|
|
|
|
|
|
|
See @ref matrix-vector for brief introduction. See also @ref Matrix (square)
|
|
|
|
|
and @ref Vector.
|
|
|
|
|
|
|
|
|
|
The data are stored in column-major order, to reflect that, all indices in
|
|
|
|
|
math formulas are in reverse order (i.e. @f$ \boldsymbol A_{ji} @f$ instead
|
|
|
|
|
of @f$ \boldsymbol A_{ij} @f$).
|
|
|
|
|
@see @ref Matrix2x3, @ref Matrix3x2, @ref Matrix2x4, @ref Matrix4x2,
|
|
|
|
|
@ref Matrix3x4, @ref Matrix4x3
|
|
|
|
|
*/
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> class RectangularMatrix {
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
static_assert(cols != 0 && rows != 0, "RectangularMatrix cannot have zero elements");
|
|
|
|
|
|
|
|
|
|
template<std::size_t, std::size_t, class> friend class RectangularMatrix;
|
|
|
|
|
|
|
|
|
|
public:
|
|
|
|
|
typedef T Type; /**< @brief Underlying data type */
|
|
|
|
|
const static std::size_t Cols = cols; /**< @brief %Matrix column count */
|
|
|
|
|
const static std::size_t Rows = rows; /**< @brief %Matrix row count */
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Size of matrix diagonal
|
|
|
|
|
*
|
|
|
|
|
* @see @ref fromDiagonal(), @ref diagonal()
|
|
|
|
|
*/
|
|
|
|
|
const static std::size_t DiagonalSize = (cols < rows ? cols : rows);
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief %Matrix from array
|
|
|
|
|
* @return Reference to the data as if it was matrix, thus doesn't
|
|
|
|
|
* perform any copying.
|
|
|
|
|
*
|
|
|
|
|
* @attention Use with caution, the function doesn't check whether the
|
|
|
|
|
* array is long enough.
|
|
|
|
|
*/
|
|
|
|
|
static RectangularMatrix<cols, rows, T>& from(T* data) {
|
|
|
|
|
return *reinterpret_cast<RectangularMatrix<cols, rows, T>*>(data);
|
|
|
|
|
}
|
|
|
|
|
/** @overload */
|
|
|
|
|
static const RectangularMatrix<cols, rows, T>& from(const T* data) {
|
|
|
|
|
return *reinterpret_cast<const RectangularMatrix<cols, rows, T>*>(data);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Construct matrix from vector
|
|
|
|
|
*
|
|
|
|
|
* Rolls the vector into matrix, i.e. first `rows` elements of the
|
|
|
|
|
* vector will make first column of resulting matrix.
|
|
|
|
|
* @see @ref toVector()
|
|
|
|
|
*/
|
|
|
|
|
static RectangularMatrix<cols, rows, T> fromVector(const Vector<cols*rows, T>& vector) {
|
|
|
|
|
return *reinterpret_cast<const RectangularMatrix<cols, rows, T>*>(vector.data());
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Construct diagonal matrix
|
|
|
|
|
*
|
|
|
|
|
* @see @ref diagonal()
|
|
|
|
|
*/
|
|
|
|
|
constexpr static RectangularMatrix<cols, rows, T> fromDiagonal(const Vector<DiagonalSize, T>& diagonal) {
|
|
|
|
|
return RectangularMatrix(typename Implementation::GenerateSequence<cols>::Type(), diagonal);
|
|
|
|
|
}
|
|
|
|
|
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
/** @brief Construct zero-filled matrix */
|
|
|
|
|
constexpr /*implicit*/ RectangularMatrix() {}
|
|
|
|
|
|
|
|
|
|
/**
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
* @brief Construct matrix from column vectors
|
|
|
|
|
* @param first First column vector
|
|
|
|
|
* @param next Next column vectors
|
|
|
|
|
*
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
* @todo Creating matrix from arbitrary combination of matrices with n rows
|
|
|
|
|
*/
|
|
|
|
|
template<class ...U> constexpr /*implicit*/ RectangularMatrix(const Vector<rows, T>& first, const U&... next): _data{first, next...} {
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
static_assert(sizeof...(next)+1 == cols, "Improper number of arguments passed to RectangularMatrix constructor");
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Construct matrix from another of different type
|
|
|
|
|
*
|
|
|
|
|
* Performs only default casting on the values, no rounding or
|
|
|
|
|
* anything else. Example usage:
|
|
|
|
|
* @code
|
|
|
|
|
* RectangularMatrix<4, 1, Float> floatingPoint(1.3f, 2.7f, -15.0f, 7.0f);
|
|
|
|
|
* RectangularMatrix<4, 1, Byte> integral(floatingPoint);
|
|
|
|
|
* // integral == {1, 2, -15, 7}
|
|
|
|
|
* @endcode
|
|
|
|
|
*/
|
|
|
|
|
#ifndef CORRADE_GCC46_COMPATIBILITY
|
|
|
|
|
template<class U> constexpr explicit RectangularMatrix(const RectangularMatrix<cols, rows, U>& other): RectangularMatrix(typename Implementation::GenerateSequence<cols>::Type(), other) {}
|
|
|
|
|
#else
|
|
|
|
|
template<class U> explicit RectangularMatrix(const RectangularMatrix<cols, rows, U>& other) {
|
|
|
|
|
*this = RectangularMatrix(typename Implementation::GenerateSequence<cols>::Type(), other);
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/** @brief Construct matrix from external representation */
|
|
|
|
|
#ifndef CORRADE_GCC46_COMPATIBILITY
|
|
|
|
|
template<class U, class V = decltype(Implementation::RectangularMatrixConverter<cols, rows, T, U>::from(std::declval<U>()))> constexpr explicit RectangularMatrix(const U& other): RectangularMatrix(Implementation::RectangularMatrixConverter<cols, rows, T, U>::from(other)) {}
|
|
|
|
|
#else
|
|
|
|
|
template<class U, class V = decltype(Implementation::RectangularMatrixConverter<cols, rows, T, U>::from(std::declval<U>()))> explicit RectangularMatrix(const U& other) {
|
|
|
|
|
*this = Implementation::RectangularMatrixConverter<cols, rows, T, U>::from(other);
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/** @brief Copy constructor */
|
|
|
|
|
constexpr RectangularMatrix(const RectangularMatrix<cols, rows, T>&) = default;
|
|
|
|
|
|
|
|
|
|
/** @brief Assignment operator */
|
|
|
|
|
RectangularMatrix<cols, rows, T>& operator=(const RectangularMatrix<cols, rows, T>&) = default;
|
|
|
|
|
|
|
|
|
|
/** @brief Convert matrix to external representation */
|
|
|
|
|
template<class U, class V = decltype(Implementation::RectangularMatrixConverter<cols, rows, T, U>::to(std::declval<RectangularMatrix<cols, rows, T>>()))> constexpr explicit operator U() const {
|
|
|
|
|
/** @bug Why this is not constexpr under GCC 4.6? */
|
|
|
|
|
return Implementation::RectangularMatrixConverter<cols, rows, T, U>::to(*this);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Raw data
|
|
|
|
|
* @return One-dimensional array of `cols*rows` length in column-major
|
|
|
|
|
* order.
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
*
|
|
|
|
|
* @see operator[]()
|
|
|
|
|
* @todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
T* data() { return _data[0].data(); }
|
|
|
|
|
constexpr const T* data() const { return _data[0].data(); } /**< @overload */
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief %Matrix column
|
|
|
|
|
*
|
|
|
|
|
* Particular elements can be accessed using Vector::operator[](),
|
|
|
|
|
* e.g.:
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
* @code
|
|
|
|
|
* RectangularMatrix<4, 3, Float> m;
|
|
|
|
|
* Float a = m[2][1];
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
* @endcode
|
|
|
|
|
*
|
|
|
|
|
* @see @ref row(), @ref data()
|
|
|
|
|
* @todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
Vector<rows, T>& operator[](std::size_t col) { return _data[col]; }
|
|
|
|
|
constexpr const Vector<rows, T>& operator[](std::size_t col) const { return _data[col]; } /**< @overload */
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief %Matrix row
|
|
|
|
|
*
|
|
|
|
|
* Consider using @ref transposed() when accessing rows frequently, as
|
|
|
|
|
* this is slower than accessing columns due to the way the matrix is
|
|
|
|
|
* stored.
|
|
|
|
|
* @see operator[]()
|
|
|
|
|
* @todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
Vector<cols, T> row(std::size_t row) const;
|
|
|
|
|
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
/** @brief Equality comparison */
|
|
|
|
|
bool operator==(const RectangularMatrix<cols, rows, T>& other) const {
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
for(std::size_t i = 0; i != cols; ++i)
|
|
|
|
|
if(_data[i] != other._data[i]) return false;
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Non-equality operator
|
|
|
|
|
*
|
|
|
|
|
* @see Vector::operator<(), Vector::operator<=(),
|
|
|
|
|
* Vector::operator>=(), Vector::operator>()
|
|
|
|
|
* @todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
bool operator!=(const RectangularMatrix<cols, rows, T>& other) const {
|
|
|
|
|
return !operator==(other);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Negated matrix
|
|
|
|
|
*
|
|
|
|
|
* The computation is done column-wise. @f[
|
|
|
|
|
* \boldsymbol B_j = -\boldsymbol A_j
|
|
|
|
|
* @f]
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T> operator-() const;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Add and assign matrix
|
|
|
|
|
*
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
* The computation is done column-wise in-place. @f[
|
|
|
|
|
* \boldsymbol A_j = \boldsymbol A_j + \boldsymbol B_j
|
|
|
|
|
* @f]
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T>& operator+=(const RectangularMatrix<cols, rows, T>& other) {
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
for(std::size_t i = 0; i != cols; ++i)
|
|
|
|
|
_data[i] += other._data[i];
|
|
|
|
|
|
|
|
|
|
return *this;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Add matrix
|
|
|
|
|
*
|
|
|
|
|
* @see operator+=()
|
|
|
|
|
* @todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T> operator+(const RectangularMatrix<cols, rows, T>& other) const {
|
|
|
|
|
return RectangularMatrix<cols, rows, T>(*this)+=other;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Subtract and assign matrix
|
|
|
|
|
*
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
* The computation is done column-wise in-place. @f[
|
|
|
|
|
* \boldsymbol A_j = \boldsymbol A_j - \boldsymbol B_j
|
|
|
|
|
* @f]
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T>& operator-=(const RectangularMatrix<cols, rows, T>& other) {
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
for(std::size_t i = 0; i != cols; ++i)
|
|
|
|
|
_data[i] -= other._data[i];
|
|
|
|
|
|
|
|
|
|
return *this;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Subtract matrix
|
|
|
|
|
*
|
|
|
|
|
* @see @ref operator-=()
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T> operator-(const RectangularMatrix<cols, rows, T>& other) const {
|
|
|
|
|
return RectangularMatrix<cols, rows, T>(*this)-=other;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Multiply matrix with number and assign
|
|
|
|
|
*
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
* The computation is done column-wise in-place. @f[
|
|
|
|
|
* \boldsymbol A_j = a \boldsymbol A_j
|
|
|
|
|
* @f]
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T>& operator*=(T number) {
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
for(std::size_t i = 0; i != cols; ++i)
|
|
|
|
|
_data[i] *= number;
|
|
|
|
|
|
|
|
|
|
return *this;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Multiply matrix with number
|
|
|
|
|
*
|
|
|
|
|
* @see operator*=(T), operator*(T, const RectangularMatrix<cols, rows, T>&)
|
|
|
|
|
* @todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T> operator*(T number) const {
|
|
|
|
|
return RectangularMatrix<cols, rows, T>(*this) *= number;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Divide matrix with number and assign
|
|
|
|
|
*
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
* The computation is done column-wise in-place. @f[
|
|
|
|
|
* \boldsymbol A_j = \frac{\boldsymbol A_j} a
|
|
|
|
|
* @f]
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T>& operator/=(T number) {
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
for(std::size_t i = 0; i != cols; ++i)
|
|
|
|
|
_data[i] /= number;
|
|
|
|
|
|
|
|
|
|
return *this;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Divide matrix with number
|
|
|
|
|
*
|
|
|
|
|
* @see operator/=(T),
|
|
|
|
|
* operator/(T, const RectangularMatrix<cols, rows, T>&)
|
|
|
|
|
* @todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<cols, rows, T> operator/(T number) const {
|
|
|
|
|
return RectangularMatrix<cols, rows, T>(*this) /= number;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Multiply matrix
|
|
|
|
|
*
|
|
|
|
|
* @f[
|
|
|
|
|
* (\boldsymbol {AB})_{ji} = \sum_{k=0}^{m-1} \boldsymbol A_{ki} \boldsymbol B_{jk}
|
|
|
|
|
* @f]
|
|
|
|
|
*/
|
|
|
|
|
template<std::size_t size> RectangularMatrix<size, rows, T> operator*(const RectangularMatrix<size, cols, T>& other) const;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Multiply vector
|
|
|
|
|
*
|
|
|
|
|
* Internally the same as multiplying with one-column matrix, but
|
|
|
|
|
* returns vector. @f[
|
|
|
|
|
* (\boldsymbol {Aa})_i = \sum_{k=0}^{m-1} \boldsymbol A_{ki} \boldsymbol a_k
|
|
|
|
|
* @f]
|
|
|
|
|
*/
|
|
|
|
|
Vector<rows, T> operator*(const Vector<cols, T>& other) const {
|
|
|
|
|
return operator*(RectangularMatrix<1, cols, T>(other))[0];
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Transposed matrix
|
|
|
|
|
*
|
|
|
|
|
* @see @ref row()
|
|
|
|
|
*/
|
|
|
|
|
RectangularMatrix<rows, cols, T> transposed() const;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Values on diagonal
|
|
|
|
|
*
|
|
|
|
|
* @see @ref fromDiagonal()
|
|
|
|
|
*/
|
|
|
|
|
constexpr Vector<DiagonalSize, T> diagonal() const;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @brief Convert matrix to vector
|
|
|
|
|
*
|
|
|
|
|
* Returns the matrix unrolled into one large vector, i.e. first column
|
|
|
|
|
* of the matrix will make first `rows` elements of resulting vector.
|
|
|
|
|
* Useful for performing vector operations with the matrix (e.g.
|
|
|
|
|
* summing the elements etc.).
|
|
|
|
|
* @see @ref fromVector()
|
|
|
|
|
*/
|
|
|
|
|
Vector<rows*cols, T> toVector() const {
|
|
|
|
|
return *reinterpret_cast<const Vector<rows*cols, T>*>(data());
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifndef DOXYGEN_GENERATING_OUTPUT
|
|
|
|
|
protected:
|
|
|
|
|
#else
|
|
|
|
|
private:
|
|
|
|
|
#endif
|
|
|
|
|
/* Implementation for RectangularMatrix<cols, rows, T>::fromDiagonal() and Matrix<size, T>(T) */
|
|
|
|
|
template<std::size_t ...sequence> constexpr explicit RectangularMatrix(Implementation::Sequence<sequence...>, const Vector<DiagonalSize, T>& diagonal);
|
|
|
|
|
|
|
|
|
|
private:
|
|
|
|
|
/* Implementation for RectangularMatrix<cols, rows, T>::RectangularMatrix(const RectangularMatrix<cols, rows, U>&) */
|
|
|
|
|
template<class U, std::size_t ...sequence> constexpr explicit RectangularMatrix(Implementation::Sequence<sequence...>, const RectangularMatrix<cols, rows, U>& matrix): _data{Vector<rows, T>(matrix[sequence])...} {}
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
|
|
|
|
|
template<std::size_t ...sequence> constexpr Vector<DiagonalSize, T> diagonalInternal(Implementation::Sequence<sequence...>) const;
|
|
|
|
|
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
Vector<rows, T> _data[cols];
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
#ifndef CORRADE_GCC46_COMPATIBILITY
|
|
|
|
|
/**
|
|
|
|
|
@brief %Matrix with 2 columns and 3 rows
|
|
|
|
|
|
|
|
|
|
Convenience alternative to <tt>%RectangularMatrix<2, 3, T></tt>. See
|
|
|
|
|
@ref RectangularMatrix for more information.
|
|
|
|
|
@note Not available on GCC < 4.7. Use <tt>%RectangularMatrix<2, 3, T></tt>
|
|
|
|
|
instead.
|
|
|
|
|
@see @ref Magnum::Matrix2x3, @ref Magnum::Matrix2x3d
|
|
|
|
|
*/
|
|
|
|
|
template<class T> using Matrix2x3 = RectangularMatrix<2, 3, T>;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
@brief %Matrix with 3 columns and 2 rows
|
|
|
|
|
|
|
|
|
|
Convenience alternative to <tt>%RectangularMatrix<3, 2, T></tt>. See
|
|
|
|
|
@ref RectangularMatrix for more information.
|
|
|
|
|
@note Not available on GCC < 4.7. Use <tt>%RectangularMatrix<3, 2, T></tt>
|
|
|
|
|
instead.
|
|
|
|
|
@see @ref Magnum::Matrix3x2, @ref Magnum::Matrix3x2d
|
|
|
|
|
*/
|
|
|
|
|
template<class T> using Matrix3x2 = RectangularMatrix<3, 2, T>;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
@brief %Matrix with 2 columns and 4 rows
|
|
|
|
|
|
|
|
|
|
Convenience alternative to <tt>%RectangularMatrix<2, 4, T></tt>. See
|
|
|
|
|
@ref RectangularMatrix for more information.
|
|
|
|
|
@note Not available on GCC < 4.7. Use <tt>%RectangularMatrix<2, 4, T></tt>
|
|
|
|
|
instead.
|
|
|
|
|
@see @ref Magnum::Matrix2x4, @ref Magnum::Matrix2x4d
|
|
|
|
|
*/
|
|
|
|
|
template<class T> using Matrix2x4 = RectangularMatrix<2, 4, T>;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
@brief %Matrix with 4 columns and 2 rows
|
|
|
|
|
|
|
|
|
|
Convenience alternative to <tt>%RectangularMatrix<4, 2, T></tt>. See
|
|
|
|
|
@ref RectangularMatrix for more information.
|
|
|
|
|
@note Not available on GCC < 4.7. Use <tt>%RectangularMatrix<4, 2, T></tt>
|
|
|
|
|
instead.
|
|
|
|
|
@see @ref Magnum::Matrix4x2, @ref Magnum::Matrix4x2d
|
|
|
|
|
*/
|
|
|
|
|
template<class T> using Matrix4x2 = RectangularMatrix<4, 2, T>;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
@brief %Matrix with 3 columns and 4 rows
|
|
|
|
|
|
|
|
|
|
Convenience alternative to <tt>%RectangularMatrix<3, 4, T></tt>. See
|
|
|
|
|
@ref RectangularMatrix for more information.
|
|
|
|
|
@note Not available on GCC < 4.7. Use <tt>%RectangularMatrix<3, 4, T></tt>
|
|
|
|
|
instead.
|
|
|
|
|
@see @ref Magnum::Matrix3x4, @ref Magnum::Matrix3x4d
|
|
|
|
|
*/
|
|
|
|
|
template<class T> using Matrix3x4 = RectangularMatrix<3, 4, T>;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
@brief %Matrix with 4 columns and 3 rows
|
|
|
|
|
|
|
|
|
|
Convenience alternative to <tt>%RectangularMatrix<4, 3, T></tt>. See
|
|
|
|
|
@ref RectangularMatrix for more information.
|
|
|
|
|
@note Not available on GCC < 4.7. Use <tt>%RectangularMatrix<4, 3, T></tt>
|
|
|
|
|
instead.
|
|
|
|
|
@see @ref Magnum::Matrix4x3, @ref Magnum::Matrix4x3d
|
|
|
|
|
*/
|
|
|
|
|
template<class T> using Matrix4x3 = RectangularMatrix<4, 3, T>;
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/** @relates RectangularMatrix
|
|
|
|
|
@brief Multiply number with matrix
|
|
|
|
|
|
|
|
|
|
Same as RectangularMatrix::operator*(T) const.
|
|
|
|
|
@todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> inline RectangularMatrix<cols, rows, T> operator*(
|
|
|
|
|
#ifdef DOXYGEN_GENERATING_OUTPUT
|
|
|
|
|
T
|
|
|
|
|
#else
|
|
|
|
|
typename std::common_type<T>::type
|
|
|
|
|
#endif
|
|
|
|
|
number, const RectangularMatrix<cols, rows, T>& matrix)
|
|
|
|
|
{
|
|
|
|
|
return matrix*number;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** @relates RectangularMatrix
|
|
|
|
|
@brief Divide matrix with number and invert
|
|
|
|
|
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
The computation is done column-wise. @f[
|
|
|
|
|
\boldsymbol B_j = \frac a {\boldsymbol A_j}
|
|
|
|
|
@f]
|
|
|
|
|
@see RectangularMatrix::operator/(T) const
|
|
|
|
|
@todoc Make explicit reference when Doxygen can handle operators
|
|
|
|
|
*/
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> inline RectangularMatrix<cols, rows, T> operator/(
|
|
|
|
|
#ifdef DOXYGEN_GENERATING_OUTPUT
|
|
|
|
|
T
|
|
|
|
|
#else
|
|
|
|
|
typename std::common_type<T>::type
|
|
|
|
|
#endif
|
|
|
|
|
number, const RectangularMatrix<cols, rows, T>& matrix)
|
|
|
|
|
{
|
|
|
|
|
RectangularMatrix<cols, rows, T> out;
|
|
|
|
|
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
for(std::size_t i = 0; i != cols; ++i)
|
|
|
|
|
out[i] = number/matrix[i];
|
|
|
|
|
|
|
|
|
|
return out;
|
|
|
|
|
}
|
|
|
|
|
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
/** @relates RectangularMatrix
|
|
|
|
|
@brief Multiply vector with rectangular matrix
|
|
|
|
|
|
|
|
|
|
Internally the same as multiplying one-column matrix with one-row matrix. @f[
|
|
|
|
|
(\boldsymbol {aA})_{ji} = \boldsymbol a_i \boldsymbol A_j
|
|
|
|
|
@f]
|
|
|
|
|
@see RectangularMatrix::operator*(const RectangularMatrix<size, cols, T>&) const
|
|
|
|
|
@todoc Make explicit reference when Doxygen can handle operators
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
*/
|
|
|
|
|
template<std::size_t size, std::size_t cols, class T> inline RectangularMatrix<cols, size, T> operator*(const Vector<size, T>& vector, const RectangularMatrix<cols, 1, T>& matrix) {
|
|
|
|
|
return RectangularMatrix<1, size, T>(vector)*matrix;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** @debugoperator{Magnum::Math::RectangularMatrix} */
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> Corrade::Utility::Debug operator<<(Corrade::Utility::Debug debug, const Magnum::Math::RectangularMatrix<cols, rows, T>& value) {
|
|
|
|
|
debug << "Matrix(";
|
|
|
|
|
debug.setFlag(Corrade::Utility::Debug::SpaceAfterEachValue, false);
|
|
|
|
|
for(std::size_t row = 0; row != rows; ++row) {
|
|
|
|
|
if(row != 0) debug << ",\n ";
|
|
|
|
|
for(std::size_t col = 0; col != cols; ++col) {
|
|
|
|
|
if(col != 0) debug << ", ";
|
|
|
|
|
debug << value[col][row];
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
debug << ")";
|
|
|
|
|
debug.setFlag(Corrade::Utility::Debug::SpaceAfterEachValue, true);
|
|
|
|
|
return debug;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifndef DOXYGEN_GENERATING_OUTPUT
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
/* Explicit instantiation for types used in OpenGL */
|
|
|
|
|
/* Square matrices */
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<2, 2, Float>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<3, 3, Float>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<4, 4, Float>&);
|
|
|
|
|
#ifndef MAGNUM_TARGET_GLES
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<2, 2, Double>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<3, 3, Double>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<4, 4, Double>&);
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/* Rectangular matrices */
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<2, 3, Float>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<3, 2, Float>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<2, 4, Float>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<4, 2, Float>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<3, 4, Float>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<4, 3, Float>&);
|
|
|
|
|
#ifndef MAGNUM_TARGET_GLES
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<2, 3, Double>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<3, 2, Double>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<2, 4, Double>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<4, 2, Double>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<3, 4, Double>&);
|
|
|
|
|
extern template Corrade::Utility::Debug MAGNUM_EXPORT operator<<(Corrade::Utility::Debug, const RectangularMatrix<4, 3, Double>&);
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
#define MAGNUM_RECTANGULARMATRIX_SUBCLASS_IMPLEMENTATION(cols, rows, ...) \
|
|
|
|
|
static __VA_ARGS__& from(T* data) { \
|
|
|
|
|
return *reinterpret_cast<__VA_ARGS__*>(data); \
|
|
|
|
|
} \
|
|
|
|
|
static const __VA_ARGS__& from(const T* data) { \
|
|
|
|
|
return *reinterpret_cast<const __VA_ARGS__*>(data); \
|
|
|
|
|
} \
|
|
|
|
|
constexpr static __VA_ARGS__ fromDiagonal(const Vector<Math::RectangularMatrix<cols, rows, T>::DiagonalSize, T>& diagonal) { \
|
|
|
|
|
return Math::RectangularMatrix<cols, rows, T>::fromDiagonal(diagonal); \
|
|
|
|
|
} \
|
|
|
|
|
\
|
|
|
|
|
__VA_ARGS__ operator-() const { \
|
|
|
|
|
return Math::RectangularMatrix<cols, rows, T>::operator-(); \
|
|
|
|
|
} \
|
|
|
|
|
__VA_ARGS__& operator+=(const Math::RectangularMatrix<cols, rows, T>& other) { \
|
|
|
|
|
Math::RectangularMatrix<cols, rows, T>::operator+=(other); \
|
|
|
|
|
return *this; \
|
|
|
|
|
} \
|
|
|
|
|
__VA_ARGS__ operator+(const Math::RectangularMatrix<cols, rows, T>& other) const { \
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
return Math::RectangularMatrix<cols, rows, T>::operator+(other); \
|
|
|
|
|
} \
|
|
|
|
|
__VA_ARGS__& operator-=(const Math::RectangularMatrix<cols, rows, T>& other) { \
|
|
|
|
|
Math::RectangularMatrix<cols, rows, T>::operator-=(other); \
|
|
|
|
|
return *this; \
|
|
|
|
|
} \
|
|
|
|
|
__VA_ARGS__ operator-(const Math::RectangularMatrix<cols, rows, T>& other) const { \
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
return Math::RectangularMatrix<cols, rows, T>::operator-(other); \
|
|
|
|
|
} \
|
|
|
|
|
__VA_ARGS__& operator*=(T number) { \
|
|
|
|
|
Math::RectangularMatrix<cols, rows, T>::operator*=(number); \
|
|
|
|
|
return *this; \
|
|
|
|
|
} \
|
|
|
|
|
__VA_ARGS__ operator*(T number) const { \
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
return Math::RectangularMatrix<cols, rows, T>::operator*(number); \
|
|
|
|
|
} \
|
|
|
|
|
__VA_ARGS__& operator/=(T number) { \
|
|
|
|
|
Math::RectangularMatrix<cols, rows, T>::operator/=(number); \
|
|
|
|
|
return *this; \
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
} \
|
|
|
|
|
__VA_ARGS__ operator/(T number) const { \
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
return Math::RectangularMatrix<cols, rows, T>::operator/(number); \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define MAGNUM_MATRIX_OPERATOR_IMPLEMENTATION(...) \
|
|
|
|
|
template<std::size_t size, class T> inline __VA_ARGS__ operator*(typename std::common_type<T>::type number, const __VA_ARGS__& matrix) { \
|
|
|
|
|
return number*static_cast<const Math::RectangularMatrix<size, size, T>&>(matrix); \
|
|
|
|
|
} \
|
|
|
|
|
template<std::size_t size, class T> inline __VA_ARGS__ operator/(typename std::common_type<T>::type number, const __VA_ARGS__& matrix) { \
|
|
|
|
|
return number/static_cast<const Math::RectangularMatrix<size, size, T>&>(matrix); \
|
|
|
|
|
} \
|
|
|
|
|
template<std::size_t size, class T> inline __VA_ARGS__ operator*(const Vector<size, T>& vector, const RectangularMatrix<size, 1, T>& matrix) { \
|
|
|
|
|
return Math::RectangularMatrix<1, size, T>(vector)*matrix; \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define MAGNUM_MATRIXn_OPERATOR_IMPLEMENTATION(size, Type) \
|
|
|
|
|
template<class T> inline Type<T> operator*(typename std::common_type<T>::type number, const Type<T>& matrix) { \
|
|
|
|
|
return number*static_cast<const Math::RectangularMatrix<size, size, T>&>(matrix); \
|
|
|
|
|
} \
|
|
|
|
|
template<class T> inline Type<T> operator/(typename std::common_type<T>::type number, const Type<T>& matrix) { \
|
|
|
|
|
return number/static_cast<const Math::RectangularMatrix<size, size, T>&>(matrix); \
|
|
|
|
|
} \
|
|
|
|
|
template<class T> inline Type<T> operator*(const Vector<size, T>& vector, const RectangularMatrix<size, 1, T>& matrix) { \
|
|
|
|
|
return Math::RectangularMatrix<1, size, T>(vector)*matrix; \
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
namespace Implementation {
|
|
|
|
|
template<std::size_t rows, std::size_t i, class T, std::size_t ...sequence> inline constexpr Vector<rows, T> diagonalMatrixColumn2(Implementation::Sequence<sequence...>, const T& number) {
|
|
|
|
|
return {(sequence == i ? number : T(0))...};
|
|
|
|
|
}
|
|
|
|
|
template<std::size_t rows, std::size_t i, class T> inline constexpr Vector<rows, T> diagonalMatrixColumn(const T& number) {
|
|
|
|
|
return diagonalMatrixColumn2<rows, i, T>(typename Implementation::GenerateSequence<rows>::Type(), number);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> template<std::size_t ...sequence> inline constexpr RectangularMatrix<cols, rows, T>::RectangularMatrix(Implementation::Sequence<sequence...>, const Vector<DiagonalSize, T>& diagonal): _data{Implementation::diagonalMatrixColumn<rows, sequence>(sequence < DiagonalSize ? diagonal[sequence] : T{})...} {}
|
|
|
|
|
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> inline Vector<cols, T> RectangularMatrix<cols, rows, T>::row(std::size_t row) const {
|
|
|
|
|
Vector<cols, T> out;
|
|
|
|
|
|
|
|
|
|
for(std::size_t i = 0; i != cols; ++i)
|
|
|
|
|
out[i] = _data[i][row];
|
|
|
|
|
|
|
|
|
|
return out;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> inline RectangularMatrix<cols, rows, T> RectangularMatrix<cols, rows, T>::operator-() const {
|
|
|
|
|
RectangularMatrix<cols, rows, T> out;
|
|
|
|
|
|
|
|
|
|
for(std::size_t i = 0; i != cols; ++i)
|
|
|
|
|
out._data[i] = -_data[i];
|
|
|
|
|
|
|
|
|
|
return out;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> template<std::size_t size> inline RectangularMatrix<size, rows, T> RectangularMatrix<cols, rows, T>::operator*(const RectangularMatrix<size, cols, T>& other) const {
|
|
|
|
|
RectangularMatrix<size, rows, T> out;
|
|
|
|
|
|
|
|
|
|
for(std::size_t col = 0; col != size; ++col)
|
|
|
|
|
for(std::size_t row = 0; row != rows; ++row)
|
|
|
|
|
for(std::size_t pos = 0; pos != cols; ++pos)
|
|
|
|
|
out[col][row] += _data[pos][row]*other._data[col][pos];
|
|
|
|
|
|
|
|
|
|
return out;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> inline RectangularMatrix<rows, cols, T> RectangularMatrix<cols, rows, T>::transposed() const {
|
|
|
|
|
RectangularMatrix<rows, cols, T> out;
|
|
|
|
|
|
|
|
|
|
for(std::size_t col = 0; col != cols; ++col)
|
|
|
|
|
for(std::size_t row = 0; row != rows; ++row)
|
|
|
|
|
out[row][col] = _data[col][row];
|
|
|
|
|
|
|
|
|
|
return out;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> inline constexpr auto RectangularMatrix<cols, rows, T>::diagonal() const -> Vector<DiagonalSize, T> { return diagonalInternal(typename Implementation::GenerateSequence<DiagonalSize>::Type()); }
|
|
|
|
|
|
|
|
|
|
#ifndef DOXYGEN_GENERATING_OUTPUT
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> template<std::size_t ...sequence> inline constexpr auto RectangularMatrix<cols, rows, T>::diagonalInternal(Implementation::Sequence<sequence...>) const -> Vector<DiagonalSize, T> {
|
|
|
|
|
return {(*this)[sequence][sequence]...};
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
}}
|
|
|
|
|
|
|
|
|
|
namespace Corrade { namespace Utility {
|
|
|
|
|
|
|
|
|
|
/** @configurationvalue{Magnum::Math::RectangularMatrix} */
|
|
|
|
|
template<std::size_t cols, std::size_t rows, class T> struct ConfigurationValue<Magnum::Math::RectangularMatrix<cols, rows, T>> {
|
|
|
|
|
ConfigurationValue() = delete;
|
|
|
|
|
|
|
|
|
|
/** @brief Writes elements separated with spaces */
|
|
|
|
|
static std::string toString(const Magnum::Math::RectangularMatrix<cols, rows, T>& value, ConfigurationValueFlags flags) {
|
|
|
|
|
std::string output;
|
|
|
|
|
|
|
|
|
|
for(std::size_t row = 0; row != rows; ++row) {
|
|
|
|
|
for(std::size_t col = 0; col != cols; ++col) {
|
|
|
|
|
if(!output.empty()) output += ' ';
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
output += ConfigurationValue<T>::toString(value[col][row], flags);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return output;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/** @brief Reads elements separated with whitespace */
|
|
|
|
|
static Magnum::Math::RectangularMatrix<cols, rows, T> fromString(const std::string& stringValue, ConfigurationValueFlags flags) {
|
|
|
|
|
Magnum::Math::RectangularMatrix<cols, rows, T> result;
|
|
|
|
|
|
|
|
|
|
std::size_t oldpos = 0, pos = std::string::npos, i = 0;
|
|
|
|
|
do {
|
|
|
|
|
pos = stringValue.find(' ', oldpos);
|
|
|
|
|
std::string part = stringValue.substr(oldpos, pos-oldpos);
|
|
|
|
|
|
|
|
|
|
if(!part.empty()) {
|
Math: matrix/vector rework, part 2: matrix as array of column vectors.
Overall architecture is simplififed with this change and also it's not
needed to use reinterpret_cast in matrix internals anymore, thus there
is no need for operator() and [][] works now always as expected without
any risk of GCC misoptimizations.
On the other side, constructing matrix from list of elements is not
possible anymore. You have to specify the elements as list of
column vectors, which might be less convenient to write, but it helps to
distinguish what is column and what is row:
Matrix<2, int> a(1, 2, // before
3, 4);
Matrix<2, int> a(Vector<2, int>(1, 2), // now
Vector<2, int>(3, 4));
For some matrix specializations (i.e. Matrix3 and Matrix4) it is
possible to use list-initialization instead of explicit type
specification:
Matrix<3, int>({1, 2, 3},
{4, 5, 6},
{7, 8, 9});
I didn't yet figure out how to properly implement the general
(constexpr) constructor to also take lists, so it's a bit ugly for now.
Matrix operations are now done column-wise, which should help with
future SIMD implementations, documentation is also updated accordingly.
I also removed forgotten remains of matrix/matrix operator*=(), which
can be confusing, as the multiplication is not commutative. Why it is
not present is explained in d9c900f076f2f87c7b7ba3f37a3179c0c0e4a02c.
13 years ago
|
|
|
result[i%cols][i/cols] = ConfigurationValue<T>::fromString(part, flags);
|
|
|
|
|
++i;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
oldpos = pos+1;
|
|
|
|
|
} while(pos != std::string::npos);
|
|
|
|
|
|
|
|
|
|
return result;
|
|
|
|
|
}
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
#ifndef DOXYGEN_GENERATING_OUTPUT
|
|
|
|
|
/* Square matrices */
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<2, 2, Magnum::Float>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<3, 3, Magnum::Float>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<4, 4, Magnum::Float>>;
|
|
|
|
|
#ifndef MAGNUM_TARGET_GLES
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<2, 2, Magnum::Double>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<3, 3, Magnum::Double>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<4, 4, Magnum::Double>>;
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/* Rectangular matrices */
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<2, 3, Magnum::Float>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<3, 2, Magnum::Float>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<2, 4, Magnum::Float>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<4, 2, Magnum::Float>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<3, 4, Magnum::Float>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<4, 3, Magnum::Float>>;
|
|
|
|
|
#ifndef MAGNUM_TARGET_GLES
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<2, 3, Magnum::Double>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<3, 2, Magnum::Double>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<2, 4, Magnum::Double>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<4, 2, Magnum::Double>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<3, 4, Magnum::Double>>;
|
|
|
|
|
extern template struct MAGNUM_EXPORT ConfigurationValue<Magnum::Math::RectangularMatrix<4, 3, Magnum::Double>>;
|
|
|
|
|
#endif
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
}}
|
|
|
|
|
|
|
|
|
|
#endif
|