1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
|
// This file is part of Eigen, a lightweight C++ template library
// for linear algebra.
//
// Copyright (C) 2009-2010 Gael Guennebaud <gael.guennebaud@inria.fr>
//
// Eigen is free software; you can redistribute it and/or
// modify it under the terms of the GNU Lesser General Public
// License as published by the Free Software Foundation; either
// version 3 of the License, or (at your option) any later version.
//
// Alternatively, you can redistribute it and/or
// modify it under the terms of the GNU General Public License as
// published by the Free Software Foundation; either version 2 of
// the License, or (at your option) any later version.
//
// Eigen is distributed in the hope that it will be useful, but WITHOUT ANY
// WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
// FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License or the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License and a copy of the GNU General Public License along with
// Eigen. If not, see <http://www.gnu.org/licenses/>.
#ifndef EIGEN_SELFCWISEBINARYOP_H
#define EIGEN_SELFCWISEBINARYOP_H
/** \class SelfCwiseBinaryOp
* \ingroup Core_Module
*
* \internal
*
* \brief Internal helper class for optimizing operators like +=, -=
*
* This is a pseudo expression class re-implementing the copyCoeff/copyPacket
* method to directly performs a +=/-= operations in an optimal way. In particular,
* this allows to make sure that the input/output data are loaded only once using
* aligned packet loads.
*
* \sa class SwapWrapper for a similar trick.
*/
template<typename BinaryOp, typename Lhs, typename Rhs>
struct ei_traits<SelfCwiseBinaryOp<BinaryOp,Lhs,Rhs> >
: ei_traits<CwiseBinaryOp<BinaryOp,Lhs,Rhs> >
{
enum {
// Note that it is still a good idea to preserve the DirectAccessBit
// so that assign can correctly align the data.
Flags = ei_traits<CwiseBinaryOp<BinaryOp,Lhs,Rhs> >::Flags | (Lhs::Flags&DirectAccessBit) | (Lhs::Flags&LvalueBit),
OuterStrideAtCompileTime = Lhs::OuterStrideAtCompileTime,
InnerStrideAtCompileTime = Lhs::InnerStrideAtCompileTime
};
};
template<typename BinaryOp, typename Lhs, typename Rhs> class SelfCwiseBinaryOp
: public ei_dense_xpr_base< SelfCwiseBinaryOp<BinaryOp, Lhs, Rhs> >::type
{
public:
typedef typename ei_dense_xpr_base<SelfCwiseBinaryOp>::type Base;
EIGEN_DENSE_PUBLIC_INTERFACE(SelfCwiseBinaryOp)
typedef typename ei_packet_traits<Scalar>::type Packet;
inline SelfCwiseBinaryOp(Lhs& xpr, const BinaryOp& func = BinaryOp()) : m_matrix(xpr), m_functor(func) {}
inline Index rows() const { return m_matrix.rows(); }
inline Index cols() const { return m_matrix.cols(); }
inline Index outerStride() const { return m_matrix.outerStride(); }
inline Index innerStride() const { return m_matrix.innerStride(); }
inline const Scalar* data() const { return m_matrix.data(); }
// note that this function is needed by assign to correctly align loads/stores
// TODO make Assign use .data()
inline Scalar& coeffRef(Index row, Index col)
{
return m_matrix.const_cast_derived().coeffRef(row, col);
}
// note that this function is needed by assign to correctly align loads/stores
// TODO make Assign use .data()
inline Scalar& coeffRef(Index index)
{
return m_matrix.const_cast_derived().coeffRef(index);
}
template<typename OtherDerived>
void copyCoeff(Index row, Index col, const DenseBase<OtherDerived>& other)
{
OtherDerived& _other = other.const_cast_derived();
ei_internal_assert(row >= 0 && row < rows()
&& col >= 0 && col < cols());
Scalar& tmp = m_matrix.coeffRef(row,col);
tmp = m_functor(tmp, _other.coeff(row,col));
}
template<typename OtherDerived>
void copyCoeff(Index index, const DenseBase<OtherDerived>& other)
{
OtherDerived& _other = other.const_cast_derived();
ei_internal_assert(index >= 0 && index < m_matrix.size());
Scalar& tmp = m_matrix.coeffRef(index);
tmp = m_functor(tmp, _other.coeff(index));
}
template<typename OtherDerived, int StoreMode, int LoadMode>
void copyPacket(Index row, Index col, const DenseBase<OtherDerived>& other)
{
OtherDerived& _other = other.const_cast_derived();
ei_internal_assert(row >= 0 && row < rows()
&& col >= 0 && col < cols());
m_matrix.template writePacket<StoreMode>(row, col,
m_functor.packetOp(m_matrix.template packet<StoreMode>(row, col),_other.template packet<LoadMode>(row, col)) );
}
template<typename OtherDerived, int StoreMode, int LoadMode>
void copyPacket(Index index, const DenseBase<OtherDerived>& other)
{
OtherDerived& _other = other.const_cast_derived();
ei_internal_assert(index >= 0 && index < m_matrix.size());
m_matrix.template writePacket<StoreMode>(index,
m_functor.packetOp(m_matrix.template packet<StoreMode>(index),_other.template packet<LoadMode>(index)) );
}
// reimplement lazyAssign to handle complex *= real
// see CwiseBinaryOp ctor for details
template<typename RhsDerived>
EIGEN_STRONG_INLINE SelfCwiseBinaryOp& lazyAssign(const DenseBase<RhsDerived>& rhs)
{
EIGEN_STATIC_ASSERT_SAME_MATRIX_SIZE(Lhs,RhsDerived)
EIGEN_CHECK_BINARY_COMPATIBILIY(BinaryOp,typename Lhs::Scalar,typename RhsDerived::Scalar);
#ifdef EIGEN_DEBUG_ASSIGN
ei_assign_traits<SelfCwiseBinaryOp, RhsDerived>::debug();
#endif
ei_assert(rows() == rhs.rows() && cols() == rhs.cols());
ei_assign_impl<SelfCwiseBinaryOp, RhsDerived>::run(*this,rhs.derived());
#ifndef EIGEN_NO_DEBUG
this->checkTransposeAliasing(rhs.derived());
#endif
return *this;
}
// overloaded to honor evaluation of special matrices
// maybe another solution would be to not use SelfCwiseBinaryOp
// at first...
SelfCwiseBinaryOp& operator=(const Rhs& _rhs)
{
typename ei_nested<Rhs>::type rhs(_rhs);
return Base::operator=(rhs);
}
protected:
Lhs& m_matrix;
const BinaryOp& m_functor;
private:
SelfCwiseBinaryOp& operator=(const SelfCwiseBinaryOp&);
};
template<typename Derived>
inline Derived& DenseBase<Derived>::operator*=(const Scalar& other)
{
typedef typename Derived::PlainObject PlainObject;
SelfCwiseBinaryOp<ei_scalar_product_op<Scalar>, Derived, typename PlainObject::ConstantReturnType> tmp(derived());
tmp = PlainObject::Constant(rows(),cols(),other);
return derived();
}
template<typename Derived>
inline Derived& DenseBase<Derived>::operator/=(const Scalar& other)
{
typedef typename ei_meta_if<NumTraits<Scalar>::IsInteger,
ei_scalar_quotient_op<Scalar>,
ei_scalar_product_op<Scalar> >::ret BinOp;
typedef typename Derived::PlainObject PlainObject;
SelfCwiseBinaryOp<BinOp, Derived, typename PlainObject::ConstantReturnType> tmp(derived());
tmp = PlainObject::Constant(rows(),cols(), NumTraits<Scalar>::IsInteger ? other : Scalar(1)/other);
return derived();
}
#endif // EIGEN_SELFCWISEBINARYOP_H
|
