[136] | 1 | // This file is part of Eigen, a lightweight C++ template library
|
---|
| 2 | // for linear algebra.
|
---|
| 3 | //
|
---|
| 4 | // Copyright (C) 2006-2008 Benoit Jacob <jacob.benoit.1@gmail.com>
|
---|
| 5 | // Copyright (C) 2008-2011 Gael Guennebaud <gael.guennebaud@inria.fr>
|
---|
| 6 | //
|
---|
| 7 | // This Source Code Form is subject to the terms of the Mozilla
|
---|
| 8 | // Public License v. 2.0. If a copy of the MPL was not distributed
|
---|
| 9 | // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
|
---|
| 10 |
|
---|
| 11 | #ifndef EIGEN_GENERAL_PRODUCT_H
|
---|
| 12 | #define EIGEN_GENERAL_PRODUCT_H
|
---|
| 13 |
|
---|
| 14 | namespace Eigen {
|
---|
| 15 |
|
---|
| 16 | /** \class GeneralProduct
|
---|
| 17 | * \ingroup Core_Module
|
---|
| 18 | *
|
---|
| 19 | * \brief Expression of the product of two general matrices or vectors
|
---|
| 20 | *
|
---|
| 21 | * \param LhsNested the type used to store the left-hand side
|
---|
| 22 | * \param RhsNested the type used to store the right-hand side
|
---|
| 23 | * \param ProductMode the type of the product
|
---|
| 24 | *
|
---|
| 25 | * This class represents an expression of the product of two general matrices.
|
---|
| 26 | * We call a general matrix, a dense matrix with full storage. For instance,
|
---|
| 27 | * This excludes triangular, selfadjoint, and sparse matrices.
|
---|
| 28 | * It is the return type of the operator* between general matrices. Its template
|
---|
| 29 | * arguments are determined automatically by ProductReturnType. Therefore,
|
---|
| 30 | * GeneralProduct should never be used direclty. To determine the result type of a
|
---|
| 31 | * function which involves a matrix product, use ProductReturnType::Type.
|
---|
| 32 | *
|
---|
| 33 | * \sa ProductReturnType, MatrixBase::operator*(const MatrixBase<OtherDerived>&)
|
---|
| 34 | */
|
---|
| 35 | template<typename Lhs, typename Rhs, int ProductType = internal::product_type<Lhs,Rhs>::value>
|
---|
| 36 | class GeneralProduct;
|
---|
| 37 |
|
---|
| 38 | enum {
|
---|
| 39 | Large = 2,
|
---|
| 40 | Small = 3
|
---|
| 41 | };
|
---|
| 42 |
|
---|
| 43 | namespace internal {
|
---|
| 44 |
|
---|
| 45 | template<int Rows, int Cols, int Depth> struct product_type_selector;
|
---|
| 46 |
|
---|
| 47 | template<int Size, int MaxSize> struct product_size_category
|
---|
| 48 | {
|
---|
| 49 | enum { is_large = MaxSize == Dynamic ||
|
---|
| 50 | Size >= EIGEN_CACHEFRIENDLY_PRODUCT_THRESHOLD,
|
---|
| 51 | value = is_large ? Large
|
---|
| 52 | : Size == 1 ? 1
|
---|
| 53 | : Small
|
---|
| 54 | };
|
---|
| 55 | };
|
---|
| 56 |
|
---|
| 57 | template<typename Lhs, typename Rhs> struct product_type
|
---|
| 58 | {
|
---|
| 59 | typedef typename remove_all<Lhs>::type _Lhs;
|
---|
| 60 | typedef typename remove_all<Rhs>::type _Rhs;
|
---|
| 61 | enum {
|
---|
| 62 | MaxRows = _Lhs::MaxRowsAtCompileTime,
|
---|
| 63 | Rows = _Lhs::RowsAtCompileTime,
|
---|
| 64 | MaxCols = _Rhs::MaxColsAtCompileTime,
|
---|
| 65 | Cols = _Rhs::ColsAtCompileTime,
|
---|
| 66 | MaxDepth = EIGEN_SIZE_MIN_PREFER_FIXED(_Lhs::MaxColsAtCompileTime,
|
---|
| 67 | _Rhs::MaxRowsAtCompileTime),
|
---|
| 68 | Depth = EIGEN_SIZE_MIN_PREFER_FIXED(_Lhs::ColsAtCompileTime,
|
---|
| 69 | _Rhs::RowsAtCompileTime),
|
---|
| 70 | LargeThreshold = EIGEN_CACHEFRIENDLY_PRODUCT_THRESHOLD
|
---|
| 71 | };
|
---|
| 72 |
|
---|
| 73 | // the splitting into different lines of code here, introducing the _select enums and the typedef below,
|
---|
| 74 | // is to work around an internal compiler error with gcc 4.1 and 4.2.
|
---|
| 75 | private:
|
---|
| 76 | enum {
|
---|
| 77 | rows_select = product_size_category<Rows,MaxRows>::value,
|
---|
| 78 | cols_select = product_size_category<Cols,MaxCols>::value,
|
---|
| 79 | depth_select = product_size_category<Depth,MaxDepth>::value
|
---|
| 80 | };
|
---|
| 81 | typedef product_type_selector<rows_select, cols_select, depth_select> selector;
|
---|
| 82 |
|
---|
| 83 | public:
|
---|
| 84 | enum {
|
---|
| 85 | value = selector::ret
|
---|
| 86 | };
|
---|
| 87 | #ifdef EIGEN_DEBUG_PRODUCT
|
---|
| 88 | static void debug()
|
---|
| 89 | {
|
---|
| 90 | EIGEN_DEBUG_VAR(Rows);
|
---|
| 91 | EIGEN_DEBUG_VAR(Cols);
|
---|
| 92 | EIGEN_DEBUG_VAR(Depth);
|
---|
| 93 | EIGEN_DEBUG_VAR(rows_select);
|
---|
| 94 | EIGEN_DEBUG_VAR(cols_select);
|
---|
| 95 | EIGEN_DEBUG_VAR(depth_select);
|
---|
| 96 | EIGEN_DEBUG_VAR(value);
|
---|
| 97 | }
|
---|
| 98 | #endif
|
---|
| 99 | };
|
---|
| 100 |
|
---|
| 101 |
|
---|
| 102 | /* The following allows to select the kind of product at compile time
|
---|
| 103 | * based on the three dimensions of the product.
|
---|
| 104 | * This is a compile time mapping from {1,Small,Large}^3 -> {product types} */
|
---|
| 105 | // FIXME I'm not sure the current mapping is the ideal one.
|
---|
| 106 | template<int M, int N> struct product_type_selector<M,N,1> { enum { ret = OuterProduct }; };
|
---|
| 107 | template<int Depth> struct product_type_selector<1, 1, Depth> { enum { ret = InnerProduct }; };
|
---|
| 108 | template<> struct product_type_selector<1, 1, 1> { enum { ret = InnerProduct }; };
|
---|
| 109 | template<> struct product_type_selector<Small,1, Small> { enum { ret = CoeffBasedProductMode }; };
|
---|
| 110 | template<> struct product_type_selector<1, Small,Small> { enum { ret = CoeffBasedProductMode }; };
|
---|
| 111 | template<> struct product_type_selector<Small,Small,Small> { enum { ret = CoeffBasedProductMode }; };
|
---|
| 112 | template<> struct product_type_selector<Small, Small, 1> { enum { ret = LazyCoeffBasedProductMode }; };
|
---|
| 113 | template<> struct product_type_selector<Small, Large, 1> { enum { ret = LazyCoeffBasedProductMode }; };
|
---|
| 114 | template<> struct product_type_selector<Large, Small, 1> { enum { ret = LazyCoeffBasedProductMode }; };
|
---|
| 115 | template<> struct product_type_selector<1, Large,Small> { enum { ret = CoeffBasedProductMode }; };
|
---|
| 116 | template<> struct product_type_selector<1, Large,Large> { enum { ret = GemvProduct }; };
|
---|
| 117 | template<> struct product_type_selector<1, Small,Large> { enum { ret = CoeffBasedProductMode }; };
|
---|
| 118 | template<> struct product_type_selector<Large,1, Small> { enum { ret = CoeffBasedProductMode }; };
|
---|
| 119 | template<> struct product_type_selector<Large,1, Large> { enum { ret = GemvProduct }; };
|
---|
| 120 | template<> struct product_type_selector<Small,1, Large> { enum { ret = CoeffBasedProductMode }; };
|
---|
| 121 | template<> struct product_type_selector<Small,Small,Large> { enum { ret = GemmProduct }; };
|
---|
| 122 | template<> struct product_type_selector<Large,Small,Large> { enum { ret = GemmProduct }; };
|
---|
| 123 | template<> struct product_type_selector<Small,Large,Large> { enum { ret = GemmProduct }; };
|
---|
| 124 | template<> struct product_type_selector<Large,Large,Large> { enum { ret = GemmProduct }; };
|
---|
| 125 | template<> struct product_type_selector<Large,Small,Small> { enum { ret = GemmProduct }; };
|
---|
| 126 | template<> struct product_type_selector<Small,Large,Small> { enum { ret = GemmProduct }; };
|
---|
| 127 | template<> struct product_type_selector<Large,Large,Small> { enum { ret = GemmProduct }; };
|
---|
| 128 |
|
---|
| 129 | } // end namespace internal
|
---|
| 130 |
|
---|
| 131 | /** \class ProductReturnType
|
---|
| 132 | * \ingroup Core_Module
|
---|
| 133 | *
|
---|
| 134 | * \brief Helper class to get the correct and optimized returned type of operator*
|
---|
| 135 | *
|
---|
| 136 | * \param Lhs the type of the left-hand side
|
---|
| 137 | * \param Rhs the type of the right-hand side
|
---|
| 138 | * \param ProductMode the type of the product (determined automatically by internal::product_mode)
|
---|
| 139 | *
|
---|
| 140 | * This class defines the typename Type representing the optimized product expression
|
---|
| 141 | * between two matrix expressions. In practice, using ProductReturnType<Lhs,Rhs>::Type
|
---|
| 142 | * is the recommended way to define the result type of a function returning an expression
|
---|
| 143 | * which involve a matrix product. The class Product should never be
|
---|
| 144 | * used directly.
|
---|
| 145 | *
|
---|
| 146 | * \sa class Product, MatrixBase::operator*(const MatrixBase<OtherDerived>&)
|
---|
| 147 | */
|
---|
| 148 | template<typename Lhs, typename Rhs, int ProductType>
|
---|
| 149 | struct ProductReturnType
|
---|
| 150 | {
|
---|
| 151 | // TODO use the nested type to reduce instanciations ????
|
---|
| 152 | // typedef typename internal::nested<Lhs,Rhs::ColsAtCompileTime>::type LhsNested;
|
---|
| 153 | // typedef typename internal::nested<Rhs,Lhs::RowsAtCompileTime>::type RhsNested;
|
---|
| 154 |
|
---|
| 155 | typedef GeneralProduct<Lhs/*Nested*/, Rhs/*Nested*/, ProductType> Type;
|
---|
| 156 | };
|
---|
| 157 |
|
---|
| 158 | template<typename Lhs, typename Rhs>
|
---|
| 159 | struct ProductReturnType<Lhs,Rhs,CoeffBasedProductMode>
|
---|
| 160 | {
|
---|
| 161 | typedef typename internal::nested<Lhs, Rhs::ColsAtCompileTime, typename internal::plain_matrix_type<Lhs>::type >::type LhsNested;
|
---|
| 162 | typedef typename internal::nested<Rhs, Lhs::RowsAtCompileTime, typename internal::plain_matrix_type<Rhs>::type >::type RhsNested;
|
---|
| 163 | typedef CoeffBasedProduct<LhsNested, RhsNested, EvalBeforeAssigningBit | EvalBeforeNestingBit> Type;
|
---|
| 164 | };
|
---|
| 165 |
|
---|
| 166 | template<typename Lhs, typename Rhs>
|
---|
| 167 | struct ProductReturnType<Lhs,Rhs,LazyCoeffBasedProductMode>
|
---|
| 168 | {
|
---|
| 169 | typedef typename internal::nested<Lhs, Rhs::ColsAtCompileTime, typename internal::plain_matrix_type<Lhs>::type >::type LhsNested;
|
---|
| 170 | typedef typename internal::nested<Rhs, Lhs::RowsAtCompileTime, typename internal::plain_matrix_type<Rhs>::type >::type RhsNested;
|
---|
| 171 | typedef CoeffBasedProduct<LhsNested, RhsNested, NestByRefBit> Type;
|
---|
| 172 | };
|
---|
| 173 |
|
---|
| 174 | // this is a workaround for sun CC
|
---|
| 175 | template<typename Lhs, typename Rhs>
|
---|
| 176 | struct LazyProductReturnType : public ProductReturnType<Lhs,Rhs,LazyCoeffBasedProductMode>
|
---|
| 177 | {};
|
---|
| 178 |
|
---|
| 179 | /***********************************************************************
|
---|
| 180 | * Implementation of Inner Vector Vector Product
|
---|
| 181 | ***********************************************************************/
|
---|
| 182 |
|
---|
| 183 | // FIXME : maybe the "inner product" could return a Scalar
|
---|
| 184 | // instead of a 1x1 matrix ??
|
---|
| 185 | // Pro: more natural for the user
|
---|
| 186 | // Cons: this could be a problem if in a meta unrolled algorithm a matrix-matrix
|
---|
| 187 | // product ends up to a row-vector times col-vector product... To tackle this use
|
---|
| 188 | // case, we could have a specialization for Block<MatrixType,1,1> with: operator=(Scalar x);
|
---|
| 189 |
|
---|
| 190 | namespace internal {
|
---|
| 191 |
|
---|
| 192 | template<typename Lhs, typename Rhs>
|
---|
| 193 | struct traits<GeneralProduct<Lhs,Rhs,InnerProduct> >
|
---|
| 194 | : traits<Matrix<typename scalar_product_traits<typename Lhs::Scalar, typename Rhs::Scalar>::ReturnType,1,1> >
|
---|
| 195 | {};
|
---|
| 196 |
|
---|
| 197 | }
|
---|
| 198 |
|
---|
| 199 | template<typename Lhs, typename Rhs>
|
---|
| 200 | class GeneralProduct<Lhs, Rhs, InnerProduct>
|
---|
| 201 | : internal::no_assignment_operator,
|
---|
| 202 | public Matrix<typename internal::scalar_product_traits<typename Lhs::Scalar, typename Rhs::Scalar>::ReturnType,1,1>
|
---|
| 203 | {
|
---|
| 204 | typedef Matrix<typename internal::scalar_product_traits<typename Lhs::Scalar, typename Rhs::Scalar>::ReturnType,1,1> Base;
|
---|
| 205 | public:
|
---|
| 206 | GeneralProduct(const Lhs& lhs, const Rhs& rhs)
|
---|
| 207 | {
|
---|
| 208 | Base::coeffRef(0,0) = (lhs.transpose().cwiseProduct(rhs)).sum();
|
---|
| 209 | }
|
---|
| 210 |
|
---|
| 211 | /** Convertion to scalar */
|
---|
| 212 | operator const typename Base::Scalar() const {
|
---|
| 213 | return Base::coeff(0,0);
|
---|
| 214 | }
|
---|
| 215 | };
|
---|
| 216 |
|
---|
| 217 | /***********************************************************************
|
---|
| 218 | * Implementation of Outer Vector Vector Product
|
---|
| 219 | ***********************************************************************/
|
---|
| 220 |
|
---|
| 221 | namespace internal {
|
---|
| 222 |
|
---|
| 223 | // Column major
|
---|
| 224 | template<typename ProductType, typename Dest, typename Func>
|
---|
| 225 | EIGEN_DONT_INLINE void outer_product_selector_run(const ProductType& prod, Dest& dest, const Func& func, const false_type&)
|
---|
| 226 | {
|
---|
| 227 | typedef typename Dest::Index Index;
|
---|
| 228 | // FIXME make sure lhs is sequentially stored
|
---|
| 229 | // FIXME not very good if rhs is real and lhs complex while alpha is real too
|
---|
| 230 | const Index cols = dest.cols();
|
---|
| 231 | for (Index j=0; j<cols; ++j)
|
---|
| 232 | func(dest.col(j), prod.rhs().coeff(0,j) * prod.lhs());
|
---|
| 233 | }
|
---|
| 234 |
|
---|
| 235 | // Row major
|
---|
| 236 | template<typename ProductType, typename Dest, typename Func>
|
---|
| 237 | EIGEN_DONT_INLINE void outer_product_selector_run(const ProductType& prod, Dest& dest, const Func& func, const true_type&) {
|
---|
| 238 | typedef typename Dest::Index Index;
|
---|
| 239 | // FIXME make sure rhs is sequentially stored
|
---|
| 240 | // FIXME not very good if lhs is real and rhs complex while alpha is real too
|
---|
| 241 | const Index rows = dest.rows();
|
---|
| 242 | for (Index i=0; i<rows; ++i)
|
---|
| 243 | func(dest.row(i), prod.lhs().coeff(i,0) * prod.rhs());
|
---|
| 244 | }
|
---|
| 245 |
|
---|
| 246 | template<typename Lhs, typename Rhs>
|
---|
| 247 | struct traits<GeneralProduct<Lhs,Rhs,OuterProduct> >
|
---|
| 248 | : traits<ProductBase<GeneralProduct<Lhs,Rhs,OuterProduct>, Lhs, Rhs> >
|
---|
| 249 | {};
|
---|
| 250 |
|
---|
| 251 | }
|
---|
| 252 |
|
---|
| 253 | template<typename Lhs, typename Rhs>
|
---|
| 254 | class GeneralProduct<Lhs, Rhs, OuterProduct>
|
---|
| 255 | : public ProductBase<GeneralProduct<Lhs,Rhs,OuterProduct>, Lhs, Rhs>
|
---|
| 256 | {
|
---|
| 257 | template<typename T> struct is_row_major : internal::conditional<(int(T::Flags)&RowMajorBit), internal::true_type, internal::false_type>::type {};
|
---|
| 258 |
|
---|
| 259 | public:
|
---|
| 260 | EIGEN_PRODUCT_PUBLIC_INTERFACE(GeneralProduct)
|
---|
| 261 |
|
---|
| 262 | GeneralProduct(const Lhs& lhs, const Rhs& rhs) : Base(lhs,rhs)
|
---|
| 263 | {
|
---|
| 264 | }
|
---|
| 265 |
|
---|
| 266 | struct set { template<typename Dst, typename Src> void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived() = src; } };
|
---|
| 267 | struct add { template<typename Dst, typename Src> void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived() += src; } };
|
---|
| 268 | struct sub { template<typename Dst, typename Src> void operator()(const Dst& dst, const Src& src) const { dst.const_cast_derived() -= src; } };
|
---|
| 269 | struct adds {
|
---|
| 270 | Scalar m_scale;
|
---|
| 271 | adds(const Scalar& s) : m_scale(s) {}
|
---|
| 272 | template<typename Dst, typename Src> void operator()(const Dst& dst, const Src& src) const {
|
---|
| 273 | dst.const_cast_derived() += m_scale * src;
|
---|
| 274 | }
|
---|
| 275 | };
|
---|
| 276 |
|
---|
| 277 | template<typename Dest>
|
---|
| 278 | inline void evalTo(Dest& dest) const {
|
---|
| 279 | internal::outer_product_selector_run(*this, dest, set(), is_row_major<Dest>());
|
---|
| 280 | }
|
---|
| 281 |
|
---|
| 282 | template<typename Dest>
|
---|
| 283 | inline void addTo(Dest& dest) const {
|
---|
| 284 | internal::outer_product_selector_run(*this, dest, add(), is_row_major<Dest>());
|
---|
| 285 | }
|
---|
| 286 |
|
---|
| 287 | template<typename Dest>
|
---|
| 288 | inline void subTo(Dest& dest) const {
|
---|
| 289 | internal::outer_product_selector_run(*this, dest, sub(), is_row_major<Dest>());
|
---|
| 290 | }
|
---|
| 291 |
|
---|
| 292 | template<typename Dest> void scaleAndAddTo(Dest& dest, const Scalar& alpha) const
|
---|
| 293 | {
|
---|
| 294 | internal::outer_product_selector_run(*this, dest, adds(alpha), is_row_major<Dest>());
|
---|
| 295 | }
|
---|
| 296 | };
|
---|
| 297 |
|
---|
| 298 | /***********************************************************************
|
---|
| 299 | * Implementation of General Matrix Vector Product
|
---|
| 300 | ***********************************************************************/
|
---|
| 301 |
|
---|
| 302 | /* According to the shape/flags of the matrix we have to distinghish 3 different cases:
|
---|
| 303 | * 1 - the matrix is col-major, BLAS compatible and M is large => call fast BLAS-like colmajor routine
|
---|
| 304 | * 2 - the matrix is row-major, BLAS compatible and N is large => call fast BLAS-like rowmajor routine
|
---|
| 305 | * 3 - all other cases are handled using a simple loop along the outer-storage direction.
|
---|
| 306 | * Therefore we need a lower level meta selector.
|
---|
| 307 | * Furthermore, if the matrix is the rhs, then the product has to be transposed.
|
---|
| 308 | */
|
---|
| 309 | namespace internal {
|
---|
| 310 |
|
---|
| 311 | template<typename Lhs, typename Rhs>
|
---|
| 312 | struct traits<GeneralProduct<Lhs,Rhs,GemvProduct> >
|
---|
| 313 | : traits<ProductBase<GeneralProduct<Lhs,Rhs,GemvProduct>, Lhs, Rhs> >
|
---|
| 314 | {};
|
---|
| 315 |
|
---|
| 316 | template<int Side, int StorageOrder, bool BlasCompatible>
|
---|
| 317 | struct gemv_selector;
|
---|
| 318 |
|
---|
| 319 | } // end namespace internal
|
---|
| 320 |
|
---|
| 321 | template<typename Lhs, typename Rhs>
|
---|
| 322 | class GeneralProduct<Lhs, Rhs, GemvProduct>
|
---|
| 323 | : public ProductBase<GeneralProduct<Lhs,Rhs,GemvProduct>, Lhs, Rhs>
|
---|
| 324 | {
|
---|
| 325 | public:
|
---|
| 326 | EIGEN_PRODUCT_PUBLIC_INTERFACE(GeneralProduct)
|
---|
| 327 |
|
---|
| 328 | typedef typename Lhs::Scalar LhsScalar;
|
---|
| 329 | typedef typename Rhs::Scalar RhsScalar;
|
---|
| 330 |
|
---|
| 331 | GeneralProduct(const Lhs& a_lhs, const Rhs& a_rhs) : Base(a_lhs,a_rhs)
|
---|
| 332 | {
|
---|
| 333 | // EIGEN_STATIC_ASSERT((internal::is_same<typename Lhs::Scalar, typename Rhs::Scalar>::value),
|
---|
| 334 | // YOU_MIXED_DIFFERENT_NUMERIC_TYPES__YOU_NEED_TO_USE_THE_CAST_METHOD_OF_MATRIXBASE_TO_CAST_NUMERIC_TYPES_EXPLICITLY)
|
---|
| 335 | }
|
---|
| 336 |
|
---|
| 337 | enum { Side = Lhs::IsVectorAtCompileTime ? OnTheLeft : OnTheRight };
|
---|
| 338 | typedef typename internal::conditional<int(Side)==OnTheRight,_LhsNested,_RhsNested>::type MatrixType;
|
---|
| 339 |
|
---|
| 340 | template<typename Dest> void scaleAndAddTo(Dest& dst, const Scalar& alpha) const
|
---|
| 341 | {
|
---|
| 342 | eigen_assert(m_lhs.rows() == dst.rows() && m_rhs.cols() == dst.cols());
|
---|
| 343 | internal::gemv_selector<Side,(int(MatrixType::Flags)&RowMajorBit) ? RowMajor : ColMajor,
|
---|
| 344 | bool(internal::blas_traits<MatrixType>::HasUsableDirectAccess)>::run(*this, dst, alpha);
|
---|
| 345 | }
|
---|
| 346 | };
|
---|
| 347 |
|
---|
| 348 | namespace internal {
|
---|
| 349 |
|
---|
| 350 | // The vector is on the left => transposition
|
---|
| 351 | template<int StorageOrder, bool BlasCompatible>
|
---|
| 352 | struct gemv_selector<OnTheLeft,StorageOrder,BlasCompatible>
|
---|
| 353 | {
|
---|
| 354 | template<typename ProductType, typename Dest>
|
---|
| 355 | static void run(const ProductType& prod, Dest& dest, const typename ProductType::Scalar& alpha)
|
---|
| 356 | {
|
---|
| 357 | Transpose<Dest> destT(dest);
|
---|
| 358 | enum { OtherStorageOrder = StorageOrder == RowMajor ? ColMajor : RowMajor };
|
---|
| 359 | gemv_selector<OnTheRight,OtherStorageOrder,BlasCompatible>
|
---|
| 360 | ::run(GeneralProduct<Transpose<const typename ProductType::_RhsNested>,Transpose<const typename ProductType::_LhsNested>, GemvProduct>
|
---|
| 361 | (prod.rhs().transpose(), prod.lhs().transpose()), destT, alpha);
|
---|
| 362 | }
|
---|
| 363 | };
|
---|
| 364 |
|
---|
| 365 | template<typename Scalar,int Size,int MaxSize,bool Cond> struct gemv_static_vector_if;
|
---|
| 366 |
|
---|
| 367 | template<typename Scalar,int Size,int MaxSize>
|
---|
| 368 | struct gemv_static_vector_if<Scalar,Size,MaxSize,false>
|
---|
| 369 | {
|
---|
| 370 | EIGEN_STRONG_INLINE Scalar* data() { eigen_internal_assert(false && "should never be called"); return 0; }
|
---|
| 371 | };
|
---|
| 372 |
|
---|
| 373 | template<typename Scalar,int Size>
|
---|
| 374 | struct gemv_static_vector_if<Scalar,Size,Dynamic,true>
|
---|
| 375 | {
|
---|
| 376 | EIGEN_STRONG_INLINE Scalar* data() { return 0; }
|
---|
| 377 | };
|
---|
| 378 |
|
---|
| 379 | template<typename Scalar,int Size,int MaxSize>
|
---|
| 380 | struct gemv_static_vector_if<Scalar,Size,MaxSize,true>
|
---|
| 381 | {
|
---|
| 382 | #if EIGEN_ALIGN_STATICALLY
|
---|
| 383 | internal::plain_array<Scalar,EIGEN_SIZE_MIN_PREFER_FIXED(Size,MaxSize),0> m_data;
|
---|
| 384 | EIGEN_STRONG_INLINE Scalar* data() { return m_data.array; }
|
---|
| 385 | #else
|
---|
| 386 | // Some architectures cannot align on the stack,
|
---|
| 387 | // => let's manually enforce alignment by allocating more data and return the address of the first aligned element.
|
---|
| 388 | enum {
|
---|
| 389 | ForceAlignment = internal::packet_traits<Scalar>::Vectorizable,
|
---|
| 390 | PacketSize = internal::packet_traits<Scalar>::size
|
---|
| 391 | };
|
---|
| 392 | internal::plain_array<Scalar,EIGEN_SIZE_MIN_PREFER_FIXED(Size,MaxSize)+(ForceAlignment?PacketSize:0),0> m_data;
|
---|
| 393 | EIGEN_STRONG_INLINE Scalar* data() {
|
---|
| 394 | return ForceAlignment
|
---|
| 395 | ? reinterpret_cast<Scalar*>((reinterpret_cast<size_t>(m_data.array) & ~(size_t(15))) + 16)
|
---|
| 396 | : m_data.array;
|
---|
| 397 | }
|
---|
| 398 | #endif
|
---|
| 399 | };
|
---|
| 400 |
|
---|
| 401 | template<> struct gemv_selector<OnTheRight,ColMajor,true>
|
---|
| 402 | {
|
---|
| 403 | template<typename ProductType, typename Dest>
|
---|
| 404 | static inline void run(const ProductType& prod, Dest& dest, const typename ProductType::Scalar& alpha)
|
---|
| 405 | {
|
---|
| 406 | typedef typename ProductType::Index Index;
|
---|
| 407 | typedef typename ProductType::LhsScalar LhsScalar;
|
---|
| 408 | typedef typename ProductType::RhsScalar RhsScalar;
|
---|
| 409 | typedef typename ProductType::Scalar ResScalar;
|
---|
| 410 | typedef typename ProductType::RealScalar RealScalar;
|
---|
| 411 | typedef typename ProductType::ActualLhsType ActualLhsType;
|
---|
| 412 | typedef typename ProductType::ActualRhsType ActualRhsType;
|
---|
| 413 | typedef typename ProductType::LhsBlasTraits LhsBlasTraits;
|
---|
| 414 | typedef typename ProductType::RhsBlasTraits RhsBlasTraits;
|
---|
| 415 | typedef Map<Matrix<ResScalar,Dynamic,1>, Aligned> MappedDest;
|
---|
| 416 |
|
---|
| 417 | ActualLhsType actualLhs = LhsBlasTraits::extract(prod.lhs());
|
---|
| 418 | ActualRhsType actualRhs = RhsBlasTraits::extract(prod.rhs());
|
---|
| 419 |
|
---|
| 420 | ResScalar actualAlpha = alpha * LhsBlasTraits::extractScalarFactor(prod.lhs())
|
---|
| 421 | * RhsBlasTraits::extractScalarFactor(prod.rhs());
|
---|
| 422 |
|
---|
| 423 | // make sure Dest is a compile-time vector type (bug 1166)
|
---|
| 424 | typedef typename conditional<Dest::IsVectorAtCompileTime, Dest, typename Dest::ColXpr>::type ActualDest;
|
---|
| 425 |
|
---|
| 426 | enum {
|
---|
| 427 | // FIXME find a way to allow an inner stride on the result if packet_traits<Scalar>::size==1
|
---|
| 428 | // on, the other hand it is good for the cache to pack the vector anyways...
|
---|
| 429 | EvalToDestAtCompileTime = (ActualDest::InnerStrideAtCompileTime==1),
|
---|
| 430 | ComplexByReal = (NumTraits<LhsScalar>::IsComplex) && (!NumTraits<RhsScalar>::IsComplex),
|
---|
| 431 | MightCannotUseDest = (ActualDest::InnerStrideAtCompileTime!=1) || ComplexByReal
|
---|
| 432 | };
|
---|
| 433 |
|
---|
| 434 | gemv_static_vector_if<ResScalar,ActualDest::SizeAtCompileTime,ActualDest::MaxSizeAtCompileTime,MightCannotUseDest> static_dest;
|
---|
| 435 |
|
---|
| 436 | bool alphaIsCompatible = (!ComplexByReal) || (numext::imag(actualAlpha)==RealScalar(0));
|
---|
| 437 | bool evalToDest = EvalToDestAtCompileTime && alphaIsCompatible;
|
---|
| 438 |
|
---|
| 439 | RhsScalar compatibleAlpha = get_factor<ResScalar,RhsScalar>::run(actualAlpha);
|
---|
| 440 |
|
---|
| 441 | ei_declare_aligned_stack_constructed_variable(ResScalar,actualDestPtr,dest.size(),
|
---|
| 442 | evalToDest ? dest.data() : static_dest.data());
|
---|
| 443 |
|
---|
| 444 | if(!evalToDest)
|
---|
| 445 | {
|
---|
| 446 | #ifdef EIGEN_DENSE_STORAGE_CTOR_PLUGIN
|
---|
| 447 | int size = dest.size();
|
---|
| 448 | EIGEN_DENSE_STORAGE_CTOR_PLUGIN
|
---|
| 449 | #endif
|
---|
| 450 | if(!alphaIsCompatible)
|
---|
| 451 | {
|
---|
| 452 | MappedDest(actualDestPtr, dest.size()).setZero();
|
---|
| 453 | compatibleAlpha = RhsScalar(1);
|
---|
| 454 | }
|
---|
| 455 | else
|
---|
| 456 | MappedDest(actualDestPtr, dest.size()) = dest;
|
---|
| 457 | }
|
---|
| 458 |
|
---|
| 459 | general_matrix_vector_product
|
---|
| 460 | <Index,LhsScalar,ColMajor,LhsBlasTraits::NeedToConjugate,RhsScalar,RhsBlasTraits::NeedToConjugate>::run(
|
---|
| 461 | actualLhs.rows(), actualLhs.cols(),
|
---|
| 462 | actualLhs.data(), actualLhs.outerStride(),
|
---|
| 463 | actualRhs.data(), actualRhs.innerStride(),
|
---|
| 464 | actualDestPtr, 1,
|
---|
| 465 | compatibleAlpha);
|
---|
| 466 |
|
---|
| 467 | if (!evalToDest)
|
---|
| 468 | {
|
---|
| 469 | if(!alphaIsCompatible)
|
---|
| 470 | dest += actualAlpha * MappedDest(actualDestPtr, dest.size());
|
---|
| 471 | else
|
---|
| 472 | dest = MappedDest(actualDestPtr, dest.size());
|
---|
| 473 | }
|
---|
| 474 | }
|
---|
| 475 | };
|
---|
| 476 |
|
---|
| 477 | template<> struct gemv_selector<OnTheRight,RowMajor,true>
|
---|
| 478 | {
|
---|
| 479 | template<typename ProductType, typename Dest>
|
---|
| 480 | static void run(const ProductType& prod, Dest& dest, const typename ProductType::Scalar& alpha)
|
---|
| 481 | {
|
---|
| 482 | typedef typename ProductType::LhsScalar LhsScalar;
|
---|
| 483 | typedef typename ProductType::RhsScalar RhsScalar;
|
---|
| 484 | typedef typename ProductType::Scalar ResScalar;
|
---|
| 485 | typedef typename ProductType::Index Index;
|
---|
| 486 | typedef typename ProductType::ActualLhsType ActualLhsType;
|
---|
| 487 | typedef typename ProductType::ActualRhsType ActualRhsType;
|
---|
| 488 | typedef typename ProductType::_ActualRhsType _ActualRhsType;
|
---|
| 489 | typedef typename ProductType::LhsBlasTraits LhsBlasTraits;
|
---|
| 490 | typedef typename ProductType::RhsBlasTraits RhsBlasTraits;
|
---|
| 491 |
|
---|
| 492 | typename add_const<ActualLhsType>::type actualLhs = LhsBlasTraits::extract(prod.lhs());
|
---|
| 493 | typename add_const<ActualRhsType>::type actualRhs = RhsBlasTraits::extract(prod.rhs());
|
---|
| 494 |
|
---|
| 495 | ResScalar actualAlpha = alpha * LhsBlasTraits::extractScalarFactor(prod.lhs())
|
---|
| 496 | * RhsBlasTraits::extractScalarFactor(prod.rhs());
|
---|
| 497 |
|
---|
| 498 | enum {
|
---|
| 499 | // FIXME find a way to allow an inner stride on the result if packet_traits<Scalar>::size==1
|
---|
| 500 | // on, the other hand it is good for the cache to pack the vector anyways...
|
---|
| 501 | DirectlyUseRhs = _ActualRhsType::InnerStrideAtCompileTime==1
|
---|
| 502 | };
|
---|
| 503 |
|
---|
| 504 | gemv_static_vector_if<RhsScalar,_ActualRhsType::SizeAtCompileTime,_ActualRhsType::MaxSizeAtCompileTime,!DirectlyUseRhs> static_rhs;
|
---|
| 505 |
|
---|
| 506 | ei_declare_aligned_stack_constructed_variable(RhsScalar,actualRhsPtr,actualRhs.size(),
|
---|
| 507 | DirectlyUseRhs ? const_cast<RhsScalar*>(actualRhs.data()) : static_rhs.data());
|
---|
| 508 |
|
---|
| 509 | if(!DirectlyUseRhs)
|
---|
| 510 | {
|
---|
| 511 | #ifdef EIGEN_DENSE_STORAGE_CTOR_PLUGIN
|
---|
| 512 | int size = actualRhs.size();
|
---|
| 513 | EIGEN_DENSE_STORAGE_CTOR_PLUGIN
|
---|
| 514 | #endif
|
---|
| 515 | Map<typename _ActualRhsType::PlainObject>(actualRhsPtr, actualRhs.size()) = actualRhs;
|
---|
| 516 | }
|
---|
| 517 |
|
---|
| 518 | general_matrix_vector_product
|
---|
| 519 | <Index,LhsScalar,RowMajor,LhsBlasTraits::NeedToConjugate,RhsScalar,RhsBlasTraits::NeedToConjugate>::run(
|
---|
| 520 | actualLhs.rows(), actualLhs.cols(),
|
---|
| 521 | actualLhs.data(), actualLhs.outerStride(),
|
---|
| 522 | actualRhsPtr, 1,
|
---|
| 523 | dest.data(), dest.col(0).innerStride(), //NOTE if dest is not a vector at compile-time, then dest.innerStride() might be wrong. (bug 1166)
|
---|
| 524 | actualAlpha);
|
---|
| 525 | }
|
---|
| 526 | };
|
---|
| 527 |
|
---|
| 528 | template<> struct gemv_selector<OnTheRight,ColMajor,false>
|
---|
| 529 | {
|
---|
| 530 | template<typename ProductType, typename Dest>
|
---|
| 531 | static void run(const ProductType& prod, Dest& dest, const typename ProductType::Scalar& alpha)
|
---|
| 532 | {
|
---|
| 533 | typedef typename Dest::Index Index;
|
---|
| 534 | // TODO makes sure dest is sequentially stored in memory, otherwise use a temp
|
---|
| 535 | const Index size = prod.rhs().rows();
|
---|
| 536 | for(Index k=0; k<size; ++k)
|
---|
| 537 | dest += (alpha*prod.rhs().coeff(k)) * prod.lhs().col(k);
|
---|
| 538 | }
|
---|
| 539 | };
|
---|
| 540 |
|
---|
| 541 | template<> struct gemv_selector<OnTheRight,RowMajor,false>
|
---|
| 542 | {
|
---|
| 543 | template<typename ProductType, typename Dest>
|
---|
| 544 | static void run(const ProductType& prod, Dest& dest, const typename ProductType::Scalar& alpha)
|
---|
| 545 | {
|
---|
| 546 | typedef typename Dest::Index Index;
|
---|
| 547 | // TODO makes sure rhs is sequentially stored in memory, otherwise use a temp
|
---|
| 548 | const Index rows = prod.rows();
|
---|
| 549 | for(Index i=0; i<rows; ++i)
|
---|
| 550 | dest.coeffRef(i) += alpha * (prod.lhs().row(i).cwiseProduct(prod.rhs().transpose())).sum();
|
---|
| 551 | }
|
---|
| 552 | };
|
---|
| 553 |
|
---|
| 554 | } // end namespace internal
|
---|
| 555 |
|
---|
| 556 | /***************************************************************************
|
---|
| 557 | * Implementation of matrix base methods
|
---|
| 558 | ***************************************************************************/
|
---|
| 559 |
|
---|
| 560 | /** \returns the matrix product of \c *this and \a other.
|
---|
| 561 | *
|
---|
| 562 | * \note If instead of the matrix product you want the coefficient-wise product, see Cwise::operator*().
|
---|
| 563 | *
|
---|
| 564 | * \sa lazyProduct(), operator*=(const MatrixBase&), Cwise::operator*()
|
---|
| 565 | */
|
---|
| 566 | template<typename Derived>
|
---|
| 567 | template<typename OtherDerived>
|
---|
| 568 | inline const typename ProductReturnType<Derived, OtherDerived>::Type
|
---|
| 569 | MatrixBase<Derived>::operator*(const MatrixBase<OtherDerived> &other) const
|
---|
| 570 | {
|
---|
| 571 | // A note regarding the function declaration: In MSVC, this function will sometimes
|
---|
| 572 | // not be inlined since DenseStorage is an unwindable object for dynamic
|
---|
| 573 | // matrices and product types are holding a member to store the result.
|
---|
| 574 | // Thus it does not help tagging this function with EIGEN_STRONG_INLINE.
|
---|
| 575 | enum {
|
---|
| 576 | ProductIsValid = Derived::ColsAtCompileTime==Dynamic
|
---|
| 577 | || OtherDerived::RowsAtCompileTime==Dynamic
|
---|
| 578 | || int(Derived::ColsAtCompileTime)==int(OtherDerived::RowsAtCompileTime),
|
---|
| 579 | AreVectors = Derived::IsVectorAtCompileTime && OtherDerived::IsVectorAtCompileTime,
|
---|
| 580 | SameSizes = EIGEN_PREDICATE_SAME_MATRIX_SIZE(Derived,OtherDerived)
|
---|
| 581 | };
|
---|
| 582 | // note to the lost user:
|
---|
| 583 | // * for a dot product use: v1.dot(v2)
|
---|
| 584 | // * for a coeff-wise product use: v1.cwiseProduct(v2)
|
---|
| 585 | EIGEN_STATIC_ASSERT(ProductIsValid || !(AreVectors && SameSizes),
|
---|
| 586 | INVALID_VECTOR_VECTOR_PRODUCT__IF_YOU_WANTED_A_DOT_OR_COEFF_WISE_PRODUCT_YOU_MUST_USE_THE_EXPLICIT_FUNCTIONS)
|
---|
| 587 | EIGEN_STATIC_ASSERT(ProductIsValid || !(SameSizes && !AreVectors),
|
---|
| 588 | INVALID_MATRIX_PRODUCT__IF_YOU_WANTED_A_COEFF_WISE_PRODUCT_YOU_MUST_USE_THE_EXPLICIT_FUNCTION)
|
---|
| 589 | EIGEN_STATIC_ASSERT(ProductIsValid || SameSizes, INVALID_MATRIX_PRODUCT)
|
---|
| 590 | #ifdef EIGEN_DEBUG_PRODUCT
|
---|
| 591 | internal::product_type<Derived,OtherDerived>::debug();
|
---|
| 592 | #endif
|
---|
| 593 | return typename ProductReturnType<Derived,OtherDerived>::Type(derived(), other.derived());
|
---|
| 594 | }
|
---|
| 595 |
|
---|
| 596 | /** \returns an expression of the matrix product of \c *this and \a other without implicit evaluation.
|
---|
| 597 | *
|
---|
| 598 | * The returned product will behave like any other expressions: the coefficients of the product will be
|
---|
| 599 | * computed once at a time as requested. This might be useful in some extremely rare cases when only
|
---|
| 600 | * a small and no coherent fraction of the result's coefficients have to be computed.
|
---|
| 601 | *
|
---|
| 602 | * \warning This version of the matrix product can be much much slower. So use it only if you know
|
---|
| 603 | * what you are doing and that you measured a true speed improvement.
|
---|
| 604 | *
|
---|
| 605 | * \sa operator*(const MatrixBase&)
|
---|
| 606 | */
|
---|
| 607 | template<typename Derived>
|
---|
| 608 | template<typename OtherDerived>
|
---|
| 609 | const typename LazyProductReturnType<Derived,OtherDerived>::Type
|
---|
| 610 | MatrixBase<Derived>::lazyProduct(const MatrixBase<OtherDerived> &other) const
|
---|
| 611 | {
|
---|
| 612 | enum {
|
---|
| 613 | ProductIsValid = Derived::ColsAtCompileTime==Dynamic
|
---|
| 614 | || OtherDerived::RowsAtCompileTime==Dynamic
|
---|
| 615 | || int(Derived::ColsAtCompileTime)==int(OtherDerived::RowsAtCompileTime),
|
---|
| 616 | AreVectors = Derived::IsVectorAtCompileTime && OtherDerived::IsVectorAtCompileTime,
|
---|
| 617 | SameSizes = EIGEN_PREDICATE_SAME_MATRIX_SIZE(Derived,OtherDerived)
|
---|
| 618 | };
|
---|
| 619 | // note to the lost user:
|
---|
| 620 | // * for a dot product use: v1.dot(v2)
|
---|
| 621 | // * for a coeff-wise product use: v1.cwiseProduct(v2)
|
---|
| 622 | EIGEN_STATIC_ASSERT(ProductIsValid || !(AreVectors && SameSizes),
|
---|
| 623 | INVALID_VECTOR_VECTOR_PRODUCT__IF_YOU_WANTED_A_DOT_OR_COEFF_WISE_PRODUCT_YOU_MUST_USE_THE_EXPLICIT_FUNCTIONS)
|
---|
| 624 | EIGEN_STATIC_ASSERT(ProductIsValid || !(SameSizes && !AreVectors),
|
---|
| 625 | INVALID_MATRIX_PRODUCT__IF_YOU_WANTED_A_COEFF_WISE_PRODUCT_YOU_MUST_USE_THE_EXPLICIT_FUNCTION)
|
---|
| 626 | EIGEN_STATIC_ASSERT(ProductIsValid || SameSizes, INVALID_MATRIX_PRODUCT)
|
---|
| 627 |
|
---|
| 628 | return typename LazyProductReturnType<Derived,OtherDerived>::Type(derived(), other.derived());
|
---|
| 629 | }
|
---|
| 630 |
|
---|
| 631 | } // end namespace Eigen
|
---|
| 632 |
|
---|
| 633 | #endif // EIGEN_PRODUCT_H
|
---|