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 (at) gmail.com> 5 // Copyright (C) 2008-2011 Gael Guennebaud <gael.guennebaud (at) inria.fr> 6 // 7 // This Source Code Form is subject to the terms of the Mozilla 8 // Public License v. 2.0. If a copy of the MPL was not distributed 9 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/. 10 11 #ifndef EIGEN_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 EIGEN_STATIC_ASSERT((internal::is_same<typename Lhs::RealScalar, typename Rhs::RealScalar>::value), 209 YOU_MIXED_DIFFERENT_NUMERIC_TYPES__YOU_NEED_TO_USE_THE_CAST_METHOD_OF_MATRIXBASE_TO_CAST_NUMERIC_TYPES_EXPLICITLY) 210 211 Base::coeffRef(0,0) = (lhs.transpose().cwiseProduct(rhs)).sum(); 212 } 213 214 /** Convertion to scalar */ 215 operator const typename Base::Scalar() const { 216 return Base::coeff(0,0); 217 } 218 }; 219 220 /*********************************************************************** 221 * Implementation of Outer Vector Vector Product 222 ***********************************************************************/ 223 224 namespace internal { 225 template<int StorageOrder> struct outer_product_selector; 226 227 template<typename Lhs, typename Rhs> 228 struct traits<GeneralProduct<Lhs,Rhs,OuterProduct> > 229 : traits<ProductBase<GeneralProduct<Lhs,Rhs,OuterProduct>, Lhs, Rhs> > 230 {}; 231 232 } 233 234 template<typename Lhs, typename Rhs> 235 class GeneralProduct<Lhs, Rhs, OuterProduct> 236 : public ProductBase<GeneralProduct<Lhs,Rhs,OuterProduct>, Lhs, Rhs> 237 { 238 public: 239 EIGEN_PRODUCT_PUBLIC_INTERFACE(GeneralProduct) 240 241 GeneralProduct(const Lhs& lhs, const Rhs& rhs) : Base(lhs,rhs) 242 { 243 EIGEN_STATIC_ASSERT((internal::is_same<typename Lhs::RealScalar, typename Rhs::RealScalar>::value), 244 YOU_MIXED_DIFFERENT_NUMERIC_TYPES__YOU_NEED_TO_USE_THE_CAST_METHOD_OF_MATRIXBASE_TO_CAST_NUMERIC_TYPES_EXPLICITLY) 245 } 246 247 template<typename Dest> void scaleAndAddTo(Dest& dest, Scalar alpha) const 248 { 249 internal::outer_product_selector<(int(Dest::Flags)&RowMajorBit) ? RowMajor : ColMajor>::run(*this, dest, alpha); 250 } 251 }; 252 253 namespace internal { 254 255 template<> struct outer_product_selector<ColMajor> { 256 template<typename ProductType, typename Dest> 257 static EIGEN_DONT_INLINE void run(const ProductType& prod, Dest& dest, typename ProductType::Scalar alpha) { 258 typedef typename Dest::Index Index; 259 // FIXME make sure lhs is sequentially stored 260 // FIXME not very good if rhs is real and lhs complex while alpha is real too 261 const Index cols = dest.cols(); 262 for (Index j=0; j<cols; ++j) 263 dest.col(j) += (alpha * prod.rhs().coeff(j)) * prod.lhs(); 264 } 265 }; 266 267 template<> struct outer_product_selector<RowMajor> { 268 template<typename ProductType, typename Dest> 269 static EIGEN_DONT_INLINE void run(const ProductType& prod, Dest& dest, typename ProductType::Scalar alpha) { 270 typedef typename Dest::Index Index; 271 // FIXME make sure rhs is sequentially stored 272 // FIXME not very good if lhs is real and rhs complex while alpha is real too 273 const Index rows = dest.rows(); 274 for (Index i=0; i<rows; ++i) 275 dest.row(i) += (alpha * prod.lhs().coeff(i)) * prod.rhs(); 276 } 277 }; 278 279 } // end namespace internal 280 281 /*********************************************************************** 282 * Implementation of General Matrix Vector Product 283 ***********************************************************************/ 284 285 /* According to the shape/flags of the matrix we have to distinghish 3 different cases: 286 * 1 - the matrix is col-major, BLAS compatible and M is large => call fast BLAS-like colmajor routine 287 * 2 - the matrix is row-major, BLAS compatible and N is large => call fast BLAS-like rowmajor routine 288 * 3 - all other cases are handled using a simple loop along the outer-storage direction. 289 * Therefore we need a lower level meta selector. 290 * Furthermore, if the matrix is the rhs, then the product has to be transposed. 291 */ 292 namespace internal { 293 294 template<typename Lhs, typename Rhs> 295 struct traits<GeneralProduct<Lhs,Rhs,GemvProduct> > 296 : traits<ProductBase<GeneralProduct<Lhs,Rhs,GemvProduct>, Lhs, Rhs> > 297 {}; 298 299 template<int Side, int StorageOrder, bool BlasCompatible> 300 struct gemv_selector; 301 302 } // end namespace internal 303 304 template<typename Lhs, typename Rhs> 305 class GeneralProduct<Lhs, Rhs, GemvProduct> 306 : public ProductBase<GeneralProduct<Lhs,Rhs,GemvProduct>, Lhs, Rhs> 307 { 308 public: 309 EIGEN_PRODUCT_PUBLIC_INTERFACE(GeneralProduct) 310 311 typedef typename Lhs::Scalar LhsScalar; 312 typedef typename Rhs::Scalar RhsScalar; 313 314 GeneralProduct(const Lhs& lhs, const Rhs& rhs) : Base(lhs,rhs) 315 { 316 // EIGEN_STATIC_ASSERT((internal::is_same<typename Lhs::Scalar, typename Rhs::Scalar>::value), 317 // YOU_MIXED_DIFFERENT_NUMERIC_TYPES__YOU_NEED_TO_USE_THE_CAST_METHOD_OF_MATRIXBASE_TO_CAST_NUMERIC_TYPES_EXPLICITLY) 318 } 319 320 enum { Side = Lhs::IsVectorAtCompileTime ? OnTheLeft : OnTheRight }; 321 typedef typename internal::conditional<int(Side)==OnTheRight,_LhsNested,_RhsNested>::type MatrixType; 322 323 template<typename Dest> void scaleAndAddTo(Dest& dst, Scalar alpha) const 324 { 325 eigen_assert(m_lhs.rows() == dst.rows() && m_rhs.cols() == dst.cols()); 326 internal::gemv_selector<Side,(int(MatrixType::Flags)&RowMajorBit) ? RowMajor : ColMajor, 327 bool(internal::blas_traits<MatrixType>::HasUsableDirectAccess)>::run(*this, dst, alpha); 328 } 329 }; 330 331 namespace internal { 332 333 // The vector is on the left => transposition 334 template<int StorageOrder, bool BlasCompatible> 335 struct gemv_selector<OnTheLeft,StorageOrder,BlasCompatible> 336 { 337 template<typename ProductType, typename Dest> 338 static void run(const ProductType& prod, Dest& dest, typename ProductType::Scalar alpha) 339 { 340 Transpose<Dest> destT(dest); 341 enum { OtherStorageOrder = StorageOrder == RowMajor ? ColMajor : RowMajor }; 342 gemv_selector<OnTheRight,OtherStorageOrder,BlasCompatible> 343 ::run(GeneralProduct<Transpose<const typename ProductType::_RhsNested>,Transpose<const typename ProductType::_LhsNested>, GemvProduct> 344 (prod.rhs().transpose(), prod.lhs().transpose()), destT, alpha); 345 } 346 }; 347 348 template<typename Scalar,int Size,int MaxSize,bool Cond> struct gemv_static_vector_if; 349 350 template<typename Scalar,int Size,int MaxSize> 351 struct gemv_static_vector_if<Scalar,Size,MaxSize,false> 352 { 353 EIGEN_STRONG_INLINE Scalar* data() { eigen_internal_assert(false && "should never be called"); return 0; } 354 }; 355 356 template<typename Scalar,int Size> 357 struct gemv_static_vector_if<Scalar,Size,Dynamic,true> 358 { 359 EIGEN_STRONG_INLINE Scalar* data() { return 0; } 360 }; 361 362 template<typename Scalar,int Size,int MaxSize> 363 struct gemv_static_vector_if<Scalar,Size,MaxSize,true> 364 { 365 #if EIGEN_ALIGN_STATICALLY 366 internal::plain_array<Scalar,EIGEN_SIZE_MIN_PREFER_FIXED(Size,MaxSize),0> m_data; 367 EIGEN_STRONG_INLINE Scalar* data() { return m_data.array; } 368 #else 369 // Some architectures cannot align on the stack, 370 // => let's manually enforce alignment by allocating more data and return the address of the first aligned element. 371 enum { 372 ForceAlignment = internal::packet_traits<Scalar>::Vectorizable, 373 PacketSize = internal::packet_traits<Scalar>::size 374 }; 375 internal::plain_array<Scalar,EIGEN_SIZE_MIN_PREFER_FIXED(Size,MaxSize)+(ForceAlignment?PacketSize:0),0> m_data; 376 EIGEN_STRONG_INLINE Scalar* data() { 377 return ForceAlignment 378 ? reinterpret_cast<Scalar*>((reinterpret_cast<size_t>(m_data.array) & ~(size_t(15))) + 16) 379 : m_data.array; 380 } 381 #endif 382 }; 383 384 template<> struct gemv_selector<OnTheRight,ColMajor,true> 385 { 386 template<typename ProductType, typename Dest> 387 static inline void run(const ProductType& prod, Dest& dest, typename ProductType::Scalar alpha) 388 { 389 typedef typename ProductType::Index Index; 390 typedef typename ProductType::LhsScalar LhsScalar; 391 typedef typename ProductType::RhsScalar RhsScalar; 392 typedef typename ProductType::Scalar ResScalar; 393 typedef typename ProductType::RealScalar RealScalar; 394 typedef typename ProductType::ActualLhsType ActualLhsType; 395 typedef typename ProductType::ActualRhsType ActualRhsType; 396 typedef typename ProductType::LhsBlasTraits LhsBlasTraits; 397 typedef typename ProductType::RhsBlasTraits RhsBlasTraits; 398 typedef Map<Matrix<ResScalar,Dynamic,1>, Aligned> MappedDest; 399 400 ActualLhsType actualLhs = LhsBlasTraits::extract(prod.lhs()); 401 ActualRhsType actualRhs = RhsBlasTraits::extract(prod.rhs()); 402 403 ResScalar actualAlpha = alpha * LhsBlasTraits::extractScalarFactor(prod.lhs()) 404 * RhsBlasTraits::extractScalarFactor(prod.rhs()); 405 406 enum { 407 // FIXME find a way to allow an inner stride on the result if packet_traits<Scalar>::size==1 408 // on, the other hand it is good for the cache to pack the vector anyways... 409 EvalToDestAtCompileTime = Dest::InnerStrideAtCompileTime==1, 410 ComplexByReal = (NumTraits<LhsScalar>::IsComplex) && (!NumTraits<RhsScalar>::IsComplex), 411 MightCannotUseDest = (Dest::InnerStrideAtCompileTime!=1) || ComplexByReal 412 }; 413 414 gemv_static_vector_if<ResScalar,Dest::SizeAtCompileTime,Dest::MaxSizeAtCompileTime,MightCannotUseDest> static_dest; 415 416 bool alphaIsCompatible = (!ComplexByReal) || (imag(actualAlpha)==RealScalar(0)); 417 bool evalToDest = EvalToDestAtCompileTime && alphaIsCompatible; 418 419 RhsScalar compatibleAlpha = get_factor<ResScalar,RhsScalar>::run(actualAlpha); 420 421 ei_declare_aligned_stack_constructed_variable(ResScalar,actualDestPtr,dest.size(), 422 evalToDest ? dest.data() : static_dest.data()); 423 424 if(!evalToDest) 425 { 426 #ifdef EIGEN_DENSE_STORAGE_CTOR_PLUGIN 427 int size = dest.size(); 428 EIGEN_DENSE_STORAGE_CTOR_PLUGIN 429 #endif 430 if(!alphaIsCompatible) 431 { 432 MappedDest(actualDestPtr, dest.size()).setZero(); 433 compatibleAlpha = RhsScalar(1); 434 } 435 else 436 MappedDest(actualDestPtr, dest.size()) = dest; 437 } 438 439 general_matrix_vector_product 440 <Index,LhsScalar,ColMajor,LhsBlasTraits::NeedToConjugate,RhsScalar,RhsBlasTraits::NeedToConjugate>::run( 441 actualLhs.rows(), actualLhs.cols(), 442 actualLhs.data(), actualLhs.outerStride(), 443 actualRhs.data(), actualRhs.innerStride(), 444 actualDestPtr, 1, 445 compatibleAlpha); 446 447 if (!evalToDest) 448 { 449 if(!alphaIsCompatible) 450 dest += actualAlpha * MappedDest(actualDestPtr, dest.size()); 451 else 452 dest = MappedDest(actualDestPtr, dest.size()); 453 } 454 } 455 }; 456 457 template<> struct gemv_selector<OnTheRight,RowMajor,true> 458 { 459 template<typename ProductType, typename Dest> 460 static void run(const ProductType& prod, Dest& dest, typename ProductType::Scalar alpha) 461 { 462 typedef typename ProductType::LhsScalar LhsScalar; 463 typedef typename ProductType::RhsScalar RhsScalar; 464 typedef typename ProductType::Scalar ResScalar; 465 typedef typename ProductType::Index Index; 466 typedef typename ProductType::ActualLhsType ActualLhsType; 467 typedef typename ProductType::ActualRhsType ActualRhsType; 468 typedef typename ProductType::_ActualRhsType _ActualRhsType; 469 typedef typename ProductType::LhsBlasTraits LhsBlasTraits; 470 typedef typename ProductType::RhsBlasTraits RhsBlasTraits; 471 472 typename add_const<ActualLhsType>::type actualLhs = LhsBlasTraits::extract(prod.lhs()); 473 typename add_const<ActualRhsType>::type actualRhs = RhsBlasTraits::extract(prod.rhs()); 474 475 ResScalar actualAlpha = alpha * LhsBlasTraits::extractScalarFactor(prod.lhs()) 476 * RhsBlasTraits::extractScalarFactor(prod.rhs()); 477 478 enum { 479 // FIXME find a way to allow an inner stride on the result if packet_traits<Scalar>::size==1 480 // on, the other hand it is good for the cache to pack the vector anyways... 481 DirectlyUseRhs = _ActualRhsType::InnerStrideAtCompileTime==1 482 }; 483 484 gemv_static_vector_if<RhsScalar,_ActualRhsType::SizeAtCompileTime,_ActualRhsType::MaxSizeAtCompileTime,!DirectlyUseRhs> static_rhs; 485 486 ei_declare_aligned_stack_constructed_variable(RhsScalar,actualRhsPtr,actualRhs.size(), 487 DirectlyUseRhs ? const_cast<RhsScalar*>(actualRhs.data()) : static_rhs.data()); 488 489 if(!DirectlyUseRhs) 490 { 491 #ifdef EIGEN_DENSE_STORAGE_CTOR_PLUGIN 492 int size = actualRhs.size(); 493 EIGEN_DENSE_STORAGE_CTOR_PLUGIN 494 #endif 495 Map<typename _ActualRhsType::PlainObject>(actualRhsPtr, actualRhs.size()) = actualRhs; 496 } 497 498 general_matrix_vector_product 499 <Index,LhsScalar,RowMajor,LhsBlasTraits::NeedToConjugate,RhsScalar,RhsBlasTraits::NeedToConjugate>::run( 500 actualLhs.rows(), actualLhs.cols(), 501 actualLhs.data(), actualLhs.outerStride(), 502 actualRhsPtr, 1, 503 dest.data(), dest.innerStride(), 504 actualAlpha); 505 } 506 }; 507 508 template<> struct gemv_selector<OnTheRight,ColMajor,false> 509 { 510 template<typename ProductType, typename Dest> 511 static void run(const ProductType& prod, Dest& dest, typename ProductType::Scalar alpha) 512 { 513 typedef typename Dest::Index Index; 514 // TODO makes sure dest is sequentially stored in memory, otherwise use a temp 515 const Index size = prod.rhs().rows(); 516 for(Index k=0; k<size; ++k) 517 dest += (alpha*prod.rhs().coeff(k)) * prod.lhs().col(k); 518 } 519 }; 520 521 template<> struct gemv_selector<OnTheRight,RowMajor,false> 522 { 523 template<typename ProductType, typename Dest> 524 static void run(const ProductType& prod, Dest& dest, typename ProductType::Scalar alpha) 525 { 526 typedef typename Dest::Index Index; 527 // TODO makes sure rhs is sequentially stored in memory, otherwise use a temp 528 const Index rows = prod.rows(); 529 for(Index i=0; i<rows; ++i) 530 dest.coeffRef(i) += alpha * (prod.lhs().row(i).cwiseProduct(prod.rhs().transpose())).sum(); 531 } 532 }; 533 534 } // end namespace internal 535 536 /*************************************************************************** 537 * Implementation of matrix base methods 538 ***************************************************************************/ 539 540 /** \returns the matrix product of \c *this and \a other. 541 * 542 * \note If instead of the matrix product you want the coefficient-wise product, see Cwise::operator*(). 543 * 544 * \sa lazyProduct(), operator*=(const MatrixBase&), Cwise::operator*() 545 */ 546 template<typename Derived> 547 template<typename OtherDerived> 548 inline const typename ProductReturnType<Derived, OtherDerived>::Type 549 MatrixBase<Derived>::operator*(const MatrixBase<OtherDerived> &other) const 550 { 551 // A note regarding the function declaration: In MSVC, this function will sometimes 552 // not be inlined since DenseStorage is an unwindable object for dynamic 553 // matrices and product types are holding a member to store the result. 554 // Thus it does not help tagging this function with EIGEN_STRONG_INLINE. 555 enum { 556 ProductIsValid = Derived::ColsAtCompileTime==Dynamic 557 || OtherDerived::RowsAtCompileTime==Dynamic 558 || int(Derived::ColsAtCompileTime)==int(OtherDerived::RowsAtCompileTime), 559 AreVectors = Derived::IsVectorAtCompileTime && OtherDerived::IsVectorAtCompileTime, 560 SameSizes = EIGEN_PREDICATE_SAME_MATRIX_SIZE(Derived,OtherDerived) 561 }; 562 // note to the lost user: 563 // * for a dot product use: v1.dot(v2) 564 // * for a coeff-wise product use: v1.cwiseProduct(v2) 565 EIGEN_STATIC_ASSERT(ProductIsValid || !(AreVectors && SameSizes), 566 INVALID_VECTOR_VECTOR_PRODUCT__IF_YOU_WANTED_A_DOT_OR_COEFF_WISE_PRODUCT_YOU_MUST_USE_THE_EXPLICIT_FUNCTIONS) 567 EIGEN_STATIC_ASSERT(ProductIsValid || !(SameSizes && !AreVectors), 568 INVALID_MATRIX_PRODUCT__IF_YOU_WANTED_A_COEFF_WISE_PRODUCT_YOU_MUST_USE_THE_EXPLICIT_FUNCTION) 569 EIGEN_STATIC_ASSERT(ProductIsValid || SameSizes, INVALID_MATRIX_PRODUCT) 570 #ifdef EIGEN_DEBUG_PRODUCT 571 internal::product_type<Derived,OtherDerived>::debug(); 572 #endif 573 return typename ProductReturnType<Derived,OtherDerived>::Type(derived(), other.derived()); 574 } 575 576 /** \returns an expression of the matrix product of \c *this and \a other without implicit evaluation. 577 * 578 * The returned product will behave like any other expressions: the coefficients of the product will be 579 * computed once at a time as requested. This might be useful in some extremely rare cases when only 580 * a small and no coherent fraction of the result's coefficients have to be computed. 581 * 582 * \warning This version of the matrix product can be much much slower. So use it only if you know 583 * what you are doing and that you measured a true speed improvement. 584 * 585 * \sa operator*(const MatrixBase&) 586 */ 587 template<typename Derived> 588 template<typename OtherDerived> 589 const typename LazyProductReturnType<Derived,OtherDerived>::Type 590 MatrixBase<Derived>::lazyProduct(const MatrixBase<OtherDerived> &other) const 591 { 592 enum { 593 ProductIsValid = Derived::ColsAtCompileTime==Dynamic 594 || OtherDerived::RowsAtCompileTime==Dynamic 595 || int(Derived::ColsAtCompileTime)==int(OtherDerived::RowsAtCompileTime), 596 AreVectors = Derived::IsVectorAtCompileTime && OtherDerived::IsVectorAtCompileTime, 597 SameSizes = EIGEN_PREDICATE_SAME_MATRIX_SIZE(Derived,OtherDerived) 598 }; 599 // note to the lost user: 600 // * for a dot product use: v1.dot(v2) 601 // * for a coeff-wise product use: v1.cwiseProduct(v2) 602 EIGEN_STATIC_ASSERT(ProductIsValid || !(AreVectors && SameSizes), 603 INVALID_VECTOR_VECTOR_PRODUCT__IF_YOU_WANTED_A_DOT_OR_COEFF_WISE_PRODUCT_YOU_MUST_USE_THE_EXPLICIT_FUNCTIONS) 604 EIGEN_STATIC_ASSERT(ProductIsValid || !(SameSizes && !AreVectors), 605 INVALID_MATRIX_PRODUCT__IF_YOU_WANTED_A_COEFF_WISE_PRODUCT_YOU_MUST_USE_THE_EXPLICIT_FUNCTION) 606 EIGEN_STATIC_ASSERT(ProductIsValid || SameSizes, INVALID_MATRIX_PRODUCT) 607 608 return typename LazyProductReturnType<Derived,OtherDerived>::Type(derived(), other.derived()); 609 } 610 611 } // end namespace Eigen 612 613 #endif // EIGEN_PRODUCT_H 614
