Home | History | Annotate | Download | only in test
      1 // This file is part of Eigen, a lightweight C++ template library
      2 // for linear algebra.
      3 //
      4 // Copyright (C) 2008-2011 Gael Guennebaud <gael.guennebaud (at) inria.fr>
      5 // Copyright (C) 2008 Daniel Gomez Ferro <dgomezferro (at) gmail.com>
      6 // Copyright (C) 2013 Dsir Nuentsa-Wakam <desire.nuentsa_wakam (at) inria.fr>
      7 //
      8 // This Source Code Form is subject to the terms of the Mozilla
      9 // Public License v. 2.0. If a copy of the MPL was not distributed
     10 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
     11 
     12 static long g_realloc_count = 0;
     13 #define EIGEN_SPARSE_COMPRESSED_STORAGE_REALLOCATE_PLUGIN g_realloc_count++;
     14 
     15 #include "sparse.h"
     16 
     17 template<typename SparseMatrixType> void sparse_basic(const SparseMatrixType& ref)
     18 {
     19   typedef typename SparseMatrixType::StorageIndex StorageIndex;
     20   typedef Matrix<StorageIndex,2,1> Vector2;
     21 
     22   const Index rows = ref.rows();
     23   const Index cols = ref.cols();
     24   //const Index inner = ref.innerSize();
     25   //const Index outer = ref.outerSize();
     26 
     27   typedef typename SparseMatrixType::Scalar Scalar;
     28   typedef typename SparseMatrixType::RealScalar RealScalar;
     29   enum { Flags = SparseMatrixType::Flags };
     30 
     31   double density = (std::max)(8./(rows*cols), 0.01);
     32   typedef Matrix<Scalar,Dynamic,Dynamic> DenseMatrix;
     33   typedef Matrix<Scalar,Dynamic,1> DenseVector;
     34   Scalar eps = 1e-6;
     35 
     36   Scalar s1 = internal::random<Scalar>();
     37   {
     38     SparseMatrixType m(rows, cols);
     39     DenseMatrix refMat = DenseMatrix::Zero(rows, cols);
     40     DenseVector vec1 = DenseVector::Random(rows);
     41 
     42     std::vector<Vector2> zeroCoords;
     43     std::vector<Vector2> nonzeroCoords;
     44     initSparse<Scalar>(density, refMat, m, 0, &zeroCoords, &nonzeroCoords);
     45 
     46     // test coeff and coeffRef
     47     for (std::size_t i=0; i<zeroCoords.size(); ++i)
     48     {
     49       VERIFY_IS_MUCH_SMALLER_THAN( m.coeff(zeroCoords[i].x(),zeroCoords[i].y()), eps );
     50       if(internal::is_same<SparseMatrixType,SparseMatrix<Scalar,Flags> >::value)
     51         VERIFY_RAISES_ASSERT( m.coeffRef(zeroCoords[i].x(),zeroCoords[i].y()) = 5 );
     52     }
     53     VERIFY_IS_APPROX(m, refMat);
     54 
     55     if(!nonzeroCoords.empty()) {
     56       m.coeffRef(nonzeroCoords[0].x(), nonzeroCoords[0].y()) = Scalar(5);
     57       refMat.coeffRef(nonzeroCoords[0].x(), nonzeroCoords[0].y()) = Scalar(5);
     58     }
     59 
     60     VERIFY_IS_APPROX(m, refMat);
     61 
     62       // test assertion
     63       VERIFY_RAISES_ASSERT( m.coeffRef(-1,1) = 0 );
     64       VERIFY_RAISES_ASSERT( m.coeffRef(0,m.cols()) = 0 );
     65     }
     66 
     67     // test insert (inner random)
     68     {
     69       DenseMatrix m1(rows,cols);
     70       m1.setZero();
     71       SparseMatrixType m2(rows,cols);
     72       bool call_reserve = internal::random<int>()%2;
     73       Index nnz = internal::random<int>(1,int(rows)/2);
     74       if(call_reserve)
     75       {
     76         if(internal::random<int>()%2)
     77           m2.reserve(VectorXi::Constant(m2.outerSize(), int(nnz)));
     78         else
     79           m2.reserve(m2.outerSize() * nnz);
     80       }
     81       g_realloc_count = 0;
     82       for (Index j=0; j<cols; ++j)
     83       {
     84         for (Index k=0; k<nnz; ++k)
     85         {
     86           Index i = internal::random<Index>(0,rows-1);
     87           if (m1.coeff(i,j)==Scalar(0))
     88             m2.insert(i,j) = m1(i,j) = internal::random<Scalar>();
     89         }
     90       }
     91 
     92       if(call_reserve && !SparseMatrixType::IsRowMajor)
     93       {
     94         VERIFY(g_realloc_count==0);
     95       }
     96 
     97       m2.finalize();
     98       VERIFY_IS_APPROX(m2,m1);
     99     }
    100 
    101     // test insert (fully random)
    102     {
    103       DenseMatrix m1(rows,cols);
    104       m1.setZero();
    105       SparseMatrixType m2(rows,cols);
    106       if(internal::random<int>()%2)
    107         m2.reserve(VectorXi::Constant(m2.outerSize(), 2));
    108       for (int k=0; k<rows*cols; ++k)
    109       {
    110         Index i = internal::random<Index>(0,rows-1);
    111         Index j = internal::random<Index>(0,cols-1);
    112         if ((m1.coeff(i,j)==Scalar(0)) && (internal::random<int>()%2))
    113           m2.insert(i,j) = m1(i,j) = internal::random<Scalar>();
    114         else
    115         {
    116           Scalar v = internal::random<Scalar>();
    117           m2.coeffRef(i,j) += v;
    118           m1(i,j) += v;
    119         }
    120       }
    121       VERIFY_IS_APPROX(m2,m1);
    122     }
    123 
    124     // test insert (un-compressed)
    125     for(int mode=0;mode<4;++mode)
    126     {
    127       DenseMatrix m1(rows,cols);
    128       m1.setZero();
    129       SparseMatrixType m2(rows,cols);
    130       VectorXi r(VectorXi::Constant(m2.outerSize(), ((mode%2)==0) ? int(m2.innerSize()) : std::max<int>(1,int(m2.innerSize())/8)));
    131       m2.reserve(r);
    132       for (Index k=0; k<rows*cols; ++k)
    133       {
    134         Index i = internal::random<Index>(0,rows-1);
    135         Index j = internal::random<Index>(0,cols-1);
    136         if (m1.coeff(i,j)==Scalar(0))
    137           m2.insert(i,j) = m1(i,j) = internal::random<Scalar>();
    138         if(mode==3)
    139           m2.reserve(r);
    140       }
    141       if(internal::random<int>()%2)
    142         m2.makeCompressed();
    143       VERIFY_IS_APPROX(m2,m1);
    144     }
    145 
    146   // test basic computations
    147   {
    148     DenseMatrix refM1 = DenseMatrix::Zero(rows, cols);
    149     DenseMatrix refM2 = DenseMatrix::Zero(rows, cols);
    150     DenseMatrix refM3 = DenseMatrix::Zero(rows, cols);
    151     DenseMatrix refM4 = DenseMatrix::Zero(rows, cols);
    152     SparseMatrixType m1(rows, cols);
    153     SparseMatrixType m2(rows, cols);
    154     SparseMatrixType m3(rows, cols);
    155     SparseMatrixType m4(rows, cols);
    156     initSparse<Scalar>(density, refM1, m1);
    157     initSparse<Scalar>(density, refM2, m2);
    158     initSparse<Scalar>(density, refM3, m3);
    159     initSparse<Scalar>(density, refM4, m4);
    160 
    161     if(internal::random<bool>())
    162       m1.makeCompressed();
    163 
    164     Index m1_nnz = m1.nonZeros();
    165 
    166     VERIFY_IS_APPROX(m1*s1, refM1*s1);
    167     VERIFY_IS_APPROX(m1+m2, refM1+refM2);
    168     VERIFY_IS_APPROX(m1+m2+m3, refM1+refM2+refM3);
    169     VERIFY_IS_APPROX(m3.cwiseProduct(m1+m2), refM3.cwiseProduct(refM1+refM2));
    170     VERIFY_IS_APPROX(m1*s1-m2, refM1*s1-refM2);
    171     VERIFY_IS_APPROX(m4=m1/s1, refM1/s1);
    172     VERIFY_IS_EQUAL(m4.nonZeros(), m1_nnz);
    173 
    174     if(SparseMatrixType::IsRowMajor)
    175       VERIFY_IS_APPROX(m1.innerVector(0).dot(refM2.row(0)), refM1.row(0).dot(refM2.row(0)));
    176     else
    177       VERIFY_IS_APPROX(m1.innerVector(0).dot(refM2.col(0)), refM1.col(0).dot(refM2.col(0)));
    178 
    179     DenseVector rv = DenseVector::Random(m1.cols());
    180     DenseVector cv = DenseVector::Random(m1.rows());
    181     Index r = internal::random<Index>(0,m1.rows()-2);
    182     Index c = internal::random<Index>(0,m1.cols()-1);
    183     VERIFY_IS_APPROX(( m1.template block<1,Dynamic>(r,0,1,m1.cols()).dot(rv)) , refM1.row(r).dot(rv));
    184     VERIFY_IS_APPROX(m1.row(r).dot(rv), refM1.row(r).dot(rv));
    185     VERIFY_IS_APPROX(m1.col(c).dot(cv), refM1.col(c).dot(cv));
    186 
    187     VERIFY_IS_APPROX(m1.conjugate(), refM1.conjugate());
    188     VERIFY_IS_APPROX(m1.real(), refM1.real());
    189 
    190     refM4.setRandom();
    191     // sparse cwise* dense
    192     VERIFY_IS_APPROX(m3.cwiseProduct(refM4), refM3.cwiseProduct(refM4));
    193     // dense cwise* sparse
    194     VERIFY_IS_APPROX(refM4.cwiseProduct(m3), refM4.cwiseProduct(refM3));
    195 //     VERIFY_IS_APPROX(m3.cwise()/refM4, refM3.cwise()/refM4);
    196 
    197     VERIFY_IS_APPROX(refM4 + m3, refM4 + refM3);
    198     VERIFY_IS_APPROX(m3 + refM4, refM3 + refM4);
    199     VERIFY_IS_APPROX(refM4 - m3, refM4 - refM3);
    200     VERIFY_IS_APPROX(m3 - refM4, refM3 - refM4);
    201     VERIFY_IS_APPROX((RealScalar(0.5)*refM4 + RealScalar(0.5)*m3).eval(), RealScalar(0.5)*refM4 + RealScalar(0.5)*refM3);
    202     VERIFY_IS_APPROX((RealScalar(0.5)*refM4 + m3*RealScalar(0.5)).eval(), RealScalar(0.5)*refM4 + RealScalar(0.5)*refM3);
    203     VERIFY_IS_APPROX((RealScalar(0.5)*refM4 + m3.cwiseProduct(m3)).eval(), RealScalar(0.5)*refM4 + refM3.cwiseProduct(refM3));
    204 
    205     VERIFY_IS_APPROX((RealScalar(0.5)*refM4 + RealScalar(0.5)*m3).eval(), RealScalar(0.5)*refM4 + RealScalar(0.5)*refM3);
    206     VERIFY_IS_APPROX((RealScalar(0.5)*refM4 + m3*RealScalar(0.5)).eval(), RealScalar(0.5)*refM4 + RealScalar(0.5)*refM3);
    207     VERIFY_IS_APPROX((RealScalar(0.5)*refM4 + (m3+m3)).eval(), RealScalar(0.5)*refM4 + (refM3+refM3));
    208     VERIFY_IS_APPROX(((refM3+m3)+RealScalar(0.5)*m3).eval(), RealScalar(0.5)*refM3 + (refM3+refM3));
    209     VERIFY_IS_APPROX((RealScalar(0.5)*refM4 + (refM3+m3)).eval(), RealScalar(0.5)*refM4 + (refM3+refM3));
    210     VERIFY_IS_APPROX((RealScalar(0.5)*refM4 + (m3+refM3)).eval(), RealScalar(0.5)*refM4 + (refM3+refM3));
    211 
    212 
    213     VERIFY_IS_APPROX(m1.sum(), refM1.sum());
    214 
    215     m4 = m1; refM4 = m4;
    216 
    217     VERIFY_IS_APPROX(m1*=s1, refM1*=s1);
    218     VERIFY_IS_EQUAL(m1.nonZeros(), m1_nnz);
    219     VERIFY_IS_APPROX(m1/=s1, refM1/=s1);
    220     VERIFY_IS_EQUAL(m1.nonZeros(), m1_nnz);
    221 
    222     VERIFY_IS_APPROX(m1+=m2, refM1+=refM2);
    223     VERIFY_IS_APPROX(m1-=m2, refM1-=refM2);
    224 
    225     if (rows>=2 && cols>=2)
    226     {
    227       VERIFY_RAISES_ASSERT( m1 += m1.innerVector(0) );
    228       VERIFY_RAISES_ASSERT( m1 -= m1.innerVector(0) );
    229       VERIFY_RAISES_ASSERT( refM1 -= m1.innerVector(0) );
    230       VERIFY_RAISES_ASSERT( refM1 += m1.innerVector(0) );
    231       m1 = m4; refM1 = refM4;
    232     }
    233 
    234     // test aliasing
    235     VERIFY_IS_APPROX((m1 = -m1), (refM1 = -refM1));
    236     VERIFY_IS_EQUAL(m1.nonZeros(), m1_nnz);
    237     m1 = m4; refM1 = refM4;
    238     VERIFY_IS_APPROX((m1 = m1.transpose()), (refM1 = refM1.transpose().eval()));
    239     VERIFY_IS_EQUAL(m1.nonZeros(), m1_nnz);
    240     m1 = m4; refM1 = refM4;
    241     VERIFY_IS_APPROX((m1 = -m1.transpose()), (refM1 = -refM1.transpose().eval()));
    242     VERIFY_IS_EQUAL(m1.nonZeros(), m1_nnz);
    243     m1 = m4; refM1 = refM4;
    244     VERIFY_IS_APPROX((m1 += -m1), (refM1 += -refM1));
    245     VERIFY_IS_EQUAL(m1.nonZeros(), m1_nnz);
    246     m1 = m4; refM1 = refM4;
    247 
    248     if(m1.isCompressed())
    249     {
    250       VERIFY_IS_APPROX(m1.coeffs().sum(), m1.sum());
    251       m1.coeffs() += s1;
    252       for(Index j = 0; j<m1.outerSize(); ++j)
    253         for(typename SparseMatrixType::InnerIterator it(m1,j); it; ++it)
    254           refM1(it.row(), it.col()) += s1;
    255       VERIFY_IS_APPROX(m1, refM1);
    256     }
    257 
    258     // and/or
    259     {
    260       typedef SparseMatrix<bool, SparseMatrixType::Options, typename SparseMatrixType::StorageIndex> SpBool;
    261       SpBool mb1 = m1.real().template cast<bool>();
    262       SpBool mb2 = m2.real().template cast<bool>();
    263       VERIFY_IS_EQUAL(mb1.template cast<int>().sum(), refM1.real().template cast<bool>().count());
    264       VERIFY_IS_EQUAL((mb1 && mb2).template cast<int>().sum(), (refM1.real().template cast<bool>() && refM2.real().template cast<bool>()).count());
    265       VERIFY_IS_EQUAL((mb1 || mb2).template cast<int>().sum(), (refM1.real().template cast<bool>() || refM2.real().template cast<bool>()).count());
    266       SpBool mb3 = mb1 && mb2;
    267       if(mb1.coeffs().all() && mb2.coeffs().all())
    268       {
    269         VERIFY_IS_EQUAL(mb3.nonZeros(), (refM1.real().template cast<bool>() && refM2.real().template cast<bool>()).count());
    270       }
    271     }
    272   }
    273 
    274   // test reverse iterators
    275   {
    276     DenseMatrix refMat2 = DenseMatrix::Zero(rows, cols);
    277     SparseMatrixType m2(rows, cols);
    278     initSparse<Scalar>(density, refMat2, m2);
    279     std::vector<Scalar> ref_value(m2.innerSize());
    280     std::vector<Index> ref_index(m2.innerSize());
    281     if(internal::random<bool>())
    282       m2.makeCompressed();
    283     for(Index j = 0; j<m2.outerSize(); ++j)
    284     {
    285       Index count_forward = 0;
    286 
    287       for(typename SparseMatrixType::InnerIterator it(m2,j); it; ++it)
    288       {
    289         ref_value[ref_value.size()-1-count_forward] = it.value();
    290         ref_index[ref_index.size()-1-count_forward] = it.index();
    291         count_forward++;
    292       }
    293       Index count_reverse = 0;
    294       for(typename SparseMatrixType::ReverseInnerIterator it(m2,j); it; --it)
    295       {
    296         VERIFY_IS_APPROX( std::abs(ref_value[ref_value.size()-count_forward+count_reverse])+1, std::abs(it.value())+1);
    297         VERIFY_IS_EQUAL( ref_index[ref_index.size()-count_forward+count_reverse] , it.index());
    298         count_reverse++;
    299       }
    300       VERIFY_IS_EQUAL(count_forward, count_reverse);
    301     }
    302   }
    303 
    304   // test transpose
    305   {
    306     DenseMatrix refMat2 = DenseMatrix::Zero(rows, cols);
    307     SparseMatrixType m2(rows, cols);
    308     initSparse<Scalar>(density, refMat2, m2);
    309     VERIFY_IS_APPROX(m2.transpose().eval(), refMat2.transpose().eval());
    310     VERIFY_IS_APPROX(m2.transpose(), refMat2.transpose());
    311 
    312     VERIFY_IS_APPROX(SparseMatrixType(m2.adjoint()), refMat2.adjoint());
    313 
    314     // check isApprox handles opposite storage order
    315     typename Transpose<SparseMatrixType>::PlainObject m3(m2);
    316     VERIFY(m2.isApprox(m3));
    317   }
    318 
    319   // test prune
    320   {
    321     SparseMatrixType m2(rows, cols);
    322     DenseMatrix refM2(rows, cols);
    323     refM2.setZero();
    324     int countFalseNonZero = 0;
    325     int countTrueNonZero = 0;
    326     m2.reserve(VectorXi::Constant(m2.outerSize(), int(m2.innerSize())));
    327     for (Index j=0; j<m2.cols(); ++j)
    328     {
    329       for (Index i=0; i<m2.rows(); ++i)
    330       {
    331         float x = internal::random<float>(0,1);
    332         if (x<0.1f)
    333         {
    334           // do nothing
    335         }
    336         else if (x<0.5f)
    337         {
    338           countFalseNonZero++;
    339           m2.insert(i,j) = Scalar(0);
    340         }
    341         else
    342         {
    343           countTrueNonZero++;
    344           m2.insert(i,j) = Scalar(1);
    345           refM2(i,j) = Scalar(1);
    346         }
    347       }
    348     }
    349     if(internal::random<bool>())
    350       m2.makeCompressed();
    351     VERIFY(countFalseNonZero+countTrueNonZero == m2.nonZeros());
    352     if(countTrueNonZero>0)
    353       VERIFY_IS_APPROX(m2, refM2);
    354     m2.prune(Scalar(1));
    355     VERIFY(countTrueNonZero==m2.nonZeros());
    356     VERIFY_IS_APPROX(m2, refM2);
    357   }
    358 
    359   // test setFromTriplets
    360   {
    361     typedef Triplet<Scalar,StorageIndex> TripletType;
    362     std::vector<TripletType> triplets;
    363     Index ntriplets = rows*cols;
    364     triplets.reserve(ntriplets);
    365     DenseMatrix refMat_sum  = DenseMatrix::Zero(rows,cols);
    366     DenseMatrix refMat_prod = DenseMatrix::Zero(rows,cols);
    367     DenseMatrix refMat_last = DenseMatrix::Zero(rows,cols);
    368 
    369     for(Index i=0;i<ntriplets;++i)
    370     {
    371       StorageIndex r = internal::random<StorageIndex>(0,StorageIndex(rows-1));
    372       StorageIndex c = internal::random<StorageIndex>(0,StorageIndex(cols-1));
    373       Scalar v = internal::random<Scalar>();
    374       triplets.push_back(TripletType(r,c,v));
    375       refMat_sum(r,c) += v;
    376       if(std::abs(refMat_prod(r,c))==0)
    377         refMat_prod(r,c) = v;
    378       else
    379         refMat_prod(r,c) *= v;
    380       refMat_last(r,c) = v;
    381     }
    382     SparseMatrixType m(rows,cols);
    383     m.setFromTriplets(triplets.begin(), triplets.end());
    384     VERIFY_IS_APPROX(m, refMat_sum);
    385 
    386     m.setFromTriplets(triplets.begin(), triplets.end(), std::multiplies<Scalar>());
    387     VERIFY_IS_APPROX(m, refMat_prod);
    388 #if (defined(__cplusplus) && __cplusplus >= 201103L)
    389     m.setFromTriplets(triplets.begin(), triplets.end(), [] (Scalar,Scalar b) { return b; });
    390     VERIFY_IS_APPROX(m, refMat_last);
    391 #endif
    392   }
    393 
    394   // test Map
    395   {
    396     DenseMatrix refMat2(rows, cols), refMat3(rows, cols);
    397     SparseMatrixType m2(rows, cols), m3(rows, cols);
    398     initSparse<Scalar>(density, refMat2, m2);
    399     initSparse<Scalar>(density, refMat3, m3);
    400     {
    401       Map<SparseMatrixType> mapMat2(m2.rows(), m2.cols(), m2.nonZeros(), m2.outerIndexPtr(), m2.innerIndexPtr(), m2.valuePtr(), m2.innerNonZeroPtr());
    402       Map<SparseMatrixType> mapMat3(m3.rows(), m3.cols(), m3.nonZeros(), m3.outerIndexPtr(), m3.innerIndexPtr(), m3.valuePtr(), m3.innerNonZeroPtr());
    403       VERIFY_IS_APPROX(mapMat2+mapMat3, refMat2+refMat3);
    404       VERIFY_IS_APPROX(mapMat2+mapMat3, refMat2+refMat3);
    405     }
    406     {
    407       MappedSparseMatrix<Scalar,SparseMatrixType::Options,StorageIndex> mapMat2(m2.rows(), m2.cols(), m2.nonZeros(), m2.outerIndexPtr(), m2.innerIndexPtr(), m2.valuePtr(), m2.innerNonZeroPtr());
    408       MappedSparseMatrix<Scalar,SparseMatrixType::Options,StorageIndex> mapMat3(m3.rows(), m3.cols(), m3.nonZeros(), m3.outerIndexPtr(), m3.innerIndexPtr(), m3.valuePtr(), m3.innerNonZeroPtr());
    409       VERIFY_IS_APPROX(mapMat2+mapMat3, refMat2+refMat3);
    410       VERIFY_IS_APPROX(mapMat2+mapMat3, refMat2+refMat3);
    411     }
    412 
    413     Index i = internal::random<Index>(0,rows-1);
    414     Index j = internal::random<Index>(0,cols-1);
    415     m2.coeffRef(i,j) = 123;
    416     if(internal::random<bool>())
    417       m2.makeCompressed();
    418     Map<SparseMatrixType> mapMat2(rows, cols, m2.nonZeros(), m2.outerIndexPtr(), m2.innerIndexPtr(), m2.valuePtr(),  m2.innerNonZeroPtr());
    419     VERIFY_IS_EQUAL(m2.coeff(i,j),Scalar(123));
    420     VERIFY_IS_EQUAL(mapMat2.coeff(i,j),Scalar(123));
    421     mapMat2.coeffRef(i,j) = -123;
    422     VERIFY_IS_EQUAL(m2.coeff(i,j),Scalar(-123));
    423   }
    424 
    425   // test triangularView
    426   {
    427     DenseMatrix refMat2(rows, cols), refMat3(rows, cols);
    428     SparseMatrixType m2(rows, cols), m3(rows, cols);
    429     initSparse<Scalar>(density, refMat2, m2);
    430     refMat3 = refMat2.template triangularView<Lower>();
    431     m3 = m2.template triangularView<Lower>();
    432     VERIFY_IS_APPROX(m3, refMat3);
    433 
    434     refMat3 = refMat2.template triangularView<Upper>();
    435     m3 = m2.template triangularView<Upper>();
    436     VERIFY_IS_APPROX(m3, refMat3);
    437 
    438     {
    439       refMat3 = refMat2.template triangularView<UnitUpper>();
    440       m3 = m2.template triangularView<UnitUpper>();
    441       VERIFY_IS_APPROX(m3, refMat3);
    442 
    443       refMat3 = refMat2.template triangularView<UnitLower>();
    444       m3 = m2.template triangularView<UnitLower>();
    445       VERIFY_IS_APPROX(m3, refMat3);
    446     }
    447 
    448     refMat3 = refMat2.template triangularView<StrictlyUpper>();
    449     m3 = m2.template triangularView<StrictlyUpper>();
    450     VERIFY_IS_APPROX(m3, refMat3);
    451 
    452     refMat3 = refMat2.template triangularView<StrictlyLower>();
    453     m3 = m2.template triangularView<StrictlyLower>();
    454     VERIFY_IS_APPROX(m3, refMat3);
    455 
    456     // check sparse-triangular to dense
    457     refMat3 = m2.template triangularView<StrictlyUpper>();
    458     VERIFY_IS_APPROX(refMat3, DenseMatrix(refMat2.template triangularView<StrictlyUpper>()));
    459   }
    460 
    461   // test selfadjointView
    462   if(!SparseMatrixType::IsRowMajor)
    463   {
    464     DenseMatrix refMat2(rows, rows), refMat3(rows, rows);
    465     SparseMatrixType m2(rows, rows), m3(rows, rows);
    466     initSparse<Scalar>(density, refMat2, m2);
    467     refMat3 = refMat2.template selfadjointView<Lower>();
    468     m3 = m2.template selfadjointView<Lower>();
    469     VERIFY_IS_APPROX(m3, refMat3);
    470 
    471     refMat3 += refMat2.template selfadjointView<Lower>();
    472     m3 += m2.template selfadjointView<Lower>();
    473     VERIFY_IS_APPROX(m3, refMat3);
    474 
    475     refMat3 -= refMat2.template selfadjointView<Lower>();
    476     m3 -= m2.template selfadjointView<Lower>();
    477     VERIFY_IS_APPROX(m3, refMat3);
    478 
    479     // selfadjointView only works for square matrices:
    480     SparseMatrixType m4(rows, rows+1);
    481     VERIFY_RAISES_ASSERT(m4.template selfadjointView<Lower>());
    482     VERIFY_RAISES_ASSERT(m4.template selfadjointView<Upper>());
    483   }
    484 
    485   // test sparseView
    486   {
    487     DenseMatrix refMat2 = DenseMatrix::Zero(rows, rows);
    488     SparseMatrixType m2(rows, rows);
    489     initSparse<Scalar>(density, refMat2, m2);
    490     VERIFY_IS_APPROX(m2.eval(), refMat2.sparseView().eval());
    491 
    492     // sparse view on expressions:
    493     VERIFY_IS_APPROX((s1*m2).eval(), (s1*refMat2).sparseView().eval());
    494     VERIFY_IS_APPROX((m2+m2).eval(), (refMat2+refMat2).sparseView().eval());
    495     VERIFY_IS_APPROX((m2*m2).eval(), (refMat2.lazyProduct(refMat2)).sparseView().eval());
    496     VERIFY_IS_APPROX((m2*m2).eval(), (refMat2*refMat2).sparseView().eval());
    497   }
    498 
    499   // test diagonal
    500   {
    501     DenseMatrix refMat2 = DenseMatrix::Zero(rows, cols);
    502     SparseMatrixType m2(rows, cols);
    503     initSparse<Scalar>(density, refMat2, m2);
    504     VERIFY_IS_APPROX(m2.diagonal(), refMat2.diagonal().eval());
    505     DenseVector d = m2.diagonal();
    506     VERIFY_IS_APPROX(d, refMat2.diagonal().eval());
    507     d = m2.diagonal().array();
    508     VERIFY_IS_APPROX(d, refMat2.diagonal().eval());
    509     VERIFY_IS_APPROX(const_cast<const SparseMatrixType&>(m2).diagonal(), refMat2.diagonal().eval());
    510 
    511     initSparse<Scalar>(density, refMat2, m2, ForceNonZeroDiag);
    512     m2.diagonal()      += refMat2.diagonal();
    513     refMat2.diagonal() += refMat2.diagonal();
    514     VERIFY_IS_APPROX(m2, refMat2);
    515   }
    516 
    517   // test diagonal to sparse
    518   {
    519     DenseVector d = DenseVector::Random(rows);
    520     DenseMatrix refMat2 = d.asDiagonal();
    521     SparseMatrixType m2(rows, rows);
    522     m2 = d.asDiagonal();
    523     VERIFY_IS_APPROX(m2, refMat2);
    524     SparseMatrixType m3(d.asDiagonal());
    525     VERIFY_IS_APPROX(m3, refMat2);
    526     refMat2 += d.asDiagonal();
    527     m2 += d.asDiagonal();
    528     VERIFY_IS_APPROX(m2, refMat2);
    529   }
    530 
    531   // test conservative resize
    532   {
    533       std::vector< std::pair<StorageIndex,StorageIndex> > inc;
    534       if(rows > 3 && cols > 2)
    535         inc.push_back(std::pair<StorageIndex,StorageIndex>(-3,-2));
    536       inc.push_back(std::pair<StorageIndex,StorageIndex>(0,0));
    537       inc.push_back(std::pair<StorageIndex,StorageIndex>(3,2));
    538       inc.push_back(std::pair<StorageIndex,StorageIndex>(3,0));
    539       inc.push_back(std::pair<StorageIndex,StorageIndex>(0,3));
    540 
    541       for(size_t i = 0; i< inc.size(); i++) {
    542         StorageIndex incRows = inc[i].first;
    543         StorageIndex incCols = inc[i].second;
    544         SparseMatrixType m1(rows, cols);
    545         DenseMatrix refMat1 = DenseMatrix::Zero(rows, cols);
    546         initSparse<Scalar>(density, refMat1, m1);
    547 
    548         m1.conservativeResize(rows+incRows, cols+incCols);
    549         refMat1.conservativeResize(rows+incRows, cols+incCols);
    550         if (incRows > 0) refMat1.bottomRows(incRows).setZero();
    551         if (incCols > 0) refMat1.rightCols(incCols).setZero();
    552 
    553         VERIFY_IS_APPROX(m1, refMat1);
    554 
    555         // Insert new values
    556         if (incRows > 0)
    557           m1.insert(m1.rows()-1, 0) = refMat1(refMat1.rows()-1, 0) = 1;
    558         if (incCols > 0)
    559           m1.insert(0, m1.cols()-1) = refMat1(0, refMat1.cols()-1) = 1;
    560 
    561         VERIFY_IS_APPROX(m1, refMat1);
    562 
    563 
    564       }
    565   }
    566 
    567   // test Identity matrix
    568   {
    569     DenseMatrix refMat1 = DenseMatrix::Identity(rows, rows);
    570     SparseMatrixType m1(rows, rows);
    571     m1.setIdentity();
    572     VERIFY_IS_APPROX(m1, refMat1);
    573     for(int k=0; k<rows*rows/4; ++k)
    574     {
    575       Index i = internal::random<Index>(0,rows-1);
    576       Index j = internal::random<Index>(0,rows-1);
    577       Scalar v = internal::random<Scalar>();
    578       m1.coeffRef(i,j) = v;
    579       refMat1.coeffRef(i,j) = v;
    580       VERIFY_IS_APPROX(m1, refMat1);
    581       if(internal::random<Index>(0,10)<2)
    582         m1.makeCompressed();
    583     }
    584     m1.setIdentity();
    585     refMat1.setIdentity();
    586     VERIFY_IS_APPROX(m1, refMat1);
    587   }
    588 
    589   // test array/vector of InnerIterator
    590   {
    591     typedef typename SparseMatrixType::InnerIterator IteratorType;
    592 
    593     DenseMatrix refMat2 = DenseMatrix::Zero(rows, cols);
    594     SparseMatrixType m2(rows, cols);
    595     initSparse<Scalar>(density, refMat2, m2);
    596     IteratorType static_array[2];
    597     static_array[0] = IteratorType(m2,0);
    598     static_array[1] = IteratorType(m2,m2.outerSize()-1);
    599     VERIFY( static_array[0] || m2.innerVector(static_array[0].outer()).nonZeros() == 0 );
    600     VERIFY( static_array[1] || m2.innerVector(static_array[1].outer()).nonZeros() == 0 );
    601     if(static_array[0] && static_array[1])
    602     {
    603       ++(static_array[1]);
    604       static_array[1] = IteratorType(m2,0);
    605       VERIFY( static_array[1] );
    606       VERIFY( static_array[1].index() == static_array[0].index() );
    607       VERIFY( static_array[1].outer() == static_array[0].outer() );
    608       VERIFY( static_array[1].value() == static_array[0].value() );
    609     }
    610 
    611     std::vector<IteratorType> iters(2);
    612     iters[0] = IteratorType(m2,0);
    613     iters[1] = IteratorType(m2,m2.outerSize()-1);
    614   }
    615 }
    616 
    617 
    618 template<typename SparseMatrixType>
    619 void big_sparse_triplet(Index rows, Index cols, double density) {
    620   typedef typename SparseMatrixType::StorageIndex StorageIndex;
    621   typedef typename SparseMatrixType::Scalar Scalar;
    622   typedef Triplet<Scalar,Index> TripletType;
    623   std::vector<TripletType> triplets;
    624   double nelements = density * rows*cols;
    625   VERIFY(nelements>=0 && nelements <  NumTraits<StorageIndex>::highest());
    626   Index ntriplets = Index(nelements);
    627   triplets.reserve(ntriplets);
    628   Scalar sum = Scalar(0);
    629   for(Index i=0;i<ntriplets;++i)
    630   {
    631     Index r = internal::random<Index>(0,rows-1);
    632     Index c = internal::random<Index>(0,cols-1);
    633     Scalar v = internal::random<Scalar>();
    634     triplets.push_back(TripletType(r,c,v));
    635     sum += v;
    636   }
    637   SparseMatrixType m(rows,cols);
    638   m.setFromTriplets(triplets.begin(), triplets.end());
    639   VERIFY(m.nonZeros() <= ntriplets);
    640   VERIFY_IS_APPROX(sum, m.sum());
    641 }
    642 
    643 
    644 void test_sparse_basic()
    645 {
    646   for(int i = 0; i < g_repeat; i++) {
    647     int r = Eigen::internal::random<int>(1,200), c = Eigen::internal::random<int>(1,200);
    648     if(Eigen::internal::random<int>(0,4) == 0) {
    649       r = c; // check square matrices in 25% of tries
    650     }
    651     EIGEN_UNUSED_VARIABLE(r+c);
    652     CALL_SUBTEST_1(( sparse_basic(SparseMatrix<double>(1, 1)) ));
    653     CALL_SUBTEST_1(( sparse_basic(SparseMatrix<double>(8, 8)) ));
    654     CALL_SUBTEST_2(( sparse_basic(SparseMatrix<std::complex<double>, ColMajor>(r, c)) ));
    655     CALL_SUBTEST_2(( sparse_basic(SparseMatrix<std::complex<double>, RowMajor>(r, c)) ));
    656     CALL_SUBTEST_1(( sparse_basic(SparseMatrix<double>(r, c)) ));
    657     CALL_SUBTEST_5(( sparse_basic(SparseMatrix<double,ColMajor,long int>(r, c)) ));
    658     CALL_SUBTEST_5(( sparse_basic(SparseMatrix<double,RowMajor,long int>(r, c)) ));
    659 
    660     r = Eigen::internal::random<int>(1,100);
    661     c = Eigen::internal::random<int>(1,100);
    662     if(Eigen::internal::random<int>(0,4) == 0) {
    663       r = c; // check square matrices in 25% of tries
    664     }
    665 
    666     CALL_SUBTEST_6(( sparse_basic(SparseMatrix<double,ColMajor,short int>(short(r), short(c))) ));
    667     CALL_SUBTEST_6(( sparse_basic(SparseMatrix<double,RowMajor,short int>(short(r), short(c))) ));
    668   }
    669 
    670   // Regression test for bug 900: (manually insert higher values here, if you have enough RAM):
    671   CALL_SUBTEST_3((big_sparse_triplet<SparseMatrix<float, RowMajor, int> >(10000, 10000, 0.125)));
    672   CALL_SUBTEST_4((big_sparse_triplet<SparseMatrix<double, ColMajor, long int> >(10000, 10000, 0.125)));
    673 
    674   // Regression test for bug 1105
    675 #ifdef EIGEN_TEST_PART_7
    676   {
    677     int n = Eigen::internal::random<int>(200,600);
    678     SparseMatrix<std::complex<double>,0, long> mat(n, n);
    679     std::complex<double> val;
    680 
    681     for(int i=0; i<n; ++i)
    682     {
    683       mat.coeffRef(i, i%(n/10)) = val;
    684       VERIFY(mat.data().allocatedSize()<20*n);
    685     }
    686   }
    687 #endif
    688 }
    689