ground/gcs/src/libs/eigen/test/redux.cpp

   1 // This file is part of Eigen, a lightweight C++ template library
   2 // for linear algebra.
   3 //
   4 // Copyright (C) 2008 Benoit Jacob <jacob.benoit.1@gmail.com>
   5 //
   6 // This Source Code Form is subject to the terms of the Mozilla
   7 // Public License v. 2.0. If a copy of the MPL was not distributed
   8 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
   9
  10 #include "main.h"
  11
  12 template<typename MatrixType> void matrixRedux(const MatrixType& m)
  13 {
  14   typedef typename MatrixType::Index Index;
  15   typedef typename MatrixType::Scalar Scalar;
  16   typedef typename MatrixType::RealScalar RealScalar;
  17
  18   Index rows = m.rows();
  19   Index cols = m.cols();
  20
  21   MatrixType m1 = MatrixType::Random(rows, cols);
  22
  23   // The entries of m1 are uniformly distributed in [0,1], so m1.prod() is very small. This may lead to test
  24   // failures if we underflow into denormals. Thus, we scale so that entires are close to 1.
  25   MatrixType m1_for_prod = MatrixType::Ones(rows, cols) + RealScalar(0.2) * m1;
  26
  27   VERIFY_IS_MUCH_SMALLER_THAN(MatrixType::Zero(rows, cols).sum(), Scalar(1));
  28   VERIFY_IS_APPROX(MatrixType::Ones(rows, cols).sum(), Scalar(float(rows*cols))); // the float() here to shut up excessive MSVC warning about int->complex conversion being lossy
  29   Scalar s(0), p(1), minc(numext::real(m1.coeff(0))), maxc(numext::real(m1.coeff(0)));
  30   for(int j = 0; j < cols; j++)
  31   for(int i = 0; i < rows; i++)
  32   {
  33     s += m1(i,j);
  34     p *= m1_for_prod(i,j);
  35     minc = (std::min)(numext::real(minc), numext::real(m1(i,j)));
  36     maxc = (std::max)(numext::real(maxc), numext::real(m1(i,j)));
  37   }
  38   const Scalar mean = s/Scalar(RealScalar(rows*cols));
  39
  40   VERIFY_IS_APPROX(m1.sum(), s);
  41   VERIFY_IS_APPROX(m1.mean(), mean);
  42   VERIFY_IS_APPROX(m1_for_prod.prod(), p);
  43   VERIFY_IS_APPROX(m1.real().minCoeff(), numext::real(minc));
  44   VERIFY_IS_APPROX(m1.real().maxCoeff(), numext::real(maxc));
  45
  46   // test slice vectorization assuming assign is ok
  47   Index r0 = internal::random<Index>(0,rows-1);
  48   Index c0 = internal::random<Index>(0,cols-1);
  49   Index r1 = internal::random<Index>(r0+1,rows)-r0;
  50   Index c1 = internal::random<Index>(c0+1,cols)-c0;
  51   VERIFY_IS_APPROX(m1.block(r0,c0,r1,c1).sum(), m1.block(r0,c0,r1,c1).eval().sum());
  52   VERIFY_IS_APPROX(m1.block(r0,c0,r1,c1).mean(), m1.block(r0,c0,r1,c1).eval().mean());
  53   VERIFY_IS_APPROX(m1_for_prod.block(r0,c0,r1,c1).prod(), m1_for_prod.block(r0,c0,r1,c1).eval().prod());
  54   VERIFY_IS_APPROX(m1.block(r0,c0,r1,c1).real().minCoeff(), m1.block(r0,c0,r1,c1).real().eval().minCoeff());
  55   VERIFY_IS_APPROX(m1.block(r0,c0,r1,c1).real().maxCoeff(), m1.block(r0,c0,r1,c1).real().eval().maxCoeff());
  56
  57   // regression for bug 1090
  58   const int R1 = MatrixType::RowsAtCompileTime>=2 ? MatrixType::RowsAtCompileTime/2 : 6;
  59   const int C1 = MatrixType::ColsAtCompileTime>=2 ? MatrixType::ColsAtCompileTime/2 : 6;
  60   if(R1<=rows-r0 && C1<=cols-c0)
  61   {
  62     VERIFY_IS_APPROX( (m1.template block<R1,C1>(r0,c0).sum()), m1.block(r0,c0,R1,C1).sum() );
  63   }
  64
  65   // test empty objects
  66   VERIFY_IS_APPROX(m1.block(r0,c0,0,0).sum(),   Scalar(0));
  67   VERIFY_IS_APPROX(m1.block(r0,c0,0,0).prod(),  Scalar(1));
  68 }
  69
  70 template<typename VectorType> void vectorRedux(const VectorType& w)
  71 {
  72   using std::abs;
  73   typedef typename VectorType::Index Index;
  74   typedef typename VectorType::Scalar Scalar;
  75   typedef typename NumTraits<Scalar>::Real RealScalar;
  76   Index size = w.size();
  77
  78   VectorType v = VectorType::Random(size);
  79   VectorType v_for_prod = VectorType::Ones(size) + Scalar(0.2) * v; // see comment above declaration of m1_for_prod
  80
  81   for(int i = 1; i < size; i++)
  82   {
  83     Scalar s(0), p(1);
  84     RealScalar minc(numext::real(v.coeff(0))), maxc(numext::real(v.coeff(0)));
  85     for(int j = 0; j < i; j++)
  86     {
  87       s += v[j];
  88       p *= v_for_prod[j];
  89       minc = (std::min)(minc, numext::real(v[j]));
  90       maxc = (std::max)(maxc, numext::real(v[j]));
  91     }
  92     VERIFY_IS_MUCH_SMALLER_THAN(abs(s - v.head(i).sum()), Scalar(1));
  93     VERIFY_IS_APPROX(p, v_for_prod.head(i).prod());
  94     VERIFY_IS_APPROX(minc, v.real().head(i).minCoeff());
  95     VERIFY_IS_APPROX(maxc, v.real().head(i).maxCoeff());
  96   }
  97
  98   for(int i = 0; i < size-1; i++)
  99   {
 100     Scalar s(0), p(1);
 101     RealScalar minc(numext::real(v.coeff(i))), maxc(numext::real(v.coeff(i)));
 102     for(int j = i; j < size; j++)
 103     {
 104       s += v[j];
 105       p *= v_for_prod[j];
 106       minc = (std::min)(minc, numext::real(v[j]));
 107       maxc = (std::max)(maxc, numext::real(v[j]));
 108     }
 109     VERIFY_IS_MUCH_SMALLER_THAN(abs(s - v.tail(size-i).sum()), Scalar(1));
 110     VERIFY_IS_APPROX(p, v_for_prod.tail(size-i).prod());
 111     VERIFY_IS_APPROX(minc, v.real().tail(size-i).minCoeff());
 112     VERIFY_IS_APPROX(maxc, v.real().tail(size-i).maxCoeff());
 113   }
 114
 115   for(int i = 0; i < size/2; i++)
 116   {
 117     Scalar s(0), p(1);
 118     RealScalar minc(numext::real(v.coeff(i))), maxc(numext::real(v.coeff(i)));
 119     for(int j = i; j < size-i; j++)
 120     {
 121       s += v[j];
 122       p *= v_for_prod[j];
 123       minc = (std::min)(minc, numext::real(v[j]));
 124       maxc = (std::max)(maxc, numext::real(v[j]));
 125     }
 126     VERIFY_IS_MUCH_SMALLER_THAN(abs(s - v.segment(i, size-2*i).sum()), Scalar(1));
 127     VERIFY_IS_APPROX(p, v_for_prod.segment(i, size-2*i).prod());
 128     VERIFY_IS_APPROX(minc, v.real().segment(i, size-2*i).minCoeff());
 129     VERIFY_IS_APPROX(maxc, v.real().segment(i, size-2*i).maxCoeff());
 130   }
 131
 132   // test empty objects
 133   VERIFY_IS_APPROX(v.head(0).sum(),   Scalar(0));
 134   VERIFY_IS_APPROX(v.tail(0).prod(),  Scalar(1));
 135   VERIFY_RAISES_ASSERT(v.head(0).mean());
 136   VERIFY_RAISES_ASSERT(v.head(0).minCoeff());
 137   VERIFY_RAISES_ASSERT(v.head(0).maxCoeff());
 138 }
 139
 140 void test_redux()
 141 {
 142   // the max size cannot be too large, otherwise reduxion operations obviously generate large errors.
 143   int maxsize = (std::min)(100,EIGEN_TEST_MAX_SIZE);
 144   TEST_SET_BUT_UNUSED_VARIABLE(maxsize);
 145   for(int i = 0; i < g_repeat; i++) {
 146     CALL_SUBTEST_1( matrixRedux(Matrix<float, 1, 1>()) );
 147     CALL_SUBTEST_1( matrixRedux(Array<float, 1, 1>()) );
 148     CALL_SUBTEST_2( matrixRedux(Matrix2f()) );
 149     CALL_SUBTEST_2( matrixRedux(Array2f()) );
 150     CALL_SUBTEST_3( matrixRedux(Matrix4d()) );
 151     CALL_SUBTEST_3( matrixRedux(Array4d()) );
 152     CALL_SUBTEST_4( matrixRedux(MatrixXcf(internal::random<int>(1,maxsize), internal::random<int>(1,maxsize))) );
 153     CALL_SUBTEST_4( matrixRedux(ArrayXXcf(internal::random<int>(1,maxsize), internal::random<int>(1,maxsize))) );
 154     CALL_SUBTEST_5( matrixRedux(MatrixXd (internal::random<int>(1,maxsize), internal::random<int>(1,maxsize))) );
 155     CALL_SUBTEST_5( matrixRedux(ArrayXXd (internal::random<int>(1,maxsize), internal::random<int>(1,maxsize))) );
 156     CALL_SUBTEST_6( matrixRedux(MatrixXi (internal::random<int>(1,maxsize), internal::random<int>(1,maxsize))) );
 157     CALL_SUBTEST_6( matrixRedux(ArrayXXi (internal::random<int>(1,maxsize), internal::random<int>(1,maxsize))) );
 158   }
 159   for(int i = 0; i < g_repeat; i++) {
 160     CALL_SUBTEST_7( vectorRedux(Vector4f()) );
 161     CALL_SUBTEST_7( vectorRedux(Array4f()) );
 162     CALL_SUBTEST_5( vectorRedux(VectorXd(internal::random<int>(1,maxsize))) );
 163     CALL_SUBTEST_5( vectorRedux(ArrayXd(internal::random<int>(1,maxsize))) );
 164     CALL_SUBTEST_8( vectorRedux(VectorXf(internal::random<int>(1,maxsize))) );
 165     CALL_SUBTEST_8( vectorRedux(ArrayXf(internal::random<int>(1,maxsize))) );
 166   }
 167 }