MPRealSupport
Go to the documentation of this file.
1 // This file is part of a joint effort between Eigen, a lightweight C++ template library
2 // for linear algebra, and MPFR C++, a C++ interface to MPFR library (http://www.holoborodko.com/pavel/)
3 //
4 // Copyright (C) 2010-2012 Pavel Holoborodko <pavel@holoborodko.com>
5 // Copyright (C) 2010 Konstantin Holoborodko <konstantin@holoborodko.com>
6 // Copyright (C) 2010 Gael Guennebaud <gael.guennebaud@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 #ifndef EIGEN_MPREALSUPPORT_MODULE_H
13 #define EIGEN_MPREALSUPPORT_MODULE_H
14 
15 #include "../../Eigen/Core"
16 #include <mpreal.h>
17 
18 namespace Eigen {
19 
62  template<> struct NumTraits<mpfr::mpreal>
63  : GenericNumTraits<mpfr::mpreal>
64  {
65  enum {
66  IsInteger = 0,
67  IsSigned = 1,
68  IsComplex = 0,
73  };
74 
75  typedef mpfr::mpreal Real;
76  typedef mpfr::mpreal NonInteger;
77 
78  static inline Real highest (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::maxval(Precision); }
79  static inline Real lowest (long Precision = mpfr::mpreal::get_default_prec()) { return -mpfr::maxval(Precision); }
80 
81  // Constants
82  static inline Real Pi (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::const_pi(Precision); }
83  static inline Real Euler (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::const_euler(Precision); }
84  static inline Real Log2 (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::const_log2(Precision); }
85  static inline Real Catalan (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::const_catalan(Precision); }
86 
87  static inline Real epsilon (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::machine_epsilon(Precision); }
88  static inline Real epsilon (const Real& x) { return mpfr::machine_epsilon(x); }
89 
90 #ifdef MPREAL_HAVE_DYNAMIC_STD_NUMERIC_LIMITS
91  static inline int digits10 (long Precision = mpfr::mpreal::get_default_prec()) { return std::numeric_limits<Real>::digits10(Precision); }
92  static inline int digits10 (const Real& x) { return std::numeric_limits<Real>::digits10(x); }
93 
94  static inline int digits () { return std::numeric_limits<Real>::digits(); }
95  static inline int digits (const Real& x) { return std::numeric_limits<Real>::digits(x); }
96 #endif
97 
98  static inline Real dummy_precision()
99  {
100  mpfr_prec_t weak_prec = ((mpfr::mpreal::get_default_prec()-1) * 90) / 100;
101  return mpfr::machine_epsilon(weak_prec);
102  }
103  };
104 
105  namespace internal {
106 
107  template<> inline mpfr::mpreal random<mpfr::mpreal>()
108  {
109  return mpfr::random();
110  }
111 
112  template<> inline mpfr::mpreal random<mpfr::mpreal>(const mpfr::mpreal& a, const mpfr::mpreal& b)
113  {
114  return a + (b-a) * random<mpfr::mpreal>();
115  }
116 
117  inline bool isMuchSmallerThan(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& eps)
118  {
119  return mpfr::abs(a) <= mpfr::abs(b) * eps;
120  }
121 
122  inline bool isApprox(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& eps)
123  {
124  return mpfr::isEqualFuzzy(a,b,eps);
125  }
126 
127  inline bool isApproxOrLessThan(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& eps)
128  {
129  return a <= b || mpfr::isEqualFuzzy(a,b,eps);
130  }
131 
132  template<> inline long double cast<mpfr::mpreal,long double>(const mpfr::mpreal& x)
133  { return x.toLDouble(); }
134 
135  template<> inline double cast<mpfr::mpreal,double>(const mpfr::mpreal& x)
136  { return x.toDouble(); }
137 
138  template<> inline long cast<mpfr::mpreal,long>(const mpfr::mpreal& x)
139  { return x.toLong(); }
140 
141  template<> inline int cast<mpfr::mpreal,int>(const mpfr::mpreal& x)
142  { return int(x.toLong()); }
143 
144  // Specialize GEBP kernel and traits for mpreal (no need for peeling, nor complicated stuff)
145  // This also permits to directly call mpfr's routines and avoid many temporaries produced by mpreal
146  template<>
147  class gebp_traits<mpfr::mpreal, mpfr::mpreal, false, false>
148  {
149  public:
150  typedef mpfr::mpreal ResScalar;
151  enum {
152  Vectorizable = false,
153  LhsPacketSize = 1,
154  RhsPacketSize = 1,
155  ResPacketSize = 1,
156  NumberOfRegisters = 1,
157  nr = 1,
158  mr = 1,
159  LhsProgress = 1,
160  RhsProgress = 1
161  };
162  typedef ResScalar LhsPacket;
163  typedef ResScalar RhsPacket;
164  typedef ResScalar ResPacket;
165  typedef LhsPacket LhsPacket4Packing;
166 
167  };
168 
169 
170 
171  template<typename Index, typename DataMapper, bool ConjugateLhs, bool ConjugateRhs>
172  struct gebp_kernel<mpfr::mpreal,mpfr::mpreal,Index,DataMapper,1,1,ConjugateLhs,ConjugateRhs>
173  {
174  typedef mpfr::mpreal mpreal;
175 
177  void operator()(const DataMapper& res, const mpreal* blockA, const mpreal* blockB,
178  Index rows, Index depth, Index cols, const mpreal& alpha,
179  Index strideA=-1, Index strideB=-1, Index offsetA=0, Index offsetB=0)
180  {
181  if(rows==0 || cols==0 || depth==0)
182  return;
183 
184  mpreal acc1(0,mpfr_get_prec(blockA[0].mpfr_srcptr())),
185  tmp (0,mpfr_get_prec(blockA[0].mpfr_srcptr()));
186 
187  if(strideA==-1) strideA = depth;
188  if(strideB==-1) strideB = depth;
189 
190  for(Index i=0; i<rows; ++i)
191  {
192  for(Index j=0; j<cols; ++j)
193  {
194  const mpreal *A = blockA + i*strideA + offsetA;
195  const mpreal *B = blockB + j*strideB + offsetB;
196 
197  acc1 = 0;
198  for(Index k=0; k<depth; k++)
199  {
200  mpfr_mul(tmp.mpfr_ptr(), A[k].mpfr_srcptr(), B[k].mpfr_srcptr(), mpreal::get_default_rnd());
201  mpfr_add(acc1.mpfr_ptr(), acc1.mpfr_ptr(), tmp.mpfr_ptr(), mpreal::get_default_rnd());
202  }
203 
204  mpfr_mul(acc1.mpfr_ptr(), acc1.mpfr_srcptr(), alpha.mpfr_srcptr(), mpreal::get_default_rnd());
205  mpfr_add(res(i,j).mpfr_ptr(), res(i,j).mpfr_srcptr(), acc1.mpfr_srcptr(), mpreal::get_default_rnd());
206  }
207  }
208  }
209  };
210  } // end namespace internal
211 }
212 
213 #endif // EIGEN_MPREALSUPPORT_MODULE_H
ArrayXXi a
SparseMatrix< double > A(n, n)
int i
MatrixXf B
IndexedView_or_VectorBlock operator()(const Indices &indices)
#define EIGEN_DONT_INLINE
cout<< "Here is the matrix m:"<< endl<< m<< endl;Matrix< ptrdiff_t, 3, 1 > res
bool isApproxOrLessThan(const Scalar &x, const Scalar &y, const typename NumTraits< Scalar >::Real &precision=NumTraits< Scalar >::dummy_precision())
bool isApprox(const Scalar &x, const Scalar &y, const typename NumTraits< Scalar >::Real &precision=NumTraits< Scalar >::dummy_precision())
bool isMuchSmallerThan(const Scalar &x, const OtherScalar &y, const typename NumTraits< Scalar >::Real &precision=NumTraits< Scalar >::dummy_precision())
: TensorContractionSycl.h, provides various tensor contraction kernel for SYCL backend
const int HugeCost
EIGEN_DEFAULT_DENSE_INDEX_TYPE Index
adouble abs(const adouble &x)
Definition: AdolcForward:74
static Real Log2(long Precision=mpfr::mpreal::get_default_prec())
Definition: MPRealSupport:84
static Real Pi(long Precision=mpfr::mpreal::get_default_prec())
Definition: MPRealSupport:82
static Real Euler(long Precision=mpfr::mpreal::get_default_prec())
Definition: MPRealSupport:83
static Real highest(long Precision=mpfr::mpreal::get_default_prec())
Definition: MPRealSupport:78
static Real lowest(long Precision=mpfr::mpreal::get_default_prec())
Definition: MPRealSupport:79
static Real epsilon(long Precision=mpfr::mpreal::get_default_prec())
Definition: MPRealSupport:87
static Real epsilon(const Real &x)
Definition: MPRealSupport:88
static Real Catalan(long Precision=mpfr::mpreal::get_default_prec())
Definition: MPRealSupport:85
Derived::Index cols
Derived::Index rows
std::ptrdiff_t j