File: SparseVector-inl.h

package info (click to toggle)
lammps 20220106.git7586adbb6a%2Bds1-2
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 348,064 kB
  • sloc: cpp: 831,421; python: 24,896; xml: 14,949; f90: 10,845; ansic: 7,967; sh: 4,226; perl: 4,064; fortran: 2,424; makefile: 1,501; objc: 238; lisp: 163; csh: 16; awk: 14; tcl: 6
file content (226 lines) | stat: -rw-r--r-- 6,240 bytes parent folder | download | duplicates (2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
// SparseVector-inl.h: provides templated functions for SparseVector in
// separate header


namespace ATC_matrix {

template<class T>
SparseVector<T> sparse_rand(INDEX n, INDEX fill, int seed=1234)
{
  srand(seed);
  const double rmax_inv = 1.0/double(RAND_MAX);
  SparseVector<T> r(n);
  while (r.size()<fill)
    r(std::rand()%r.nRows())=double(std::rand()*rmax_inv);
  return r;
}

// Multiplies a Matrix by a SparseVector (M*v) and returns a DenseVector.
template<class T>
DenseVector<T> operator*(const Matrix<T> &M, const SparseVector<T> &v)
{
  DenseVector<T> y(M.nRows());
  STORE::const_iterator it=v.data_.begin();
  for (; it!=v.data_.end(); it++) {
    const INDEX j  = it->first;
    const T& vj = it->second;
     for (INDEX i=0; i<M.nRows(); i++) y(i)+=M(i,j)*vj;
  }
  return y;
}

// Multiplies a SparseVector by a Matrix (M'*v) and returns a DenseVector.
template<class T>
DenseVector<T> operator*(const SparseVector<T> &v, const Matrix<T> &M)
{
  DenseVector<T> y(M.nCols());
  STORE::const_iterator it=v.data_.begin();
  for (; it!=v.data_.end(); it++) {
    const INDEX i  = it->first;
    const T& vi = it->second;
     for (INDEX j=0; j<M.nCols(); j++) y(j)+=vi*M(i,j);
  }
  return y;
}

// Computes the dot product between two SparseVectors of equal length.
template<class T>
T dot(const SparseVector<T> &a, const SparseVector<T> &b)
{
  T v = 0.0;
  for (STORE::const_iterator ai=a.data_.begin(); ai!=a.data_.end(); ai++) {
    STORE::const_iterator bi=b.data_.find(ai->first);
    if (bi == b.data_.end()) continue;
    v += ai->second * bi->second;
  }
  return v;
}
// Computes the product of a SparseMatrix transpose with a SparseVector (M'*v).
template<class T>
SparseVector<T> operator*(const SparseMatrix<T> &M, const SparseVector<T> &v)
{
  SparseVector<T> y(M.nRows());
  for (INDEX i=0; i<M.nRows(); i++) {
    double yi=0.0;
    for (INDEX ij=M._ia[i]; ij<M._ia[i+1]; ij++) {
      const INDEX j = M._ja[ij];
      STORE::const_iterator it = v._data.find(j);
      if (it == v._data.end()) continue;
      yi += M._v[ij] * it->second;
    }
    if (yi!=0.0) y(i)+=yi;
  }
  return y;
}

// computes the product of a SparseMatrix transpose with a SparseVector (M'*v).
template<class T>
SparseVector<T> operator*(const SparseVector<T> &v, const SparseMatrix<T> &M)
{
  SparseVector<T> y(M.nCols());
  for (INDEX i=0; i<M.nRows(); i++) {
    STORE::const_iterator it = v._data.find(i);
    if (it == v._data.end()) continue;
    for (INDEX ij=M._ia[i]; ij<M._ia[i+1]; ij++)
      y(M._ja[ij]) += it->second * M._v[ij];
  }
  return y;
}

// General constructor - sets the vector length.
template<class T>
SparseVector<T>::SparseVector(INDEX n) : length_(n){}

// Outputs the vector to string
template<class T>
std::string SparseVector<T>::to_string() const
{
  if (size() > nRows()/2) return Vector<T>::to_string();
  STORE::const_iterator it=data_.begin();
  std::string str;
  using ATC_Utility::to_string;
  for (; it!=data_.end(); it++)
    str += to_string(it->first)+": "+to_string(it->second)+'\n';
  return str;
}

// Indexes the ith component of the vector or returns zero if not found.
template<class T>
T SparseVector<T>::operator()(INDEX i, INDEX /* j */) const
{
  STORE::const_iterator it = data_.find(i);
  if (it == data_.end()) return 0.0;
  return it->second;
}

// Indexes the ith component of the vector or returns zero if not found.
template<class T>
T& SparseVector<T>::operator()(INDEX i, INDEX /* j */)
{
  return data_[i];
}

// Indexes the ith component of the vector or returns zero if not found.
template<class T> T SparseVector<T>::operator[](INDEX i) const
{
  return (*this)(i);
}

// Indexes the ith component of the vector or returns zero if not found.
template<class T> T& SparseVector<T>::operator[](INDEX i)
{
  return (*this)(i);
}

// Returns a pair (index, value) for a nonzero in the vector.
template<class T>
std::pair<INDEX, T> SparseVector<T>::pair(INDEX i) const
{
  STORE::const_iterator it=data_.begin() + i;
  return std::pair<INDEX, T>(it->first, it->second);
}

//* Adds SparseVector x, scaled by s to this one.  Can be different sparcity.
template<class T>
void SparseVector<T>::add_scaled(SparseVector<T>& x, const T& s)
{
  STORE::const_iterator it;
  for (it=x.data_.begin(); it!=x.data_.end(); it++)  {
   data_[it->first] += it->second * s;
  }
}

// Returns the number of nonzeros in the sparse vector.
template<class T> inline INDEX SparseVector<T>::size() const
{ return data_.size(); }

// Returns the number of nonzeros in the sparse vector.
template<class T> inline INDEX SparseVector<T>::nRows() const
{ return length_; }

// Copy constructor for sparse vector.
template<typename T>
SparseVector<T>::SparseVector(const SparseVector<T> &c)
{
  length_ = c.length_;
  data_ = c.data_;
}

// operator equal for sparse vector.
template<class T>
SparseVector<T>& SparseVector<T>::operator=(const SparseVector<T> &c)
{
  length_ = c.length_;
  data_ = c.data_;
  return *this;
}

// Changes the size of the SparseVector
template<class T>
void SparseVector<T>::resize(INDEX nRows, INDEX /* nCols */, bool copy)
{
  length_ = nRows;
  STORE::iterator it;
  for (it=data_.begin(); it!=data_.end(); it++) {
    if (it->second >= length_) data_.erase(it);
    else if (!copy) it->second=T(0);
  }
}
// same as resize, but zero rather than copy
template<class T>
void SparseVector<T>::reset(INDEX nRows, INDEX nCols, bool zero)
{
  resize(nRows, nCols, !zero);
}
// sets all elements to zero but preserves sparcity
template<class T>
void SparseVector<T>::zero()
{
  STORE::iterator it;
  for (it=data_.begin(); it!=data_.end(); it++) it->second=T(0);
}


template<class T>
void SparseVector<T>::copy(const T* /* ptr */, INDEX /* nRows */, INDEX /* nCols */)
{
}


template<class T>
void SparseVector<T>::write_restart(FILE */* F */) const
{
}

// writes a stream to a matlab script to recreate this variable
template<class T>
void SparseVector<T>::matlab(std::ostream &o, const std::string &s) const
{
  o << s << "=sparse(" << nRows() << ",1);\n";
  o << std::showbase << std::scientific;
  STORE::const_iterator it;
  for (it=data_.begin(); it!=data_.end(); it++)
    o << s << "(" << it->first+1 << ") = " << it->second << ";\n";
}

} // end namespace