/usr/include/linbox/matrix/matrixdomain/plain-domain.h is in liblinbox-dev 1.4.2-5build1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 | #ifndef __LINBOX_plain_domain_h
#define __LINBOX_plain_domain_h
#include "linbox/util/error.h"
#include "linbox/matrix/plain-matrix.h"
namespace LinBox {
/*
PlainDomain is a reference implementation of a matrix domain.
It makes a domain from any field by contributing very basic matrix functions using
the getEntry/setEntry functions of the DenseMatrix interface.
PlainDomain has a constructor from field,
it's PlainDomain::Matrix type meets the dense submatrix concept
and it is the interface for working with dense matrices. When you need to allocate a new matrix use
PlainDomain::NewMatrix.
A Blackbox over PlainDomain<Field> is any matrix type that has apply and applyTranspose applicable to PlainDomain<Field>::Matrix. That is to say the signature of apply is
Matrix& Blackbox::apply(Matrix& Y, const Matrix& X)
PlainDomain provides, for dense Submatrices A,B,C
mul(C, A, B) // C = A * B
axpyin(C, A, B) // C += A * B
addin(A, B) // A += B
functions on Submatrix
add sub neg mul div inv axpy (both scalar and matrix
inplace forms of those
areEqual isZero
alternative functions
gemm (rplace axpy) (trans)
additional functions
trsm (trans, uplo) trmm(trans, uplo) symm (trans, uplo)
(in particular triangular solving must be included
*/
template<class Field_>
struct PlainDomain : public Field_
{
typedef PlainDomain<Field_> Self_t;
typedef Field_ Father_t;
typedef size_t Index;
// A domain provides distinct types: Scalar, Matrix, and Blackbox.
typedef typename Father_t::Element Scalar;
typedef typename Father_t::Element Element;
typedef PlainSubmatrix<Self_t> Submatrix;
typedef PlainMatrix<Self_t> Matrix;
// but matrix should admit triangular forms
// constructors and assignment
//PlainDomain (size_t p = 0, size_t e = 1): Field(p) {}
PlainDomain()
{}
PlainDomain(const Element& p)
: Father_t(p) {}
PlainDomain(const Father_t& F)
: Father_t(F) {}
PlainDomain(const PlainDomain<Father_t>& D)
: Father_t(D) {}
//~PlainDomain()
// {} // default dstor is fine.
using Father_t::operator=;
const Father_t& field() const // transitional
{ return *this; }
// A domain provides field functions and (dense) matrix functions.
// field functions
using Father_t::init;
using Father_t::add;
using Father_t::sub;
using Father_t::neg;
using Father_t::mul;
using Father_t::div;
using Father_t::inv;
using Father_t::axpy;
using Father_t::addin;
using Father_t::subin;
using Father_t::negin;
using Father_t::mulin;
using Father_t::divin;
using Father_t::invin;
using Father_t::axpyin;
using Father_t::isZero;
using Father_t::isOne;
using Father_t::areEqual;
using Father_t::cardinality;
using Father_t::characteristic;
using Father_t::write;
using Father_t::read;
// matrix arithmetic functions
Submatrix& add (Submatrix& C, const Submatrix& A, const Submatrix& B) const // C = A + B, where A, B, C have the same shape.
{ C = B; return addin(C,A); }
Submatrix& neg (Submatrix& B, const Submatrix& A) const // B = -A, where B and A have the same shape.
{ B = A; return negin(B); }
Submatrix& sub (Submatrix& C, const Submatrix& A, const Submatrix& B) const // C = A - B, where A, B, C have the same shape.
{ C = B; return subin(C,A); }
Submatrix& smul (Submatrix& B, const Scalar& a, const Submatrix& A) const // B = a*A.
{ B = A; return smulin(B,a); }
Submatrix& saxpy (Submatrix& C, const Scalar& a, const Submatrix& A, const Submatrix& B) const // C = a*A + B.
{ C = B; return saxpyin(C,a,A); }
Submatrix& mul (Submatrix& C, const Submatrix& A, const Submatrix& B) const // C = A*B, conformal shapes required.
{ for (Index i = 0; i < C.rowdim(); ++i)
for (Index j = 0; j < C.coldim(); ++j)
{ Scalar x,y,z; field().assign(x, field().zero); field().assign(y,field().zero); field().assign(z,field().zero);
for (Index k = 0; k < B.coldim(); ++k)
field().axpyin(x, A.getEntry(y,i,k), B.getEntry(z,k,j));
C.setEntry(i,j,x);
}
return C;
}
Submatrix& inv (Submatrix& B, const Submatrix& A) const // B = A^{-1}
{ B = A; return invin(B); }
/*?*/Submatrix& div (Submatrix& C, const Submatrix& A, const Submatrix& B) const // C = A/B, B nonsingular required.
{ throw(LinboxError("PlainDomain - what is div?")); return C; }
Submatrix& axpy( Submatrix& D, const Submatrix& C, const Submatrix& A, const Submatrix &B) const // D = C + A*B, conformal shapes required.
{ D = C; return axpyin(D,A,B); }
Submatrix& addin (Submatrix& B, const Submatrix& A) const // B += A, where B and A have the same shape.
{ for (Index i = 0; i < B.rowdim(); ++i)
for (Index j = 0; j < B.coldim(); ++j)
{ Scalar x = field().zero, y = field().zero;
A.getEntry(x,i,j); B.getEntry(y,i,j);
B.setEntry(i,j, field().addin(x, y));
}
return B;
}
Submatrix& negin (Submatrix& A) const // A = -A.
{ for (Index i = 0; i < A.rowdim(); ++i)
for (Index j = 0; j < A.coldim(); ++j)
{ Scalar x = field().zero;
A.getEntry(x,i,j);
A.setEntry(i,j,field().negin(x));
}
return A;
}
Submatrix& subin (Submatrix& B, const Submatrix& A) const // B -= A, where B and A have the same shape.
{ for (Index i = 0; i < B.rowdim(); ++i)
for (Index j = 0; j < B.coldim(); ++j)
{ Scalar x = field().zero, y = field().zero;
A.getEntry(x,i,j);
B.setEntry(i,j, field().subin(x, y));
}
return B;
}
Submatrix& smulin (Submatrix& B, const Scalar& a) const // B = aB
{ for (Index i = 0; i < B.rowdim(); ++i)
for (Index j = 0; j < B.coldim(); ++j)
{ Scalar x; init(x);
B.getEntry(x,i,j);
B.setEntry(i,j, mulin(x, a));
}
// this could use a _private_ refEntry...
return B;
}
Submatrix& saxpyin( Submatrix& A, const Scalar& a, const Submatrix &B) const // A += a*B, shapes must conform.
{ for (Index i = 0; i < B.rowdim(); ++i)
for (Index j = 0; j < B.coldim(); ++j)
{ Scalar x = field().zero, y = field().zero;
A.getEntry(x,i,j); B.getEntry(y,i,j);
A.setEntry(i,j, field().axpyin(x, a, y));
}
// this could use a _private_ refEntry...
return A;
}
Submatrix& mulin_left (Submatrix& A, const Submatrix& B) const // A = A*B, for square, same dim A and B.
{ Matrix C(*this, B.rowdim(), A.coldim());
mul(C,A,B); return A = C;
}
Submatrix& mulin_right (const Submatrix& A, Submatrix& B) const // B = A*B, for square, same dim A and B.
{ Matrix C(*this, B.rowdim(), A.coldim());
mul(C,A,B); return B = C;
}
/*?*/Submatrix& invin (Submatrix& A) const // A = A^{-1}
{ throw(LinboxError("PlainDomain invin for nonsing Submatrix not yet impl.")); return A; }
/*?*/Submatrix& divin (Submatrix& A, const Submatrix& B) const // A /= B, B nonsingular required.
{ throw(LinboxError("PlainDomain what is divin?")); return A; }
Submatrix& axpyin( Submatrix& C, const Submatrix& A, const Submatrix &B) const // C += A*B, shapes must conform.
{ for (Index i = 0; i < C.rowdim(); ++i)
for (Index j = 0; j < C.coldim(); ++j)
{ Scalar x,y,z; field().assign(x, field().zero); field().assign(y,field().zero); field().assign(z,field().zero);
C.getEntry(x,i,j);
for (Index k = 0; k < B.coldim(); ++k)
field().axpyin(x, A.getEntry(y,i,k), B.getEntry(z,k,j));
C.setEntry(i,j,x);
}
return C;
}
//Submatrix& copy(Submatrix &dst, const Submatrix& src) // deep copy, same as assignment
bool areEqual(const Submatrix& A, const Submatrix &B) const // A == B, same shape not required
{
Scalar a = field().zero, b = field().zero;
for (size_t i = 0; i < A.rowdim(); ++i)
for (size_t j = 0; j < A.coldim(); ++j)
{ A.getEntry(a, i, j);
B.getEntry(b, i, j);
if (not field().areEqual(a, b) )
return false;
}
return true;
}
// simple write
std::ostream& write(std::ostream& out, const Submatrix& A) const
{
Scalar a;
//out << "%%MatrixMarketExtended array" << std::endl;
out << A.rowdim() << " " << A.coldim() << std::endl;
for (size_t i = 0; i < A.rowdim(); ++i){
for (size_t j = 0; j < A.coldim(); ++j)
write(out, A.getEntry(a, i, j)) << " ";
out << std::endl;
}
return out << std::endl;
}
// simple read. should use matrix reader
std::istream& read(std::istream& in, Submatrix& A) const
{
Scalar a;
size_t r, c;
in >> r >> c;
A.init(r, c);
for (size_t i = 0; i < A.rowdim(); ++i)
for (size_t j = 0; j < A.coldim(); ++j){
read(in, a);
A.setEntry(i, j, a);
}
return in;
}
}; // PlainDomain
} // LinBox
#endif // __LINBOX_plain_domain_h
|