⭐ 欢迎来到虫虫下载站! | 📦 资源下载 📁 资源专辑 ℹ️ 关于我们
⭐ 虫虫下载站

📄 mvmd.cc

📁 sparselib库
💻 CC
字号:
/*+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++*//*                                                                           *//*                                                                           *//*                   MV++ Numerical Matrix/Vector C++ Library                *//*                             MV++ Version 1.5                              *//*                                                                           *//*                                  R. Pozo                                  *//*               National Institute of Standards and Technology              *//*                                                                           *//*                                  NOTICE                                   *//*                                                                           *//* Permission to use, copy, modify, and distribute this software and         *//* its documentation for any purpose and without fee is hereby granted       *//* provided that this permission notice appear in all copies and             *//* supporting documentation.                                                 *//*                                                                           *//* Neither the Institution (National Institute of Standards and Technology)  *//* nor the author makes any representations about the suitability of this    *//* software for any purpose.  This software is provided ``as is''without     *//* expressed or implied warranty.                                            *//*                                                                           *//*+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++*/////          Basic matrix class (double precision)//#include <iostream>#include "mvmd.h" int MV_ColMat_double::dim(int i) const {    if (i==0) return dim0_;    if (i==1) return dim1_;    else    {     std::cerr << "Called MV_ColMat::dim(" << i << ")  must be 0 or 1 " << "\n";     exit(1);    }    // never should be here, but many compilers warn about not    // returning a value    return 0;}// NOTE: null construct have ref_ flag turned OFF, otherwise, we can//          never reset the dim of matrix....MV_ColMat_double::MV_ColMat_double()  : v_(), dim0_(0), dim1_(0) , lda_(0), ref_(0){}                                                                MV_ColMat_double::MV_ColMat_double( int m,  int n) : v_(m*n),        dim0_(m), dim1_(n), lda_(m), ref_(0) {}MV_ColMat_double::MV_ColMat_double( int m,  int n, const double &s) : v_(m*n),        dim0_(m), dim1_(n), lda_(m), ref_(0) {    operator=(s);}// operators and member functionsMV_ColMat_double& MV_ColMat_double::operator=(const double & s) {    int M = dim(0);    int N = dim(1);    if (lda_ == M)      // if continuous, then just assign as a ?        v_ =  s;        // single long vector.    else    {    // this should run much faster than the just accessing each (i,j)    // element individually     //        MV_VecIndex I(0,M-1);        for (int j=0; j<N; j++)        {            v_(I) = s;            I += lda_;        }    }    return *this;}MV_ColMat_double& MV_ColMat_double::newsize( int M,  int N){    v_.newsize(M*N);    dim0_ = M;    dim1_ = N;    lda_ = M;    return *this;}MV_ColMat_double& MV_ColMat_double::operator=(const MV_ColMat_double & m) {    int lM = dim0_;     // left hand arg  (this)    int lN = dim1_;    int rM = m.dim0_;   // right hand arg (m)    int rN = m.dim1_;    // if the left-hand side is a matrix reference, the we copy the    // elements of m *into* the region specfied by the reference.    // i.e. inject().    if (ref_)    {        // check conformance,               if (lM != rM  || lN != rN)              {            std::cerr << "MV_ColMatRef::operator=  non-conformant assignment.\n";            exit(1);        }    }    else    {        newsize(rM,rN);    }    // at this point the left hand and right hand sides are conformant    // this should run much faster than the just accessing each (i,j)    // element individually     // if both sides are contigous, then just copy as one vector    if ( lM == lda_ && rM == m.lda_)    {        MV_VecIndex I(0,rM*rN-1);        v_(I) = m.v_(I);    }    else    {        // slower way...        MV_VecIndex I(0,rM-1);        MV_VecIndex K(0,rM-1);        for (int j=0; j<rN; j++)        {            v_(I) = m.v_(K);            I += lda_;            K += m.lda_;        }    }    return *this;   }MV_ColMat_double::MV_ColMat_double(const MV_ColMat_double & m) :         v_(m.dim0_*m.dim1_), dim0_(m.dim0_),        dim1_(m.dim1_), lda_(m.dim0_), ref_(0){    int M = m.dim0_;    int N = m.dim1_;    // this should run much faster than the just accessing each (i,j)    // element individually     MV_VecIndex I(0,M-1);    MV_VecIndex K(0,M-1);    for (int j=0; j<N; j++)    {        v_(I) = m.v_(K);        I += lda_;        K += m.lda_;    }}MV_ColMat_double::MV_ColMat_double(double* d,  int m,  int n) :    v_(m*n), dim0_(m), dim1_(n), lda_(m), ref_(0){    int mn = m*n;    // d is contiguous, so just copy 1-d vector    for (int i=0; i< mn; i++)            v_[i] = d[i];}MV_ColMat_double::MV_ColMat_double(double* d,  int m,  int n,          int lda) :    v_(m*n), dim0_(m), dim1_(n), lda_(lda), ref_(0){    for (int j=0; j< n; j++)        for (int i=0; i<m; i++)            operator()(i,j) = d[j*lda + i];   // could be made faster!!}MV_ColMat_double MV_ColMat_double::operator()(const MV_VecIndex &I, const MV_VecIndex &J){    // check that index is not out of bounds    //    if (I.end() >= dim0_  || J.end() >= dim1_)    {        std::cerr << "Matrix index: (" << I.start() << ":" << I.end()               << "," << J.start() << ":" << J.end()                << ") not a subset of (0:" << dim0_ - 1 << ", 0:"              << dim1_-1 << ") " << "\n";        exit(1);    }    // this automatically returns a reference    //     return MV_ColMat_double(&v_[J.start()*lda_ + I.start()],             I.end() - I.start() + 1,             J.end() - J.start() + 1, lda_, MV_Matrix_::ref);}const MV_ColMat_double MV_ColMat_double::operator()(const MV_VecIndex &I,     const MV_VecIndex &J) const{    std::cerr << "Const operator()(MV_VecIndex, MV_VecIndex) called " << "\n";    // check that index is not out of bounds    //    if (I.end() >= dim0_  || J.end() >= dim1_)    {        std::cerr << "Matrix index: (" << I.start() << ":" << I.end()               << "," << J.start() << ":" << J.end()                << ") not a subset of (0:" << dim0_ - 1 << ", 0:"              << dim1_-1 << ") " << "\n";        exit(1);    }    // this automatically returns a reference.  we need to     // "cast away" constness here, so the &v_[] arg will    // not cause a compiler error.    //    MV_ColMat_double *t =  (MV_ColMat_double*) this;    return MV_ColMat_double(&(t->v_[J.start()*lda_ + I.start()]),             I.end() - I.start() + 1,             J.end() - J.start() + 1, lda_, MV_Matrix_::ref);}MV_ColMat_double::~MV_ColMat_double() {}std::ostream&   operator<<(std::ostream& s, const MV_ColMat_double& V){    int M = V.dim(0);    int N = V.dim(1);    for (int i=0; i<M; i++)    {        for (int j=0; j<N; j++)            s << V(i,j) << " " ;        s << "\n";    }        return s;}

⌨️ 快捷键说明

复制代码 Ctrl + C
搜索代码 Ctrl + F
全屏模式 F11
切换主题 Ctrl + Shift + D
显示快捷键 ?
增大字号 Ctrl + =
减小字号 Ctrl + -