LA_library/vec.h

1740 lines
56 KiB
C
Raw Normal View History

2010-09-08 18:27:58 +02:00
/* vim: set ts=8 sw=8 sts=8 noexpandtab cindent: */
/*******************************************************************************
2008-02-26 14:55:23 +01:00
LA: linear algebra C++ interface library
Copyright (C) 2008 Jiri Pittner <jiri.pittner@jh-inst.cas.cz> or <jiri@pittnerovi.com>
complex versions written by Roman Curik <roman.curik@jh-inst.cas.cz>
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
2010-09-08 18:27:58 +02:00
*******************************************************************************/
2004-03-17 04:07:21 +01:00
#ifndef _LA_VEC_H_
#define _LA_VEC_H_
2005-02-14 01:10:07 +01:00
#include "la_traits.h"
2009-11-12 22:01:19 +01:00
namespace LA {
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* forward declarations
******************************************************************************/
template <typename T> void lawritemat(FILE *file, const T *a, int r, int c,
const char *form0, int nodim, int modulo, int issym);
2004-03-17 04:07:21 +01:00
template <typename T> class NRPerm;
2021-05-14 17:39:22 +02:00
template <typename T> class CyclePerm;
/***************************************************************************//**
* auxiliary macro to avoid compilation errors for some types
******************************************************************************/
template <typename T>
inline typename LA_traits<T>::normtype MYABS(const T &x) {return abs(x);}
template <> inline unsigned char MYABS(const unsigned char &x) {return x;}
template <> inline unsigned short MYABS(const unsigned short &x) {return x;}
template <> inline unsigned int MYABS(const unsigned int &x) {return x;}
template <> inline unsigned long MYABS(const unsigned long &x) {return x;}
template <> inline unsigned long long MYABS(const unsigned long long &x) {return x;}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* static constants used in several cblas-routines
******************************************************************************/
2021-04-21 15:04:37 +02:00
const static std::complex<double> CONE = 1.0, CMONE = -1.0, CZERO = 0.0;
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
const static cuDoubleComplex CUONE = {1.,0.}, CUMONE = {-1.,0.}, CUZERO = {0.,0.};
#endif
2004-03-17 04:07:21 +01:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* macros to construct binary operators +,-,*, from +=, -=, *=
* for 3 cases: X + a, a + X, X + Y
******************************************************************************/
2004-03-17 04:07:21 +01:00
#define NRVECMAT_OPER(E,X) \
template<class T> \
inline const NR##E<T> NR##E<T>::operator X(const T &a) const \
{ return NR##E(*this) X##= a; } \
\
template<class T> \
inline const NR##E<T> operator X(const T &a, const NR##E<T> &rhs) \
{ return NR##E<T>(rhs) X##= a; }
#define NRVECMAT_OPER2(E,X) \
template<class T> \
2010-06-25 17:28:19 +02:00
inline const NR##E<T> NR##E<T>::operator X(const NR##E<T> &a) const \
2004-03-17 04:07:21 +01:00
{ return NR##E(*this) X##= a; }
2004-03-17 17:39:07 +01:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* \brief NRVec<T> class template implementing the vector interface
* @see NRMat<T>, NRSMat<T>
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
class NRVec {
protected:
2010-09-08 18:27:58 +02:00
int nn;//!< size of the vector
T *v;//!< pointer to the underlying data structure
int *count;//!< pointer to the reference-counter
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
GPUID location;//!< determines the memory address space of this object (CPU/GPU)
2010-06-25 17:28:19 +02:00
#endif
2004-03-17 04:07:21 +01:00
public:
friend class NRSMat<T>;
friend class NRMat<T>;
2021-04-21 15:04:37 +02:00
template <typename U> friend NRVec<std::complex<U> > complexify(const NRVec<U>&);
2004-03-17 04:07:21 +01:00
2010-09-08 18:27:58 +02:00
typedef T ROWTYPE;
//! standard destructor
~NRVec();
2010-06-25 17:28:19 +02:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* inlined constructor creating zero vector of general type <code>T</code>
******************************************************************************/
inline NRVec(): nn(0), v(0), count(0) {
#ifdef CUDALA
location = DEFAULT_LOC;
#endif
};
/***************************************************************************//**
* Explicit inlined constructor creating vector of given size and location.
* Because of performance reasons, no incialization is done.
* @param[in] n vector size (count of elements)
* @param[in] loc location of the underlying data (CPU/GPU)
******************************************************************************/
explicit inline NRVec(const int n, const GPUID loc = undefined): nn(n), count(new int(1)) {
#ifdef CUDALA
location = (loc == undefined)?DEFAULT_LOC:loc;
if(location == cpu){
#endif
v = new T[n];
#ifdef CUDALA
}else{
v = (T*) gpualloc(n*sizeof(T));
}
#endif
};
//! inlined constructor creating vector of given size filled with prescribed value
2004-03-17 04:07:21 +01:00
inline NRVec(const T &a, const int n);
2010-09-08 18:27:58 +02:00
//! inlined constructor creating vector of given size filled with data located at given memory location
2006-09-12 01:07:22 +02:00
inline NRVec(const T *a, const int n);
2010-09-08 18:27:58 +02:00
//! inlined constructor creating vector of given size filled with data located at given memory location
2006-09-12 01:07:22 +02:00
inline NRVec(T *a, const int n, bool skeleton);
2010-09-08 18:27:58 +02:00
//! inlined copy constructor
2004-03-17 04:07:21 +01:00
inline NRVec(const NRVec &rhs);
2010-09-08 18:27:58 +02:00
//! complexifying constructor
NRVec(const typename LA_traits_complex<T>::NRVec_Noncomplex_type &rhs, bool imagpart=false);//construct complex from real
//! explicit inlined constructor converting symmetric matrix into a vector
2004-03-17 04:07:21 +01:00
inline explicit NRVec(const NRSMat<T> & S);
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
******************************************************************************/
#ifdef MATPTR
explicit NRVec(const NRMat<T> &rhs): NRVec(&rhs[0][0], rhs.nrows()*rhs.ncols()) {};
#else
explicit NRVec(const NRMat<T> &rhs);
#endif
/***************************************************************************//**
* routines for CUDA related stuff
* \li <code>getlocation()</code> gets the protected data member location
* \li <code>moveto(const GPUID)</code> moves underlying data between CPU/GPU memory
******************************************************************************/
#ifdef CUDALA
inline GPUID getlocation() const { return location; }
void moveto(const GPUID dest);
#else
inline GPUID getlocation() const { return cpu; }
void moveto(const GPUID dest) {};
#endif
//! create separate copy of the data corresponding to this vector
2013-11-04 15:56:39 +01:00
void copyonwrite(bool detachonly=false);
2010-09-08 18:27:58 +02:00
//! purge this vector
2021-05-24 18:45:58 +02:00
void clear() { copyonwrite(LA_traits<T>::is_plaindata()); LA_traits<T>::clear(v, nn); };
2010-09-08 18:27:58 +02:00
//! assignment operator assigns given vector
NRVec& operator=(const NRVec &rhs);
//! assigment operator assigns given scalar to each element of this vector
NRVec& operator=(const T &a);
//! fills in this vector with pseudo-random numbers generated using uniform distribution
2009-09-04 10:09:32 +02:00
void randomize(const typename LA_traits<T>::normtype &x);
2010-09-08 18:27:58 +02:00
//! perform deep-copy of given vector
NRVec& operator|=(const NRVec &rhs);
//! relational operators
2021-06-24 17:08:55 +02:00
const bool operator!=(const NRVec &rhs) const
{
if(nn!=rhs.nn) return 1;
if(LA_traits<T>::is_plaindata()) return LA_traits<T>::gencmp(v,rhs.v,nn);
else
{
for(int i=0; i<nn; ++i) if(v[i]!=rhs.v[i]) return 1;
return 0;
}
}
2005-02-14 01:10:07 +01:00
const bool operator==(const NRVec &rhs) const {return !(*this != rhs);};
2005-09-06 17:55:07 +02:00
const bool operator>(const NRVec &rhs) const;
const bool operator<(const NRVec &rhs) const;
const bool operator>=(const NRVec &rhs) const {return !(*this < rhs);};
const bool operator<=(const NRVec &rhs) const {return !(*this > rhs);};
2010-09-08 18:27:58 +02:00
//! unary minus
2004-03-17 04:07:21 +01:00
const NRVec operator-() const;
2010-09-08 18:27:58 +02:00
//! bunch of vector-vector arithmetic operators defined element-wise
inline NRVec& operator+=(const NRVec &rhs);
inline NRVec& operator-=(const NRVec &rhs);
inline NRVec& operator*=(const NRVec &rhs);
inline NRVec& operator/=(const NRVec &rhs);
2004-03-17 04:07:21 +01:00
inline const NRVec operator+(const NRVec &rhs) const;
inline const NRVec operator-(const NRVec &rhs) const;
2010-09-08 18:27:58 +02:00
//! bunch of scalar-vector arithmetic operators defined element-wise
inline NRVec& operator+=(const T &a);
inline NRVec& operator-=(const T &a);
inline NRVec& operator*=(const T &a);
inline NRVec& operator/=(const T &a);
2010-09-08 18:27:58 +02:00
2004-03-17 04:07:21 +01:00
inline const NRVec operator+(const T &a) const;
inline const NRVec operator-(const T &a) const;
inline const NRVec operator*(const T &a) const;
inline const NRVec operator/(const T &a) const;
2010-09-08 18:27:58 +02:00
//! determine the actual value of the reference counter
inline int getcount() const {return count?*count:0;}
//! compute the Euclidean inner product (with conjugation in complex case)
inline const T operator*(const NRVec &rhs) const;
inline const T dot(const NRVec &rhs) const {return *this * rhs;};
//! compute the Euclidean inner product (with conjugation in complex case) with a stride-vector
inline const T dot(const T *a, const int stride = 1) const;
2005-02-18 23:08:15 +01:00
void gemv(const T beta, const NRMat<T> &a, const char trans, const T alpha, const NRVec &x);
2010-09-08 18:27:58 +02:00
void gemv(const T beta, const NRSMat<T> &a, const char trans /**< just for compatibility reasons */, const T alpha, const NRVec &x);
void gemv(const T beta, const SparseMat<T> &a, const char trans, const T alpha, const NRVec &x,const bool treat_as_symmetric = false);
void gemv( const typename LA_traits_complex<T>::Component_type beta,
const typename LA_traits_complex<T>::NRMat_Noncomplex_type &a,
const char trans,
const typename LA_traits_complex<T>::Component_type alpha,
const NRVec &x);
void gemv( const typename LA_traits_complex<T>::Component_type beta,
const typename LA_traits_complex<T>::NRSMat_Noncomplex_type &a,
const char trans,
const typename LA_traits_complex<T>::Component_type alpha, const NRVec &x);
//! multiply given matrix with this vector from left
const NRVec operator*(const NRMat<T> &mat) const {
SAME_LOC(*this, mat);
NRVec<T> result(mat.ncols(), mat.getlocation());
result.gemv((T)0, mat, 't', (T)1, *this);
return result;
};
//! multiply given symmetric matrix in packed form with this vector from left
const NRVec operator*(const NRSMat<T> &mat) const {
SAME_LOC(*this, mat);
NRVec<T> result(mat.ncols(), mat.getlocation());
result.gemv((T)0, mat, 't', (T)1, *this);
return result;
};
//! multiply given sparse matrix with this vector from left
const NRVec operator*(const SparseMat<T> &mat) const {
NOT_GPU(*this);
NRVec<T> result(mat.ncols());
result.gemv((T)0, mat, 't', (T)1, *this);
return result;
};
//! compute the outer product of two vectors
const NRMat<T> otimes(const NRVec<T> &rhs, const bool conjugate = false, const T &scale = 1) const;
//! opeartor for outer product computation
inline const NRMat<T> operator|(const NRVec<T> &rhs) const { return otimes(rhs,true); };
//! compute the sum of the vector elements
inline const T sum() const {
2021-05-14 17:39:22 +02:00
T sum(v[0]);
for(register int i=1; i<nn; i++){ sum += v[i];}
2010-09-08 18:27:58 +02:00
return sum;
};
2021-05-14 17:39:22 +02:00
//! compute the product of the vector elements
inline const T prod() const {
T prod(v[0]);
for(register int i=1; i<nn; i++){ prod *= v[i];}
return prod;
};
//! permute vector elements
2021-05-19 22:29:47 +02:00
const NRVec permuted(const NRPerm<int> &p, const bool inverse=false) const;
void permuteme(const CyclePerm<int> &p); //in place
2010-09-08 18:27:58 +02:00
//! compute the sum of the absolute values of the elements of this vector
inline const typename LA_traits<T>::normtype asum() const;
//! indexing operator - index running from zero
2004-03-17 04:07:21 +01:00
inline T & operator[](const int i);
inline const T & operator[](const int i) const;
2010-09-08 18:27:58 +02:00
//! dummy routine
inline void setcoldim(int i) {};
//! get the pointer to the underlying data structure
inline operator T*();
2021-04-21 15:04:37 +02:00
2010-09-08 18:27:58 +02:00
//! get the constant pointer to the underlying data structure
inline operator const T*() const;
//! add up a scalar multiple of a given vector
void axpy(const T alpha, const NRVec &x);
//! add up a scalar multiple of a given vector with given stride
void axpy(const T alpha, const T *x, const int stride=1);
//! determine the number of elements
2004-03-17 04:07:21 +01:00
inline int size() const;
2010-09-08 18:27:58 +02:00
//! resize the current vector, optionally preserving data
void resize(const int n, const bool preserve=false);
2010-09-08 18:27:58 +02:00
2011-01-18 15:37:05 +01:00
//!deallocate the current vector
void dealloc(void) {resize(0);}
2010-09-08 18:27:58 +02:00
//! determine the norm of this vector
2009-09-04 10:09:32 +02:00
inline const typename LA_traits<T>::normtype norm() const;
2010-09-08 18:27:58 +02:00
//! normalize this vector and optionally save the norm
NRVec& normalize(typename LA_traits<T>::normtype* norm = 0);
//! get normalized copy of this vector
2004-03-17 04:07:21 +01:00
inline const NRVec unitvector() const;
2010-09-08 18:27:58 +02:00
//! determine the maximal element (in the absolute value) of this vector
inline const T amax() const;
//! determine the minimal element (in the absolute value) of this vector
inline const T amin() const;
2021-05-19 22:29:47 +02:00
//! determine the maximal element of this vector
const T max() const;
//! determine the minimal element of this vector
const T min() const;
2010-09-08 18:27:58 +02:00
//! routine for formatted output
2004-03-17 04:07:21 +01:00
void fprintf(FILE *f, const char *format, const int modulo) const;
2010-09-08 18:27:58 +02:00
//! routine for unformatted output
void put(int fd, bool dimensions=1, bool transp=0) const;
//! routine for formatted input
2004-03-17 04:07:21 +01:00
void fscanf(FILE *f, const char *format);
2010-09-08 18:27:58 +02:00
//! routine for unformatted input
void get(int fd, bool dimensions=1, bool transp=0);
//! constructor creating vector from sparse matrix
explicit NRVec(const SparseMat<T> &rhs);
//! routine for compatibility with sparse types
inline void simplify() {};
//! determine whether the i<sup>th</sup> element is bigger than the j<sup>th</sup> element
bool bigger(int i, int j) const {
NOT_GPU(*this);
return LA_traits<T>::bigger(v[i], v[j]);
};
//! determine whether the i<sup>th</sup> element is bigger than the j<sup>th</sup> element
bool smaller(int i, int j) const {
NOT_GPU(*this);
return LA_traits<T>::smaller(v[i], v[j]);
};
//! swap the i<sup>th</sup> and j<sup>th</sup> element
void swap(int i, int j) {
const T tmp(v[i]);
v[i] = v[j];
v[j] = tmp;
};
//! sort by default in ascending order and return the parity of corresponding permutation resulting to this order
int sort(int direction = 0, int from = 0, int to = -1, int *perm = NULL);
2021-05-19 22:29:47 +02:00
int sort(int direction, NRPerm<int> &perm);
2010-09-08 18:27:58 +02:00
//! apply given function to each element
NRVec& call_on_me(T (*_F)(const T &) ){
NOT_GPU(*this);
copyonwrite();
for(int i=0; i<nn; ++i) v[i] = _F(v[i]);
return *this;
};
void swap(NRVec &rhs) //more efficient swap than via tmp and constructors and operator=
{
int tmpnn=nn; nn=rhs.nn; rhs.nn=tmpnn;
T *tmpv=v; v=rhs.v; rhs.v=tmpv;
int *tmpcount=count; count=rhs.count; rhs.count=tmpcount;
#ifdef CUDALA
GPUID tmplocation=location; location=rhs.location; rhs.location=tmplocation;
#endif
}
NRVec diffabs(const NRVec &rhs) const; //difference of absolute values
2004-03-17 04:07:21 +01:00
};
2005-02-18 23:08:15 +01:00
2010-06-25 17:28:19 +02:00
/***************************************************************************//**
* implements \c NRVec<T> functionality with indexing from 1
* all possible constructors have to be given explicitly, other stuff is inherited
* with exception of the operator() which differs
******************************************************************************/
template<typename T>
class NRVec_from1 : public NRVec<T> {
public:
NRVec_from1(): NRVec<T>() {};
explicit NRVec_from1(const int n): NRVec<T>(n) {};
NRVec_from1(const NRVec<T> &rhs): NRVec<T>(rhs) {};//!< be able to convert the parent class transparently to this
NRVec_from1(const T &a, const int n): NRVec<T>(a, n) {};
NRVec_from1(const T *a, const int n): NRVec<T>(a, n) {};
inline const T& operator[] (const int i) const;
inline T& operator[] (const int i);
};
}//namespace
//due to mutual includes this has to be after full class declaration
#include "mat.h"
#include "smat.h"
#include "sparsemat.h"
#include "sparsesmat.h"
//needs NRVec_from1
#include "permutation.h"
namespace LA {
/***************************************************************************//**
* indexing operator giving the element at given position with range checking in
* the DEBUG mode
* @param[in] i position of the required vector element (starting from 0)
* @return reference to the requested element
******************************************************************************/
template <typename T>
inline T& NRVec_from1<T>::operator[](const int i) {
#ifdef DEBUG
if(_LA_count_check && *NRVec<T>::count != 1) laerror("possible use of NRVec[] with count>1 as l-value");
if(i < 1 || i > NRVec<T>::nn) laerror("out of range");
if(!NRVec<T>::v) laerror("unallocated NRVec");
#endif
NOT_GPU(*this);
return NRVec<T>::v[i-1];
}
/***************************************************************************//**
* indexing operator giving the element at given position with range checking in
* the DEBUG mode
* @param[in] i position of the required vector element (starting from 0)
* @return constant reference to the requested element
******************************************************************************/
template <typename T>
inline const T& NRVec_from1<T>::operator[](const int i) const {
#ifdef DEBUG
if(i < 1 || i > NRVec<T>::nn) laerror("out of range");
if(!NRVec<T>::v) laerror("unallocated NRVec");
#endif
NOT_GPU(*this);
return NRVec<T>::v[i-1];
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* output operator
* @param[in,out] s output stream
* @param[in] x vector of general type intended for output
* @return modified stream
******************************************************************************/
2006-10-21 17:32:53 +02:00
template <typename T>
2010-09-08 18:27:58 +02:00
std::ostream & operator<<(std::ostream &s, const NRVec<T> &x) {
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(x.getlocation() == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
const int n = x.size();
s << n << std::endl;
for(register int i = 0; i<n; i++){
s << (typename LA_traits_io<T>::IOtype)x[i] << (i == n-1 ? '\n' : ' ');
}
return s;
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
NRVec<T> tmp(x);
tmp.moveto(cpu);
return s << tmp;
}
2010-06-25 17:28:19 +02:00
#endif
2006-10-21 17:32:53 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* input operator
* @param[in,out] s input stream
* @param[in] x vector of general type intended for input
* @return modified stream
******************************************************************************/
2006-10-21 17:32:53 +02:00
template <typename T>
2010-09-08 18:27:58 +02:00
std::istream & operator>>(std::istream &s, NRVec<T> &x) {
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(x.getlocation() == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
int i,n;
s >> n;
x.resize(n);
typename LA_traits_io<T>::IOtype tmp;
for(i=0; i<n; i++){
s >> tmp;
x[i] = tmp;
}
return s;
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
NRVec<T> tmp;
tmp.moveto(cpu);
s >> tmp;
tmp.moveto(x.getlocation());
x = tmp;
return s;
}
2010-06-25 17:28:19 +02:00
#endif
2006-10-21 17:32:53 +02:00
}
2004-03-17 04:07:21 +01:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* inline constructor creating vector of given size filled with prescribed value
* @param[in] a value to be assigned to all vector elements
* @param[in] n required vector size
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>::NRVec(const T& a, const int n): nn(n), count(new int) {
2004-03-17 04:07:21 +01:00
*count = 1;
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
location = DEFAULT_LOC;
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
v = new T[n];
if(a != (T)0){
for(register int i=0; i<n; i++) v[i] = a;
}else{
memset(v, 0, nn*sizeof(T));
}
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
v = (T*) gpualloc(n*sizeof(T));
smart_gpu_set(n, a, v);
2010-06-25 17:28:19 +02:00
}
#endif
2004-03-17 04:07:21 +01:00
}
2010-06-25 17:28:19 +02:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* inline constructor creating vector of given size filled with given data
* @param[in] a pointer to the data
* @param[in] n required vector size
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>::NRVec(const T *a, const int n): nn(n), count(new int) {
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
location = DEFAULT_LOC;
if(location == cpu) {
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
v = new T[n];
*count = 1;
memcpy(v, a, n*sizeof(T));
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
v = (T*) gpualloc(n*sizeof(T));
cublasSetVector(n, sizeof(T), a, 1, v, 1);
TEST_CUBLAS("cublasSetVector");
}
2010-06-25 17:28:19 +02:00
#endif
2006-09-12 01:07:22 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* inline constructor creating vector of given size filled with given data
* @param[in] a pointer to the data
* @param[in] n required vector size
* @param[in] skeleton if equal to true, only the internal data pointer is modified
* and reference counter is set to two, i.e. no data deallocation occurs in destructor
******************************************************************************/
2006-09-12 01:07:22 +02:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>::NRVec(T *a, const int n, bool skeleton) : nn(n), count(new int) {
if(!skeleton){
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
location = DEFAULT_LOC;
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
v = new T[n];
*count = 1;
memcpy(v, a, n*sizeof(T));
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
v= (T*) gpualloc(n*sizeof(T));
cublasSetVector(n, sizeof(T), a, 1, v, 1);
TEST_CUBLAS("cublasSetVector");
2006-09-12 01:07:22 +02:00
}
2010-09-08 18:27:58 +02:00
#endif
}else{
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location != cpu) laerror("NRVec() with skeleton option cannot be on GPU");
2010-06-25 17:28:19 +02:00
#endif
2006-09-12 01:07:22 +02:00
*count = 2;
2010-09-08 18:27:58 +02:00
v = a;
}
2004-03-17 04:07:21 +01:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* inline copy constructor
* @param[in] rhs reference vector being copied
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>::NRVec(const NRVec<T> &rhs) {
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
location = rhs.location;
2010-06-25 17:28:19 +02:00
#endif
2004-03-17 04:07:21 +01:00
v = rhs.v;
nn = rhs.nn;
count = rhs.count;
if(count) (*count)++;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* inline constructor interpreting symmetric matrix of order \f$n\f$ stored in packed form
* as a linear vector consisting of \f$n(n+1)/2\f$ elements
* @param[in] rhs symmetric matrix of type <code>NRSMat<T></code>
* @see NRSMat<T>
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>::NRVec(const NRSMat<T> &rhs) {
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
location = rhs.location;
2010-06-25 17:28:19 +02:00
#endif
2004-03-17 04:07:21 +01:00
nn = rhs.nn;
2010-09-08 18:27:58 +02:00
//! using macro NN2 defined in smat.h
2004-03-17 04:07:21 +01:00
nn = NN2;
v = rhs.v;
count = rhs.count;
(*count)++;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* adds given scalar value of type T to all vector elements
* @param[in] a scalar value being added
* @return reference to the modified vector
******************************************************************************/
2004-03-17 06:34:59 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T> & NRVec<T>::operator+=(const T &a) {
2010-06-25 17:28:19 +02:00
NOT_GPU(*this);
2010-09-08 18:27:58 +02:00
2004-03-17 06:34:59 +01:00
copyonwrite();
2010-09-08 18:27:58 +02:00
if(a != (T)0){ for(register int i=0; i<nn; ++i) v[i] += a; }
2004-03-17 06:34:59 +01:00
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* subtracts given scalar value of type T from all vector elements
* @param[in] a scalar value being subtracted
* @return reference to the modified vector
******************************************************************************/
2004-03-17 06:34:59 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>& NRVec<T>::operator-=(const T &a) {
2010-06-25 17:28:19 +02:00
NOT_GPU(*this);
2004-03-17 06:34:59 +01:00
2010-09-08 18:27:58 +02:00
copyonwrite();
if(a != (T)0){ for(register int i=0; i<nn; ++i) v[i] -= a; }
return *this;
}
2004-03-17 06:34:59 +01:00
2004-03-17 04:07:21 +01:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* adds a vector \f$\vec{y}\f$ of general type <code>T</code> to this vector \f$\vec{x}\f$
* \f[\vec{x}\leftarrow\vec{x}+\vec{y}\f]
* @param[in] rhs vector \f$\vec{y}\f$ of type <code>T</code>
* @return reference to the modified vector
******************************************************************************/
2004-03-17 06:34:59 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>& NRVec<T>::operator+=(const NRVec<T> &rhs) {
2004-03-17 06:34:59 +01:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
2004-03-17 06:34:59 +01:00
#endif
2010-09-08 18:27:58 +02:00
NOT_GPU(*this);
NOT_GPU(rhs);
copyonwrite();
for(register int i=0; i<nn; ++i) v[i] += rhs.v[i];
return *this;
2004-03-17 06:34:59 +01:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* multiplies this vector \f$\vec{y}\f$ componentwise by general vector \f$\vec{x}\f$
* \f[\vec{x}_i = \vec{x}_i\times\vec{y}_i\f]
* @param[in] rhs general vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
2006-04-06 23:45:51 +02:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>& NRVec<T>::operator*=(const NRVec<T>& rhs) {
2006-04-06 23:45:51 +02:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
2006-04-06 23:45:51 +02:00
#endif
2010-09-08 18:27:58 +02:00
NOT_GPU(*this);
NOT_GPU(rhs);
copyonwrite();
for(register int i=0; i<nn; ++i) v[i] *= rhs.v[i];
2006-04-06 23:45:51 +02:00
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* divides this vector \f$\vec{y}\f$ componentwise by general vector \f$\vec{x}\f$
* \f[\vec{x}_i = \vec{x}_i\slash\vec{y}_i\f]
* @param[in] rhs general vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
2006-04-06 23:45:51 +02:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T> & NRVec<T>::operator/=(const NRVec<T> &rhs) {
2006-04-06 23:45:51 +02:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
2006-04-06 23:45:51 +02:00
#endif
2010-09-08 18:27:58 +02:00
NOT_GPU(*this);
NOT_GPU(rhs);
2006-04-06 23:45:51 +02:00
copyonwrite();
2010-09-08 18:27:58 +02:00
for(register int i=0; i<nn; ++i) v[i] /= rhs.v[i];
2006-04-06 23:45:51 +02:00
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* subtracts given vector \f$\vec{y}\f$ from this vector \f$\vec{x}\f$
* \f[\vec{x}_i = \vec{x}_i-\vec{y}_i\f]
* @param[in] rhs vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
2004-03-17 06:34:59 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T> & NRVec<T>::operator-=(const NRVec<T> &rhs) {
2004-03-17 06:34:59 +01:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
2004-03-17 06:34:59 +01:00
#endif
2010-09-08 18:27:58 +02:00
NOT_GPU(*this);
NOT_GPU(rhs);
copyonwrite();
for(register int i=0; i<nn; ++i) v[i] -= rhs.v[i];
2004-03-17 06:34:59 +01:00
return *this;
}
/***************************************************************************//**
* difference of elements of two vectors in absolute values
* \f[\vec{z}_i = \vec{x}_i-\vec{y}_i\f]
* @param[in] rhs vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
template <typename T>
NRVec<T> NRVec<T>::diffabs(const NRVec<T> &rhs) const {
#ifdef DEBUG
if (nn != rhs.nn) laerror("incompatible dimensions");
#endif
NOT_GPU(*this);
NOT_GPU(rhs);
NRVec<T> r(nn);
for(int i=0; i<nn; ++i) r[i] = MYABS(v[i]) - MYABS(rhs.v[i]);
return r;
}
2004-03-17 06:34:59 +01:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* multiply this general vector \f$\vec{x}\f$ by scalar value \f$\lambda\f$
* \f[\vec{x}_i \leftarrow \lambda\vec{x}_i\f]
* @param[in] a scalar value \f$\lambda\f$
* @return reference to the modified vector
******************************************************************************/
2004-03-17 06:34:59 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T> & NRVec<T>::operator*=(const T &a) {
NOT_GPU(*this);
copyonwrite();
2004-03-17 06:34:59 +01:00
2010-09-08 18:27:58 +02:00
for(register int i=0; i<nn; ++i) v[i] *= a;
return *this;
}
2004-03-17 06:34:59 +01:00
template <typename T>
inline NRVec<T> & NRVec<T>::operator/=(const T &a) {
NOT_GPU(*this);
copyonwrite();
for(register int i=0; i<nn; ++i) v[i] /= a;
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* compute scalar product \f$d\f$ of this vector \f$\vec{x}\f$ of general type <code>T</code>
* with given vector \f$\vec{y}\f$ of type <code>T</code> and order \f$N\f$
* \f[d = \sum_{i=1}^N\vec{x}_i\cdot\vec{y}_i\f]
* @param[in] rhs general vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
2005-02-18 23:08:15 +01:00
template<typename T>
2010-09-08 18:27:58 +02:00
inline const T NRVec<T>::operator*(const NRVec<T> &rhs) const {
2005-02-18 23:08:15 +01:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
2005-02-18 23:08:15 +01:00
#endif
2010-09-08 18:27:58 +02:00
NOT_GPU(*this);
NOT_GPU(rhs);
T dot(0);
for(register int i=0; i<nn; ++i) dot += v[i]*rhs.v[i];
2005-02-18 23:08:15 +01:00
return dot;
2004-03-17 04:07:21 +01:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* indexing operator giving the element at given position with range checking in
* the DEBUG mode
* @param[in] i position of the required vector element (starting from 0)
* @return reference to the requested element
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline T& NRVec<T>::operator[](const int i) {
2004-03-17 04:07:21 +01:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if(_LA_count_check && *count != 1) laerror("possible use of NRVec[] with count>1 as l-value");
if(i < 0 || i >= nn) laerror("out of range");
if(!v) laerror("unallocated NRVec");
2004-03-17 04:07:21 +01:00
#endif
2010-09-08 18:27:58 +02:00
NOT_GPU(*this);
2004-03-17 04:07:21 +01:00
return v[i];
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* indexing operator giving the element at given position with range checking in
* the DEBUG mode
* @param[in] i position of the required vector element (starting from 0)
* @return constant reference to the requested element
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline const T& NRVec<T>::operator[](const int i) const {
2004-03-17 04:07:21 +01:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if(i < 0 || i >= nn) laerror("out of range");
if(!v) laerror("unallocated NRVec");
2004-03-17 04:07:21 +01:00
#endif
2010-09-08 18:27:58 +02:00
NOT_GPU(*this);
2004-03-17 04:07:21 +01:00
return v[i];
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* determine the number of elements of this vector
* @return length of this vector
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline int NRVec<T>::size() const {
2004-03-17 04:07:21 +01:00
return nn;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* get the pointer to the underlying data of this vector
* @return pointer to the first vector element
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>::operator T*() {
2004-03-17 04:07:21 +01:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if(!v) laerror("unallocated NRVec");
2004-03-17 04:07:21 +01:00
#endif
return v;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* get the constant pointer to the underlying data of this vector
* @return constant pointer to the first vector element
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline NRVec<T>::operator const T*() const {
2004-03-17 04:07:21 +01:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if(!v) laerror("unallocated NRVec");
2004-03-17 04:07:21 +01:00
#endif
return v;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* create normalized copy of this vector
* @return copy of this vector after normalization
* @see NRVec<T>::normalize()
******************************************************************************/
2004-03-17 04:07:21 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
inline const NRVec<T> NRVec<T>::unitvector() const {
2004-03-17 04:07:21 +01:00
return NRVec<T>(*this).normalize();
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* generate operators involving vector and scalar
******************************************************************************/
2004-03-17 04:07:21 +01:00
NRVECMAT_OPER(Vec,+)
NRVECMAT_OPER(Vec,-)
NRVECMAT_OPER(Vec,*)
NRVECMAT_OPER(Vec,/)
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* generate operators involving vector and vector
******************************************************************************/
2004-03-17 04:07:21 +01:00
NRVECMAT_OPER2(Vec,+)
NRVECMAT_OPER2(Vec,-)
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* destructor for general vector decreases the reference count and performs
* deallocation if neccessary
******************************************************************************/
2004-03-17 17:39:07 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
NRVec<T>::~NRVec() {
if(!count) return;
if(--(*count) <= 0) {
if(v){
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
delete[] v;
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{ gpufree(v); }
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
}
2004-03-17 17:39:07 +01:00
delete count;
}
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* make own copy of the underlying data connected with this vector
******************************************************************************/
2004-03-17 17:39:07 +01:00
template <typename T>
2013-11-04 15:56:39 +01:00
void NRVec<T>::copyonwrite(bool detachonly) {
2010-09-08 18:27:58 +02:00
if(!count) laerror("copyonwrite of an undefined vector");
if(*count > 1) {
(*count)--;
count = new int;
*count = 1;
T *newv;
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
newv = new T[nn];
if(!detachonly)
{
if(LA_traits<T>::is_plaindata()) memcpy(newv, v, nn*sizeof(T));
else
{
for(int i=0; i<nn; ++i) {newv[i]=v[i]; LA_traits<T>::copyonwrite(newv[i]);}
}
}
2010-09-08 18:27:58 +02:00
#ifdef CUDALA
}else{
if(!LA_traits<T>::is_plaindata()) laerror("nested types not supported on gpu memory");
2010-09-08 18:27:58 +02:00
newv = (T *) gpualloc(nn*sizeof(T));
if(sizeof(T)%sizeof(float) != 0) laerror("memory alignment problem in NRVec<T>::copyonwrite()");
2013-11-04 15:56:39 +01:00
if(!detachonly) cublasScopy(nn*sizeof(T)/sizeof(float), (const float *) v, 1, (float *)newv, 1);
2010-09-08 18:27:58 +02:00
TEST_CUBLAS("cublasScopy");//"NRVec<T>::copyonwrite()"
}
#endif
v = newv;
}
2004-03-17 17:39:07 +01:00
}
2010-06-25 17:28:19 +02:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* assigns general vector \f$\vec{y}\f$ to this vector \f$\vec{x}\f$
* \li checks for self-assignment
* \li decreases the reference count and performs deallocation if neccesary
* \li links the internal data structures with corresponding properties of vector \f$\vec{y}\f$
* \li updates the reference count properly
******************************************************************************/
2004-03-17 17:39:07 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
NRVec<T> & NRVec<T>::operator=(const NRVec<T> &rhs) {
//check for self-assignment
if(this != &rhs){
if(count){
if(--(*count) == 0){
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
delete[] v;
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
gpufree(v);
}
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
delete count;
}
}
v = rhs.v;
nn = rhs.nn;
count = rhs.count;
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
location = rhs.location;
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
if(count){ (*count)++; }
}
return *this;
2004-03-17 17:39:07 +01:00
}
2010-06-25 17:28:19 +02:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* resizes this vector
* @param[in] n requested size
******************************************************************************/
2004-03-17 17:39:07 +01:00
template <typename T>
void NRVec<T>::resize(const int n, const bool preserve)
{
2004-03-17 17:39:07 +01:00
#ifdef DEBUG
if(n < 0) laerror("illegal dimension in NRVec::resize");
2004-03-17 17:39:07 +01:00
#endif
T *vold=0;
int nnold=0;
bool preserved=false;
bool do_delete=false;
if(count) //we are allocated
{
if(n == 0) //just deallocate
{
if(--(*count) <= 0)
{
if(v)
{
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
delete[] (v);
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
gpufree(v);
}
2010-06-25 17:28:19 +02:00
#endif
}
2010-09-08 18:27:58 +02:00
delete count;
}
2010-09-08 18:27:58 +02:00
count = 0;
nn = 0;
v = 0;
return;
}
if(*count > 1) //detach from shared data
{
2010-09-08 18:27:58 +02:00
(*count)--;
count = 0;
vold=v;
2010-09-08 18:27:58 +02:00
v = 0;
nnold=nn;
2010-09-08 18:27:58 +02:00
nn = 0;
preserved=true;
}
2010-09-08 18:27:58 +02:00
}
if(!count) //we were not allocated or we just detached
{
2010-09-08 18:27:58 +02:00
count = new int;
*count = 1;
nn = n;
#ifdef CUDALA
if(location == cpu)
#endif
v = new T[nn];
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
else
v = (T*) gpualloc(nn*sizeof(T));
#endif
if(preserved && preserve) goto do_preserve;
2010-09-08 18:27:58 +02:00
return;
}
// *count == 1 in this branch
if (n == nn) return; //nothing to do
nnold=nn;
nn = n;
2010-09-08 18:27:58 +02:00
#ifdef CUDALA
if(location == cpu)
{
2010-09-08 18:27:58 +02:00
#endif
if(preserve) {vold=v; preserved= do_delete=true;} else delete[] v;
2010-09-08 18:27:58 +02:00
v = new T[nn];
#ifdef CUDALA
}
else
{
if(preserve) {vold=v; d preserved= o_delete=true;} else gpufree(v);
2010-09-08 18:27:58 +02:00
v = (T*) gpualloc(nn*sizeof(T));
}
2010-06-25 17:28:19 +02:00
#endif
if(!preserve) return;
//copy data from old location and zero excess allocated memory
do_preserve:
if(!preserve || !preserved) laerror("assertion failed in NRVec::resize");
// omit this check since we would need to have traits for presently unknown user defined classes
// if(!LA_traits<T>::is_plaindata()) laerror("do not know how to preserve non-plain data");
2021-06-09 22:59:19 +02:00
int nnmin=nnold;
if(nn<nnmin) nnmin=nn;
#ifdef CUDALA
if(location == cpu)
{
#endif
2021-06-09 22:59:19 +02:00
for(int i=0; i<nnmin; ++i) v[i]=vold[i]; //preserve even non-plain data classes
if(nn>nnold) memset(v+nnold,0,(nn-nnold)*sizeof(T)); //just zero the new memory
if(do_delete) delete[] vold;
#ifdef CUDALA
}
else
{
//!!!works only with plain data
2021-06-09 22:59:19 +02:00
cublasSetVector(nnmin, sizeof(T), vold, 1, v, 1);
TEST_CUBLAS("cublasSetVector");
T a(0);
2021-06-09 22:59:19 +02:00
if(nn>nnold) smart_gpu_set(nn-nnold, a, v+nnold);
if(do_delete) gpufree(vold);
}
#endif
return;
2004-03-17 17:39:07 +01:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* perfrom deep copy
* @param[in] rhs vector being copied
* @see NRVec<T>::copyonwrite()
******************************************************************************/
2004-03-17 17:39:07 +01:00
template <typename T>
2010-09-08 18:27:58 +02:00
NRVec<T> & NRVec<T>::operator|=(const NRVec<T> &rhs) {
2004-03-17 17:39:07 +01:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if(!rhs.v) laerror("unallocated vector");
2004-03-17 17:39:07 +01:00
#endif
2010-09-08 18:27:58 +02:00
if(this == &rhs) return *this;
2010-06-25 17:28:19 +02:00
*this = rhs;
this->copyonwrite();
return *this;
2004-03-17 17:39:07 +01:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* complexify given vector of general type <code>T</code>, i.e. convert its
* elements to type <code>complex<T></code>
* @param[in] rhs vector being complexified
* @see NRVec<T>::copyonwrite()
******************************************************************************/
2006-04-01 06:48:01 +02:00
template<typename T>
2021-04-21 15:04:37 +02:00
NRVec<std::complex<T> > complexify(const NRVec<T> &rhs) {
2010-09-08 18:27:58 +02:00
NOT_GPU(rhs);
2006-04-01 06:48:01 +02:00
2021-04-21 15:04:37 +02:00
NRVec<std::complex<T> > r(rhs.size(), rhs.getlocation());
2010-09-08 18:27:58 +02:00
for(register int i=0; i<rhs.size(); ++i) r[i] = rhs[i];
return r;
}
2021-04-21 15:04:37 +02:00
template<> NRVec<std::complex<double> > complexify<double>(const NRVec<double> &rhs);
2010-06-25 17:28:19 +02:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* routine for moving vector data between CPU and GPU memory
* @param[in] dest required location
* @see NRVec<T>::location, NRVec<T>::getlocation()
******************************************************************************/
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
template<typename T>
2010-09-08 18:27:58 +02:00
void NRVec<T>::moveto(const GPUID dest) {
if(location == dest) return;
CPU_GPU(location, dest);
location = dest;
if(v && !count) laerror("internal");
if (!count) return;
if(v && *count == 0) laerror("internal");
if(!v) return;
T *vold = v;
if(dest == cpu){ // moving from GPU to CPU
v = new T[nn];
gpuget(nn,sizeof(T),vold,v);
if(*count == 1) gpufree(vold);
else {--(*count); count = new int(1);}
}else{ // moving from CPU to GPU
v = (T *) gpualloc(nn*sizeof(T));
gpuput(nn,sizeof(T),vold,v);
if(*count == 1) delete[] vold;
else {--(*count); count = new int(1);}
}
2010-06-25 17:28:19 +02:00
}
#endif
2021-05-19 22:29:47 +02:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* adds a real scalar value \f$\alpha\f$ to all elements of this real vector \f$\vec{x}\f$
* \f[\vec{x}_i\leftarrow\vec{x}_i+\alpha\f]
* @param[in] a real scalar value \f$\alpha\f$ being added
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline NRVec<double>& NRVec<double>::operator+=(const double &a) {
2010-06-25 17:28:19 +02:00
copyonwrite();
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
cblas_daxpy(nn, 1.0, &a, 0, v, 1);
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
double *d = gpuputdouble(a);
2010-06-25 17:28:19 +02:00
cublasDaxpy(nn, 1.0, d, 0, v, 1);
2010-09-08 18:27:58 +02:00
TEST_CUBLAS("cublasDaxpy");
2010-06-25 17:28:19 +02:00
gpufree(d);
2010-09-08 18:27:58 +02:00
}
2010-06-25 17:28:19 +02:00
#endif
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* adds a complex scalar value \f$\alpha\f$ to all elements of this complex vector \f$\vec{x}\f$
* \f[\vec{x}_i\leftarrow\vec{x}_i+\alpha\f]
* @param[in] a complex scalar value \f$\alpha\f$ being added
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline NRVec<std::complex<double> >& NRVec<std::complex<double> >::operator+=(const std::complex<double> &a) {
2010-06-25 17:28:19 +02:00
copyonwrite();
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
cblas_zaxpy(nn, &CONE, &a, 0, v, 1);
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
2021-04-21 15:04:37 +02:00
std::complex<double> *d = gpuputcomplex(a);
2010-06-25 17:28:19 +02:00
cublasZaxpy(nn, CUONE, (cuDoubleComplex *)d, 0, (cuDoubleComplex *)v, 1);
2010-09-08 18:27:58 +02:00
TEST_CUBLAS("cublasZaxpy");
2010-06-25 17:28:19 +02:00
gpufree(d);
2010-09-08 18:27:58 +02:00
}
2010-06-25 17:28:19 +02:00
#endif
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* subtracts a real scalar value \f$\alpha\f$ from all elements of this real vector \f$\vec{x}\f$
* \f[\vec{x}_i\leftarrow\vec{x}_i-\alpha\f]
* @param[in] a real scalar value \f$\alpha\f$ being subtracted
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline NRVec<double>& NRVec<double>::operator-=(const double &a) {
2010-06-25 17:28:19 +02:00
copyonwrite();
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
cblas_daxpy(nn, -1.0, &a, 0, v, 1);
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
double *d = gpuputdouble(a);
2010-06-25 17:28:19 +02:00
cublasDaxpy(nn, -1.0, d, 0, v, 1);
2010-09-08 18:27:58 +02:00
TEST_CUBLAS("cublasDaxpy");
2010-06-25 17:28:19 +02:00
gpufree(d);
2010-09-08 18:27:58 +02:00
}
2010-06-25 17:28:19 +02:00
#endif
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* subtracts a complex scalar value \f$\alpha\f$ from all elements of this complex vector \f$\vec{x}\f$
* \f[\vec{x}_i\leftarrow\vec{x}_i-\alpha\f]
* @param[in] a complex scalar value \f$\alpha\f$ being subtracted
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline NRVec<std::complex<double> >& NRVec<std::complex<double> >::operator-=(const std::complex<double> &a) {
2010-06-25 17:28:19 +02:00
copyonwrite();
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
cblas_zaxpy(nn, &CMONE, &a, 0, v, 1);
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
2021-04-21 15:04:37 +02:00
std::complex<double> *d = gpuputcomplex(a);
2010-09-08 18:27:58 +02:00
cublasZaxpy(nn, CUMONE, (cuDoubleComplex *)d, 0, (cuDoubleComplex *)v, 1);
TEST_CUBLAS("cublasZaxpy");
gpufree(d);
}
2010-06-25 17:28:19 +02:00
#endif
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* adds a real vector \f$\vec{y}\f$ to this real vector \f$\vec{x}\f$
* \f[\vec{x}\leftarrow\vec{x}+\vec{y}\f]
* @param[in] rhs real vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline NRVec<double>& NRVec<double>::operator+=(const NRVec<double> &rhs) {
2010-06-25 17:28:19 +02:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
SAME_LOC(*this, rhs);
2010-06-25 17:28:19 +02:00
copyonwrite();
2010-09-08 18:27:58 +02:00
#ifdef CUDALA
if(location == cpu){
#endif
cblas_daxpy(nn, 1.0, rhs.v, 1, v, 1);
#ifdef CUDALA
}else{
cublasDaxpy(nn, 1.0, rhs.v, 1, v, 1);
TEST_CUBLAS("cubasDaxpy");
}
#endif
2010-06-25 17:28:19 +02:00
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* adds a complex vector \f$\vec{y}\f$ to this complex vector \f$\vec{x}\f$
* \f[\vec{x}\leftarrow\vec{x}+\vec{y}\f]
* @param[in] rhs complex vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline NRVec<std::complex<double> >& NRVec<std::complex<double> >::operator+=(const NRVec<std::complex<double> > &rhs) {
2010-06-25 17:28:19 +02:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
SAME_LOC(*this, rhs);
2010-06-25 17:28:19 +02:00
copyonwrite();
2010-09-08 18:27:58 +02:00
#ifdef CUDALA
if(location == cpu){
#endif
cblas_zaxpy(nn, &CONE, rhs.v, 1, v, 1);
#ifdef CUDALA
}else{
cublasZaxpy(nn, CUONE, (cuDoubleComplex*)rhs.v, 1, (cuDoubleComplex*)v, 1);
TEST_CUBLAS("cublasZaxpy");
}
#endif
return *this;
2010-06-25 17:28:19 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* subtracts a real vector \f$\vec{y}\f$ from this real vector \f$\vec{x}\f$
* \f[\vec{x}\leftarrow\vec{x}-\vec{y}\f]
* @param[in] rhs real vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline NRVec<double> & NRVec<double>::operator-=(const NRVec<double> &rhs) {
2010-06-25 17:28:19 +02:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
SAME_LOC(*this,rhs);
copyonwrite();
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
cblas_daxpy(nn, -1.0, rhs.v, 1, v, 1);
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
}else{
2010-06-25 17:28:19 +02:00
cublasDaxpy(nn, -1.0, rhs.v, 1, v, 1);
2010-09-08 18:27:58 +02:00
TEST_CUBLAS("cubasDaxpy");
}
2010-06-25 17:28:19 +02:00
#endif
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* subtracts a complex vector \f$\vec{y}\f$ from this complex vector \f$\vec{x}\f$
* \f[\vec{x}\leftarrow\vec{x}-\vec{y}\f]
* @param[in] rhs double-precision complex vector \f$\vec{y}\f$
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline NRVec<std::complex<double> >& NRVec<std::complex<double> >::operator-=(const NRVec<std::complex<double> > &rhs) {
2010-06-25 17:28:19 +02:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if (nn != rhs.nn) laerror("incompatible dimensions");
#endif
SAME_LOC(*this, rhs);
copyonwrite();
#ifdef CUDALA
if(location == cpu){
#endif
cblas_zaxpy(nn, &CMONE, rhs.v, 1, v, 1);
#ifdef CUDALA
}else{
cublasZaxpy(nn, CUMONE, (cuDoubleComplex*)rhs.v, 1, (cuDoubleComplex*)v, 1);
TEST_CUBLAS("cublasZaxpy");
}
2010-06-25 17:28:19 +02:00
#endif
return *this;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* multiplies this real vector \f$\vec{x}\f$ by a real scalar value \f$\alpha\f$
* \f[\vec{x}_i\leftarrow\alpha\vec{x}_i\f]
* @param[in] a real scalar value \f$\alpha\f$
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline NRVec<double>& NRVec<double>::operator*=(const double &a) {
2010-06-25 17:28:19 +02:00
copyonwrite();
2010-09-08 18:27:58 +02:00
#ifdef CUDALA
if(location == cpu){
#endif
cblas_dscal(nn, a, v, 1);
#ifdef CUDALA
}else{
cublasDscal(nn, a, v, 1);
TEST_CUBLAS("cublasDscal");
}
#endif
2010-06-25 17:28:19 +02:00
return *this;
}
template<>
inline NRVec<double>& NRVec<double>::operator/=(const double &a) {return *this *= (1./a);}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* multiplies this complex vector \f$\vec{x}\f$ by a complex scalar value \f$\alpha\f$
* \f[\vec{x}_i\leftarrow\alpha\vec{x}_i\f]
* @param[in] a complex scalar value \f$\alpha\f$
* @return reference to the modified vector
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline NRVec<std::complex<double> >& NRVec<std::complex<double> >::operator*=(const std::complex<double> &a) {
2010-06-25 17:28:19 +02:00
copyonwrite();
2010-09-08 18:27:58 +02:00
#ifdef CUDALA
if(location == cpu){
#endif
cblas_zscal(nn, &a, v, 1);
#ifdef CUDALA
}else{
const cuDoubleComplex alpha = make_cuDoubleComplex(a.real(), a.imag());
cublasZscal(nn, alpha, (cuDoubleComplex*)v, 1);
TEST_CUBLAS("cublasZscal");
}
#endif
2010-06-25 17:28:19 +02:00
return *this;
}
template<>
inline NRVec<std::complex<double> >& NRVec<std::complex<double> >::operator/=(const std::complex<double> &a) {return *this *= (1./a);}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* computes the inner product of this real vector \f$\vec{x}\f$ with given real vector \f$\vec{y]\f$
* @param[in] rhs real vector \f$\vec{y}\f$
* @return \f$\sum_{i=1}^N\vec{x}_i\cdot\vec{y}_i\f$
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline const double NRVec<double>::operator*(const NRVec<double> &rhs) const {
double ret(0.0);
2010-06-25 17:28:19 +02:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if(nn != rhs.nn) laerror("incompatible dimensions");
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
SAME_LOC(*this, rhs);
#ifdef CUDALA
if(location == cpu){
#endif
ret = cblas_ddot(nn, v, 1, rhs.v, 1);
#ifdef CUDALA
}else{
ret = cublasDdot(nn, v, 1, rhs.v, 1);
TEST_CUBLAS("cublasDdot");
}
#endif
return ret;
2010-06-25 17:28:19 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* computes the inner product of this complex vector \f$\vec{x}\f$ with given complex vector \f$\vec{y}\f$
* taking conjugation of vector \f$\vec{x}\f$ into account
* @param[in] rhs complex vector \f$\vec{y}\f$
* @return \f$\sum_{i=1}^N\overbar{\vec{x}_i}\cdot\vec{y}_i\f$
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline const std::complex<double> NRVec<std::complex<double> >::operator*(const NRVec< std::complex<double> > &rhs) const {
2010-06-25 17:28:19 +02:00
#ifdef DEBUG
2010-09-08 18:27:58 +02:00
if(nn != rhs.nn) laerror("incompatible dimensions");
2010-06-25 17:28:19 +02:00
#endif
2021-04-21 15:04:37 +02:00
std::complex<double> dot;
2010-09-08 18:27:58 +02:00
SAME_LOC(*this, rhs);
#ifdef CUDALA
if(location == cpu){
#endif
cblas_zdotc_sub(nn, v, 1, rhs.v, 1, &dot);
#ifdef CUDALA
}else{
const cuDoubleComplex val = cublasZdotc(nn, (cuDoubleComplex*)v, 1, (cuDoubleComplex*)rhs.v, 1);
TEST_CUBLAS("cublasZdotc");
2021-04-21 15:04:37 +02:00
dot = std::complex<double>(cuCreal(val), cuCimag(val));
2010-09-08 18:27:58 +02:00
}
#endif
2010-06-25 17:28:19 +02:00
return dot;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* computes the inner product of this real vector \f$\vec{x}\f$ with given real data
* @param[in] y pointer to the double-precision real array (sufficient length assumed)
* @param[in] stride specifies the stride regarding the data pointe to by <tt>y</tt>
* @return \f$\sum_{i=1}^N\vec{x}_{i}\cdot y_{\mathrm{stride}\cdot(i-1) + 1}\f$
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline const double NRVec<double>::dot(const double *y, const int stride) const {
NOT_GPU(*this);
return cblas_ddot(nn, y, stride, v, 1);
2010-06-25 17:28:19 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* computes the inner product of this complex vector \f$\vec{x}\f$ with given complex data
* @param[in] y pointer to the double-precision complex array (sufficient length assumed)
* @param[in] stride specifies the stride regarding the data pointe to by <tt>y</tt>
* @return \f$\sum_{i=1}^N\vec{x}_{i}\cdot \overbar{y_{\mathrm{stride}\cdot(i-1) + 1}}\f$
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline const std::complex<double> NRVec<std::complex<double> >::dot(const std::complex<double> *y, const int stride) const {
std::complex<double> dot;
2010-09-08 18:27:58 +02:00
NOT_GPU(*this);
2010-06-25 17:28:19 +02:00
cblas_zdotc_sub(nn, y, stride, v, 1, &dot);
return dot;
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* computes the sum of the absolute values of the elements of this real vector \f$\vec{x}\f$
* @return \f$\sum_{i=1}^N\left|\vec{x}_i\right|\f$
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline const double NRVec<double>::asum() const {
double ret(0.0);
2010-06-25 17:28:19 +02:00
#ifdef CUDALA
2010-09-08 18:27:58 +02:00
if(location == cpu){
2010-06-25 17:28:19 +02:00
#endif
2010-09-08 18:27:58 +02:00
ret = cblas_dasum(nn, v, 1);
#ifdef CUDALA
}else{
ret = cublasDasum(nn, v, 1);
TEST_CUBLAS("cublasDasum");
}
#endif
return ret;
2010-06-25 17:28:19 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* for this complex vector \f$\vec{x}\f$ compute the expression
* \f[\sum_{i=1}^N\left|\Re{}\vec{x}_i\right| + \left|\Im{}\vec{x}_i\right|\f]
* @return the value of this sum
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline const double NRVec<std::complex<double> >::asum() const {
2010-09-08 18:27:58 +02:00
double ret(0.0);
#ifdef CUDALA
if(location == cpu){
#endif
ret = cblas_dzasum(nn, v, 1);
#ifdef CUDALA
}else{
ret = cublasDzasum(nn, (cuDoubleComplex*)v, 1);
TEST_CUBLAS("cublasDzasum");
}
#endif
return ret;
}
/***************************************************************************//**
* for this real vector \f$\vec{x}\f$ (of \f$N\f$ elements) determine the Frobenius norm
* @return \f$\sum_{i=1}^N\left|\vec{x}_i\right|^2\f$
******************************************************************************/
template<>
inline const double NRVec<double>::norm() const {
double ret(0.);
#ifdef CUDALA
if(location == cpu){
#endif
ret = cblas_dnrm2(nn, v, 1);
#ifdef CUDALA
}else{
ret = cublasDnrm2(nn, v, 1);
TEST_CUBLAS("cublasDnrm2");
}
#endif
return ret;
2010-06-25 17:28:19 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* for this complex vector \f$\vec{x}\f$ (of \f$N\f$ elements) determine the Frobenius norm
* @return \f$\sum_{i=1}^N\left|\vec{x}_i\right|^2\f$
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2021-04-21 15:04:37 +02:00
inline const double NRVec< std::complex<double> >::norm() const {
2010-09-08 18:27:58 +02:00
double ret(0.);
#ifdef CUDALA
if(location == cpu){
#endif
ret = cblas_dznrm2(nn, v, 1);
#ifdef CUDALA
}else{
ret = cublasDznrm2(nn, (cuDoubleComplex*)v, 1);
TEST_CUBLAS("cublasDzrm2");
}
#endif
return ret;
2010-06-25 17:28:19 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* for this real vector \f$\vec{x}\f$ determine the element with largest absolute value
* @return \f$\vec{x}_i\f$ where \f$\left|\vec{x]_i\right|=\mathrm{max}_{j}\left|\vec{x}_{j}\right|\f$
******************************************************************************/
2010-06-25 17:28:19 +02:00
template<>
2010-09-08 18:27:58 +02:00
inline const double NRVec<double>::amax() const {
double ret(0.0);
#ifdef CUDALA
if(location == cpu){
#endif
ret = v[cblas_idamax(nn, v, 1) - 1];
#ifdef CUDALA
}else{
const int pozice = cublasIdamax(nn, v, 1) - 1;
TEST_CUBLAS("cublasIdamax");
gpuget(1, sizeof(double), v + pozice, &ret);
}
#endif
return ret;
2010-06-25 17:28:19 +02:00
}
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* for this real vector \f$\vec{x}\f$ determine the element with smallest absolute value
* @return \f$\vec{x}_i\f$ where \f$\left|\vec{x]_i\right|=\mathrm{min}_{j}\left|\vec{x}_{j}\right|\f$
******************************************************************************/
template<>
inline const double NRVec<double>::amin() const {
double ret(std::numeric_limits<double>::max());
#ifdef CUDALA
if(location == cpu){
#endif
//BLAS routine idamin seems no to be supported
double val(0.0);
int index(-1);
for(register int i = 0; i < nn; i++){
val = std::abs(v[i]);
if(val < ret){ index = i; ret = val; }
}
ret = v[index];
#ifdef CUDALA
}else{
const int pozice = cublasIdamin(nn, v, 1) - 1;
TEST_CUBLAS("cublasIdamin");
gpuget(1, sizeof(double), v + pozice, &ret);
}
#endif
return ret;
}
2010-06-25 17:28:19 +02:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* for a given complex vector \f$\vec{v}\f$, determine the smallest index of the maximum
* magnitude element, i.e. maximal element in the 1-norm
* @return \f$\vec{v}_{j}\f$ which maximizes \f$\left\{\left|\Re{}\vec{v}_{i}\right|+\left|\Im{}\vec{v}_{i}\right|\right}\f$
******************************************************************************/
template<>
2021-04-21 15:04:37 +02:00
inline const std::complex<double> NRVec<std::complex<double> >::amax() const {
std::complex<double> ret(0., 0.);
2010-09-08 18:27:58 +02:00
#ifdef CUDALA
if(location == cpu){
#endif
ret = v[cblas_izamax(nn, v, 1) - 1];
#ifdef CUDALA
}else{
const int pozice = cublasIzamax(nn, (cuDoubleComplex*)v, 1) - 1;
TEST_CUBLAS("cublasIzamax");
2021-04-21 15:04:37 +02:00
gpuget(1, sizeof(std::complex<double>), v + pozice, &ret);
2010-09-08 18:27:58 +02:00
}
#endif
return ret;
}
2010-06-25 17:28:19 +02:00
2010-09-08 18:27:58 +02:00
/***************************************************************************//**
* for a given complex vector \f$\vec{v}\f$, determine the smallest index of the minimum
* magnitude element, i.e. minimal element in the 1-norm
* @return \f$\vec{v}_{j}\f$ which minimizes \f$\left\{\left|\Re{}\vec{v}_{i}\right|+\left|\Im{}\vec{v}_{i}\right|\right}\f$
******************************************************************************/
template<>
2021-04-21 15:04:37 +02:00
inline const std::complex<double> NRVec<std::complex<double> >::amin() const {
std::complex<double> ret(0., 0.);
2010-09-08 18:27:58 +02:00
#ifdef CUDALA
if(location == cpu){
#endif
// izamin seems not to be supported
int index(0);
double val(0.0), min_val(std::numeric_limits<double>::max());
2021-04-21 15:04:37 +02:00
std::complex<double> z_val(0.0, 0.0);
2010-09-08 18:27:58 +02:00
for(register int i=0; i < nn; i++){
z_val = v[i];
val = std::abs(z_val.real()) + std::abs(z_val.imag());
if(val < min_val){ index = i; min_val = val; }
}
ret = v[index];
#ifdef CUDALA
}else{
const int pozice = cublasIzamin(nn, (cuDoubleComplex*)v, 1) - 1;
TEST_CUBLAS("cublasIzamin");
2021-04-21 15:04:37 +02:00
gpuget(1, sizeof(std::complex<double>), v + pozice, &ret);
2010-09-08 18:27:58 +02:00
}
#endif
return ret;
}
2009-11-12 22:01:19 +01:00
}//namespace
2004-03-17 04:07:21 +01:00
#endif /* _LA_VEC_H_ */