A vector-like data storage class that supports both internal and external storage mechanisms. More...

#include <tensor.h>

Inheritance diagram for maia::tensor::detail_::TensorStorage< T >:

Public Types
typedef MLong	size_type

typedef T	value_type

typedef value_type *	pointer

typedef const value_type *	const_pointer

typedef value_type &	reference

typedef const value_type &	const_reference

typedef pointer	iterator

typedef const_pointer	const_iterator

Public Member Functions
	TensorStorage ()
	Default constructor does not allocate any memory but creates a zero-sized storage container. More...

	TensorStorage (const TensorStorage< T > &ts)
	Copy constructor creates a copy of the provided object. More...

	TensorStorage (size_type n)
	Creates a new TensorStorage object using internal storage. More...

	TensorStorage (size_type n, T *data)
	Creates a new TensorStorage object using external storage, i.e. no memory is allocated. More...

	~TensorStorage ()
	The destructor only calls clear(). More...

size_type	size () const
	Returns the current size of the data container. More...

void	resize (size_type n)
	Resizes the container by first dropping all existing elements and then reallocating new memory. More...

void	resize (size_type n, T *data)
	Resizes the container by first dropping all existing elements and then using the external data pointer for the new storage. More...

void	clear ()
	Clears the data storage by deallocating any previously allocated memory and setting the size to zero. More...

void	swap (TensorStorage< T > &other)
	Swap the contents of the existing TensorStorage object with another one. More...

TensorStorage< T > &	operator= (TensorStorage< T > copy)
	Assignment operator copies another object to the current one. More...

T &	operator[] (size_type n)
	Returns a reference to the element at position n (non-const version). More...

const T &	operator[] (size_type n) const
	Returns a reference to the element at position n (const version). More...

Private Member Functions
MBool	isAllocated ()
	Returns true if this class has allocated its own memory, and false if external memory is used. More...

void	reallocate (size_type n)
	Clears any existing internal data storage and then allocates new memory to store as many elements as specified. More...

Private Attributes
T *	m_data

size_type	m_size

MBool	m_isAllocated

Friends
template<class TT >
void	swap (TensorStorage< TT > &a, TensorStorage< TT > &b)
	Swap the contents between two TensorStorage objects. More...

Detailed Description

template<class T>
class maia::tensor::detail_::TensorStorage< T >

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Template Parameters

T	Scalar data type of the tensor elements.

By default, an internal storage mechanism is used, i.e. memory is allocated and deallocated automatically. However, there is also the possibility to provide a pointer to an existing storage location, which allows this class to be used as a "viewport" to data in memory. For the design of the class, a close look was taken at std::vector, in order to support most of the dynamic reallocation methods and other operations that std::vector has as well.

This class is mainly used by Tensor as its internal storage.

Definition at line 780 of file tensor.h.

Member Typedef Documentation

◆ const_iterator

template<class T >

typedef const_pointer maia::tensor::detail_::TensorStorage< T >::const_iterator

Definition at line 790 of file tensor.h.

◆ const_pointer

template<class T >

typedef const value_type* maia::tensor::detail_::TensorStorage< T >::const_pointer

Definition at line 786 of file tensor.h.

◆ const_reference

template<class T >

typedef const value_type& maia::tensor::detail_::TensorStorage< T >::const_reference

Definition at line 788 of file tensor.h.

◆ iterator

template<class T >

typedef pointer maia::tensor::detail_::TensorStorage< T >::iterator

Definition at line 789 of file tensor.h.

◆ pointer

template<class T >

typedef value_type* maia::tensor::detail_::TensorStorage< T >::pointer

Definition at line 785 of file tensor.h.

◆ reference

template<class T >

typedef value_type& maia::tensor::detail_::TensorStorage< T >::reference

Definition at line 787 of file tensor.h.

◆ size_type

template<class T >

typedef MLong maia::tensor::detail_::TensorStorage< T >::size_type

Definition at line 783 of file tensor.h.

◆ value_type

template<class T >

typedef T maia::tensor::detail_::TensorStorage< T >::value_type

Definition at line 784 of file tensor.h.

Constructor & Destructor Documentation

◆ TensorStorage() [1/4]

template<class T >

maia::tensor::detail_::TensorStorage< T >::TensorStorage

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Definition at line 830 of file tensor.h.

                                       : m_data(0), m_size(0), m_isAllocated(false) {
  // Nothing here
}

◆ TensorStorage() [2/4]

template<class T >

maia::tensor::detail_::TensorStorage< T >::TensorStorage ( const TensorStorage< T > & ts )

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in] ts Object that is copied.

Note: The copy constructor will always allocate memory, even if the original object used external storage.

Definition at line 847 of file tensor.h.

                                                                 : m_data(0), m_size(0), m_isAllocated(false) {
  // Allocate new memory and copy the data members
  reallocate(ts.size());
  std::copy(&ts.m_data[0], &ts.m_data[0] + ts.size(), m_data);
}

◆ TensorStorage() [3/4]

template<class T >

maia::tensor::detail_::TensorStorage< T >::TensorStorage ( size_type n )

inlineexplicit

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in] n Size of the new container.

Definition at line 863 of file tensor.h.

                                                  : m_data(0), m_size(0), m_isAllocated(false) {
  resize(n);
}

◆ TensorStorage() [4/4]

template<class T >

maia::tensor::detail_::TensorStorage< T >::TensorStorage	(	size_type	n,
		T *	data
	)

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in]	n	Size of the storage in memory.
[in]	data	Pointer to existing memory location.

Since this is supposed to be a very lightweight and fast way to access memory through TensorStorage, no data is changed to some initial value when this constructor is called. Furthermore, the user has to ensure that the data location that is pointed to does not get deallocated before the TensorStorage object is destroyed.

Definition at line 883 of file tensor.h.

                                                           : m_data(data), m_size(n), m_isAllocated(false) {
  ASSERT(n >= 0, "New storage size must be greater or equal to zero.");
  ASSERT(data != 0, "Data pointer may not be a null pointer.");
}

◆ ~TensorStorage()

template<class T >

maia::tensor::detail_::TensorStorage< T >::~TensorStorage

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Definition at line 896 of file tensor.h.

                                        {
  clear();
}

Member Function Documentation

◆ clear()

template<class T >

void maia::tensor::detail_::TensorStorage< T >::clear

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Definition at line 959 of file tensor.h.

                                    {
  if(isAllocated()) {
    delete[] m_data;
  }
 
  m_data = 0;
  m_size = 0;
  m_isAllocated = false;
}

◆ isAllocated()

template<class T >

MBool maia::tensor::detail_::TensorStorage< T >::isAllocated

inlineprivate

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2013-05-08

Definition at line 1064 of file tensor.h.

                                           {
  return m_isAllocated;
}

◆ operator=()

template<class T >

TensorStorage< T > & maia::tensor::detail_::TensorStorage< T >::operator= ( TensorStorage< T > copy )

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in] copy The object that is copied.

Returns: The assigned object (*this)

Definition at line 1014 of file tensor.h.

                                                                          {
  swap(copy);
 
  return *this;
}

◆ operator[]() [1/2]

template<class T >

T & maia::tensor::detail_::TensorStorage< T >::operator[] ( size_type n )

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in] n Position of the element in the data storage.

Returns: The element at the specified position.

Definition at line 1032 of file tensor.h.

                                                  {
  ASSERT(n >= 0 && n < size(), "Index out of bounds");
 
  return m_data[n];
}

◆ operator[]() [2/2]

template<class T >

const T & maia::tensor::detail_::TensorStorage< T >::operator[] ( size_type n ) const

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in] n Position of the element in the data storage.

Returns: The element at the specified position.

Definition at line 1050 of file tensor.h.

                                                              {
  ASSERT(n >= 0 && n < size(), "Index out of bounds");
 
  return m_data[n];
}

◆ reallocate()

template<class T >

void maia::tensor::detail_::TensorStorage< T >::reallocate ( size_type n )

private

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in] n The number of elements that can be stored in the new container.

Definition at line 1080 of file tensor.h.

                                             {
  clear();
 
  m_data = (n <= 0) ? nullptr : new T[std::min(n, PTRDIFF_MAX)];
  m_size = n;
  m_isAllocated = true;
}

◆ resize() [1/2]

template<class T >

void maia::tensor::detail_::TensorStorage< T >::resize ( size_type n )

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in] n New size of the data container.

Definition at line 924 of file tensor.h.

                                                {
  ASSERT(n >= 0, "New storage size must be greater or equal to zero.");
 
  reallocate(n);
}

◆ resize() [2/2]

template<class T >

void maia::tensor::detail_::TensorStorage< T >::resize	(	size_type	n,
		T *	data
	)

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in]	n	New size of the data container.
[in]	data	Pointer to existing memory location.

Definition at line 942 of file tensor.h.

                                                         {
  ASSERT(n >= 0, "New storage size must be greater or equal to zero.");
  ASSERT(data != 0, "Data pointer may not be a null pointer.");
 
  clear();
  m_size = n;
  m_data = data;
}

◆ size()

template<class T >

TensorStorage< T >::size_type maia::tensor::detail_::TensorStorage< T >::size

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Returns: Length of the data array.

Definition at line 910 of file tensor.h.

                                                                       {
  return m_size;
}

◆ swap()

template<class T >

void maia::tensor::detail_::TensorStorage< T >::swap ( TensorStorage< T > & other )

inline

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Parameters

[in] other The object with which the contents are exchanged.

Definition at line 979 of file tensor.h.

                                                          {
  using std::swap;
  swap(m_data, other.m_data);
  swap(m_size, other.m_size);
  swap(m_isAllocated, other.m_isAllocated);
}

Friends And Related Function Documentation

◆ swap

template<class T >

template<class TT >

void swap	(	TensorStorage< TT > &	a,
		TensorStorage< TT > &	b
	)

friend

Author: Michael Schlottke (mic) mic@a.nosp@m.ia.r.nosp@m.wth-a.nosp@m.ache.nosp@m.n.de

Date: 2012-11-14

Template Parameters

TT

Parameters

[in]	a	First object.
[in]	b	Second object.

Definition at line 998 of file tensor.h.

                                                             {
  a.swap(b);
}

Member Data Documentation

◆ m_data

template<class T >

T* maia::tensor::detail_::TensorStorage< T >::m_data

private

Definition at line 817 of file tensor.h.

◆ m_isAllocated

template<class T >

MBool maia::tensor::detail_::TensorStorage< T >::m_isAllocated

private

Definition at line 819 of file tensor.h.

◆ m_size

template<class T >

size_type maia::tensor::detail_::TensorStorage< T >::m_size

private

Definition at line 818 of file tensor.h.

The documentation for this class was generated from the following file:

/home/gitlab-runner/scratch/builds/oxpnswJ6/1/aia/m-AIA/m-AIA/src/UTIL/tensor.h

Public Types

Public Member Functions

Private Member Functions

Private Attributes

Friends

Detailed Description

Member Typedef Documentation

◆ const_iterator

◆ const_pointer

◆ const_reference

◆ iterator

◆ pointer

◆ reference

◆ size_type

◆ value_type

Constructor & Destructor Documentation

◆ TensorStorage() [1/4]

◆ TensorStorage() [2/4]

◆ TensorStorage() [3/4]

◆ TensorStorage() [4/4]

◆ ~TensorStorage()

Member Function Documentation

◆ clear()

◆ isAllocated()

◆ operator=()

◆ operator[]() [1/2]

◆ operator[]() [2/2]

◆ reallocate()

◆ resize() [1/2]

◆ resize() [2/2]

◆ size()

◆ swap()

Friends And Related Function Documentation

◆ swap

Member Data Documentation

◆ m_data

◆ m_isAllocated

◆ m_size