Template class for the emulation. More...

#include <floating_point_emulation.hpp>

Public Types
enum	{ shift = S }

enum	{ factor = 1 << S }

typedef T	value_type

Public Member Functions
	tfloat ()

	tfloat (const int value)

	tfloat (const double value)

tfloat &	operator= (const tfloat rhs)

tfloat &	operator= (const int rhs)

tfloat &	operator= (const double rhs)

bool	operator< (const tfloat rhs) const

bool	operator<= (const tfloat rhs) const

bool	operator> (const tfloat rhs) const

bool	operator>= (const tfloat rhs) const

bool	operator== (const tfloat rhs) const

bool	operator== (const int rhs) const

bool	operator== (const double rhs) const

bool	operator!= (const tfloat rhs) const

bool	operator!= (const int rhs) const

bool	operator!= (const double rhs) const

bool	operator! () const

tfloat	operator- () const

tfloat	operator+ () const

tfloat &	operator*= (const tfloat rhs)
	Multiply. More...

tfloat &	operator*= (const int rhs)
	Multiply. More...

tfloat &	operator*= (const double rhs)
	Multiply. More...

tfloat &	operator/= (const tfloat rhs)
	Divide. More...

tfloat &	operator/= (const int rhs)
	Divide. More...

tfloat &	operator/= (const double rhs)
	Divide. More...

tfloat &	operator+= (const tfloat rhs)

tfloat &	operator+= (const int rhs)

tfloat &	operator+= (const double rhs)

tfloat &	operator-= (const tfloat rhs)

tfloat &	operator-= (const int rhs)

tfloat &	operator-= (const double rhs)

int	to_int () const

double	to_double () const

int	floor ()

T	get_value () const

template<>
tfloat< Sint32, 8 > &	operator*= (const tfloat< Sint32, 8 > rhs)

Private Member Functions
tfloat &	negative ()
	Changes the object to its negative value. More...

Private Attributes
T	value_
	The value of the emulation. More...

Friends
template<class TT , unsigned SS>
struct	detail::tidiv

Detailed Description

template<class T, unsigned S>
class floating_point_emulation::tfloat< T, S >

Template class for the emulation.

Note: Operators that are not documented do the expected thing.; For conversions it would be nice to use the `operator int()' instead of to_int(). Unfortunately C++98 doesn't support the explicit for conversion operators and the implicit conversion is evil.

Template Parameters

T	The type used for the emulation. At the moment the following types are supported: `double`

Definition at line 143 of file floating_point_emulation.hpp.

Member Typedef Documentation

template<class T, unsigned S>

typedef T floating_point_emulation::tfloat< T, S >::value_type

Definition at line 584 of file floating_point_emulation.hpp.

Member Enumeration Documentation

template<class T, unsigned S>

anonymous enum

Enumerator
shift

Definition at line 586 of file floating_point_emulation.hpp.

template<class T, unsigned S>

anonymous enum

Enumerator
factor

Definition at line 587 of file floating_point_emulation.hpp.

Constructor & Destructor Documentation

template<class T, unsigned S>

floating_point_emulation::tfloat< T, S >::tfloat ( )

inline

Definition at line 591 of file floating_point_emulation.hpp.

Referenced by floating_point_emulation::tfloat< T, S >::operator+=(), floating_point_emulation::tfloat< T, S >::operator-(), floating_point_emulation::tfloat< T, S >::operator-=(), floating_point_emulation::tfloat< T, S >::operator=(), and floating_point_emulation::tfloat< T, S >::operator==().

template<class T, unsigned S>

floating_point_emulation::tfloat< T, S >::tfloat ( const int value )

inlineexplicit

Definition at line 596 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK.

template<class T, unsigned S>

floating_point_emulation::tfloat< T, S >::tfloat ( const double value )

inlineexplicit

Definition at line 602 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK.

Member Function Documentation

template<class T, unsigned S>

int floating_point_emulation::tfloat< T, S >::floor ( )

inline

Definition at line 897 of file floating_point_emulation.hpp.

References floating_point_emulation::floor(), and floating_point_emulation::tfloat< T, S >::value_.

Referenced by floating_point_emulation::floor(), and test_floor().

template<class T, unsigned S>

T floating_point_emulation::tfloat< T, S >::get_value ( ) const

inline

Definition at line 905 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::negative ( )

inlineprivate

Changes the object to its negative value.

There is an issue with operator-(). There is an constructor that takes value_type as parameter. When shift != 0 the issue occurs. Implementing operator-() as return tfloat(-value_) returns a scaled value. Scaling twice might loose resolution and cause some overhead. So just add this function to solve the issue.

Returns: *this, but it's value is negated.

Definition at line 933 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator! ( ) const

inline

Definition at line 701 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator!= ( const tfloat< T, S > rhs ) const

inline

Definition at line 676 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator==().

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator!= ( const int rhs ) const

inline

Definition at line 682 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator==().

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator!= ( const double rhs ) const

inline

Definition at line 688 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator==().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator*= ( const tfloat< T, S > rhs )

inline

Multiply.

Keep in mind that:

THIS * RHS = this * sf * rhs * sf = result * sf * sf = RESULT * sf

Thus in order to get RESULT there needs to be divided by sf:

RESULT = THIS * RHS / sf

Definition at line 735 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK, and floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator*= ( const int rhs )

inline

Multiply.

No extra adjustment needed since:

RESULT = THIS * rhs

Definition at line 758 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK, and floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator*= ( const double rhs )

inline

Multiply.

No extra adjustment needed since:

RESULT = THIS * rhs

Definition at line 773 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK, and floating_point_emulation::tfloat< T, S >::value_.

template<>

tfloat< Sint32, 8 > & floating_point_emulation::tfloat< Sint32, 8 >::operator*= ( const tfloat< Sint32, 8 > rhs )

inline

Specialized for the Sint32 with a shift of 8.

Instead of figuring out the optimal shift before multiplying simply multiply as a 64-bit value and then perform the shift. This is rather cheap on the Pandora and also keeps the code short on that platform (only two extra instructions on the ARM v7; a `logical shift right' instruction followed by a `logical left shifted or' instruction.)

Definition at line 988 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

tfloat floating_point_emulation::tfloat< T, S >::operator+ ( ) const

inline

Definition at line 713 of file floating_point_emulation.hpp.

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator+= ( const tfloat< T, S > rhs )

inline

Definition at line 839 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK, and floating_point_emulation::tfloat< T, S >::value_.

Referenced by floating_point_emulation::tfloat< T, S >::operator+=().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator+= ( const int rhs )

inline

Definition at line 847 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator+=(), and floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator+= ( const double rhs )

inline

Definition at line 853 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator+=(), and floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

tfloat floating_point_emulation::tfloat< T, S >::operator- ( ) const

inline

Definition at line 707 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator-= ( const tfloat< T, S > rhs )

inline

Definition at line 861 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK, and floating_point_emulation::tfloat< T, S >::value_.

Referenced by floating_point_emulation::tfloat< T, S >::operator-=().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator-= ( const int rhs )

inline

Definition at line 869 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator-=(), and floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator-= ( const double rhs )

inline

Definition at line 875 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator-=(), and floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator/= ( const tfloat< T, S > rhs )

inline

Divide.

Keep in mind that:

THIS * RHS = this * sf * rhs * sf = result * sf * sf = RESULT * sf

THIS this * sf this -— = ------— = -— = result RHS rhs * sf rhs

Thus in order to get RESULT there needs to be multiplied by sf:

     THIS

RESULT = -— * sf RHS

Definition at line 800 of file floating_point_emulation.hpp.

References floating_point_emulation::detail::tidiv< T, S >::idiv().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator/= ( const int rhs )

inline

Divide.

No extra adjustment needed since:

RESULT = THIS / rhs

Definition at line 813 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK, and floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator/= ( const double rhs )

inline

Divide.

No extra adjustment needed since:

RESULT = THIS / rhs

Definition at line 828 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK, and floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator< ( const tfloat< T, S > rhs ) const

inline

Definition at line 632 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator<= ( const tfloat< T, S > rhs ) const

inline

Definition at line 638 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator= ( const tfloat< T, S > rhs )

inline

Definition at line 610 of file floating_point_emulation.hpp.

References FLOATING_POINT_EMULATION_RANGE_CHECK, and floating_point_emulation::tfloat< T, S >::value_.

Referenced by floating_point_emulation::tfloat< T, S >::operator=().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator= ( const int rhs )

inline

Definition at line 618 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator=(), and floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

tfloat& floating_point_emulation::tfloat< T, S >::operator= ( const double rhs )

inline

Definition at line 624 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator=(), and floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator== ( const tfloat< T, S > rhs ) const

inline

Definition at line 657 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

Referenced by floating_point_emulation::tfloat< T, S >::operator!=(), and floating_point_emulation::tfloat< T, S >::operator==().

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator== ( const int rhs ) const

inline

Definition at line 663 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator==(), and floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator== ( const double rhs ) const

inline

Definition at line 669 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::operator==(), and floating_point_emulation::tfloat< T, S >::tfloat().

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator> ( const tfloat< T, S > rhs ) const

inline

Definition at line 644 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

bool floating_point_emulation::tfloat< T, S >::operator>= ( const tfloat< T, S > rhs ) const

inline

Definition at line 650 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

template<class T, unsigned S>

double floating_point_emulation::tfloat< T, S >::to_double ( ) const

inline

Definition at line 889 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

Referenced by test_to_double().

template<class T, unsigned S>

int floating_point_emulation::tfloat< T, S >::to_int ( ) const

inline

Definition at line 883 of file floating_point_emulation.hpp.

References floating_point_emulation::tfloat< T, S >::value_.

Referenced by scale_surface_sharp(), and test_to_int().

Friends And Related Function Documentation

template<class T, unsigned S>

template<class TT , unsigned SS>

friend struct detail::tidiv

friend

Definition at line 579 of file floating_point_emulation.hpp.

Member Data Documentation

template<class T, unsigned S>

T floating_point_emulation::tfloat< T, S >::value_

private

The value of the emulation.

What the value represents is not really defined for the user.

Definition at line 919 of file floating_point_emulation.hpp.

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

floating_point_emulation.hpp

Public Types

Public Member Functions

Private Member Functions

Private Attributes

Friends

Detailed Description

template<class T, unsigned S> class floating_point_emulation::tfloat< T, S >

Member Typedef Documentation

Member Enumeration Documentation

Constructor & Destructor Documentation

Member Function Documentation

Friends And Related Function Documentation

Member Data Documentation

template<class T, unsigned S>
class floating_point_emulation::tfloat< T, S >