itpp::Fix Class Reference

Fixed-point data type. More...

#include <itpp/fixed/fix.h>

Inheritance diagram for itpp::Fix:

Public Member Functions
	Fix (double x=0.0, int s=0, int w=MAX_WORDLEN, e_mode e=TC, o_mode o=WRAP, q_mode q=TRN, Stat *ptr=0)
	Default constructor.
	Fix (const Fix_Factory &f)
	Constructor.
	Fix (fixrep r, int s, int, int)
	Constructor for internal use. No restrictions are applied. The dummies help to avoid ambiguities.
	Fix (const Fix &x, int w=MAX_WORDLEN, e_mode e=TC, o_mode o=WRAP, q_mode q=TRN, Stat *ptr=0)
	Copy constructor.
virtual	~Fix ()
	Destructor.
Fix &	operator= (const Fix &x)
	Assignment from Fix.
Fix &	operator= (const int x)
	Assignment from int.
Fix &	operator+= (const Fix &x)
	Addition of Fix.
Fix &	operator+= (const int x)
	Addition of int.
Fix &	operator-= (const Fix &x)
	Subtraction of Fix.
Fix &	operator-= (const int x)
	Subtraction of int.
Fix &	operator*= (const Fix &x)
	Multiplication with Fix.
Fix &	operator*= (const int x)
	Multiplication with int.
Fix &	operator/= (const Fix &x)
	Division with Fix using quantization mode TRN.
Fix &	operator/= (const int x)
	Division with int using quantization mode TRN.
Fix	operator- () const
	Unary negative of Fix.
Fix &	operator<<= (const int n)
	Left shift n bits.
Fix &	operator>>= (const int n)
	Right shift n bits using quantization mode qmode (constructor argument)
void	set (double x, int n)
	Set to x * pow2(n) using quantization mode qmode (constructor argument)
void	set (double x, int n, q_mode q)
	Set to x * pow2(n) using quantization mode q (function argument)
void	set_re (fixrep x)
	Set data representation (mainly for internal use since it reveals the representation type)
void	lshift (int n)
	Left shift n bits.
void	rshift (int n)
	Right shift n bits using quantization mode qmode (constructor argument)
void	rshift (int n, q_mode q)
	Right shift n bits using quantization mode q (function argument)
virtual void	print () const
	Print restrictions.
fixrep	get_re () const
	Get data representation (mainly for internal use since it reveals the representation type)
double	unfix () const
	Conversion to double.
	operator double () const
	Conversion to double.
void	set_shift (int s)
	Set shift (without shifting)
int	get_shift () const
	Get shift.
int	get_wordlen () const
	Get word length.
e_mode	get_e_mode () const
	Get sign encoding mode.
o_mode	get_o_mode () const
	Get overflow mode.
q_mode	get_q_mode () const
	Get quantization mode.
output_mode	get_output_mode () const
	Get output mode.
fixrep	get_max () const
	Get maximum value of data representation.
fixrep	get_min () const
	Get minimum value of data representation.

Static Public Member Functions
static void	set_output_mode (output_mode o)
	Set output mode to OUTPUT_FIX, OUTPUT_FIX_SHIFT, OUTPUT_FLOAT or OUTPUT_FLOAT_SHIFT. Static member function.
static void	set_output_mode (std::string o)
	Set output mode to "OUTPUT_FIX", "OUTPUT_FIX_SHIFT", "OUTPUT_FLOAT" or "OUTPUT_FLOAT_SHIFT". Static member function.

Protected Member Functions
void	init ()
	Calculate help variables min, max and n_unused_bits.
fixrep	apply_o_mode (fixrep x) const
	Handle overflows using overflow mode omode and make call to statistics object (if any)
fixrep	scale_and_apply_modes (double x) const
	Convert from double to fixrep using shift and quantization mode qmode, then call limit()
fixrep	scale_and_apply_modes (double x, q_mode q) const
	Convert from double to fixrep using shift and quantization mode q, then call limit()
fixrep	rshift_and_apply_q_mode (fixrep x, int n) const
	Right shift n bits using quantization mode qmode and make call to statistics object (if any)
fixrep	rshift_and_apply_q_mode (fixrep x, int n, q_mode q) const
	Right shift n bits using quantization mode q and make call to statistics object (if any)

Protected Attributes
fixrep	re
	Data representation.
int	shift
	Accumulated bitshift (positive means left-shifted, negative means right-shifted)
int	wordlen
	Word length.
e_mode	emode
	Sign encoding mode.
o_mode	omode
	Overflow mode.
q_mode	qmode
	Quantization mode.
Stat *	stat_ptr
	Pointer to statistics object.
fixrep	min
	Minimum allowed value (help variable to speed up calculations)
fixrep	max
	Maximum allowed value (help variable to speed up calculations)
int	n_unused_bits
	Number of unused (MSB) bits (help variable to speed up calculations)

Friends
class	CFix
template<int , e_mode , o_mode , q_mode >
class	Fixed
template<int , e_mode , o_mode , q_mode >
class	CFixed
ITPP_EXPORT int	assert_shifts (const CFix &x, const Fix &y)
	Check that x.shift==y.shift OR x==0 OR y==0 and return the shift (for the non-zero argument)
ITPP_EXPORT int	assert_shifts (const Fix &x, const Fix &y)
	Check that x.shift==y.shift OR x==0 OR y==0 and return the shift (for the non-zero argument)
ITPP_EXPORT int	assert_shifts (const Fix &x, int y)
	Check that x.shift==0 OR x==0 OR y==0 and return x.shift.

Detailed Description

Fixed-point data type.

See the Detailed Description in the Fixed-point Module module.

Definition at line 51 of file fix.h.

---

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

• itpp/fixed/fix.h
• itpp/fixed/fix.cpp