hgps::core::MathHelper Class Reference

Additional mathematical functions and determines the parameters of the floating point representation. More...

#include <math_util.h>

Public Member Functions
	MathHelper ()=delete

Static Public Member Functions
static int	radix () noexcept
	Gets the machine radix used by floating-point numbers. More...
static double	machine_precision () noexcept
	Gets the largest positive value which, when added to 1.0, yields 0. More...
static double	default_numerical_precision () noexcept
	Gets the typical meaningful precision for numerical calculations. More...
static bool	equal (double left, double right) noexcept
	Compares two floating-point numbers for relative equality using the default numerical precision. More...
static bool	equal (double left, double right, double precision) noexcept
	Compares two floating-point numbers for relative equality. More...

Detailed Description

Additional mathematical functions and determines the parameters of the floating point representation.

References:

• William Cody, Algorithm 665: MACHAR, a subroutine to dynamically determine machine parameters, ACM Transactions on Mathematical Software, Volume 14, Number 4, December 1988, pages 303-311.
• Didier H. Besset, Object-Oriented Implementation of Numerical Methods An Introduction with Smalltalk, Morgan Kaufmann, November 2000.

Constructor & Destructor Documentation

MathHelper()

hgps::core::MathHelper::MathHelper ( )

delete

Member Function Documentation

default_numerical_precision()

double hgps::core::MathHelper::default_numerical_precision ( )

staticnoexcept

Gets the typical meaningful precision for numerical calculations.

Returns

The default precision for numerical calculations

Here is the call graph for this function:

equal() [1/2]

bool hgps::core::MathHelper::equal ( double  left, double  right  )

staticnoexcept

Compares two floating-point numbers for relative equality using the default numerical precision.

Parameters

left	The left double to compare.
right	The right double to compare.

Returns

true if the number are equal, otherwise. false

equal() [2/2]

bool hgps::core::MathHelper::equal ( double  left, double  right, double  precision  )

staticnoexcept

Compares two floating-point numbers for relative equality.

Since the floating-point representation is keeping a constant relative precision, comparison must be made using relative error.

Let a and b be the two numbers to be compared, we build the following expression:

x = |a - b| / max(|a|, |b|)

The two numbers can be considered equal if x is smaller than a given number xmax. If the denominator of the fraction on the above equation is less than xmax, than the two numbers can be considered as being equal. For lack of information on how the numbers a and b have been obtained, one uses for xmax the DefaultNumericalPrecision. If one can determine the precision of each number, then this relative equality method can be used.

Parameters

left	The left double to compare.
right	The right double to compare.
precision	The comparison precision.

Returns

true if the number are equal, otherwise. false

machine_precision()

double hgps::core::MathHelper::machine_precision ( )

staticnoexcept

Gets the largest positive value which, when added to 1.0, yields 0.

Returns

The machine precision value

Here is the caller graph for this function:

radix()

int hgps::core::MathHelper::radix ( )

staticnoexcept

Gets the machine radix used by floating-point numbers.

Returns

The machine radix value

---

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

• HealthGPS.Core/math_util.h
• HealthGPS.Core/math_util.cpp