Fuzzy comparisions

Compare floating point numbers match within a tolerance. More...

Macros
#define	asserteq(a, b)   assert(fuzzcmp(a, b, 1.000010))

Functions
int	fuzzcmp (double a, double b, double delta)
int	fuzzcmpf (float a, float b, float delta)
int	fuzzcmpe (double a, double b, double e)
int	fuzzcmpef (float a, float b, float e)
int	fuzzcmpb (double a, double b, int bits)
int	fuzzcmpbf (float a, float b, int bits)

Detailed Description

Compare floating point numbers match within a tolerance.

The tolerance can be specified in absolute terms or it can be relative to the magnitude of the numbers being compared.

Macro Definition Documentation

#define asserteq	(		a,
			b
	)		assert(fuzzcmp(a, b, 1.000010))

Assert two values are approximately equal.

Currently this checks the values differ by no more than 10 parts per million.

Definition at line 65 of file fuzz.h.

Function Documentation

int fuzzcmp	(	double	a,
		double	b,
		double	delta
	)

Test whether two values are equal to within a relative tolerance.

Returns

True, if a and b are approximately equal.

Definition at line 21 of file fuzz.c.

int fuzzcmpb	(	double	a,
		double	b,
		int	bits
	)

Test whether two values are equal within the first N bits.

Definition at line 73 of file fuzz.c.

int fuzzcmpbf	(	float	a,
		float	b,
		int	bits
	)

Single precision equivalence test within the first N bits.

Definition at line 79 of file fuzz.c.

int fuzzcmpe	(	double	a,
		double	b,
		double	e
	)

Test whether two values are equal to within an absolute tolerance.

Definition at line 61 of file fuzz.c.

int fuzzcmpef	(	float	a,
		float	b,
		float	e
	)

Single precision equivalence test using an absolute tolerance.

See fuzzcmpe

Definition at line 67 of file fuzz.c.

int fuzzcmpf	(	float	a,
		float	b,
		float	delta
	)

Single precision equivalence test using a relative tolerance.

See fuzzcmp

Definition at line 41 of file fuzz.c.