LCRand Class Reference

Implements a linear congruential pseudorandom number generator with period 2⁶⁴. More...

#include <geomc/random/LCRand.h>

Inheritance diagram for LCRand:

Public Member Functions
	LCRand (uint64_t seed)
virtual uint32_t	rand32 ()
virtual void	rseed (uint64_t seed)
	Reset the generator's state with the bits of seed as its source of entropy. More...
uint64_t	rand64 ()
template<typename T >
T	rand ()
template<typename T >
T	rand (T max)
template<typename T >
T	rand (T lo, T hi)
template<>
bool	rand ()
template<>
unsigned int	rand ()
template<>
int	rand ()
template<>
unsigned long	rand ()
template<>
long	rand ()
template<>
unsigned long long	rand ()
template<>
long long	rand ()
template<>
float	rand ()
template<>
double	rand ()
template<>
unsigned int	rand (unsigned int hi)
template<>
int	rand (int hi)
template<>
unsigned long	rand (unsigned long hi)
template<>
long	rand (long hi)
template<>
unsigned long long	rand (unsigned long long hi)
template<>
long long	rand (long long hi)
template<>
float	rand (float hi)
template<>
double	rand (double hi)
template<>
int	rand (int lo, int hi)
template<>
long long	rand (long long lo, long long hi)
template<>
float	rand (float lo, float hi)
template<>
double	rand (double lo, double hi)
template<>
bool	rand ()
template<>
unsigned int	rand ()
template<>
int	rand ()
template<>
unsigned long	rand ()
template<>
long	rand ()
template<>
unsigned long long	rand ()
template<>
long long	rand ()
template<>
float	rand ()
template<>
double	rand ()
template<>
unsigned int	rand (unsigned int hi)
template<>
int	rand (int hi)
template<>
unsigned long	rand (unsigned long hi)
template<>
long	rand (long hi)
template<>
unsigned long long	rand (unsigned long long hi)
template<>
long long	rand (long long hi)
template<>
float	rand (float hi)
template<>
double	rand (double hi)
template<>
int	rand (int lo, int hi)
template<>
long long	rand (long long lo, long long hi)
template<>
float	rand (float lo, float hi)
template<>
double	rand (double lo, double hi)

Protected Attributes
uint32_t	_bitpool
unsigned int	_bitsleft

Detailed Description

Implements a linear congruential pseudorandom number generator with period 2⁶⁴.

The parameters are the same as those chosen for Donald Knuth's MMIX language:

multiplier: 6364136223846793005
increment:  1442695040888963407

The underlying state is 64 bits; rand32() returns the 32 most significant digits.

The results of this generator are very fast and generally higher-quality and longer-period than those returned by std::rand().

This generator is NOT cryptographically secure; do NOT use for gambling or cryptography. An attacker may predict all future numbers by observing one or two results.

Member Function Documentation

rand() [1/24]

T rand ( )

inlineinherited

Produces a (pseudo-) random T with uniform distribution over the unit interval, if the unit interval is populated by T; or with uniform distribution over the entire space of possible T otherwise.

For example, rand<float>() returns a random float between 0 and 1.0, while rand<unsigned int>() returns a random uint between 0 and UINT_MAX.

rand() [2/24]

bool rand ( )

inherited

Returns

A random boolean value.

rand() [3/24]

unsigned int rand ( )

inherited

Returns

An unsigned int between 0 and UINT_MAX.

rand() [4/24]

int rand ( )

inherited

Returns

An int between INT_MIN and INT_MAX.

rand() [5/24]

unsigned long rand ( )

inherited

Returns

An unsigned long between 0 and ULONG_MAX.

rand() [6/24]

long rand ( )

inherited

Returns

A long between LONG_MIN and LONG_MAX.

rand() [7/24]

unsigned long long rand ( )

inherited

Returns

An unsigned long long betwen 0 and ULLONG_MAX.

rand() [8/24]

long long rand ( )

inherited

Returns

A long long between LLONG_MIN and LLONG_MAX.

rand() [9/24]

float rand ( )

inherited

Generate a floating-point value with uniform distribution between 0 and 1.0.

This method is preferable to use over the common practice of rand<int>() / (float)INT_MAX, as the latter excludes a large fraction of the representable floating point numbers between 0.0 and 1.0 (particularly those near zero), resulting in reduced entropy.

This method produces a number by carefully picking the bits of the exponent and mantissa explicitly, in such a way that the distibution remains uniform.

rand() [10/24]

double rand ( )

inherited

Generate a floating-point value with uniform distribution between 0 and 1.0.

This method is preferable to use over the common practice of rand<int>() / (float)INT_MAX, as the latter excludes a large fraction of the representable floating point numbers between 0.0 and 1.0 (particularly those near zero), resulting in reduced entropy.

This method produces a number by carefully picking the bits of the exponent and mantissa explicitly, in such a way that the distibution remains uniform.

rand() [11/24]

double rand ( double  hi )

inherited

Parameters

hi	Upper or lower bound of random number.

Returns

A double between 0 and hi, regardless of the sign of hi.

rand() [12/24]

double rand ( double  lo, double  hi  )

inherited

Parameters

lo	Lower bound of random number.
hi	Upper bound of random number.

Returns

An double between lo and hi.

rand() [13/24]

float rand ( float  hi )

inherited

Parameters

hi	Upper or lower bound of random number.

Returns

A float between 0 and hi, regardless of the sign of hi.

rand() [14/24]

float rand ( float  lo, float  hi  )

inherited

Parameters

lo	Lower bound of random number.
hi	Upper bound of random number.

Returns

An float between lo and hi.

rand() [15/24]

int rand ( int  hi )

inherited

Parameters

hi	Upper or lower bound of random number.

Returns

A number between 0 and hi (regardless of the sign of hi).

rand() [16/24]

int rand ( int  lo, int  hi  )

inherited

Parameters

lo	Lower bound of random number.
hi	Upper bound of random number.

Returns

An int between lo and hi.

rand() [17/24]

long rand ( long  hi )

inherited

Parameters

hi	Upper or lower bound of random number.

Returns

A long between 0 and hi, regardless of the sign of hi.

rand() [18/24]

long long rand ( long long  hi )

inherited

Parameters

hi	Upper or lower bound of random number.

Returns

A long long between 0 and hi, regardless of the sign of hi.

rand() [19/24]

long long rand ( long long  lo, long long  hi  )

inherited

Parameters

lo	Lower bound of random number.
hi	Upper bound of random number.

Returns

A long long between lo and hi.

rand() [20/24]

T rand ( T  lo, T  hi  )

inlineinherited

Produces a (pseudo-) random T with uniform distribution between lo and hi.

Parameters

lo	Lower bound for possible T samples.
hi	Upper bound for possible T samples.

rand() [21/24]

T rand ( T  max )

inlineinherited

Produces a (pseudo-) random T with uniform distribution between the T analogue of 0 and max.

Parameters

max	Upper bound of possible T samples.

rand() [22/24]

unsigned int rand ( unsigned int  hi )

inherited

Parameters

hi	Upper bound of random number.

Returns

An unsigned int between 0 and hi.

rand() [23/24]

unsigned long rand ( unsigned long  hi )

inherited

Parameters

hi	Upper bound of random number.

Returns

An unsigned long between 0 and hi.

rand() [24/24]

unsigned long long rand ( unsigned long long  hi )

inherited

Parameters

hi	Upper bound of random number.

Returns

An unsigned long long between 0 and hi.

rand32()

uint32_t rand32 ( )

virtual

Returns

No fewer than 32 (pseudo-) random bits.

Implements Random.

rand64()

uint64_t rand64 ( )

inline

Returns

64 random bits.

rseed()

void rseed ( uint64_t  seed )

virtual

Reset the generator's state with the bits of seed as its source of entropy.

Implements Random.

---

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

• geomc/random/LCRand.h
• geomc/random/LCRand.cpp