A random number generator. More...

#include <random_number_generator.hpp>

Public Member Functions
T	gaussian (const T &std, const T &mu=0)

	RandomNumberGenerator (const unsigned int randSeed)

	RandomNumberGenerator ()

T	uniform (const T &range)

T	uniform ()

Static Private Member Functions
static void	seed ()
	Seed the system random number generator. More...

Detailed Description

template<typename T>
class ecl::RandomNumberGenerator< T >

A random number generator.

Definition at line 41 of file random_number_generator.hpp.

Constructor & Destructor Documentation

template<typename T>

ecl::RandomNumberGenerator< T >::RandomNumberGenerator ( const unsigned int randSeed )

inline

Definition at line 44 of file random_number_generator.hpp.

template<typename T>

ecl::RandomNumberGenerator< T >::RandomNumberGenerator ( )

inline

Definition at line 49 of file random_number_generator.hpp.

Member Function Documentation

template<typename T>

T ecl::RandomNumberGenerator< T >::gaussian	(	const T &	std,
		const T &	mu = `0`
	)

inline

Definition at line 64 of file random_number_generator.hpp.

template<typename T>

static void ecl::RandomNumberGenerator< T >::seed ( )

inlinestaticprivate

Seed the system random number generator.

Here we create a seed by hashing from three independent sources. This is a more robust way to get a reliable seed that allows you to get the typically para-optimal traits

1) fine grained (e.g. you want to call srand two times in a very short interval) 2) low chance of collisions over a long period

It also avoids having to use esoteric linux functions lick clock_gettime that aren't available on all platforms.

References:

-h ttp://stackoverflow.com/questions/322938/recommended-way-to-initialize-srand