Implements a random number generator using the Marsaglia method. More...

#include "sst/core/rng/marsaglia.h"

Inheritance diagram for SST::RNG::MarsagliaRNG:

Public Member Functions
	MarsagliaRNG (unsigned int initial_z, unsigned int initial_w)
	Creates a new Marsaglia RNG using the initial seeds. More...

	MarsagliaRNG ()
	Creates a new Marsaglia RNG using random initial seeds (which are read from the system clock). More...

void	restart (unsigned int new_z, unsigned int new_w)
	Restart the random number generator with new seeds. More...

double	nextUniform () override
	Generates the next random number as a double in the range 0 to 1.

uint32_t	generateNextUInt32 () override
	Generates the next random number as an unsigned 32-bit integer.

uint64_t	generateNextUInt64 () override
	Generates the next random number as an unsigned 64-bit integer.

int64_t	generateNextInt64 () override
	Generates the next number as a signed 64-bit integer.

int32_t	generateNextInt32 () override
	Generates the next number as a signed 32-bit integer.

void	seed (uint64_t newSeed)
	Seed the XOR RNG.

Public Member Functions inherited from SST::RNG::Random
virtual	~Random ()
	Destroys the random number generator.

Detailed Description

Implements a random number generator using the Marsaglia method.

This method is computationally cheap and provides a reasonable distribution of random numbers. If you need additional strength in the random numbers you may want to consider the Mersenne RNG.

For more information see the Multiply-with-carry Random Number Generator article

at Wikipedia (http://en.wikipedia.org/wiki/Multiply-with-carry).

Constructor & Destructor Documentation

◆ MarsagliaRNG() [1/2]

MarsagliaRNG::MarsagliaRNG	(	unsigned int	initial_z,
		unsigned int	initial_w
	)

Creates a new Marsaglia RNG using the initial seeds.

Parameters

[in]	initial_z	One of the random seed pairs
[in]	initial_w	One of the random seed pairs.

◆ MarsagliaRNG() [2/2]

MarsagliaRNG::MarsagliaRNG ( )

Creates a new Marsaglia RNG using random initial seeds (which are read from the system clock).

Note that these will create variation between runs and between platforms.

Member Function Documentation

◆ restart()

void MarsagliaRNG::restart	(	unsigned int	new_z,
		unsigned int	new_w
	)

Restart the random number generator with new seeds.