org.apache.spark.util.random

Class XORShiftRandom

Object

java.util.Random

org.apache.spark.util.random.XORShiftRandom

All Implemented Interfaces:

java.io.Serializable

public class XORShiftRandom
extends java.util.Random

This class implements a XORShift random number generator algorithm Source: Marsaglia, G. (2003). Xorshift RNGs. Journal of Statistical Software, Vol. 8, Issue 14.

See Also:

Paper This implementation is approximately 3.5 times faster than {@link java.util.Random java.util.Random}, partly because of the algorithm, but also due to renouncing thread safety. JDK's implementation uses an AtomicLong seed, this class uses a regular Long. We can forgo thread safety since we use a new instance of the RNG for each thread., Serialized Form

Constructor Summary

Constructors

Constructor and Description
XORShiftRandom()
XORShiftRandom(long init)

Method Summary

Methods

Modifier and Type	Method and Description
static scala.collection.immutable.Map<String,Object>	benchmark(int numIters)
static void	main(String[] args) Main method for running benchmark
void	setSeed(long s)

Methods inherited from class java.util.Random

nextBoolean, nextBytes, nextDouble, nextFloat, nextGaussian, nextInt, nextInt, nextLong

Methods inherited from class Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

XORShiftRandom

public XORShiftRandom(long init)

XORShiftRandom

public XORShiftRandom()

Method Detail

main

public static void main(String[] args)

Main method for running benchmark

Parameters:

args - takes one argument - the number of random numbers to generate

benchmark

public static scala.collection.immutable.Map<String,Object> benchmark(int numIters)

Parameters:

numIters - Number of random numbers to generate while running the benchmark

Returns:

Map of execution times for java.util.Random and XORShift

setSeed

public void setSeed(long s)

Overrides:

setSeed in class java.util.Random