Class LongJumpDistances

java.lang.Object

org.apache.commons.rng.core.source64.LongJumpDistances

final class LongJumpDistances
extends Object

Utility for working with positive integer jump distances represented as 64-bit integers.

Since:

1.7

Field Summary

Fields

Modifier and Type	Field	Description
private static final long	MANTISSA_MASK	Mask to extract the 52-bit mantissa from a long representation of a double.
private static final int	NAN_EXP	The value of the exponent for NaN when the exponent is extracted and scaled to account for multiplying the floating-point 52-bit mantissa to an integer.
private static final int	SMALL_SHIFT	Small shift of a 53-bit significand where it remains within the same long.
private static final double	TWO_POW_63	2^63.

Constructor Summary

Constructors

Modifier	Constructor	Description
private	LongJumpDistances()	Class contains only static methods.

Method Summary

Modifier and Type	Method	Description
(package private) static void	validateJump(double distance, double period)	Validate the jump distance is valid within the given period of the cycle.
(package private) static void	validateJumpPowerOfTwo(int logDistance, int logPeriod)	Validate the jump distance of 2logDistance is valid within the given cycle period of 2logPeriod.
(package private) static void	writeUnsignedInteger(double value, long[] result)	Write the value to an unsigned integer representation.

Methods inherited from class Object

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

Field Details

TWO_POW_63

private static final double TWO_POW_63

2^63.

See Also:

• Constant Field Values

NAN_EXP

private static final int NAN_EXP

The value of the exponent for NaN when the exponent is extracted and scaled to account for multiplying the floating-point 52-bit mantissa to an integer. 1024 - 52 = 972.

See Also:

• Constant Field Values

MANTISSA_MASK

private static final long MANTISSA_MASK

Mask to extract the 52-bit mantissa from a long representation of a double.

See Also:

• Constant Field Values

SMALL_SHIFT

private static final int SMALL_SHIFT

Small shift of a 53-bit significand where it remains within the same long.

See Also:

• Constant Field Values

Constructor Details

LongJumpDistances

private LongJumpDistances()

Class contains only static methods.

Method Details

validateJump

static void validateJump(double distance,
 double period)

Validate the jump distance is valid within the given period of the cycle.

Parameters:

distance - Jump distance.

period - Cycle period.

Throws:

IllegalArgumentException - if distance is negative, or is greater than the period.

validateJumpPowerOfTwo

static void validateJumpPowerOfTwo(int logDistance,
 int logPeriod)

Validate the jump distance of 2^logDistance is valid within the given cycle period of 2^logPeriod.

Note: Negative logDistance is allowed as the distance is >= 0.

Parameters:

logDistance - Base-2 logarithm of the distance to jump forward with the state cycle.

logPeriod - Base-2 logarithm of the cycle period.

Throws:

IllegalArgumentException - if logDistance >= logPeriod.

writeUnsignedInteger

static void writeUnsignedInteger(double value,
 long[] result)

Write the value to an unsigned integer representation. Any fractional part of the floating-point value is discarded. The integer part of the value must be positive or an exception is raised.

The 53-bit significand of the value is converted to an integer, shifted to align with a 64-bit boundary and written to the provided result, ordered with the least significant bits first.

No bounds checking is performed. The result array is assumed to have a length of at least 1 to store an unshifted 53-bit significand. The required array size for values larger than 2^53 can be computed using:

1 + Math.getExponent(value) / 64

Parts of the array not used by the significand are unchanged. It is recommended to use a newly created array for the result.

Note: For jump distances validated as less than the period of an output cycle the result array length can be preallocated based on the maximum possible length.

Parameters:

value - Value

result - the integer representation

Throws:

IllegalArgumentException - if the value is negative or non-finite.