Package ptolemy.math

Class FixPointQuantization

java.lang.Object

ptolemy.math.Quantization

ptolemy.math.FixPointQuantization

All Implemented Interfaces:

java.lang.Cloneable

public class FixPointQuantization
extends Quantization

The FixPointQuantization class defines the mapping of numeric values with unlimited precision to the finite precision supported by arithmetic using the FixPoint class.

It comprises a

• Precision: to define the accuracy of the finite precision numbers.
• Overflow: to define the treatment of out-of-range numbers.
• Rounding: to define the loss of precision for in-range numbers.

The active class functionality is provided by the quantize method, which is normally invoked from FixPoint.quantize to enforce quantization constraints upon the result of an unconstrained computation.

An instance of the class is immutable, meaning that its value is set in the constructor and cannot then be modified.

Since:

Ptolemy II 2.1

Version:

$Id$

Author:

Ed Willink

See Also:

FixPoint, Precision, Overflow, Rounding

Pt.AcceptedRating:

Red

Pt.ProposedRating:

Red (Ed.Willink)

Field Summary

Fields

Modifier and Type	Field	Description
protected Precision	_precision	The precision.

Fields inherited from class ptolemy.math.Quantization

_overflow, _rounding

Constructor Summary

Constructors

Constructor	Description
FixPointQuantization(java.lang.String string)	Construct a FixPointQuantization object based on the provided string.
FixPointQuantization(Precision precision, Overflow overflow, Rounding rounding)	Construct a FixPointQuantization with specified precision, overflow and rounding.

Method Summary

Modifier and Type	Method	Description
boolean	equals(java.lang.Object object)	Return true if the given object describes the same mapping to quantized values as this object.
Precision	getMantissaPrecision()	Return the precision of the mantissa of the value.
Precision	getPrecision()	Return the precision.
int	hashCode()	Return a hash code value for this FixPointQuantization.
FixPointQuantization	setOverflow(Overflow overflow)	Return a FixPointQuantization with a changed overflow strategy.
FixPointQuantization	setPrecision(Precision precision)	Return a FixPointQuantization with a changed precision.
FixPointQuantization	setRounding(Rounding rounding)	Return a FixPointQuantization with a changed rounding strategy.
java.lang.String	toString()	Return a string representing this quantization.

Methods inherited from class ptolemy.math.Quantization

clone, getEpsilonValue, getExactOverflow, getExactRounding, getExponentBitLength, getFractionBitLength, getIntegerBitLength, getMantissaBitLength, getMaximumUnscaledValue, getMaximumValue, getMinimumUnscaledValue, getMinimumValue, getModuloUnscaledValue, getNumberOfBits, getNumberOfLevels, getOverflow, getRounding, getTinyValue

Methods inherited from class java.lang.Object

finalize, getClass, notify, notifyAll, wait, wait, wait

Field Detail

_precision

protected Precision _precision

The precision.

Constructor Detail

FixPointQuantization

public FixPointQuantization(java.lang.String string)
                     throws java.lang.IllegalArgumentException

Construct a FixPointQuantization object based on the provided string. The string may consist of just precision or precision, overflow or precision, overflow, rounding, and may optionally be enclosed in parentheses.

precision must be one of the Precision formats; integer-bits.fraction-bits or total-bits/integer-bits.

overflow must be one of the Overflow strategies; e.g. saturate or modulo or trap. The default is saturate.

rounding must be one of the Rounding strategies; e.g. up or half_even or unnecessary. The default is nearest.

Parameters:

string - The string representing the quantization specification.

Throws:

java.lang.IllegalArgumentException - If the precision string supplied, overflow strategy or rounding strategy does not match one of the known formats.

FixPointQuantization

public FixPointQuantization(Precision precision,
                            Overflow overflow,
                            Rounding rounding)

Construct a FixPointQuantization with specified precision, overflow and rounding.

Parameters:

precision - The precision.

overflow - The overflow.

rounding - The rounding.

Method Detail

equals

public boolean equals(java.lang.Object object)

Return true if the given object describes the same mapping to quantized values as this object.

Overrides:

equals in class Quantization

Parameters:

object - The FixPointQuantization object to use for equality checking.

Returns:

True if the quantizations are equal.

getMantissaPrecision

public Precision getMantissaPrecision()

Return the precision of the mantissa of the value.

Specified by:

getMantissaPrecision in class Quantization

Returns:

The precision.

getPrecision

public Precision getPrecision()

Return the precision.

Specified by:

getPrecision in class Quantization

Returns:

Precision object.

See Also:

setPrecision(Precision)

hashCode

public int hashCode()

Return a hash code value for this FixPointQuantization. This method returns the bitwise xor of the hashcode of the superclass (Quantization) and the hashcode of the precision.

Overrides:

hashCode in class Quantization

Returns:

A hash code value for this Quantization.

setOverflow

public FixPointQuantization setOverflow(Overflow overflow)

Return a FixPointQuantization with a changed overflow strategy.

Parameters:

overflow - The new overflow strategy.

Returns:

A new quantization.

setPrecision

public FixPointQuantization setPrecision(Precision precision)

Return a FixPointQuantization with a changed precision.

Parameters:

precision - The new precision.

Returns:

A new quantization.

See Also:

getPrecision()

setRounding

public FixPointQuantization setRounding(Rounding rounding)

Return a FixPointQuantization with a changed rounding strategy.

Parameters:

rounding - The new rounding.strategy.

Returns:

A new quantization.

toString

public java.lang.String toString()

Return a string representing this quantization. The string is expressed as "(m.n, o, p)", where m indicates the number of integer bits, n represents the number of fractional bits and o represents the overflow strategy and r represents the rounding strategy.

Specified by:

toString in class Quantization

Returns:

A string representing this quantization.