IEEE floating-point classes #
IEEE 754-2019 §5.7.2 specifies the same ten classification results for binary and decimal formats. Each format determines normality and its supported special values; the result type is independent of radix, precision, bias, and storage.
The ten mutually exclusive floating-point classes of IEEE 754-2019 §5.7.2.
- signalingNaN : IEEEClass
A NaN whose use by signaling operations raises invalid.
- quietNaN : IEEEClass
A quiet NaN.
- negativeInfinity : IEEEClass
Negative infinity.
- negativeNormal : IEEEClass
A negative finite normal number.
- negativeSubnormal : IEEEClass
A negative finite subnormal number.
- negativeZero : IEEEClass
Zero with its sign bit set.
- positiveZero : IEEEClass
Zero with its sign bit clear.
- positiveSubnormal : IEEEClass
A positive finite subnormal number.
- positiveNormal : IEEEClass
A positive finite normal number.
- positiveInfinity : IEEEClass
Positive infinity.
Instances For
@[instance_reducible]