Struct Norm

struct Norm<F: Float> {
    m: F::Int,
    e: i32,
    neg: bool,
}

Representation of F that has handled subnormals.

Fields

m: F::Int

Normalized significand with one guard bit, unsigned.

e: i32

Exponent of the mantissa such that m * 2^e = x. Accounts for the shift in the mantissa and the guard bit; that is, 1.0 will normalize as m = 1 << 53 and e = -53.

neg: bool

Implementations

impl<F: Float> Norm<F>

const EXP_UNBIAS: u32

Unbias the exponent and account for the mantissa’s precision, including the guard bit.

const ZERO_INF_NAN: u32

Values greater than this had a saturated exponent (infinity or NaN), OR were zero and we adjusted the exponent such that it exceeds this threashold.

fn from_float(x: F) -> Self

fn is_zero_nan_inf(self) -> bool

True if the value was zero, infinity, or NaN.

fn is_zero(self) -> bool

The only value we have

Trait Implementations

impl<F: Clone + Float> Clone for Norm<F>

where F::Int: Clone,

fn clone(&self) -> Norm<F>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<F: Debug + Float> Debug for Norm<F>

where F::Int: Debug,

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<F: Copy + Float> Copy for Norm<F>

where F::Int: Copy,