Trait RealValue 

pub trait RealValue:
    Copy
    + Add<Output = Self>
    + Sub<Output = Self>
    + Mul<Output = Self>
    + Div<Output = Self>
    + Neg<Output = Self>
    + MulAddExt<Self, Self, Output = Self> {
    const VAL_ZERO: Self;
    const VAL_ONE: Self;

    // Required method
    fn val_trunc(self) -> Self;
}

A value usable as the real/imaginary storage of a Complex.

Implemented for f32/f64 and for every Thermite float Vector. The arithmetic below is written once against it and serves both the element level (Complex<f32>) and the vector level (Complex<Vector<R>>). The math library wants the stronger RealFloatVector.

Required Associated Constants

const VAL_ZERO: Self

The additive identity in this value type.

const VAL_ONE: Self

The multiplicative identity in this value type.

Required Methods

fn val_trunc(self) -> Self

Truncate towards zero. Used to give Complex a (componentwise) Rem.

Dyn Compatibility

This trait is not dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementations on Foreign Types

impl RealValue for f32

const VAL_ZERO: Self = 0.0

const VAL_ONE: Self = 1.0

fn val_trunc(self) -> Self

impl RealValue for f64

const VAL_ZERO: Self = 0.0

const VAL_ONE: Self = 1.0

fn val_trunc(self) -> Self

impl<R: FloatRegister> RealValue for Vector<R>

const VAL_ZERO: Self = <Self as thermite::prelude::NumericVector>::ZERO

const VAL_ONE: Self = <Self as thermite::prelude::NumericVector>::ONE

fn val_trunc(self) -> Self

impl<V: DualValue, const N: usize> RealValue for Dual<V, N>

Available on crate feature dual only.

Complex<Dual<V, N>>: a complex number whose parts each carry N derivative components, giving forward-mode AD through the complex functions.

Everything here is written against RealValue, which Dual satisfies, so this impl is all it takes. Seeded along the real axis (dz = 1), the dual parts of f(z) are f'(z) for holomorphic f.

const VAL_ZERO: Self = Self::ZERO

const VAL_ONE: Self = Self::ONE

fn val_trunc(self) -> Self

impl<V: ScalarValue> RealValue for Compensated<V>

Available on crate feature compensated only.

Complex<Compensated<V>>: a complex number whose parts are each a double-double, roughly doubling the mantissa of the complex arithmetic and of every kernel built on it.

Compensated carries no error term of its own through a complex multiply; the compensation is per component, and the cross terms of (a + bi)(c + di) are summed in double-double, which is where the precision comes from.

const VAL_ZERO: Self

const VAL_ONE: Self

fn val_trunc(self) -> Self

Implementors