Trait SimdFloatElement 

pub trait SimdFloatElement:
    SimdElement
    + Neg<Output = Self>
    + From<i8>
    + From<u8> {
    type Native<S: Simd>: SimdFloat<S, Element = Self>;
}

Types that can be used as elements in float SIMD vectors.

The scalar conversion bounds are limited to types that every floating-point element can represent losslessly, including f16 for forward-compatibility.

Self::Native selects the native-width vector for this scalar type and a SIMD backend, so a float-generic caller can use crate::dispatch!:

use fearless_simd::{dispatch, prelude::*, Level};

#[inline(always)]
fn sqrt_simd<S: Simd, T: SimdFloatElement>(simd: S, value: T) -> T {
    T::Native::<S>::splat(simd, value).sqrt()[0]
}

fn sqrt<T: SimdFloatElement>(level: Level, value: T) -> T {
    dispatch!(level, simd => sqrt_simd(simd, value))
}

let level = Level::new();
assert_eq!(sqrt(level, 4.0_f32), 2.0);
assert_eq!(sqrt(level, 4.0_f64), 2.0);

Required Associated Types

type Native<S: Simd>: SimdFloat<S, Element = Self>

The native-width floating-point vector for this scalar type and backend S.

Its lane count depends on both the scalar type and the backend, and is available as T::Native::<S>::LEN with SimdBase in scope.

Dyn Compatibility

This trait is not dyn compatible.

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

Implementations on Foreign Types

impl SimdFloatElement for f32

type Native<S: Simd> = <S as Simd>::f32s

impl SimdFloatElement for f64

type Native<S: Simd> = <S as Simd>::f64s

Implementors