Trait simba::simd::SimdValue [−][src]
Base trait for every SIMD types.
Associated Types
type Element: SimdValue<Element = Self::Element, SimdBool = bool>[src]
The type of the elements of each lane of this SIMD value.
type SimdBool: SimdBool[src]
Type of the result of comparing two SIMD values like self.
Required methods
fn lanes() -> usize[src]
The number of lanes of this SIMD value.
fn splat(val: Self::Element) -> Self[src]
Initializes an SIMD value with each lanes set to val.
fn extract(&self, i: usize) -> Self::Element[src]
Extracts the i-th lane of self.
Panics if i >= Self::lanes().
unsafe fn extract_unchecked(&self, i: usize) -> Self::Element[src]
Extracts the i-th lane of self without bound-checking.
fn replace(&mut self, i: usize, val: Self::Element)[src]
Replaces the i-th lane of self by val.
Panics if i >= Self::lanes().
unsafe fn replace_unchecked(&mut self, i: usize, val: Self::Element)[src]
Replaces the i-th lane of self by val without bound-checking.
fn select(self, cond: Self::SimdBool, other: Self) -> Self[src]
Merges self and other depending on the lanes of cond.
For each lane of cond with bits set to 1, the result’s will contain the value of the lane of self.
For each lane of cond with bits set to 0, the result’s will contain the value of the lane of other.
Provided methods
fn map_lanes(self, f: impl Fn(Self::Element) -> Self::Element) -> Self where
Self: Clone, [src]
Self: Clone,
Applies a function to each lane of self.
Note that, while convenient, this method can be extremely slow as this
requires to extract each lane of self and then combine them again into
a new SIMD value.
fn zip_map_lanes(
self,
b: Self,
f: impl Fn(Self::Element, Self::Element) -> Self::Element
) -> Self where
Self: Clone, [src]
self,
b: Self,
f: impl Fn(Self::Element, Self::Element) -> Self::Element
) -> Self where
Self: Clone,
Applies a function to each lane of self paired with the corresponding lane of b.
Note that, while convenient, this method can be extremely slow as this
requires to extract each lane of self and then combine them again into
a new SIMD value.