pub trait SimdFloat: Copy + Sealed {
type Mask;
type Scalar;
type Bits;
type Cast<T: SimdElement>;
Show 24 methods
// Required methods
fn cast<T>(self) -> Self::Cast<T>
where T: SimdCast;
unsafe fn to_int_unchecked<I>(self) -> Self::Cast<I>
where I: SimdCast,
Self::Scalar: FloatToInt<I>;
fn to_bits(self) -> Self::Bits;
fn from_bits(bits: Self::Bits) -> Self;
fn abs(self) -> Self;
fn recip(self) -> Self;
fn to_degrees(self) -> Self;
fn to_radians(self) -> Self;
fn is_sign_positive(self) -> Self::Mask;
fn is_sign_negative(self) -> Self::Mask;
fn is_nan(self) -> Self::Mask;
fn is_infinite(self) -> Self::Mask;
fn is_finite(self) -> Self::Mask;
fn is_subnormal(self) -> Self::Mask;
fn is_normal(self) -> Self::Mask;
fn signum(self) -> Self;
fn copysign(self, sign: Self) -> Self;
fn simd_min(self, other: Self) -> Self;
fn simd_max(self, other: Self) -> Self;
fn simd_clamp(self, min: Self, max: Self) -> Self;
fn reduce_sum(self) -> Self::Scalar;
fn reduce_product(self) -> Self::Scalar;
fn reduce_max(self) -> Self::Scalar;
fn reduce_min(self) -> Self::Scalar;
}portable_simd #86656)Expand description
Operations on SIMD vectors of floats.
Required Associated Types§
sourcetype Mask
type Mask
portable_simd #86656)Mask type used for manipulating this SIMD vector type.
sourcetype Scalar
type Scalar
portable_simd #86656)Scalar type contained by this SIMD vector type.
sourcetype Bits
type Bits
portable_simd #86656)Bit representation of this SIMD vector type.
sourcetype Cast<T: SimdElement>
type Cast<T: SimdElement>
portable_simd #86656)A SIMD vector with a different element type.
Required Methods§
sourcefn cast<T>(self) -> Self::Cast<T>where
T: SimdCast,
fn cast<T>(self) -> Self::Cast<T>where
T: SimdCast,
portable_simd #86656)Performs elementwise conversion of this vector’s elements to another SIMD-valid type.
This follows the semantics of Rust’s as conversion for floats (truncating or saturating
at the limits) for each element.
Example
let floats: Simd<f32, 4> = Simd::from_array([1.9, -4.5, f32::INFINITY, f32::NAN]);
let ints = floats.cast::<i32>();
assert_eq!(ints, Simd::from_array([1, -4, i32::MAX, 0]));
// Formally equivalent, but `Simd::cast` can optimize better.
assert_eq!(ints, Simd::from_array(floats.to_array().map(|x| x as i32)));
// The float conversion does not round-trip.
let floats_again = ints.cast();
assert_ne!(floats, floats_again);
assert_eq!(floats_again, Simd::from_array([1.0, -4.0, 2147483647.0, 0.0]));Runsourceunsafe fn to_int_unchecked<I>(self) -> Self::Cast<I>
unsafe fn to_int_unchecked<I>(self) -> Self::Cast<I>
portable_simd #86656)Rounds toward zero and converts to the same-width integer type, assuming that the value is finite and fits in that type.
Safety
The value must:
- Not be NaN
- Not be infinite
- Be representable in the return type, after truncating off its fractional part
If these requirements are infeasible or costly, consider using the safe function cast, which saturates on conversion.
sourcefn to_bits(self) -> Self::Bits
fn to_bits(self) -> Self::Bits
portable_simd #86656)Raw transmutation to an unsigned integer vector type with the same size and number of elements.
sourcefn from_bits(bits: Self::Bits) -> Self
fn from_bits(bits: Self::Bits) -> Self
portable_simd #86656)Raw transmutation from an unsigned integer vector type with the same size and number of elements.
sourcefn abs(self) -> Self
fn abs(self) -> Self
portable_simd #86656)Produces a vector where every element has the absolute value of the
equivalently-indexed element in self.
sourcefn recip(self) -> Self
fn recip(self) -> Self
portable_simd #86656)Takes the reciprocal (inverse) of each element, 1/x.
sourcefn to_degrees(self) -> Self
fn to_degrees(self) -> Self
portable_simd #86656)Converts each element from radians to degrees.
sourcefn to_radians(self) -> Self
fn to_radians(self) -> Self
portable_simd #86656)Converts each element from degrees to radians.
sourcefn is_sign_positive(self) -> Self::Mask
fn is_sign_positive(self) -> Self::Mask
portable_simd #86656)Returns true for each element if it has a positive sign, including
+0.0, NaNs with positive sign bit and positive infinity.
sourcefn is_sign_negative(self) -> Self::Mask
fn is_sign_negative(self) -> Self::Mask
portable_simd #86656)Returns true for each element if it has a negative sign, including
-0.0, NaNs with negative sign bit and negative infinity.
sourcefn is_nan(self) -> Self::Mask
fn is_nan(self) -> Self::Mask
portable_simd #86656)Returns true for each element if its value is NaN.
sourcefn is_infinite(self) -> Self::Mask
fn is_infinite(self) -> Self::Mask
portable_simd #86656)Returns true for each element if its value is positive infinity or negative infinity.
sourcefn is_finite(self) -> Self::Mask
fn is_finite(self) -> Self::Mask
portable_simd #86656)Returns true for each element if its value is neither infinite nor NaN.
sourcefn is_subnormal(self) -> Self::Mask
fn is_subnormal(self) -> Self::Mask
portable_simd #86656)Returns true for each element if its value is subnormal.
sourcefn is_normal(self) -> Self::Mask
fn is_normal(self) -> Self::Mask
portable_simd #86656)Returns true for each element if its value is neither zero, infinite,
subnormal, nor NaN.
sourcefn signum(self) -> Self
fn signum(self) -> Self
portable_simd #86656)Replaces each element with a number that represents its sign.
1.0if the number is positive,+0.0, orINFINITY-1.0if the number is negative,-0.0, orNEG_INFINITYNANif the number isNAN
sourcefn copysign(self, sign: Self) -> Self
fn copysign(self, sign: Self) -> Self
portable_simd #86656)Returns each element with the magnitude of self and the sign of sign.
For any element containing a NAN, a NAN with the sign of sign is returned.
sourcefn simd_min(self, other: Self) -> Self
fn simd_min(self, other: Self) -> Self
portable_simd #86656)Returns the minimum of each element.
If one of the values is NAN, then the other value is returned.
sourcefn simd_max(self, other: Self) -> Self
fn simd_max(self, other: Self) -> Self
portable_simd #86656)Returns the maximum of each element.
If one of the values is NAN, then the other value is returned.
sourcefn simd_clamp(self, min: Self, max: Self) -> Self
fn simd_clamp(self, min: Self, max: Self) -> Self
portable_simd #86656)Restrict each element to a certain interval unless it is NaN.
For each element in self, returns the corresponding element in max if the element is
greater than max, and the corresponding element in min if the element is less
than min. Otherwise returns the element in self.
sourcefn reduce_sum(self) -> Self::Scalar
fn reduce_sum(self) -> Self::Scalar
portable_simd #86656)sourcefn reduce_product(self) -> Self::Scalar
fn reduce_product(self) -> Self::Scalar
portable_simd #86656)sourcefn reduce_max(self) -> Self::Scalar
fn reduce_max(self) -> Self::Scalar
portable_simd #86656)Returns the maximum element in the vector.
Returns values based on equality, so a vector containing both 0. and -0. may
return either.
This function will not return NaN unless all elements are NaN.
Examples
let v = f32x2::from_array([1., 2.]);
assert_eq!(v.reduce_max(), 2.);
// NaN values are skipped...
let v = f32x2::from_array([1., f32::NAN]);
assert_eq!(v.reduce_max(), 1.);
// ...unless all values are NaN
let v = f32x2::from_array([f32::NAN, f32::NAN]);
assert!(v.reduce_max().is_nan());Runsourcefn reduce_min(self) -> Self::Scalar
fn reduce_min(self) -> Self::Scalar
portable_simd #86656)Returns the minimum element in the vector.
Returns values based on equality, so a vector containing both 0. and -0. may
return either.
This function will not return NaN unless all elements are NaN.
Examples
let v = f32x2::from_array([3., 7.]);
assert_eq!(v.reduce_min(), 3.);
// NaN values are skipped...
let v = f32x2::from_array([1., f32::NAN]);
assert_eq!(v.reduce_min(), 1.);
// ...unless all values are NaN
let v = f32x2::from_array([f32::NAN, f32::NAN]);
assert!(v.reduce_min().is_nan());Run