diff --git a/crates/core_simd/src/first.rs b/crates/core_simd/src/first.rs
index 50602829d48..f3f88f7aa44 100644
--- a/crates/core_simd/src/first.rs
+++ b/crates/core_simd/src/first.rs
@@ -3,26 +3,54 @@ macro_rules! impl_vector {
     { $name:ident, $type:ty } => {
         impl<const LANES: usize> $name<LANES> where Self: crate::LanesAtMost32 {
             /// Construct a SIMD vector by setting all lanes to the given value.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 1.0, 1.0, 1.0]);
+            /// let b = f32x4::splat(1.0);
+            /// assert_eq!(a, b);
+            /// ```
             pub const fn splat(value: $type) -> Self {
                 Self([value; LANES])
             }
 
             /// Returns a slice containing the entire SIMD vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 1.0, 1.0, 1.0]);
+            /// let b = a.as_slice();
+            /// ```
             pub const fn as_slice(&self) -> &[$type] {
                 &self.0
             }
 
             /// Returns a mutable slice containing the entire SIMD vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let mut a = f32x4::from_array([1.0, 1.0, 1.0, 1.0]);
+            /// let b = a.as_mut_slice();
+            /// ```
             pub fn as_mut_slice(&mut self) -> &mut [$type] {
                 &mut self.0
             }
 
             /// Converts an array to a SIMD vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 1.0, 1.0, 1.0]);
+            /// let b = f32x4::splat(1.0);
+            /// assert_eq!(a, b);
+            /// ```
             pub const fn from_array(array: [$type; LANES]) -> Self {
                 Self(array)
             }
 
             /// Converts a SIMD vector to an array.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 1.0, 1.0, 1.0]);
+            /// let b = [1.0, 1.0, 1.0, 1.0];
+            /// assert_eq!(f32x4::to_array(a), b);
+            /// ```
             pub const fn to_array(self) -> [$type; LANES] {
                 // workaround for rust-lang/rust#80108
                 // TODO fix this
@@ -47,6 +75,13 @@ macro_rules! impl_vector {
         impl<const LANES: usize> Copy for $name<LANES> where Self: crate::LanesAtMost32 {}
 
         impl<const LANES: usize> Clone for $name<LANES> where Self: crate::LanesAtMost32 {
+            /// Clone a SIMD vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 1.0, 1.0, 1.0]);
+            /// assert_eq!(a, f32x4::clone(&a));
+            /// assert_eq!(a, a.clone());
+            /// ```
             #[inline]
             fn clone(&self) -> Self {
                 *self
@@ -61,6 +96,12 @@ macro_rules! impl_vector {
         }
 
         impl<const LANES: usize> PartialEq for $name<LANES> where Self: crate::LanesAtMost32 {
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::splat(1.0);
+            /// let b = f32x4::splat(1.0);
+            /// assert!(a == b);
+            /// ```
             #[inline]
             fn eq(&self, other: &Self) -> bool {
                 // TODO use SIMD equality
diff --git a/crates/core_simd/src/reduction.rs b/crates/core_simd/src/reduction.rs
index 8687d1af516..1c053c4fde9 100644
--- a/crates/core_simd/src/reduction.rs
+++ b/crates/core_simd/src/reduction.rs
@@ -5,12 +5,26 @@ macro_rules! impl_integer_reductions {
             Self: crate::LanesAtMost32
         {
             /// Horizontal wrapping add.  Returns the sum of the lanes of the vector, with wrapping addition.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = u32x4::from_array([1, 2, 3, 4]).horizontal_sum();
+            /// assert_eq!(10, a);
+            /// let b = u32x4::from_array([u32::MAX, 1, 0, 0]).horizontal_sum();
+            /// assert_eq!(0, b);
+            /// ```
             #[inline]
             pub fn horizontal_sum(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_add_ordered(self, 0) }
             }
 
             /// Horizontal wrapping multiply.  Returns the product of the lanes of the vector, with wrapping multiplication.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = u32x4::from_array([2, 2, 2, 2]).horizontal_product();
+            /// assert_eq!(16, a);
+            /// let b = u32x4::from_array([u32::MAX, 2, 1, 1]).horizontal_product();
+            /// assert_eq!(u32::MAX, b + 1); // Because 2*u32::MAX wraps, and is 1 off from the u32::MAX
+            /// ```
             #[inline]
             pub fn horizontal_product(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_mul_ordered(self, 1) }
@@ -18,6 +32,13 @@ macro_rules! impl_integer_reductions {
 
             /// Horizontal bitwise "and".  Returns the cumulative bitwise "and" across the lanes of
             /// the vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = u32x4::from_array([3, 3, 3, 3]).horizontal_and();
+            /// assert_eq!(3, a);
+            /// let b = u32x4::from_array([1, 1, 0, 0]).horizontal_and();
+            /// assert_eq!(0, b);
+            /// ```
             #[inline]
             pub fn horizontal_and(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_and(self) }
@@ -25,6 +46,11 @@ macro_rules! impl_integer_reductions {
 
             /// Horizontal bitwise "or".  Returns the cumulative bitwise "or" across the lanes of
             /// the vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = u32x4::from_array([1, 2, 0, 0]).horizontal_or();
+            /// assert_eq!(3, a);
+            /// ```
             #[inline]
             pub fn horizontal_or(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_or(self) }
@@ -32,18 +58,33 @@ macro_rules! impl_integer_reductions {
 
             /// Horizontal bitwise "xor".  Returns the cumulative bitwise "xor" across the lanes of
             /// the vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = u32x4::from_array([5, 5, 5, 0]).horizontal_xor();
+            /// assert_eq!(5, a);
+            /// ```
             #[inline]
             pub fn horizontal_xor(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_xor(self) }
             }
 
             /// Horizontal maximum.  Returns the maximum lane in the vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = u32x4::from_array([1, 2, 42, 0]).horizontal_max();
+            /// assert_eq!(42, a);
+            /// ```
             #[inline]
             pub fn horizontal_max(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_max(self) }
             }
 
             /// Horizontal minimum.  Returns the minimum lane in the vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = u32x4::from_array([1, 2, 42, 0]).horizontal_min();
+            /// assert_eq!(0, a);
+            /// ```
             #[inline]
             pub fn horizontal_min(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_min(self) }
@@ -59,7 +100,14 @@ macro_rules! impl_float_reductions {
             Self: crate::LanesAtMost32
         {
 
-            /// Horizontal add.  Returns the sum of the lanes of the vector.
+            /// Horizontal add.  Returns the sum of the lanes of the vector, with saturating addition.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 2.0, 3.0, 4.0]).horizontal_sum();
+            /// assert_eq!(10.0, a);
+            /// let b = f32x4::from_array([f32::MAX, 2.0, 0.0, 0.0]).horizontal_sum();
+            /// assert_eq!(f32::MAX, b);
+            /// ```
             #[inline]
             pub fn horizontal_sum(self) -> $scalar {
                 // LLVM sum is inaccurate on i586
@@ -71,6 +119,13 @@ macro_rules! impl_float_reductions {
             }
 
             /// Horizontal multiply.  Returns the product of the lanes of the vector.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 2.0, 3.0, 4.0]).horizontal_product();
+            /// assert_eq!(24.0, a);
+            /// let b = f32x4::from_array([f32::MAX, 2.0, 1.0, 1.0]).horizontal_product();
+            /// assert_eq!(f32::MAX, b);
+            /// ```
             #[inline]
             pub fn horizontal_product(self) -> $scalar {
                 // LLVM product is inaccurate on i586
@@ -85,6 +140,13 @@ macro_rules! impl_float_reductions {
             ///
             /// Returns values based on equality, so a vector containing both `0.` and `-0.` may
             /// return either.  This function will not return `NaN` unless all lanes are `NaN`.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 2.0, 3.0, 4.0]).horizontal_max();
+            /// assert_eq!(4.0, a);
+            /// let b = f32x4::from_array([f32::NAN, 2.0, 0.0, 0.0]).horizontal_max();
+            /// assert_eq!(2.0, b);
+            /// ```
             #[inline]
             pub fn horizontal_max(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_max(self) }
@@ -94,6 +156,13 @@ macro_rules! impl_float_reductions {
             ///
             /// Returns values based on equality, so a vector containing both `0.` and `-0.` may
             /// return either.  This function will not return `NaN` unless all lanes are `NaN`.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([1.0, 2.0, 3.0, 4.0]).horizontal_min();
+            /// assert_eq!(1.0, a);
+            /// let b = f32x4::from_array([f32::NAN, 2.0, 0.0, 0.0]).horizontal_min();
+            /// assert_eq!(0.0, b);
+            /// ```
             #[inline]
             pub fn horizontal_min(self) -> $scalar {
                 unsafe { crate::intrinsics::simd_reduce_min(self) }
diff --git a/crates/core_simd/src/round.rs b/crates/core_simd/src/round.rs
index 281851c68ac..7bd97f300ba 100644
--- a/crates/core_simd/src/round.rs
+++ b/crates/core_simd/src/round.rs
@@ -8,6 +8,13 @@ macro_rules! implement {
             Self: crate::LanesAtMost32,
         {
             /// Returns the smallest integer greater than or equal to each lane.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([-3.3, -0.7, 2.9, 0.2]);
+            /// let b = f32x4::from_array([-3.0, 0.0, 3.0, 1.0]);
+            /// let c = f32x4::ceil(a);
+            /// assert_eq!(c, b);
+            /// ```
             #[must_use = "method returns a new vector and does not mutate the original value"]
             #[inline]
             pub fn ceil(self) -> Self {
@@ -15,13 +22,27 @@ macro_rules! implement {
             }
 
             /// Returns the largest integer value less than or equal to each lane.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([-3.3, -0.7, 2.3, 1.3]);
+            /// let b = f32x4::from_array([-4.0, -1.0, 2.0, 1.0]);
+            /// let c = f32x4::floor(a);
+            /// assert_eq!(c, b);
+            /// ```
             #[must_use = "method returns a new vector and does not mutate the original value"]
             #[inline]
             pub fn floor(self) -> Self {
                 unsafe { crate::intrinsics::simd_floor(self) }
             }
 
-            /// Rounds to the nearest integer value. Ties round toward zero.
+            /// Rounds to the nearest integer value. Ties round away from zero.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([-3.6, -0.5, 2.5, 0.6]);
+            /// let b = f32x4::from_array([-4.0, -1.0, 3.0, 1.0]);
+            /// let c = f32x4::round(a);
+            /// assert_eq!(c, b);
+            /// ```
             #[must_use = "method returns a new vector and does not mutate the original value"]
             #[inline]
             pub fn round(self) -> Self {
@@ -29,6 +50,13 @@ macro_rules! implement {
             }
 
             /// Returns the floating point's integer value, with its fractional part removed.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([-3.47, -0.1234, 2.3, 0.1234]);
+            /// let b = f32x4::from_array([-3.0, -0.0, 2.0, 0.0]);
+            /// let c = f32x4::trunc(a);
+            /// assert_eq!(c, b);
+            /// ```
             #[must_use = "method returns a new vector and does not mutate the original value"]
             #[inline]
             pub fn trunc(self) -> Self {
@@ -36,6 +64,13 @@ macro_rules! implement {
             }
 
             /// Returns the floating point's fractional value, with its integer part removed.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([-3.25, -0.75, 2.5, 10.0]);
+            /// let b = f32x4::from_array([-0.25, -0.75, 0.5, 0.0]);
+            /// let c = f32x4::fract(a);
+            /// assert_eq!(c, b);
+            /// ```
             #[must_use = "method returns a new vector and does not mutate the original value"]
             #[inline]
             pub fn fract(self) -> Self {
diff --git a/crates/core_simd/src/vector/float.rs b/crates/core_simd/src/vector/float.rs
index c4565a9dd90..73a834b9bb2 100644
--- a/crates/core_simd/src/vector/float.rs
+++ b/crates/core_simd/src/vector/float.rs
@@ -31,6 +31,13 @@ macro_rules! impl_float_vector {
 
             /// Produces a vector where every lane has the absolute value of the
             /// equivalently-indexed lane in `self`.
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([-3.0, -0.0, 2.0, 0.0]);
+            /// let b = f32x4::from_array([3.0, 0.0, 2.0, 0.0]);
+            /// let c = f32x4::abs(a);
+            /// assert_eq!(c, b);
+            /// ```
             #[inline]
             pub fn abs(self) -> Self {
                 unsafe { crate::intrinsics::simd_fabs(self) }
@@ -38,6 +45,13 @@ macro_rules! impl_float_vector {
 
             /// Produces a vector where every lane has the square root value
             /// of the equivalently-indexed lane in `self`
+            /// ```
+            /// # use core_simd::*;
+            /// let a = f32x4::from_array([9.0, 16.0, 4.0, 0.0]);
+            /// let b = f32x4::from_array([3.0, 4.0, 2.0, 0.0]);
+            /// let c = f32x4::sqrt(a);
+            /// assert_eq!(c, b);
+            /// ```
             #[inline]
             #[cfg(feature = "std")]
             pub fn sqrt(self) -> Self {