Struct U8x32

pub struct U8x32(/* private fields */);

[u8; 32] as a vector.

Implementations

impl U8x32

pub const fn from_array(arr: [u8; 32]) -> Self

Create a vector from an array.

Unlike the From trait function, the from_array function is const.

Example

const MY_EXTREMELY_FUN_VALUE: U8x32 = U8x32::from_array([
    0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22,
    23, 24, 25, 26, 27, 28, 29, 30, 31,
]);
for (i, value) in MY_EXTREMELY_FUN_VALUE.as_array().iter().copied().enumerate() {
    assert_eq!(i as u8, value);
}

impl U8x32

pub fn shuffle(&self, order: U8x32) -> U8x32

Scalar Equivalent:

let mut arr = [0; 32];
for (lane_dst, (lane_src, order)) in
    arr.chunks_exact_mut(16).zip(
        self.as_array().chunks_exact(16)
        .zip(order.as_array().chunks_exact(16))
    )
{
    for (dst, idx) in lane_dst.iter_mut().zip(order) {
        let idx = *idx;
        *dst = if (idx >> 7) == 1 {
            0
        } else {
            lane_src[(idx as usize) % 16]
        };
    }
}
arr.into()

Avx2

• _mm256_shuffle_epi8

  • VPSHUFB ymm, ymm, ymm

Trait Implementations

impl Add for U8x32

fn add(self, rhs: U8x32) -> U8x32

Perform a pairwise wrapping_add

Scalar Equivalent

U8x32::from([
    self.as_array()[0].wrapping_add(rhs.as_array()[0]),
    self.as_array()[1].wrapping_add(rhs.as_array()[1]),
    self.as_array()[2].wrapping_add(rhs.as_array()[2]),
    self.as_array()[3].wrapping_add(rhs.as_array()[3]),
    self.as_array()[4].wrapping_add(rhs.as_array()[4]),
    self.as_array()[5].wrapping_add(rhs.as_array()[5]),
    self.as_array()[6].wrapping_add(rhs.as_array()[6]),
    self.as_array()[7].wrapping_add(rhs.as_array()[7]),
    self.as_array()[8].wrapping_add(rhs.as_array()[8]),
    self.as_array()[9].wrapping_add(rhs.as_array()[9]),
    self.as_array()[10].wrapping_add(rhs.as_array()[10]),
    self.as_array()[11].wrapping_add(rhs.as_array()[11]),
    self.as_array()[12].wrapping_add(rhs.as_array()[12]),
    self.as_array()[13].wrapping_add(rhs.as_array()[13]),
    self.as_array()[14].wrapping_add(rhs.as_array()[14]),
    self.as_array()[15].wrapping_add(rhs.as_array()[15]),
    self.as_array()[16].wrapping_add(rhs.as_array()[16]),
    self.as_array()[17].wrapping_add(rhs.as_array()[17]),
    self.as_array()[18].wrapping_add(rhs.as_array()[18]),
    self.as_array()[19].wrapping_add(rhs.as_array()[19]),
    self.as_array()[20].wrapping_add(rhs.as_array()[20]),
    self.as_array()[21].wrapping_add(rhs.as_array()[21]),
    self.as_array()[22].wrapping_add(rhs.as_array()[22]),
    self.as_array()[23].wrapping_add(rhs.as_array()[23]),
    self.as_array()[24].wrapping_add(rhs.as_array()[24]),
    self.as_array()[25].wrapping_add(rhs.as_array()[25]),
    self.as_array()[26].wrapping_add(rhs.as_array()[26]),
    self.as_array()[27].wrapping_add(rhs.as_array()[27]),
    self.as_array()[28].wrapping_add(rhs.as_array()[28]),
    self.as_array()[29].wrapping_add(rhs.as_array()[29]),
    self.as_array()[30].wrapping_add(rhs.as_array()[30]),
    self.as_array()[31].wrapping_add(rhs.as_array()[31]),
])

AVX2 Intrinsics Used

• _mm256_add_epi8
  • VPADDB ymm, ymm, ymm

Neon Intrinsics Used

• vaddq_u8
  • This intrinsic compiles to the following instructions:
    • ADD Vd.16B,Vn.16B,Vm.16B

type Output = U8x32

The resulting type after applying the + operator.

impl AddAssign for U8x32

fn add_assign(&mut self, other: U8x32)

Performs the += operation.

impl AsMut<[u8]> for U8x32

fn as_mut(&mut self) -> &mut [u8]

Converts this type into a mutable reference of the (usually inferred) input type.

impl AsRef<[u8]> for U8x32

fn as_ref(&self) -> &[u8]

Converts this type into a shared reference of the (usually inferred) input type.

impl BitAnd for U8x32

fn bitand(self, rhs: U8x32) -> U8x32

Perform a pairwise bitwise and

Scalar Equivalent

U8x32::from([
    self.as_array()[0] & rhs.as_array()[0],
    self.as_array()[1] & rhs.as_array()[1],
    self.as_array()[2] & rhs.as_array()[2],
    self.as_array()[3] & rhs.as_array()[3],
    self.as_array()[4] & rhs.as_array()[4],
    self.as_array()[5] & rhs.as_array()[5],
    self.as_array()[6] & rhs.as_array()[6],
    self.as_array()[7] & rhs.as_array()[7],
    self.as_array()[8] & rhs.as_array()[8],
    self.as_array()[9] & rhs.as_array()[9],
    self.as_array()[10] & rhs.as_array()[10],
    self.as_array()[11] & rhs.as_array()[11],
    self.as_array()[12] & rhs.as_array()[12],
    self.as_array()[13] & rhs.as_array()[13],
    self.as_array()[14] & rhs.as_array()[14],
    self.as_array()[15] & rhs.as_array()[15],
    self.as_array()[16] & rhs.as_array()[16],
    self.as_array()[17] & rhs.as_array()[17],
    self.as_array()[18] & rhs.as_array()[18],
    self.as_array()[19] & rhs.as_array()[19],
    self.as_array()[20] & rhs.as_array()[20],
    self.as_array()[21] & rhs.as_array()[21],
    self.as_array()[22] & rhs.as_array()[22],
    self.as_array()[23] & rhs.as_array()[23],
    self.as_array()[24] & rhs.as_array()[24],
    self.as_array()[25] & rhs.as_array()[25],
    self.as_array()[26] & rhs.as_array()[26],
    self.as_array()[27] & rhs.as_array()[27],
    self.as_array()[28] & rhs.as_array()[28],
    self.as_array()[29] & rhs.as_array()[29],
    self.as_array()[30] & rhs.as_array()[30],
    self.as_array()[31] & rhs.as_array()[31],
])

AVX2 Intrinsics Used

• _mm256_and_si256
  • VPAND ymm, ymm, ymm

Neon Intrinsics Used

• vandq_u8
  • This intrinsic compiles to the following instructions:
    • AND Vd.16B,Vn.16B,Vm.16B

type Output = U8x32

The resulting type after applying the & operator.

impl BitAndAssign for U8x32

fn bitand_assign(&mut self, other: U8x32)

Performs the &= operation.

impl BitOr for U8x32

fn bitor(self, rhs: U8x32) -> U8x32

Perform a pairwise bitwise or

Scalar Equivalent

U8x32::from([
    self.as_array()[0] | rhs.as_array()[0],
    self.as_array()[1] | rhs.as_array()[1],
    self.as_array()[2] | rhs.as_array()[2],
    self.as_array()[3] | rhs.as_array()[3],
    self.as_array()[4] | rhs.as_array()[4],
    self.as_array()[5] | rhs.as_array()[5],
    self.as_array()[6] | rhs.as_array()[6],
    self.as_array()[7] | rhs.as_array()[7],
    self.as_array()[8] | rhs.as_array()[8],
    self.as_array()[9] | rhs.as_array()[9],
    self.as_array()[10] | rhs.as_array()[10],
    self.as_array()[11] | rhs.as_array()[11],
    self.as_array()[12] | rhs.as_array()[12],
    self.as_array()[13] | rhs.as_array()[13],
    self.as_array()[14] | rhs.as_array()[14],
    self.as_array()[15] | rhs.as_array()[15],
    self.as_array()[16] | rhs.as_array()[16],
    self.as_array()[17] | rhs.as_array()[17],
    self.as_array()[18] | rhs.as_array()[18],
    self.as_array()[19] | rhs.as_array()[19],
    self.as_array()[20] | rhs.as_array()[20],
    self.as_array()[21] | rhs.as_array()[21],
    self.as_array()[22] | rhs.as_array()[22],
    self.as_array()[23] | rhs.as_array()[23],
    self.as_array()[24] | rhs.as_array()[24],
    self.as_array()[25] | rhs.as_array()[25],
    self.as_array()[26] | rhs.as_array()[26],
    self.as_array()[27] | rhs.as_array()[27],
    self.as_array()[28] | rhs.as_array()[28],
    self.as_array()[29] | rhs.as_array()[29],
    self.as_array()[30] | rhs.as_array()[30],
    self.as_array()[31] | rhs.as_array()[31],
])

AVX2 Intrinsics Used

• _mm256_or_si256
  • VPOR ymm, ymm, ymm

Neon Intrinsics Used

• vorrq_u8
  • This intrinsic compiles to the following instructions:
    • ORR Vd.16B,Vn.16B,Vm.16B

type Output = U8x32

The resulting type after applying the | operator.

impl BitOrAssign for U8x32

fn bitor_assign(&mut self, other: U8x32)

Performs the |= operation.

impl BitXor for U8x32

fn bitxor(self, rhs: U8x32) -> U8x32

Perform a pairwise bitwise xor

Scalar Equivalent

U8x32::from([
    self.as_array()[0] ^ rhs.as_array()[0],
    self.as_array()[1] ^ rhs.as_array()[1],
    self.as_array()[2] ^ rhs.as_array()[2],
    self.as_array()[3] ^ rhs.as_array()[3],
    self.as_array()[4] ^ rhs.as_array()[4],
    self.as_array()[5] ^ rhs.as_array()[5],
    self.as_array()[6] ^ rhs.as_array()[6],
    self.as_array()[7] ^ rhs.as_array()[7],
    self.as_array()[8] ^ rhs.as_array()[8],
    self.as_array()[9] ^ rhs.as_array()[9],
    self.as_array()[10] ^ rhs.as_array()[10],
    self.as_array()[11] ^ rhs.as_array()[11],
    self.as_array()[12] ^ rhs.as_array()[12],
    self.as_array()[13] ^ rhs.as_array()[13],
    self.as_array()[14] ^ rhs.as_array()[14],
    self.as_array()[15] ^ rhs.as_array()[15],
    self.as_array()[16] ^ rhs.as_array()[16],
    self.as_array()[17] ^ rhs.as_array()[17],
    self.as_array()[18] ^ rhs.as_array()[18],
    self.as_array()[19] ^ rhs.as_array()[19],
    self.as_array()[20] ^ rhs.as_array()[20],
    self.as_array()[21] ^ rhs.as_array()[21],
    self.as_array()[22] ^ rhs.as_array()[22],
    self.as_array()[23] ^ rhs.as_array()[23],
    self.as_array()[24] ^ rhs.as_array()[24],
    self.as_array()[25] ^ rhs.as_array()[25],
    self.as_array()[26] ^ rhs.as_array()[26],
    self.as_array()[27] ^ rhs.as_array()[27],
    self.as_array()[28] ^ rhs.as_array()[28],
    self.as_array()[29] ^ rhs.as_array()[29],
    self.as_array()[30] ^ rhs.as_array()[30],
    self.as_array()[31] ^ rhs.as_array()[31],
])

AVX2 Intrinsics Used

• _mm256_xor_si256
  • VPXOR ymm, ymm, ymm

Neon Intrinsics Used

• veorq_u8
  • This intrinsic compiles to the following instructions:
    • EOR Vd.16B,Vn.16B,Vm.16B

type Output = U8x32

The resulting type after applying the ^ operator.

impl BitXorAssign for U8x32

fn bitxor_assign(&mut self, other: U8x32)

Performs the ^= operation.

impl Clone for U8x32

fn clone(&self) -> U8x32

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl ConditionallySelectable for U8x32

fn conditional_select(a: &Self, b: &Self, choice: Choice) -> Self

Select a or b according to choice.

fn conditional_assign(&mut self, other: &Self, choice: Choice)

Conditionally assign other to self, according to choice.

fn conditional_swap(a: &mut Self, b: &mut Self, choice: Choice)

Conditionally swap self and other if choice == 1; otherwise, reassign both unto themselves.

impl ConstantTimeEq for U8x32

fn ct_eq(&self, other: &Self) -> Choice

Determine if two items are equal.

fn ct_ne(&self, other: &Self) -> Choice

Determine if two items are NOT equal.

impl Debug for U8x32

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for U8x32

fn default() -> Self

The zero vector.

impl<'de> Deserialize<'de> for U8x32

fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>

where D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Distribution<U8x32> for Standard

fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> U8x32

Generate a random value of T, using rng as the source of randomness.

fn sample_iter<R>(self, rng: R) -> DistIter<Self, R, T>

where R: Rng, Self: Sized,

Create an iterator that generates random values of T, using rng as the source of randomness.

fn map<F, S>(self, func: F) -> DistMap<Self, F, T, S>

where F: Fn(T) -> S, Self: Sized,

Create a distribution of values of ‘S’ by mapping the output of Self through the closure F

impl From<[U8x16; 2]> for U8x32

fn from(vectors: [U8x16; 2]) -> U8x32

Scalar Equivalent:

let mut out = [0; 32];
out[0..16].copy_from_slice(&vectors[0].as_array());
out[16..].copy_from_slice(&vectors[1].as_array());
U8x32::from(out)

Avx2

• _mm256_set_m128i

  • VINSERTF128 ymm, ymm, xmm, imm8

impl From<[u8; 32]> for U8x32

fn from(arr: [u8; 32]) -> U8x32

Converts to this type from the input type.

impl From<I16x16> for U8x32

fn from(x: I16x16) -> U8x32

This cast is 100% free. It reinterprets the little-endinan bits of I16x16 as little endian bits of U8x32.

impl From<I32x8> for U8x32

fn from(x: I32x8) -> U8x32

This cast is 100% free. It reinterprets the little-endinan bits of I32x8 as little endian bits of U8x32.

impl From<I64x4> for U8x32

fn from(x: I64x4) -> U8x32

This cast is 100% free. It reinterprets the little-endinan bits of I64x4 as little endian bits of U8x32.

impl From<I8x32> for U8x32

fn from(x: I8x32) -> U8x32

This cast is 100% free. It reinterprets the little-endinan bits of I8x32 as little endian bits of U8x32.

impl From<U16x16> for U8x32

fn from(x: U16x16) -> U8x32

This cast is 100% free. It reinterprets the little-endinan bits of U16x16 as little endian bits of U8x32.

impl From<U32x8> for U8x32

fn from(x: U32x8) -> U8x32

This cast is 100% free. It reinterprets the little-endinan bits of U32x8 as little endian bits of U8x32.

impl From<U64x4> for U8x32

fn from(x: U64x4) -> U8x32

This cast is 100% free. It reinterprets the little-endinan bits of U64x4 as little endian bits of U8x32.

impl From<U8x16> for U8x32

fn from(vector: U8x16) -> U8x32

NOTE: this will zero the upper bits of the destination. Other intrinsics are more effcient, but leave the upper bits undefined. At present, these more effcient intrinsics are not exposed.

Scalar Equivalent:

let mut out = [0; 32];
out[0..16].copy_from_slice(&vector.as_array());
U8x32::from(out)

Avx2

• _mm256_zextsi128_si256

impl From<U8x32> for [U8x16; 2]

fn from(vector: U8x32) -> [U8x16; 2]

Scalar Equivalent:

let mut lo = [0; 16];
let mut hi = [0; 16];
lo.copy_from_slice(&vector.as_array()[0..16]);
hi.copy_from_slice(&vector.as_array()[16..]);
[U8x16::from(lo), U8x16::from(hi)]

Avx2

• _mm256_extracti128_si256

  • VEXTRACTI128 xmm, ymm, imm8

impl From<U8x32> for [u8; 32]

fn from(arr: U8x32) -> [u8; 32]

Converts to this type from the input type.

impl From<U8x32> for I16x16

fn from(x: U8x32) -> I16x16

This cast is 100% free. It reinterprets the little-endinan bits of U8x32 as little endian bits of I16x16.

impl From<U8x32> for I32x8

fn from(x: U8x32) -> I32x8

This cast is 100% free. It reinterprets the little-endinan bits of U8x32 as little endian bits of I32x8.

impl From<U8x32> for I64x4

fn from(x: U8x32) -> I64x4

This cast is 100% free. It reinterprets the little-endinan bits of U8x32 as little endian bits of I64x4.

impl From<U8x32> for I8x32

fn from(x: U8x32) -> I8x32

This cast is 100% free. It reinterprets the little-endinan bits of U8x32 as little endian bits of I8x32.

impl From<U8x32> for U16x16

fn from(x: U8x32) -> U16x16

This cast is 100% free. It reinterprets the little-endinan bits of U8x32 as little endian bits of U16x16.

impl From<U8x32> for U32x8

fn from(x: U8x32) -> U32x8

This cast is 100% free. It reinterprets the little-endinan bits of U8x32 as little endian bits of U32x8.

impl From<U8x32> for U64x4

fn from(x: U8x32) -> U64x4

This cast is 100% free. It reinterprets the little-endinan bits of U8x32 as little endian bits of U64x4.

impl Hash for U8x32

fn hash<H: Hasher>(&self, state: &mut H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl PartialEq for U8x32

fn eq(&self, other: &Self) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for U8x32

fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>

where S: Serializer,

Serialize this value into the given Serde serializer.

impl Shl<u64> for U8x32

fn shl(self, amount: u64) -> U8x32

Scalar Equivalent:

if amount >= 8 {
    U8x32::ZERO
} else {
    U8x32::from([
        self.as_array()[0] << amount,
        self.as_array()[1] << amount,
        self.as_array()[2] << amount,
        self.as_array()[3] << amount,
        self.as_array()[4] << amount,
        self.as_array()[5] << amount,
        self.as_array()[6] << amount,
        self.as_array()[7] << amount,
        self.as_array()[8] << amount,
        self.as_array()[9] << amount,
        self.as_array()[10] << amount,
        self.as_array()[11] << amount,
        self.as_array()[12] << amount,
        self.as_array()[13] << amount,
        self.as_array()[14] << amount,
        self.as_array()[15] << amount,
        self.as_array()[16] << amount,
        self.as_array()[17] << amount,
        self.as_array()[18] << amount,
        self.as_array()[19] << amount,
        self.as_array()[20] << amount,
        self.as_array()[21] << amount,
        self.as_array()[22] << amount,
        self.as_array()[23] << amount,
        self.as_array()[24] << amount,
        self.as_array()[25] << amount,
        self.as_array()[26] << amount,
        self.as_array()[27] << amount,
        self.as_array()[28] << amount,
        self.as_array()[29] << amount,
        self.as_array()[30] << amount,
        self.as_array()[31] << amount,
    ])
}

Avx2

WARNING: this implementation is a polyfill which executes the scalar implemenation.

type Output = U8x32

The resulting type after applying the << operator.

impl Shl for U8x32

fn shl(self, amount: U8x32) -> U8x32

Scalar Equivalent:

let mut out = self.as_array();
for (x, amm) in out.iter_mut().zip(amount.as_array().iter().copied()) {
    *x = if (0..8).contains(&amm) {
        *x << amm
    }  else {
        0
    };
}
U8x32::from(out)

Avx2

WARNING: this implementation is a polyfill which executes the scalar implemenation.

type Output = U8x32

The resulting type after applying the << operator.

impl ShlAssign<u64> for U8x32

fn shl_assign(&mut self, amount: u64)

Performs the <<= operation.

impl ShlAssign for U8x32

fn shl_assign(&mut self, amount: U8x32)

Performs the <<= operation.

impl Shr<u64> for U8x32

fn shr(self, amount: u64) -> U8x32

Scalar Equivalent:

if amount >= 8 {
    U8x32::ZERO
} else {
    U8x32::from([
        self.as_array()[0] >> amount,
        self.as_array()[1] >> amount,
        self.as_array()[2] >> amount,
        self.as_array()[3] >> amount,
        self.as_array()[4] >> amount,
        self.as_array()[5] >> amount,
        self.as_array()[6] >> amount,
        self.as_array()[7] >> amount,
        self.as_array()[8] >> amount,
        self.as_array()[9] >> amount,
        self.as_array()[10] >> amount,
        self.as_array()[11] >> amount,
        self.as_array()[12] >> amount,
        self.as_array()[13] >> amount,
        self.as_array()[14] >> amount,
        self.as_array()[15] >> amount,
        self.as_array()[16] >> amount,
        self.as_array()[17] >> amount,
        self.as_array()[18] >> amount,
        self.as_array()[19] >> amount,
        self.as_array()[20] >> amount,
        self.as_array()[21] >> amount,
        self.as_array()[22] >> amount,
        self.as_array()[23] >> amount,
        self.as_array()[24] >> amount,
        self.as_array()[25] >> amount,
        self.as_array()[26] >> amount,
        self.as_array()[27] >> amount,
        self.as_array()[28] >> amount,
        self.as_array()[29] >> amount,
        self.as_array()[30] >> amount,
        self.as_array()[31] >> amount,
    ])
}

Avx2

WARNING: this implementation is a polyfill which executes the scalar implemenation.

type Output = U8x32

The resulting type after applying the >> operator.

impl Shr for U8x32

fn shr(self, amount: U8x32) -> U8x32

Scalar Equivalent:

let mut out = self.as_array();
for (x, amm) in out.iter_mut().zip(amount.as_array().iter().copied()) {
    *x = if (0..8).contains(&amm) {
        *x >> amm
    }  else {
        0
    };
}
U8x32::from(out)

Avx2

WARNING: this implementation is a polyfill which executes the scalar implemenation.

type Output = U8x32

The resulting type after applying the >> operator.

impl ShrAssign<u64> for U8x32

fn shr_assign(&mut self, amount: u64)

Performs the >>= operation.

impl ShrAssign for U8x32

fn shr_assign(&mut self, amount: U8x32)

Performs the >>= operation.

impl SimdBase for U8x32

fn is_zero(&self) -> bool

Scalar Equivalent:

self.as_array().iter().all(|x| *x == 0)

Avx2

• _mm256_testz_si256

  • VPTEST ymm, ymm

fn set_lo(scalar: u8) -> U8x32

Scalar Equivalent:

let mut out = [0; 32];
out[0] = scalar;
U8x32::from(out)

Avx2

• _mm256_set_epi8

  Instruction sequence.

fn extract<const I: usize>(&self) -> u8

Scalar Equivalent:

self.as_array()[I]

Avx2

• _mm256_extract_epi8

  Instruction sequence.

fn broadcast(scalar: u8) -> U8x32

Scalar Equivalent:

U8x32::from([scalar; 32])

Avx2

• _mm256_set1_epi8

  Instruction sequence.

fn broadcast_lo(vector: U8x16) -> U8x32

Scalar Equivalent:

U8x32::from([vector.as_array()[0]; 32])

Avx2

• _mm256_broadcastb_epi8

  • VPBROADCASTB ymm, xmm

fn cmp_eq(&self, other: U8x32) -> U8x32

Scalar Equivalent:

U8x32::from([
    if self.as_array()[0] == other.as_array()[0] {  u8::MAX  } else { 0 },
    if self.as_array()[1] == other.as_array()[1] {  u8::MAX  } else { 0 },
    if self.as_array()[2] == other.as_array()[2] {  u8::MAX  } else { 0 },
    if self.as_array()[3] == other.as_array()[3] {  u8::MAX  } else { 0 },
    if self.as_array()[4] == other.as_array()[4] {  u8::MAX  } else { 0 },
    if self.as_array()[5] == other.as_array()[5] {  u8::MAX  } else { 0 },
    if self.as_array()[6] == other.as_array()[6] {  u8::MAX  } else { 0 },
    if self.as_array()[7] == other.as_array()[7] {  u8::MAX  } else { 0 },
    if self.as_array()[8] == other.as_array()[8] {  u8::MAX  } else { 0 },
    if self.as_array()[9] == other.as_array()[9] {  u8::MAX  } else { 0 },
    if self.as_array()[10] == other.as_array()[10] {  u8::MAX  } else { 0 },
    if self.as_array()[11] == other.as_array()[11] {  u8::MAX  } else { 0 },
    if self.as_array()[12] == other.as_array()[12] {  u8::MAX  } else { 0 },
    if self.as_array()[13] == other.as_array()[13] {  u8::MAX  } else { 0 },
    if self.as_array()[14] == other.as_array()[14] {  u8::MAX  } else { 0 },
    if self.as_array()[15] == other.as_array()[15] {  u8::MAX  } else { 0 },
    if self.as_array()[16] == other.as_array()[16] {  u8::MAX  } else { 0 },
    if self.as_array()[17] == other.as_array()[17] {  u8::MAX  } else { 0 },
    if self.as_array()[18] == other.as_array()[18] {  u8::MAX  } else { 0 },
    if self.as_array()[19] == other.as_array()[19] {  u8::MAX  } else { 0 },
    if self.as_array()[20] == other.as_array()[20] {  u8::MAX  } else { 0 },
    if self.as_array()[21] == other.as_array()[21] {  u8::MAX  } else { 0 },
    if self.as_array()[22] == other.as_array()[22] {  u8::MAX  } else { 0 },
    if self.as_array()[23] == other.as_array()[23] {  u8::MAX  } else { 0 },
    if self.as_array()[24] == other.as_array()[24] {  u8::MAX  } else { 0 },
    if self.as_array()[25] == other.as_array()[25] {  u8::MAX  } else { 0 },
    if self.as_array()[26] == other.as_array()[26] {  u8::MAX  } else { 0 },
    if self.as_array()[27] == other.as_array()[27] {  u8::MAX  } else { 0 },
    if self.as_array()[28] == other.as_array()[28] {  u8::MAX  } else { 0 },
    if self.as_array()[29] == other.as_array()[29] {  u8::MAX  } else { 0 },
    if self.as_array()[30] == other.as_array()[30] {  u8::MAX  } else { 0 },
    if self.as_array()[31] == other.as_array()[31] {  u8::MAX  } else { 0 },
])

Avx2

• _mm256_cmpeq_epi8

  • VPCMPEQB ymm, ymm, ymm

fn and_not(&self, other: U8x32) -> U8x32

Scalar Equivalent:

U8x32::from([
    self.as_array()[0] & (!other.as_array()[0]),
    self.as_array()[1] & (!other.as_array()[1]),
    self.as_array()[2] & (!other.as_array()[2]),
    self.as_array()[3] & (!other.as_array()[3]),
    self.as_array()[4] & (!other.as_array()[4]),
    self.as_array()[5] & (!other.as_array()[5]),
    self.as_array()[6] & (!other.as_array()[6]),
    self.as_array()[7] & (!other.as_array()[7]),
    self.as_array()[8] & (!other.as_array()[8]),
    self.as_array()[9] & (!other.as_array()[9]),
    self.as_array()[10] & (!other.as_array()[10]),
    self.as_array()[11] & (!other.as_array()[11]),
    self.as_array()[12] & (!other.as_array()[12]),
    self.as_array()[13] & (!other.as_array()[13]),
    self.as_array()[14] & (!other.as_array()[14]),
    self.as_array()[15] & (!other.as_array()[15]),
    self.as_array()[16] & (!other.as_array()[16]),
    self.as_array()[17] & (!other.as_array()[17]),
    self.as_array()[18] & (!other.as_array()[18]),
    self.as_array()[19] & (!other.as_array()[19]),
    self.as_array()[20] & (!other.as_array()[20]),
    self.as_array()[21] & (!other.as_array()[21]),
    self.as_array()[22] & (!other.as_array()[22]),
    self.as_array()[23] & (!other.as_array()[23]),
    self.as_array()[24] & (!other.as_array()[24]),
    self.as_array()[25] & (!other.as_array()[25]),
    self.as_array()[26] & (!other.as_array()[26]),
    self.as_array()[27] & (!other.as_array()[27]),
    self.as_array()[28] & (!other.as_array()[28]),
    self.as_array()[29] & (!other.as_array()[29]),
    self.as_array()[30] & (!other.as_array()[30]),
    self.as_array()[31] & (!other.as_array()[31]),
])

Avx2

• _mm256_andnot_si256

  • VPANDN ymm, ymm, ymm

fn cmp_gt(&self, other: U8x32) -> U8x32

Scalar Equivalent:

U8x32::from([
    if self.as_array()[0] > other.as_array()[0] {  u8::MAX  } else { 0 },
    if self.as_array()[1] > other.as_array()[1] {  u8::MAX  } else { 0 },
    if self.as_array()[2] > other.as_array()[2] {  u8::MAX  } else { 0 },
    if self.as_array()[3] > other.as_array()[3] {  u8::MAX  } else { 0 },
    if self.as_array()[4] > other.as_array()[4] {  u8::MAX  } else { 0 },
    if self.as_array()[5] > other.as_array()[5] {  u8::MAX  } else { 0 },
    if self.as_array()[6] > other.as_array()[6] {  u8::MAX  } else { 0 },
    if self.as_array()[7] > other.as_array()[7] {  u8::MAX  } else { 0 },
    if self.as_array()[8] > other.as_array()[8] {  u8::MAX  } else { 0 },
    if self.as_array()[9] > other.as_array()[9] {  u8::MAX  } else { 0 },
    if self.as_array()[10] > other.as_array()[10] {  u8::MAX  } else { 0 },
    if self.as_array()[11] > other.as_array()[11] {  u8::MAX  } else { 0 },
    if self.as_array()[12] > other.as_array()[12] {  u8::MAX  } else { 0 },
    if self.as_array()[13] > other.as_array()[13] {  u8::MAX  } else { 0 },
    if self.as_array()[14] > other.as_array()[14] {  u8::MAX  } else { 0 },
    if self.as_array()[15] > other.as_array()[15] {  u8::MAX  } else { 0 },
    if self.as_array()[16] > other.as_array()[16] {  u8::MAX  } else { 0 },
    if self.as_array()[17] > other.as_array()[17] {  u8::MAX  } else { 0 },
    if self.as_array()[18] > other.as_array()[18] {  u8::MAX  } else { 0 },
    if self.as_array()[19] > other.as_array()[19] {  u8::MAX  } else { 0 },
    if self.as_array()[20] > other.as_array()[20] {  u8::MAX  } else { 0 },
    if self.as_array()[21] > other.as_array()[21] {  u8::MAX  } else { 0 },
    if self.as_array()[22] > other.as_array()[22] {  u8::MAX  } else { 0 },
    if self.as_array()[23] > other.as_array()[23] {  u8::MAX  } else { 0 },
    if self.as_array()[24] > other.as_array()[24] {  u8::MAX  } else { 0 },
    if self.as_array()[25] > other.as_array()[25] {  u8::MAX  } else { 0 },
    if self.as_array()[26] > other.as_array()[26] {  u8::MAX  } else { 0 },
    if self.as_array()[27] > other.as_array()[27] {  u8::MAX  } else { 0 },
    if self.as_array()[28] > other.as_array()[28] {  u8::MAX  } else { 0 },
    if self.as_array()[29] > other.as_array()[29] {  u8::MAX  } else { 0 },
    if self.as_array()[30] > other.as_array()[30] {  u8::MAX  } else { 0 },
    if self.as_array()[31] > other.as_array()[31] {  u8::MAX  } else { 0 },
])

Avx2

NOTE: this implementation uses an efficient vector polyfill, though this operation is not natively supported.

// Based on https://stackoverflow.com/a/33173643 and https://git.io/JmghK
let sign_bit = Self::broadcast(1 << 7);
Self::from(I8x32::from(*self ^ sign_bit).cmp_gt(
    I8x32::from(other ^ sign_bit)
))

fn shift_left<const BITS: usize>(&self) -> U8x32

Scalar Equivalent:

let mut out = self.as_array();
for x in out.iter_mut() {
    *x <<= BITS;
}
U8x32::from(out)

Avx2

WARNING: this implementation is a polyfill which executes the scalar implemenation.

fn shift_right<const BITS: usize>(&self) -> U8x32

Scalar Equivalent:

let mut out = self.as_array();
for x in out.iter_mut() {
    *x >>= BITS;
}
U8x32::from(out)

Avx2

WARNING: this implementation is a polyfill which executes the scalar implemenation.

fn unpack_lo(&self, other: U8x32) -> U8x32

Scalar Equivalent:

U8x32::from([
    // Lane# 0
    self.as_array()[0],
    other.as_array()[0],
    self.as_array()[1],
    other.as_array()[1],
    self.as_array()[2],
    other.as_array()[2],
    self.as_array()[3],
    other.as_array()[3],
    self.as_array()[4],
    other.as_array()[4],
    self.as_array()[5],
    other.as_array()[5],
    self.as_array()[6],
    other.as_array()[6],
    self.as_array()[7],
    other.as_array()[7],
    // Lane# 1
    self.as_array()[16],
    other.as_array()[16],
    self.as_array()[17],
    other.as_array()[17],
    self.as_array()[18],
    other.as_array()[18],
    self.as_array()[19],
    other.as_array()[19],
    self.as_array()[20],
    other.as_array()[20],
    self.as_array()[21],
    other.as_array()[21],
    self.as_array()[22],
    other.as_array()[22],
    self.as_array()[23],
    other.as_array()[23],
])

Avx2

• _mm256_unpacklo_epi8

  • VPUNPCKLBW ymm, ymm, ymm

fn unpack_hi(&self, other: U8x32) -> U8x32

Scalar Equivalent:

U8x32::from([
    // Lane# 0
    self.as_array()[8],
    other.as_array()[8],
    self.as_array()[9],
    other.as_array()[9],
    self.as_array()[10],
    other.as_array()[10],
    self.as_array()[11],
    other.as_array()[11],
    self.as_array()[12],
    other.as_array()[12],
    self.as_array()[13],
    other.as_array()[13],
    self.as_array()[14],
    other.as_array()[14],
    self.as_array()[15],
    other.as_array()[15],
    // Lane# 1
    self.as_array()[24],
    other.as_array()[24],
    self.as_array()[25],
    other.as_array()[25],
    self.as_array()[26],
    other.as_array()[26],
    self.as_array()[27],
    other.as_array()[27],
    self.as_array()[28],
    other.as_array()[28],
    self.as_array()[29],
    other.as_array()[29],
    self.as_array()[30],
    other.as_array()[30],
    self.as_array()[31],
    other.as_array()[31],
])

Avx2

• _mm256_unpackhi_epi8

  • VPUNPCKHBW ymm, ymm, ymm

fn max(&self, other: U8x32) -> U8x32

Scalar Equivalent:

U8x32::from([
    self.as_array()[0].max(other.as_array()[0]),
    self.as_array()[1].max(other.as_array()[1]),
    self.as_array()[2].max(other.as_array()[2]),
    self.as_array()[3].max(other.as_array()[3]),
    self.as_array()[4].max(other.as_array()[4]),
    self.as_array()[5].max(other.as_array()[5]),
    self.as_array()[6].max(other.as_array()[6]),
    self.as_array()[7].max(other.as_array()[7]),
    self.as_array()[8].max(other.as_array()[8]),
    self.as_array()[9].max(other.as_array()[9]),
    self.as_array()[10].max(other.as_array()[10]),
    self.as_array()[11].max(other.as_array()[11]),
    self.as_array()[12].max(other.as_array()[12]),
    self.as_array()[13].max(other.as_array()[13]),
    self.as_array()[14].max(other.as_array()[14]),
    self.as_array()[15].max(other.as_array()[15]),
    self.as_array()[16].max(other.as_array()[16]),
    self.as_array()[17].max(other.as_array()[17]),
    self.as_array()[18].max(other.as_array()[18]),
    self.as_array()[19].max(other.as_array()[19]),
    self.as_array()[20].max(other.as_array()[20]),
    self.as_array()[21].max(other.as_array()[21]),
    self.as_array()[22].max(other.as_array()[22]),
    self.as_array()[23].max(other.as_array()[23]),
    self.as_array()[24].max(other.as_array()[24]),
    self.as_array()[25].max(other.as_array()[25]),
    self.as_array()[26].max(other.as_array()[26]),
    self.as_array()[27].max(other.as_array()[27]),
    self.as_array()[28].max(other.as_array()[28]),
    self.as_array()[29].max(other.as_array()[29]),
    self.as_array()[30].max(other.as_array()[30]),
    self.as_array()[31].max(other.as_array()[31]),
])

Avx2

• _mm256_max_epu8

  • VPMAXUB ymm, ymm, ymm

fn min(&self, other: U8x32) -> U8x32

Scalar Equivalent:

U8x32::from([
    self.as_array()[0].min(other.as_array()[0]),
    self.as_array()[1].min(other.as_array()[1]),
    self.as_array()[2].min(other.as_array()[2]),
    self.as_array()[3].min(other.as_array()[3]),
    self.as_array()[4].min(other.as_array()[4]),
    self.as_array()[5].min(other.as_array()[5]),
    self.as_array()[6].min(other.as_array()[6]),
    self.as_array()[7].min(other.as_array()[7]),
    self.as_array()[8].min(other.as_array()[8]),
    self.as_array()[9].min(other.as_array()[9]),
    self.as_array()[10].min(other.as_array()[10]),
    self.as_array()[11].min(other.as_array()[11]),
    self.as_array()[12].min(other.as_array()[12]),
    self.as_array()[13].min(other.as_array()[13]),
    self.as_array()[14].min(other.as_array()[14]),
    self.as_array()[15].min(other.as_array()[15]),
    self.as_array()[16].min(other.as_array()[16]),
    self.as_array()[17].min(other.as_array()[17]),
    self.as_array()[18].min(other.as_array()[18]),
    self.as_array()[19].min(other.as_array()[19]),
    self.as_array()[20].min(other.as_array()[20]),
    self.as_array()[21].min(other.as_array()[21]),
    self.as_array()[22].min(other.as_array()[22]),
    self.as_array()[23].min(other.as_array()[23]),
    self.as_array()[24].min(other.as_array()[24]),
    self.as_array()[25].min(other.as_array()[25]),
    self.as_array()[26].min(other.as_array()[26]),
    self.as_array()[27].min(other.as_array()[27]),
    self.as_array()[28].min(other.as_array()[28]),
    self.as_array()[29].min(other.as_array()[29]),
    self.as_array()[30].min(other.as_array()[30]),
    self.as_array()[31].min(other.as_array()[31]),
])

Avx2

• _mm256_min_epu8

  • VPMINUB ymm, ymm, ymm

const LANES: usize = 32usize

The number of elements of this vector.

const ZERO: Self

A vector where every element is zero.

type Scalar = u8

The scalar that this value holds.

type Array = [u8; 32]

The equivalent array type of this vector.

type Signed = I8x32

The signed version of this vector.

type Unsigned = U8x32

The unsigned version of this vector.

type BroadcastLoInput = U8x16

A vector of [Self::Scalar; 128 / (8 * std::mem::size_of::<Self::Scalar>())]

fn as_array(&self) -> Self::Array

Convert the vector to an array.

impl SimdBase8 for U8x32

fn shift_bytes_left<const AMOUNT: usize>(&self) -> U8x32

Scalar Equivalent:

let mut out = [0; 32];
for (out_lane, src_lane) in out
    .chunks_exact_mut(16)
    .zip(self.as_array().chunks_exact(16))
{
    out_lane[AMOUNT..].copy_from_slice(&src_lane[0..16 - AMOUNT]);
}
U8x32::from(out)

Avx2

• _mm256_slli_si256

  • VPSLLDQ ymm, ymm, imm8

fn shift_bytes_right<const AMOUNT: usize>(&self) -> U8x32

Scalar Equivalent:

let mut out = [0; 32];
for (out_lane, src_lane) in out
    .chunks_exact_mut(16)
    .zip(self.as_array().chunks_exact(16))
{
    out_lane[0..16 - AMOUNT].copy_from_slice(&src_lane[AMOUNT..]);
}
U8x32::from(out)

Avx2

• _mm256_srli_si256

  • VPSRLDQ ymm, ymm, imm8

fn most_significant_bits(&self) -> u32

Scalar Equivalent:

let mut out: u32 = 0;
for (i, value) in self.as_array().iter().copied().enumerate() {
    out |= u32::from((value as u8) >> 7) << i;
}
out

Avx2

• _mm256_movemask_epi8

  • VPMOVMSKB r32, ymm

impl SimdSaturatingArithmetic for U8x32

fn saturating_add(&self, other: U8x32) -> U8x32

Scalar Equivalent:

let mut out = self.as_array();
for (dst, src) in out.iter_mut().zip(other.as_array().iter()) {
    *dst = dst.saturating_add(*src);
}
Self::from(out)

Avx2

• _mm256_adds_epu8

  • VPADDUSB ymm, ymm, ymm

fn saturating_sub(&self, other: U8x32) -> U8x32

Scalar Equivalent:

let mut out = self.as_array();
for (dst, src) in out.iter_mut().zip(other.as_array().iter()) {
    *dst = dst.saturating_sub(*src);
}
Self::from(out)

Avx2

• _mm256_subs_epu8

  • VPSUBUSB ymm, ymm, ymm

impl Sub for U8x32

fn sub(self, rhs: U8x32) -> U8x32

Perform a pairwise wrapping_sub

Scalar Equivalent

U8x32::from([
    self.as_array()[0].wrapping_sub(rhs.as_array()[0]),
    self.as_array()[1].wrapping_sub(rhs.as_array()[1]),
    self.as_array()[2].wrapping_sub(rhs.as_array()[2]),
    self.as_array()[3].wrapping_sub(rhs.as_array()[3]),
    self.as_array()[4].wrapping_sub(rhs.as_array()[4]),
    self.as_array()[5].wrapping_sub(rhs.as_array()[5]),
    self.as_array()[6].wrapping_sub(rhs.as_array()[6]),
    self.as_array()[7].wrapping_sub(rhs.as_array()[7]),
    self.as_array()[8].wrapping_sub(rhs.as_array()[8]),
    self.as_array()[9].wrapping_sub(rhs.as_array()[9]),
    self.as_array()[10].wrapping_sub(rhs.as_array()[10]),
    self.as_array()[11].wrapping_sub(rhs.as_array()[11]),
    self.as_array()[12].wrapping_sub(rhs.as_array()[12]),
    self.as_array()[13].wrapping_sub(rhs.as_array()[13]),
    self.as_array()[14].wrapping_sub(rhs.as_array()[14]),
    self.as_array()[15].wrapping_sub(rhs.as_array()[15]),
    self.as_array()[16].wrapping_sub(rhs.as_array()[16]),
    self.as_array()[17].wrapping_sub(rhs.as_array()[17]),
    self.as_array()[18].wrapping_sub(rhs.as_array()[18]),
    self.as_array()[19].wrapping_sub(rhs.as_array()[19]),
    self.as_array()[20].wrapping_sub(rhs.as_array()[20]),
    self.as_array()[21].wrapping_sub(rhs.as_array()[21]),
    self.as_array()[22].wrapping_sub(rhs.as_array()[22]),
    self.as_array()[23].wrapping_sub(rhs.as_array()[23]),
    self.as_array()[24].wrapping_sub(rhs.as_array()[24]),
    self.as_array()[25].wrapping_sub(rhs.as_array()[25]),
    self.as_array()[26].wrapping_sub(rhs.as_array()[26]),
    self.as_array()[27].wrapping_sub(rhs.as_array()[27]),
    self.as_array()[28].wrapping_sub(rhs.as_array()[28]),
    self.as_array()[29].wrapping_sub(rhs.as_array()[29]),
    self.as_array()[30].wrapping_sub(rhs.as_array()[30]),
    self.as_array()[31].wrapping_sub(rhs.as_array()[31]),
])

AVX2 Intrinsics Used

• _mm256_sub_epi8
  • VPSUBB ymm, ymm, ymm

Neon Intrinsics Used

• vsubq_u8
  • This intrinsic compiles to the following instructions:
    • SUB Vd.16B,Vn.16B,Vm.16B

type Output = U8x32

The resulting type after applying the - operator.

impl SubAssign for U8x32

fn sub_assign(&mut self, other: U8x32)

Performs the -= operation.

impl Zeroable for U8x32

fn zeroed() -> Self

Calls zeroed.

impl Copy for U8x32

impl Eq for U8x32

impl Pod for U8x32