Function safe_arch::sum_of_u8_abs_diff_m256i

#[must_use]pub fn sum_of_u8_abs_diff_m256i(a: m256i, b: m256i) -> m256i

This is supported with target feature avx2 only.

Compute "sum of u8 absolute differences".

• u8 lanewise abs(a - b), producing u8 intermediate values.
• Sum the first eight and second eight values.
• Place into the low 16 bits of four u64 lanes.

let a = m256i::from([
  0_u8, 11, 2, 13, 4, 15, 6, 17, 8, 19, 20, 21, 22, 23, 24, 127, 0, 11, 2,
  13, 4, 15, 6, 17, 8, 19, 20, 21, 22, 23, 24, 127,
]);
let b = m256i::from([
  20_u8, 110, 250, 103, 34, 105, 60, 217, 8, 19, 210, 201, 202, 203, 204,
  127, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17,
]);
let c: [u64; 4] = sum_of_u8_abs_diff_m256i(a, b).into();
assert_eq!(c, [831_u64, 910, 40, 160]);