pub unsafe fn _mm256_subs_epi8(a: __m256i, b: __m256i) -> __m256i
Subtract packed 8-bit integers in b from packed 8-bit integers in a using saturation.
b
a
Intel’s documentation