Function core::arch::x86_64::_mm_subs_epu16

pub unsafe fn _mm_subs_epu16(a: __m128i, b: __m128i) -> __m128i

This is supported on x86-64 and target feature sse2 only.

Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation.

Intel’s documentation