pub unsafe fn _mm_sub_pd(a: __m128d, b: __m128d) -> __m128d
Subtract packed double-precision (64-bit) floating-point elements in b from a.
b
a
Intel’s documentation