1.27.0[][src]Function boolean_enums::lstd::arch::x86::_mm_hsub_ps

pub unsafe fn _mm_hsub_ps(a: __m128, b: __m128) -> __m128
This is supported on x86 and target feature sse3 only.

Horizontally add adjacent pairs of single-precision (32-bit) floating-point elements in a and b, and pack the results.

Intel's documentation