pub fn _mm256_broadcast_f32x2(a: __m128) -> __m256
Broadcasts the lower 2 packed single-precision (32-bit) floating-point elements from a to all elements of dst.
Intel’s Documentation