horizontal sum of 8 packed 32bit floats [duplicate]

Answer 1

While Stephen Canon's answer is probably ideal for finding the horizontal maximum/minimum I think a better solution can be found for the horizontal sum.

float horizontal_add (__m256 a) {
    __m256 t1 = _mm256_hadd_ps(a,a);
    __m256 t2 = _mm256_hadd_ps(t1,t1);
    __m128 t3 = _mm256_extractf128_ps(t2,1);
    __m128 t4 = _mm_add_ss(_mm256_castps256_ps128(t2),t3);
    return _mm_cvtss_f32(t4);        
}