#pragma omp parallel for
for (i = 0; i < n; i++) {
__m128d sum = _mm_setzero_pd();
int j;
for (j = 0; j < n; j += 2) {
__m128d b = _mm_set_pd(v[j],v[j+1]);
__m128d a = _mm_set_pd(A(i,j), A(i,j+1));
sum = _mm_add_pd(sum, _mm_div_pd(b, a));
}
out[i] = sum[0] + sum[1];
}
}