Function hadd_s

Source/astcenc_vecmathlib_sse_4.h:861–870

* @brief Return the horizontal sum of a vector as a scalar. */

 * @brief Return the horizontal sum of a vector as a scalar.
 */
ASTCENC_SIMD_INLINE float hadd_s(vfloat4 a)
{
	// Add top and bottom halves, lane 1/0
	__m128 t = _mm_add_ps(a.m, _mm_movehl_ps(a.m, a.m));

	// Add top and bottom halves, lane 0 (_mm_hadd_ps exists but slow)
	t = _mm_add_ss(t, _mm_shuffle_ps(t, t, 0x55));

	return _mm_cvtss_f32(t);
}

/**
 * @brief Return the sqrt of the lanes in the vector.