Adds the four single-precision, floating-point values of a and b (see _mm_add_ps)
在文件 AkSimd.h 第 182 行定义.