Computes the minima of the four single-precision, floating-point values of a and b (see _mm_min_ps)
在文件 AkSimd.h 第 210 行定义.