Computes the maximums of the two single-precision, floating-point values of a and b
在文件 AkSimd.h 第 453 行定义.