Convert the single-precision float value x to a half-precision floating point value represented in unsigned short format, in round-to-nearest-even mode.
x
unsigned short