Math.f16round() returns the nearest 16-bit single precision floating point representation of a number: