Each floating point operation that combines two values will finish by rounding off the result to the nearest representable value.

• 每个结合两个浮点值的浮点操作都会将完成时的结果四舍五入到可被精确表示的那一位。