float类型占32bit空间，只有7位精度，下面的程序
public class FloatPrecision
{
public static void Main()
{
Console.WriteLine((int)123456789F);
}
}
输出结果为123456792.

double占64bit空间，15~16位精度，下面的程序
public class DoublePrecision
{
public static void Main()
{
Console.WriteLine((long)123456789012345678D);
}
}
输出结果为123456789012345680

decimal占128bit空间，28位精度，MSDN中描述如下：
This type is useful for applications (such as accounting) where rounding errors must be avoided.