Detect if two integers have opposite signs

最新推荐文章于 2020-05-23 17:00:24 发布

yingsun

最新推荐文章于 2020-05-23 17:00:24 发布

阅读量1.1k

点赞数

分类专栏： bit magic

bit magic 专栏收录该内容

1 篇文章

订阅专栏

Given two signed integers, write a function that returns true if the signs of given integers are different, otherwise false. For example, the function should return true for -1 and +100, and should return false for -100 and -200. The function should not use any of the arithmetic operators（算术运算符，不包括逻辑运算符）.

Let the given integers be x and y. The sign bit is 1 in negative numbers, and 0 in positive numbers. The XOR of x and y will have the sign bit as 1 iff they have opposite sign. In other words, XOR of x and y will be negative number number iff x and y have opposite signs. The following code use this logic.

 
        #include<stdio.h>
 
booloppositeSigns(intx, inty)
{
    return((x ^ y) < 0);
}
 
intmain()
{
    intx = 100, y = -100;
    if(oppositeSigns(x, y) == true)
       printf("Signs are opposite");
    else
      printf("Signs are not opposite");
    return0;
}

Output:

Signs are opposite

We can also solve this by using two comparison operators. See the following code.

booloppositeSigns(intx, inty)
{
    return(x < 0)? (y >= 0): (y < 0);
}

The first method is more efficient. The first method uses a bitwise XOR and a comparison operator. The second method uses two comparison operators and a bitwise XOR operation is more efficient compared to a comparison operation.

booloppositeSigns(intx, inty)
{
 return((x ^ y) >> 31);
}

We can also use following method. It doesn’t use any comparison operator. The method is suggested by Hongliang and improved by gaurav.

The function is written only for compilers where size of an integer is 32 bit. The expression basically checks sign of (x^y) using bitwise operator ‘>>’. As mentioned above, the sign bit for negative numbers is always 1. The sign bit is the leftmost bit in binary representation. So we need to checks whether the 32th bit (or leftmost bit) of x^y is 1 or not. We do it by right shifting the value of x^y by 31, so that the sign bit becomes the least significant bit. If sign bit is 1, then the value of (x^y)>>31 will be 1, otherwise 0.