I want it to crash

Short solution

As we see here , INT_MIN $=-2147483648/-1$ yields a SIGFPE signal.

Long solution

But, why does that particular combination give us an error. First we need to understand how 2's complement work.

2's complement

In our everyday lives, we deal with negative numbers, in various scenarios like owing someone money, temperature below 0 degrees, and more. It's easy for us, we just prepend a minus sign to the front of the number. But, how does a computer store negative numbers?

The following examples will be shown in a 4-bit binary representation (using only 4 bits to represent a binary number).

First, we need to know what's the 1's complement. Simply put, it's just the result of flipping all the bits of a number. For example, we have $7$ , whose binary representation is $0111$ , hence in a 4-bit binary representation, we can store $-7$ as its 1's complement, which is $1000$ . Another example is if we have $-5$ , whose binary representation is $1010$ in a 4-bit binary representation, we can get $5$ as its 1's complement, $0101$ .

As a result, if our most significant bit (leftmost bit) is equal to $1$ it is a negative number and, and in a n-bit binary representation our maximum value is $2^{n-1}-1$ , while our minimum value is $-2^{n-1}$ .

So far so good, we are able to now store and distinguish between positive and negative numbers. However, we have a new problem. Notice that 0 in a 4-bit binary representation is equal to $0000$ , but what about its 1's complement? It's $1111$ .

It does not make any sense. If we were to compare these two binary numbers, they should have the same value, which is $0$ , but here we can see that $0$ and $-0$ have different binary values. This poses a problem for us as we need to perform extra checks every time we compare between $0$ and $-0$ , rendering the system to be less efficient.

To solve this problem, we simply add $1$ to the 1's complement, and that's our 2's complement. We have $-5$ to be stored as the 2's complement of $5_{10}=0101_2$ , so $-5_{10}=1010_2+1_2=1011_2$ , and $-0_{10}=1111_2+1_2=10000_2=0000_2$ (note that we are using a 4-bit binary representation so the extra bit was truncated).

INT_MIN

If you remember from earlier, there is a maximum and minimum value for storing integers. For integer in particular, we have $4$ bytes, which is $32$ bits, whose minimum value is $-2^{32-1}=-2147483648$ and the maximum value is $2^{32-1}-1=2147483647$ .

From here, we can see that $\frac{-2147483648}{-1}=2147483648$ , which exceeds the maximum value of an integer. This causes there to be an integer overflow which triggers SIGFPE.

Note: A more proper/rigorous solution involves looking at abs() and seeing what it does to our number, that causes an integer overflow to occur. I'll leave it to you to read up.