Preface

We know that the remainder also needs to be called div Assembly instruction , however div If the instruction cycle is too long, compilers often perform arithmetic optimization , To avoid calling div Instructions .

The divisor is 2 Multiple

int main(int argc, char* args[])
{
    
	return   argc%8;
}

In order to understand this formula, we first understand some basic logic ：
We divide by 2 The displacement can be calculated , So for the remainder, it is also the same as the theoretical optimization thought .

We use 58 % 8 = 2 For example

Binary system ：
58 -> 0011 1010
8 -> 0000 1000
2 -> 0000 0010

This formula is correct in unsigned, such as the following assembly code

Corresponding source code

We just need to do the shift number and The result can be obtained by calculation .

For negative division, we need to do the following optimization ：

tip: The above figure 58>>3 error Change it to -15>>3

The above formula can be used when the remainder is not , If 0 Front complement 1 Obviously, it will lead to mistakes .
The compiler uses a trick to avoid branch judgment , Subtract one from the moved binary , Then complete the front 1, Finally, add 1.

The two extra steps above are for the remainder 0 Quite useful as follows :

According to the previous theory, we should now be able to understand the following compilation meaning

Non divisor is 2 Seeking remainder

int main( int argc, char* args[])
{
    

	return  argc % 7;
}