The sqrt function works using floating points. So maybe precision error varies among compilers.

The same thing happened to me too XD

Bro, the 2 solutions are the exactly same except for one being C++ 17 and the other one being C++ 20.

Let's say that I didn't calculate the time complexity.

But still, the other one is accepted.

Are you discussing this solution .

void solve()
{
    ll k;
    cin >> k;
    ll n = sqrtl(k);
    n += k;
    ll c = sqrtl(k);
    // cout << n -(int)sqrt(n)<< " ";
 
    while (true)
    {
        if (n - (ll)sqrtl(n) != k)
        {
            n++;
            c++;
        }
        else
        {
            break;
        }
        // n++;
        // c++;
    }
    cout << c + k << "\n";
}

Maybe the complexity is taken down by adding sqrt(k) in the first time of answer as it is necessary to be added, and then n++ would work as there would be a lot of difference between the next two consecutive perfect square numbers when we go for large values.

nah, I still use sqrt(x) but I've solved the precision error by integer multiplication.