#include <cmath>
#include <cstdio>
#include <iostream>

using namespace std;
typedef long long LL;

int main()
{
    LL N,A,P,up;
    cin>>N;
    up=(LL)ceil((1.00+sqrt(1.00+8.00*(double)N))/2.00);
    P=up;
    while (((2*N)%P)!=0)
    {
        P--;
    }

    A=(((2*N)/P)-P+1)/2;

    cout<<A<<" "<<P<<endl;
}


This is my code...but this is giving WA on #2. I can't understand the reason. It seems to give correct answer on the tests I do on my computer..?!

At first sight, I strongly suggest you to avoid operating repeteadly with floating points, the little errors will accummulate, and WA comes to the order.

at second sight the logic is just not right. Look, I got the same algorithm that a dude with TLE at test9, but with a less complex break condition.

private static int sq(long A,long N){
    double aux= Math.sqrt( ((A*A)<<2) - (A<<2) + 1 + 8*N);
    int r = (int)aux;
    if ((double)r==aux)
        return r;
    else
        return -1;
    }
    public static void main (String[]args){
        Scanner s=new Scanner(System.in);
        int N= s.nextInt();
        long A=1,P=1,root=0;
        while (true){
            root= sq(A,N);
            if (root==-1)
                A++;
            else{
                P= (1-(A<<1)+root)/2;
                break;
            }
        }
        System.out.println(A+" "+(P));
    }