演算法原理請wiki：Tonelli–Shanks algorithm，迅速深入理解是不太可能的，與cipola演算法相比，shanks解法更數論一點。
（這個演算法是正常的，但是還是tle）
大概流程是：

#include<cstdio>
#include<cstdlib>
#include<algorithm>
#include<cstring>
using namespace std;
#define ll long long
ll QP(ll a, ll b, ll p)
{
    a %=
 
 p;
    ll tmp = 1;
    while (b)
    {
        if (b & 1)tmp = tmp*a%p;
        b >>= 1;
        a = a*a%p;
    }
    return tmp;
}
ll le(ll n, ll p)
{
    return QP(n, (p - 1) >> 1, p);
}
ll solve(ll n, ll p)
{
    ll tp = p - 1;
    ll s = 0;
    while (!(tp & 1))
    {
        s
 
++;
        tp >>= 1;//tp 也就是 Q
    }
    if (s == 1)
    {
        return QP(n, (p + 1) / 4, p);
    }
    ll z;
    while (true)
    {
        z = rand() % p;
        if (le(z, p) + 1 == p)
            break;
    }
    ll c = QP(z, tp, p);
    ll R = QP(n, (tp + 1) >> 1, p), t = QP(n, tp, p), M = s;
    while (true)
    {
        if (t == 1)return (R%p + p) % p;
        ll i = 1, j = 2;
        while (QP(t, j, p) != 1)
        {
            i++;
            j = 1 << j;
        }
        ll po = 1;
        for (int i = 1; i <= M - i - 1; i++)po <<= 1;
        ll b = QP(c, po, p);
        //      ll b=QP(c,1<<(M-i-1),p);
        R = R*b%p;
        t = t*b*b%p;
        c = b*b%p;
        M = i;
    }
    return 0;
}
int main()
{
    int num;
    scanf("%d", &num);
    for (int i = 0; i < num; i++)
    {
        ll n, p;
        scanf("%lld%lld", &n, &p);
        if (n == 2)
        {
            if (n%p == 1)printf("1\n");
            else printf("No root\n");
            continue;
        }
        if (le(n, p) + 1 == p)
        {
            printf("No root\n");
            continue;
        }
        ll a = solve(n, p);
        ll b = p - a;
        if (a > b)std::swap(a, b);
        if (a == b)
        {
            printf("%lld\n", a);
        }
        else printf("%lld %lld\n", a, b);
    }
    return 0;
}