D. Power Tower

time limit per test

4.5 seconds

memory limit per test

256 megabytes

input

standard input

output

standard output

Priests of the Quetzalcoatl cult want to build a tower to represent a power of their god. Tower is usually made of power-charged rocks. It is built with the help of rare magic by levitating the current top of tower and adding rocks at its bottom. If top, which is built from k - 1 rocks, possesses power p and we want to add the rock charged with power w**k then value of power of a new tower will be {w**k}p.

Rocks are added from the last to the first. That is for sequence w1, ..., w**m value of power will be

After tower is built, its power may be extremely large. But still priests want to get some information about it, namely they want to know a number called cumulative power which is the true value of power taken modulo m. Priests have n rocks numbered from 1 to n. They ask you to calculate which value of cumulative power will the tower possess if they will build it from rocks numbered l, l + 1, ..., r.

Input

First line of input contains two integers n (1 ≤ n ≤ 105) and m (1 ≤ m ≤ 109).

Second line of input contains n integers w**k (1 ≤ w**k ≤ 109) which is the power of rocks that priests have.

Third line of input contains single integer q (1 ≤ q ≤ 105) which is amount of queries from priests to you.

k**th of next q lines contains two integers l**k and r**k (1 ≤ l**k ≤ r**k ≤ n).

Output

Output q integers. k-th of them must be the amount of cumulative power the tower will have if is built from rocks l**k, l**k + 1, ..., r**k.

Example

input

Copy

6 10000000001 2 2 3 3 381 11 62 22 32 44 44 54 6

output

Copy

1124256327597484987

Note

327 = 7625597484987

思路:

因为euler( euler(x) ) <= x/2 所以在log(x)次内欧拉函数值就会降为1,并且一直为1.而任何数对1取模的答案都是0,所以我们可以遇见模数为1时就可以结束迭代,

  • 因此每次询问最多迭代log(m)次,每一次迭代只需要一个快速幂的时间复杂度,也是log(m)

  • 因此对于每一个询问综合的时间复杂度是O(log(m)^2)

    注意,在指数循环节中快速幂时,需要在ans>=mod时,取模后再加上mod,以此才满足欧拉降幂定理。

细节见代码:

#include <iostream>
#include <cstdio>
#include <cstring>
#include <algorithm>
#include <cmath>
#include <queue>
#include <stack>
#include <map>
#include <set>
#include <vector>
#include <iomanip>
#define ALL(x) (x).begin(), (x).end()
#define sz(a) int(a.size())
#define all(a) a.begin(), a.end()
#define rep(i,x,n) for(int i=x;i<n;i++)
#define repd(i,x,n) for(int i=x;i<=n;i++)
#define pii pair<int,int>
#define pll pair<long long ,long long>
#define gbtb ios::sync_with_stdio(false),cin.tie(0),cout.tie(0)
#define MS0(X) memset((X), 0, sizeof((X)))
#define MSC0(X) memset((X), '\0', sizeof((X)))
#define pb push_back
#define mp make_pair
#define fi first
#define se second
#define eps 1e-6
#define gg(x) getInt(&x)
#define chu(x) cout<<"["<<#x<<" "<<(x)<<"]"<<endl
using namespace std;
typedef long long ll;
ll gcd(ll a, ll b) {return b ? gcd(b, a % b) : a;}
ll lcm(ll a, ll b) {return a / gcd(a, b) * b;}
inline void getInt(int* p);
const int maxn = 1000010;
const int inf = 0x3f3f3f3f;
/*** TEMPLATE CODE * * STARTS HERE ***/
ll mod(ll x, ll m)
{
return x >= m ? x % m + m : x;
}
ll powmod(ll a, ll b, ll MOD)
{
ll ans = 1;
while (b)
{
if (b % 2)
ans = mod(ans * a, MOD);
// ans = ans * a % MOD;
// a = a * a % MOD;
a = mod(a * a, MOD);
b /= 2;
}
return ans;
} ll m;
int n;
int q;
ll a[maxn];
map<ll, ll> vis;
ll euler(ll n) { //log(n)时间内求一个数的欧拉值
if (vis.count(n))
{
return vis[n];
}
ll ans = n;
for (ll i = 2; i * i <= n; i++) {
if (n % i == 0)
{
ans -= ans / i;
while (n % i == 0) n /= i;
}
}
if (n > 1) ans -= ans / n;
vis[n] = ans;
return ans;
} ll solve(int l, int r, ll m)
{
if (l == r || m == 1)
return mod(a[r], m);
return powmod(a[l], solve(l + 1, r, euler(m)), m);
}
int main()
{
//freopen("D:\\common_text\\code_stream\\in.txt","r",stdin);
//freopen("D:\\common_text\\code_stream\\out.txt","w",stdout);
// gbtb;
// cin >> n >> m;
scanf("%d%lld", &n, &m);
repd(i, 1, n)
{
scanf("%lld", &a[i]);
// cin >> a[i];
}
// cin >> q;
scanf("%d", &q);
int l, r;
while (q--)
{
scanf("%d %d", &l, &r);
printf("%lld\n", solve(l, r, m) % m);
// cin >> l >> r;
// cout << solve(l, r, m) % m << endl;
} return 0;
} inline void getInt(int* p) {
char ch;
do {
ch = getchar();
} while (ch == ' ' || ch == '\n');
if (ch == '-') {
*p = -(getchar() - '0');
while ((ch = getchar()) >= '0' && ch <= '9') {
*p = *p * 10 - ch + '0';
}
}
else {
*p = ch - '0';
while ((ch = getchar()) >= '0' && ch <= '9') {
*p = *p * 10 + ch - '0';
}
}
}

Codeforces Round #454 D. Power Tower (广义欧拉降幂)的更多相关文章

  1. CodeForces - 906D Power Tower(欧拉降幂定理)

    Power Tower CodeForces - 906D 题目大意:有N个数字,然后给你q个区间,要你求每一个区间中所有的数字从左到右依次垒起来的次方的幂对m取模之后的数字是多少. 用到一个新知识, ...

  2. Power Tower(广义欧拉降幂)

    题意:https://codeforc.es/contest/906/problem/D 计算区间的: ai ^ ai+1 ^ ai+2.......ar . 思路: 广义欧拉降幂: 注意是自下而上递 ...

  3. ACM-数论-广义欧拉降幂

    https://www.cnblogs.com/31415926535x/p/11447033.html 曾今一时的懒,造就今日的泪 记得半年前去武大参加的省赛,当时的A题就是一个广义欧拉降幂的板子题 ...

  4. 广义欧拉降幂(欧拉定理)——bzoj3884,fzu1759

    广义欧拉降幂对于狭义欧拉降幂任然适用 https://blog.csdn.net/qq_37632935/article/details/81264965?tdsourcetag=s_pctim_ai ...

  5. CF思维联系– CodeForces -CodeForces - 992C Nastya and a Wardrobe(欧拉降幂+快速幂)

    Nastya received a gift on New Year - a magic wardrobe. It is magic because in the end of each month ...

  6. Codeforces 906D Power Tower(欧拉函数 + 欧拉公式)

    题目链接  Power Tower 题意  给定一个序列,每次给定$l, r$ 求$w_{l}^{w_{l+1}^{w_{l+2}^{...^{w_{r}}}}}$  对m取模的值 根据这个公式 每次 ...

  7. The Preliminary Contest for ICPC Asia Nanjing 2019 B. super_log (广义欧拉降幂)

    In Complexity theory, some functions are nearly O(1)O(1), but it is greater then O(1)O(1). For examp ...

  8. BZOJ 3884——欧拉降幂和广义欧拉降幂

    理论部分 欧拉定理:若 $a,n$ 为正整数,且 $a,n$ 互质,则 $a^{\varphi (n)} \equiv 1(mod \ n)$. 降幂公式: $$a^b=\begin{cases}a^ ...

  9. Codeforces Round #454 (Div. 1) CodeForces 906D Power Tower (欧拉降幂)

    题目链接:http://codeforces.com/contest/906/problem/D 题目大意:给定n个整数w[1],w[2],……,w[n],和一个数m,然后有q个询问,每个询问给出一个 ...

随机推荐

  1. 【HANA系列】SAP HANA SQL IFNULL和NULLIF用法与区别

    公众号:SAP Technical 本文作者:matinal 原文出处:http://www.cnblogs.com/SAPmatinal/ 原文链接:[HANA系列]SAP HANA SQL IFN ...

  2. 纹理特征描述之灰度差分统计特征(平均值 对比度 熵) 计算和比较两幅纹理图像的灰度差分统计特征 matlab代码实现

    灰度差分统计特征有: 平均值:​ 对比度:​ 熵:​ i表示某一灰度值,p(i)表示图像取这一灰度值的概率 close all;clear all;clc; % 纹理图像的灰度差分统计特征 J = i ...

  3. docker清理

    # 删除退出的容器docker rm $(docker ps -qa --no-trunc --filter "status=exited") # 删除悬挂镜像docker rmi ...

  4. 论文阅读 | Probing Neural Network Understanding of Natural Language Arguments

    [code&data] [pdf] ARCT 任务是 Habernal 等人在 NACCL 2018 中提出的,即在给定的前提(premise)下,对于某个陈述(claim),相反的两个依据( ...

  5. NLP 对抗方法整理

    NLP中对抗应用 1. 分词 , 可以用GAN来做,消除不同分词器的差异性 2. 风格迁移, 这个在图像中应用较多,在NLP中同样可行 3. 提高问答系统/阅读理解的性能. 4. 机器翻译应该也可以做 ...

  6. day26 封装、多态、内置函数、反射、动态导入

    今日内容 1.封装 什么是封装? 封装从字面意思上看就只将某种东西封起来装好,当我们代码中的某些方法与属性不想让外界进行访问时,就对这些属性进行特殊的处理,使这种属性或者方法不能被外界直接进行访问或者 ...

  7. Linux系列(4):入门之文件权限与目录配置

    众所周知,Linux是多用户多任务的操作系统.那么如何解决自己文件不被其他用户访问呢?这就需要引入权限管理了. Linux根据文件的所属者分为3个类别:owner.group.others,且每个类别 ...

  8. AC自动机练习2:修改串

    这道题的话用到了dp,一个比较简单的dp方程 1466: [AC自动机]修改串 poj3691 时间限制: 1 Sec  内存限制: 128 MB提交: 18  解决: 14[提交] [状态] [讨论 ...

  9. Elasticsearch5.x安装及常见错误的解决方法

    Elasticsearch是基于java开发的,机器上必须要先java环境,elasticsearch5.x建议用jdk8的最新版本.下面介绍elasticsearch5.x的安装步骤: 一.安装El ...

  10. Digester库

    在之前所学习关于启动简单的Tomcat部分实现的代码中,我们使用一个启动类Bootstrap类 来实例化连接器.servlet容器.wrapper实例.和其他组件,然后调用各个对象的set方法将他们关 ...