codeforces 100633J. Ceizenpok’s formula （扩展Lucas定理+中国剩余定理）

J. Ceizenpok’s formula

time limit per test

2 seconds

memory limit per test

256 megabytes

input

standard input

output

standard output

Dr. Ceizenp'ok from planet i1c5l became famous across the whole Universe thanks to his recent discovery — the Ceizenpok’s formula. This formula has only three arguments: n, k and m, and its value is a number of k-combinations of a set of n modulo m.

While the whole Universe is trying to guess what the formula is useful for, we need to automate its calculation.

Input

Single line contains three integers n, k, m, separated with spaces (1 ≤ n ≤ 1018, 0 ≤ k ≤ n, 2 ≤ m ≤ 1 000 000).

Output

Write the formula value for given arguments n, k, m.

Examples

input

2 1 3

output

2

input

4 2 5

output

1

题解：扩展Lucas定理+中国剩余定理

快速乘要少用，死慢啊！！

各种地方炸掉，我也是醉了。。。

#include<iostream>
#include<cstdio>
#include<cstring>
#include<algorithm>
#include<cmath>
#define N 20
#define LL long long
using namespace std;
LL n,m,M,a[N];
LL p[1000003],mod[1000003],cnt;
void get_prime(LL k)
{
	LL x=k;
	for (LL i=2;i*i<=x;i++) 
	 if (x%i==0) {
	    ++cnt; p[cnt]=i; mod[cnt]=1;
	    while (x%i==0) {
	    	x/=i;
	    	mod[cnt]*=i;
		}
	 }
	if (x>1) {
		p[++cnt]=x;
		mod[cnt]=x;
	}
}
LL mul(LL a,LL b,LL mod)
{
	LL res=0;
	while (b) {
		if (b&1) res=(res+a)%mod;
		a=(a+a)%mod;
		b>>=1;
	}
	return res;
}
LL quickpow(LL num,LL x,LL p)
{
	LL base=num%p; LL ans=1;
	while (x) {
		if (x&1) ans=(ans*base)%p;
		x>>=1;
		base=(base*base)%p;
	}
	return ans;
}
void exgcd(LL a,LL b,LL &x,LL &y)
{
	if (b==0) {
		x=1; y=0; return;
	}
	exgcd(b,a%b,x,y);
	LL t=y; 
	y=x-(a/b)*y;
	x=t;
}
LL inverse(LL a,LL b)
{
	LL x,y;
	exgcd(a,b,x,y);
	x=(x%b+b)%b;
	if (!x) x+=b; 
	return x;
}
LL solve(int k1,LL n)
{
	if (n==0) return 1;
	LL ans=1;
	for (LL i=2;i<mod[k1];i++) 
	 if (i%p[k1]) ans=ans*i%mod[k1];
	ans=quickpow(ans,n/mod[k1],mod[k1]);
	for (LL i=2;i<=n%mod[k1];i++)
	 if (i%p[k1]) ans=ans*i%mod[k1];
	return ans*solve(k1,n/p[k1])%mod[k1];
}
LL calc(int k,LL n,LL m)
{
	if (m>n) return 0;
    LL a=solve(k,n),b=solve(k,m),c=solve(k,n-m);
    LL k1=0,ans;
    for (LL i=n;i;i/=p[k]) k1+=i/p[k];
    for (LL i=m;i;i/=p[k]) k1-=i/p[k];
    for (LL i=n-m;i;i/=p[k]) k1-=i/p[k];
    ans=a*inverse(b,mod[k])%mod[k]*inverse(c,mod[k])%mod[k]*quickpow(p[k],k1,mod[k])%mod[k];
    return ans;
}
LL china()
{
	LL x,y,ans=0;
    for (int i=1;i<=cnt;i++){
    	LL r=M/mod[i];
    	ans=(ans+r*inverse(r,mod[i])%M*a[i]%M)%M;
	}
	return ans%M;	
}
int main()
{
//	freopen("a.in","r",stdin);
//	freopen("my.out","w",stdout);
	cin>>n>>m>>M;
	get_prime(M);
	//for (int i=1;i<=cnt;i++) cout<<p[i]<<" "<<mod[i]<<endl;
	for (int i=1;i<=cnt;i++)
	 a[i]=calc(i,n,m);
	printf("%I64d\n",china());
}