結果

問題 No.215 素数サイコロと合成数サイコロ (3-Hard)
ユーザー 👑 hitonanodehitonanode
提出日時 2020-06-06 01:23:34
言語 C++17
(gcc 12.3.0 + boost 1.83.0)
結果
AC  
実行時間 2,371 ms / 4,000 ms
コード長 11,037 bytes
コンパイル時間 2,111 ms
コンパイル使用メモリ 119,172 KB
実行使用メモリ 7,492 KB
最終ジャッジ日時 2023-08-22 20:46:15
合計ジャッジ時間 9,868 ms
ジャッジサーバーID
(参考情報)
judge15 / judge12
このコードへのチャレンジ
(要ログイン)

テストケース

テストケース表示
入力 結果 実行時間
実行使用メモリ
testcase_00 AC 2,365 ms
7,492 KB
testcase_01 AC 2,371 ms
7,452 KB
権限があれば一括ダウンロードができます

ソースコード

diff #

#define PROBLEM "https://yukicoder.me/problems/no/215"
#include <iostream>
#include <vector>
#include <set>

// CUT begin
template <int mod>
struct ModInt
{
    using lint = long long;
    static int get_mod() { return mod; }
    static int get_primitive_root() {
        static int primitive_root = 0;
        if (!primitive_root) {
            primitive_root = [&](){
                std::set<int> fac;
                int v = mod - 1;
                for (lint i = 2; i * i <= v; i++) while (v % i == 0) fac.insert(i), v /= i;
                if (v > 1) fac.insert(v);
                for (int g = 1; g < mod; g++) {
                    bool ok = true;
                    for (auto i : fac) if (ModInt(g).power((mod - 1) / i) == 1) { ok = false; break; }
                    if (ok) return g;
                }
                return -1;
            }();
        }
        return primitive_root;
    }
    int val;
    constexpr ModInt() : val(0) {}
    constexpr ModInt &_setval(lint v) { val = (v >= mod ? v - mod : v); return *this; }
    constexpr ModInt(lint v) { _setval(v % mod + mod); }
    explicit operator bool() const { return val != 0; }
    constexpr ModInt operator+(const ModInt &x) const { return ModInt()._setval((lint)val + x.val); }
    constexpr ModInt operator-(const ModInt &x) const { return ModInt()._setval((lint)val - x.val + mod); }
    constexpr ModInt operator*(const ModInt &x) const { return ModInt()._setval((lint)val * x.val % mod); }
    constexpr ModInt operator/(const ModInt &x) const { return ModInt()._setval((lint)val * x.inv() % mod); }
    constexpr ModInt operator-() const { return ModInt()._setval(mod - val); }
    constexpr ModInt &operator+=(const ModInt &x) { return *this = *this + x; }
    constexpr ModInt &operator-=(const ModInt &x) { return *this = *this - x; }
    constexpr ModInt &operator*=(const ModInt &x) { return *this = *this * x; }
    constexpr ModInt &operator/=(const ModInt &x) { return *this = *this / x; }
    friend constexpr ModInt operator+(lint a, const ModInt &x) { return ModInt()._setval(a % mod + x.val); }
    friend constexpr ModInt operator-(lint a, const ModInt &x) { return ModInt()._setval(a % mod - x.val + mod); }
    friend constexpr ModInt operator*(lint a, const ModInt &x) { return ModInt()._setval(a % mod * x.val % mod); }
    friend constexpr ModInt operator/(lint a, const ModInt &x) { return ModInt()._setval(a % mod * x.inv() % mod); }
    constexpr bool operator==(const ModInt &x) const { return val == x.val; }
    constexpr bool operator!=(const ModInt &x) const { return val != x.val; }
    bool operator<(const ModInt &x) const { return val < x.val; }  // To use std::map<ModInt, T>
    friend std::istream &operator>>(std::istream &is, ModInt &x) { lint t; is >> t; x = ModInt(t); return is; }
    friend std::ostream &operator<<(std::ostream &os, const ModInt &x) { os << x.val;  return os; }
    constexpr lint power(lint n) const {
        lint ans = 1, tmp = this->val;
        while (n) {
            if (n & 1) ans = ans * tmp % mod;
            tmp = tmp * tmp % mod;
            n /= 2;
        }
        return ans;
    }
    constexpr lint inv() const { return this->power(mod - 2); }
    constexpr ModInt operator^(lint n) const { return ModInt(this->power(n)); }
    constexpr ModInt &operator^=(lint n) { return *this = *this ^ n; }

    inline ModInt fac() const {
        static std::vector<ModInt> facs;
        int l0 = facs.size();
        if (l0 > this->val) return facs[this->val];

        facs.resize(this->val + 1);
        for (int i = l0; i <= this->val; i++) facs[i] = (i == 0 ? ModInt(1) : facs[i - 1] * ModInt(i));
        return facs[this->val];
    }

    ModInt doublefac() const {
        lint k = (this->val + 1) / 2;
        if (this->val & 1) return ModInt(k * 2).fac() / ModInt(2).power(k) / ModInt(k).fac();
        else return ModInt(k).fac() * ModInt(2).power(k);
    }

    ModInt nCr(const ModInt &r) const {
        if (this->val < r.val) return ModInt(0);
        return this->fac() / ((*this - r).fac() * r.fac());
    }

    ModInt sqrt() const {
        if (val == 0) return 0;
        if (mod == 2) return val;
        if (power((mod - 1) / 2) != 1) return 0;
        ModInt b = 1;
        while (b.power((mod - 1) / 2) == 1) b += 1;
        int e = 0, m = mod - 1;
        while (m % 2 == 0) m >>= 1, e++;
        ModInt x = power((m - 1) / 2), y = (*this) * x * x;
        x *= (*this);
        ModInt z = b.power(m);
        while (y != 1) {
            int j = 0;
            ModInt t = y;
            while (t != 1) j++, t *= t;
            z = z.power(1LL << (e - j - 1));
            x *= z, z *= z, y *= z;
            e = j;
        }
        return ModInt(std::min(x.val, mod - x.val));
    }
};
#include <algorithm>
#include <array>
#include <cassert>
#include <vector>
using namespace std;

// CUT begin
// Integer convolution for arbitrary mod
// with NTT (and Garner's algorithm) for ModInt / ModIntRuntime class.
// We skip Garner's algorithm if `skip_garner` is true or mod is in `nttprimes`.
// input: a (size: n), b (size: m)
// return: vector (size: n + m - 1)
template <typename MODINT>
vector<MODINT> nttconv(vector<MODINT> a, vector<MODINT> b, bool skip_garner = false);

constexpr int nttprimes[3] = {998244353, 167772161, 469762049};

// Integer FFT (Fast Fourier Transform) for ModInt class
// (Also known as Number Theoretic Transform, NTT)
// is_inverse: inverse transform
// ** Input size must be 2^n **
template <typename MODINT>
void ntt(vector<MODINT> &a, bool is_inverse = false)
{
    int n = a.size();
    assert(__builtin_popcount(n) == 1);
    MODINT h = MODINT(MODINT::get_primitive_root()).power((MODINT::get_mod() - 1) / n);
    if (is_inverse) h = 1 / h;

    int i = 0;
    for (int j = 1; j < n - 1; j++) {
        for (int k = n >> 1; k > (i ^= k); k >>= 1);
        if (j < i) swap(a[i], a[j]);
    }

    for (int m = 1; m < n; m *= 2) {
        int m2 = 2 * m;
        MODINT base = h.power(n / m2), w = 1;
        for (int x = 0; x < m; x++) {
            for (int s = x; s < n; s += m2) {
                MODINT u = a[s], d = a[s + m] * w;
                a[s] = u + d, a[s + m] = u - d;
            }
            w *= base;
        }
    }
    if (is_inverse) {
        long long int n_inv = MODINT(n).inv();
        for (auto &v : a) v *= n_inv;
    }
}
template<int MOD>
vector<ModInt<MOD>> nttconv_(const vector<int> &a, const vector<int> &b) {
    int sz = a.size();
    assert(a.size() == b.size() and __builtin_popcount(sz) == 1);
    vector<ModInt<MOD>> ap(sz), bp(sz);
    for (int i = 0; i < sz; i++) ap[i] = a[i], bp[i] = b[i];
    if (a == b) {
        ntt(ap, false);
        bp = ap;
    }
    else {
        ntt(ap, false);
        ntt(bp, false);
    }
    for (int i = 0; i < sz; i++) ap[i] *= bp[i];
    ntt(ap, true);
    return ap;
}
long long int extgcd_ntt_(long long int a, long long int b, long long int &x, long long int &y)
{
    long long int d = a;
    if (b != 0) d = extgcd_ntt_(b, a % b, y, x), y -= (a / b) * x;
    else x = 1, y = 0;
    return d;
}
long long int modinv_ntt_(long long int a, long long int m)
{
    long long int x, y;
    extgcd_ntt_(a, m, x, y);
    return (m + x % m) % m;
}
long long int garner_ntt_(int r0, int r1, int r2, int mod)
{
    array<long long int, 4> rs = {r0, r1, r2, 0};
    vector<long long int> coffs(4, 1), constants(4, 0);
    for (int i = 0; i < 3; i++) {
        long long int v = (rs[i] - constants[i]) * modinv_ntt_(coffs[i], nttprimes[i]) % nttprimes[i];
        if (v < 0) v += nttprimes[i];
        for (int j = i + 1; j < 4; j++) {
            (constants[j] += coffs[j] * v) %= (j < 3 ? nttprimes[j] : mod);
            (coffs[j] *= nttprimes[i]) %= (j < 3 ? nttprimes[j] : mod);
        }
    }
    return constants.back();
}
template <typename MODINT>
vector<MODINT> nttconv(vector<MODINT> a, vector<MODINT> b, bool skip_garner)
{
    int sz = 1, n = a.size(), m = b.size();
    while (sz < n + m) sz <<= 1;
    if (sz <= 16) {
        vector<MODINT> ret(n + m - 1);
        for (int i = 0; i < n; i++) {
            for (int j = 0; j < m; j++) ret[i + j] += a[i] * b[j];
        }
        return ret;
    }
    int mod = MODINT::get_mod();
    if (skip_garner or find(begin(nttprimes), end(nttprimes), mod) != end(nttprimes)) {
        a.resize(sz), b.resize(sz);
        if (a == b) { ntt(a, false); b = a; }
        else ntt(a, false), ntt(b, false);
        for (int i = 0; i < sz; i++) a[i] *= b[i];
        ntt(a, true);
        a.resize(n + m - 1);
    }
    else {
        vector<int> ai(sz), bi(sz);
        for (int i = 0; i < n; i++) ai[i] = a[i].val;
        for (int i = 0; i < m; i++) bi[i] = b[i].val;
        auto ntt0 = nttconv_<nttprimes[0]>(ai, bi);
        auto ntt1 = nttconv_<nttprimes[1]>(ai, bi);
        auto ntt2 = nttconv_<nttprimes[2]>(ai, bi);
        a.resize(n + m - 1);
        for (int i = 0; i < n + m - 1; i++) {
            a[i] = garner_ntt_(ntt0[i].val, ntt1[i].val, ntt2[i].val, mod);
        }
    }
    return a;
}

#include <vector>

// Calculate [x^N](num(x) / den(x))
template <typename Tp>
Tp coefficient_of_rational_function(long long N, std::vector<Tp> num, std::vector<Tp> den)
{
    assert(N >= 0);
    while (den.size() and den.back() == 0) den.pop_back();
    assert(den.size());
    int h = 0;
    while (den[h] == 0) h++;
    N += h;
    den.erase(den.begin(), den.begin() + h);
    
    if (den.size() == 1)
    {
        assert(N < int(num.size()));
        return num[N] / den[0];
    }

    while (N)
    {
        std::vector<Tp> g = den;
        for (size_t i = 1; i < g.size(); i += 2)
        {
            g[i] = -g[i];
        }
        auto conv_num_g = nttconv(num, g);
        num.resize((conv_num_g.size() + 1 - (N & 1)) / 2);
        for (size_t i = 0; i < num.size(); i++)
        {
            num[i] = conv_num_g[i * 2 + (N & 1)];
        }
        auto conv_den_g = nttconv(den, g);
        for (size_t i = 0; i < den.size(); i++)
        {
            den[i] = conv_den_g[i * 2];
        }
        N >>= 1;
    }
    return num[0] / den[0];
}

using mint = ModInt<1000000007>;

#include <iostream>

using namespace std;
vector<mint> gen_dp(vector<int> v, int n)
{
    vector<vector<mint>> dp(n + 1, vector<mint>(v.back() * n + 1));
    dp[0][0] = 1;
    for (auto x : v)
    {
        for (int i = n - 1; i >= 0; i--)
        {
            for (int j = 0; j < dp[i].size(); j++) if (dp[i][j])
            {
                for (int k = 1; i + k <= n; k++) dp[i + k][j + x * k] += dp[i][j];
            }
        }
    }
    return dp.back();
}

int main()
{
    long long N;
    int P, C;
    cin >> N >> P >> C;
    vector<mint> primes = gen_dp({2, 3, 5, 7, 11, 13}, P), composites = gen_dp({4, 6, 8, 9, 10, 12}, C);
    vector<mint> f = nttconv(primes, composites);
    vector<mint> denom = f;
    for (auto &x : denom) x = -x;
    denom[0] = 1;
    for (int i = f.size() - 1; i > 1; i--) f[i - 1] += f[i];
    cout << coefficient_of_rational_function(N, f, denom) << '\n';
}
0