#include <cassert>
#include <cmath>
#include <cstdint>
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <algorithm>
#include <bitset>
#include <complex>
#include <deque>
#include <functional>
#include <iostream>
#include <map>
#include <numeric>
#include <queue>
#include <set>
#include <sstream>
#include <string>
#include <unordered_map>
#include <unordered_set>
#include <utility>
#include <vector>

using namespace std;

using Int = long long;

template <class T1, class T2> ostream &operator<<(ostream &os, const pair<T1, T2> &a) { return os << "(" << a.first << ", " << a.second << ")"; };
template <class T> ostream &operator<<(ostream &os, const vector<T> &as) { const int sz = as.size(); os << "["; for (int i = 0; i < sz; ++i) { if (i >= 256) { os << ", ..."; break; } if (i > 0) { os << ", "; } os << as[i]; } return os << "]"; }
template <class T> void pv(T a, T b) { for (T i = a; i != b; ++i) cerr << *i << " "; cerr << endl; }
template <class T> bool chmin(T &t, const T &f) { if (t > f) { t = f; return true; } return false; }
template <class T> bool chmax(T &t, const T &f) { if (t < f) { t = f; return true; } return false; }

////////////////////////////////////////////////////////////////////////////////
struct ModLong {
  static unsigned long long M;
  static long double INV_M;
  static void setM(unsigned long long m) { M = m; INV_M = 1.0L / M; }
  unsigned long long x;
  ModLong() : x(0ULL) {}
  ModLong(unsigned x_) : x(x_ % M) {}
  ModLong(unsigned long long x_) : x(x_ % M) {}
  ModLong(int x_) : x(((x_ %= static_cast<long long>(M)) < 0) ? (x_ + static_cast<long long>(M)) : x_) {}
  ModLong(long long x_) : x(((x_ %= static_cast<long long>(M)) < 0) ? (x_ + static_cast<long long>(M)) : x_) {}
  ModLong &operator+=(const ModLong &a) { x = ((x += a.x) >= M) ? (x - M) : x; return *this; }
  ModLong &operator-=(const ModLong &a) { x = ((x -= a.x) >= M) ? (x + M) : x; return *this; }
  ModLong &operator*=(const ModLong &a) {
    // https://github.com/kth-competitive-programming/kactl/blob/main/doc/modmul-proof.pdf
    // This works at least for M < 2^62, assuming the usual 80-bit long double.
    const long long y = x * a.x - M * static_cast<unsigned long long>(INV_M * x * a.x);
    x = (y < 0LL) ? (y + M) : (y >= static_cast<long long>(M)) ? (y - M) : y;
    return *this;
  }
  ModLong &operator/=(const ModLong &a) { return (*this *= a.inv()); }
  ModLong pow(long long e) const {
    if (e < 0) return inv().pow(-e);
    ModLong a = *this, b = 1ULL; for (; e; e >>= 1) { if (e & 1) b *= a; a *= a; } return b;
  }
  ModLong inv() const {
    unsigned long long a = M, b = x; long long y = 0, z = 1;
    for (; b; ) { const unsigned long long q = a / b; const unsigned long long c = a - q * b; a = b; b = c; const long long w = y - static_cast<long long>(q) * z; y = z; z = w; }
    assert(a == 1ULL); return ModLong(y);
  }
  ModLong operator+() const { return *this; }
  ModLong operator-() const { ModLong a; a.x = x ? (M - x) : 0ULL; return a; }
  ModLong operator+(const ModLong &a) const { return (ModLong(*this) += a); }
  ModLong operator-(const ModLong &a) const { return (ModLong(*this) -= a); }
  ModLong operator*(const ModLong &a) const { return (ModLong(*this) *= a); }
  ModLong operator/(const ModLong &a) const { return (ModLong(*this) /= a); }
  template <class T> friend ModLong operator+(T a, const ModLong &b) { return (ModLong(a) += b); }
  template <class T> friend ModLong operator-(T a, const ModLong &b) { return (ModLong(a) -= b); }
  template <class T> friend ModLong operator*(T a, const ModLong &b) { return (ModLong(a) *= b); }
  template <class T> friend ModLong operator/(T a, const ModLong &b) { return (ModLong(a) /= b); }
  explicit operator bool() const { return x; }
  bool operator==(const ModLong &a) const { return (x == a.x); }
  bool operator!=(const ModLong &a) const { return (x != a.x); }
  friend std::ostream &operator<<(std::ostream &os, const ModLong &a) { return os << a.x; }
};
unsigned long long ModLong::M;
long double ModLong::INV_M;
////////////////////////////////////////////////////////////////////////////////

using Mint = ModLong;


using INT = __int128;

INT divFloor(INT a, INT b) {
  return a / b - (((a ^ b) < 0 && a % b != 0) ? 1 : 0);
}
INT divCeil(INT a, INT b) {
  return a / b + (((a ^ b) > 0 && a % b != 0) ? 1 : 0);
}


constexpr int H = 6;
constexpr int NUM_CANDS = 100;
constexpr int WIDTH = 10;

int M;
Int P;
vector<Mint> A, C;
Int maxErr;

bool check(Mint x) {
  for (int i = 0; i < M; ++i) {
    const Mint e = C[i] - A[i] * x;
    if (!((Int)e.x <= maxErr || P - maxErr <= (Int)e.x)) {
      return false;
    }
  }
  return true;
}

int main() {
  for (int numCases; ~scanf("%d", &numCases); ) { for (int caseId = 1; caseId <= numCases; ++caseId) {
    scanf("%d%lld", &M, &P);
    Mint::setM(P);
    A.resize(M);
    C.resize(M);
    for (int i = 0; i < M; ++i) {
      scanf("%llu%llu", &A[i].x, &C[i].x);
    }
    
    maxErr = P / 200;
    
    // A[0] x + y = C[0]  (mod P)
    const Mint invA0 = A[0].inv();
    
    vector<pair<Mint, Mint>> acss[H + 1];
    for (int i = 1; i < M; ++i) {
      acss[0].emplace_back(-invA0 * A[i], C[i] - invA0 * A[i] * C[0]);
    }
    for (int h = 0; ; ++h) {
      sort(acss[h].begin(), acss[h].end(), [&](const pair<Mint, Mint> &ac0, const pair<Mint, Mint> &ac1) -> bool {
        return (ac0.first.x < ac1.first.x);
      });
      acss[h].erase(unique(acss[h].begin(), acss[h].end()), acss[h].end());
      if ((int)acss[h].size() > NUM_CANDS) {
        acss[h].resize(NUM_CANDS);
      }
// cerr<<"acss["<<h<<"] = ";for(int i=0;i<10;++i)cerr<<acss[h][i]<<" ";cerr<<"..."<<endl;
      if (h == H) {
        break;
      }
      for (int i = 0; i < (int)acss[h].size(); ++i) {
        for (int j = i + 1; j < (int)acss[h].size() && j < i + WIDTH; ++j) {
          const Mint ai = acss[h][i].first;
          const Mint ci = acss[h][i].second;
          const Mint aj = acss[h][j].first;
          const Mint cj = acss[h][j].second;
          if (ai != aj) {
            acss[h + 1].emplace_back(aj - ai, cj - ci);
          }
        }
      }
    }
    
    // possible intervals for y
    vector<pair<Int, Int>> fs;
    fs.emplace_back(-maxErr, +maxErr);
    
    for (int h = H; h >= 0; --h) {
      const INT err = maxErr << h;
      for (const auto &ac : acss[h]) {
        const INT a = ac.first.x;
        const INT c = ac.second.x;
        vector<pair<Int, Int>> gs;
        for (const auto &f : fs) {
          const INT lb = a * f.first - err;
          const INT ub = a * f.second + err;
          for (INT cc = c + P * divCeil(lb - c, P); cc <= ub; cc += P) {
            Int mn = divCeil(cc - err, a);
            Int mx = divFloor(cc + err, a);
            chmax(mn, f.first);
            chmin(mx, f.second);
            if (mn <= mx) {
              gs.emplace_back(mn, mx);
            }
          }
        }
        fs.swap(gs);
      }
// cerr<<"fs = "<<fs<<endl;
    }
    
    assert((int)fs.size() == 1);
    assert(fs[0].first == fs[0].second);
    const Mint y = fs[0].first;
    const Mint x = invA0 * (C[0] - y);
    assert(check(x));
    printf("%llu\n", x.x);
  }
#ifndef LOCAL
  break;
#endif
  }
  return 0;
}