#include <bits/stdc++.h>
using namespace std;
const long long INF = 1000000000000000;
template <typename T>
struct binary_indexed_tree{
  int N;
  vector<T> BIT;
  binary_indexed_tree(){
  }
  binary_indexed_tree(int N): N(N), BIT(N + 1, 0){
  }
  void add(int i, T x){
    i++;
    while (i <= N){
      BIT[i] += x;
      i += i & -i;
    }
  }
  T sum(int i){
    T ans = 0;
    while (i > 0){
      ans += BIT[i];
      i -= i & -i;
    }
    return ans;
  }
  T sum(int L, int R){
    return sum(R) - sum(L);
  }
};
int main(){
  int M, K;
  cin >> M >> K;
  vector<int> A(M * K);
  for (int i = 0; i < M * K; i++){
    cin >> A[i];
  }
  vector<int> cnt(M, 0);
  vector<int> p(M * K);
  vector<vector<int>> B(K);
  for (int i = 0; i < M * K; i++){
    p[i] = cnt[A[i]];
    B[cnt[A[i]]].push_back(A[i]);
    cnt[A[i]]++;
  }
  long long ans1 = 0;
  binary_indexed_tree<int> BIT1(K);
  for (int i = 0; i < M * K; i++){
    ans1 += BIT1.sum(p[i] + 1, K);
    BIT1.add(p[i], 1);
  }
  vector<long long> op(M, 0);
  for (int i = 0; i < K; i++){
    vector<int> pos(M);
    for (int j = 0; j < M; j++){
      pos[B[i][j]] = j;
    }
    long long C = 0;
    binary_indexed_tree<int> BIT2(M); 
    for (int j = 0; j < M; j++){
      C += BIT2.sum(B[i][j] + 1, M);
      BIT2.add(B[i][j], 1);
    }
    op[0] += C;
    for (int j = 0; j < M - 1; j++){
      C -= pos[j];
      C += M - 1 - pos[j];
      op[j + 1] += C;
    }
  }
  long long ans2 = INF;
  for (int i = 0; i < M; i++){
    ans2 = min(ans2, op[i]);
  }
  cout << ans1 + ans2 << endl;
}