#include <iostream>
#include <cstdio>
#include <cstring>
#include <vector>
#include <deque>
#include <queue>
#include <array>
#include <set>
#include <map>
#include <cmath>
#include <algorithm>
#include <numeric>
#include <cassert>
#include <utility>
#include <functional>
#include <bitset>
#include <cstdint>

using namespace std;
using i64 = int64_t;
using i32 = int32_t;
template<class T, class U> void init_n(vector<T>& v, size_t n, U x) 
{ v = vector<T>(n, x); }
template<class T> void init_n(vector<T>& v, size_t n) { init_n(v, n, T()); }
template<class T> void read_n(vector<T>& v, size_t n, size_t o = 0) 
{ v = vector<T>(n+o); for (size_t i=o; i<n+o; ++i) cin >> v[i]; }
template<class T> void read_n(T a[], size_t n, size_t o = 0)
{ for (size_t i=o; i<n+o; ++i) cin >> a[i]; }
template<class T> T gabs(const T& x) { return max(x, -x); }
#define abs gabs

i64 n;
vector<i64> y, dp[3];
int main() {
  cin >> n;
  read_n(y, n);
  sort(begin(y), end(y));
  y.insert(begin(y), 0);

  for (i64 i = 0; i < 3; ++i) init_n(dp[i], n + 1);
  dp[1][1] = dp[2][1] = 1ll << 50;
  for (i64 i = 1; i <= n; ++i) {
    if (i < n) {
      dp[0][i] = min(dp[1][i - 1], dp[2][i - 1]) + y[i + 1] - y[i]; // to succ
    }
    if (i > 1) {
      dp[1][i] = dp[0][i - 1]; // stay
      dp[2][i] = dp[1][i - 1] + y[i] - y[i - 1]; // to prev
    }
  }

  cout << min(dp[1][n], dp[2][n]) << '\n';
  return 0;
}