#pragma GCC optimize("O3,unroll-loops")
#pragma GCC target("avx2,bmi,bmi2,lzcnt,popcnt")
#include <bits/stdc++.h>
using namespace std;
using ll = long long;
template<class T> bool cmin(T &i, T j) { return i > j ? i=j,true:false; }
template<class T> bool cmax(T &i, T j) { return i < j ? i=j,true:false; }
mt19937 mrand(chrono::steady_clock::now().time_since_epoch().count());
uniform_int_distribution<int> ui(0, 1 << 30);
int rng() {
return ui(mrand);
}
const int N=51;
int dp[N][N][N][N];
int n,m;
int grid[N][N];
const int inf=1e9;
int get_sum(int x1,int y1,int x2,int y2) {
return grid[x2][y2]-(x1>0?grid[x1-1][y2]:0)-(y1>0?grid[x2][y1-1]:0)+(x1>0&&y1>0?grid[x1-1][y1-1]:0);
}
int solve(int x1,int y1,int x2,int y2) {
int &now = dp[x1][y1][x2][y2];
if (now < inf)
return now;
for (int x=x1;x<x2;x++) {
cmin(now,solve(x1,y1,x,y2)+solve(x+1,y1,x2,y2));
}
for (int y=y1;y<y2;y++) {
cmin(now,solve(x1,y1,x2,y)+solve(x1,y+1,x2,y2));
}
now+=get_sum(x1,y1,x2,y2);
return now;
}
int main() {
ios_base::sync_with_stdio(false);
cin.tie(nullptr);
cin >> n >> m;
memset(dp,0x3f,sizeof(dp));
for (int i=0;i<n;i++)
for (int j=0;j<m;j++) {
cin >> grid[i][j];
dp[i][j][i][j]=0;
}
for (int i=0;i<n;i++)
for (int j=0;j<m;j++)
grid[i][j]+=(i > 0?grid[i-1][j]:0)+(j>0?grid[i][j-1]:0)-(i>0&&j>0?grid[i-1][j-1]:0);
cout << solve(0,0,n-1,m-1) << "\n";
}
