#include <bits/stdc++.h>

using namespace std;

typedef long long I64;

const I64 NMAX = 50000;
const I64 KMAX = 25;

I64 v[NMAX+2];
I64 loc[NMAX+2][2], fx[KMAX+2];
I64 hs[NMAX+2][KMAX+1];
bool good[NMAX+2];
I64 N, K;

int main() {
    cin >> N >> K;
    for( I64 i = 1;  i <= N;  ++i ) {
        for( I64 j = 1;  j <= K;  ++j ) {
            cin >> hs[i][j];
        }
    }
    for( I64 j = 1;  j <= K;  ++j ) {
        for( I64 i = 1;  i <= N;  ++i ) v[i] = hs[i][j];
        sort(  v+1, v+N+1 );
        if( N % 2 == 1 ) {
            fx[j] = 1;
            loc[j][0] = v[ N/2 ];
        }
        else {
            I64 x = v[ N/2 ];
            I64 y = v[ N/2 + 1 ];
            if( x == y ) {
                fx[j] = 1;
                loc[j][0] = x;
            }
            else {
                loc[j][0] = x;
                loc[j][1] = y;
            }
        }
    }
    I64 Cost = 0;
    for( I64 i = 1;  i <= N;  ++i ) {
        for( I64 j = 1;  j <= K;  ++j ) {
           /// if( fx[j] ) {
                Cost += max( hs[i][j] - loc[j][0], loc[j][0]-hs[i][j] );
          ///  }
          ///  else {
           ///     I64 c1 = max( hs[i][j] - loc[j][0], loc[j][0]-hs[i][j] );
            ///    I64 c2 = max( hs[i][j] - loc[j][1], loc[j][1]-hs[i][j] );
            ///    Cost += min( c1, c2 );
           /// }
        }
    }
    cout << Cost << '\n';

    vector < vector<I64> > sol;
    for( I64 i = 1;  i <= N;  ++i ) {
        good[i] = 1;
        vector<I64> line;
        for( I64 j = 1;  j <= K;  ++j ) {
            if( fx[j] ) {
                if( hs[i][j] != loc[j][0] ) good[i] = 0;
            }
            else {
                if( hs[i][j] != loc[j][0] && hs[i][j] != loc[j][1] ) good[i] = 0;
            }
            line.push_back( hs[i][j] );
        }

        if( good[i] ) sol.push_back( line );
    }
    cout << sol.size() << '\n';
    for( auto line : sol ) {
        for( auto x : line ) {
            cout << x << ' ';
        }
        cout << '\n';
    }
    return 0;
}