#include <bits/stdc++.h>
using namespace std;
#define rep(i, n)  for(long long i=0;i<(long long)(n);i++)
#define REP(i,k,n) for(long long i=k;i<(long long)(n);i++)
#define pb emplace_back
#define lb(v,k) (lower_bound(all(v),(k))-v.begin())
#define ub(v,k) (upper_bound(all(v),(k))-v.begin())
#define fi first
#define se second
#define pi M_PI
#define PQ(T) priority_queue<T>
#define SPQ(T) priority_queue<T,vector<T>,greater<T>>
#define dame(a) {out(a);return 0;}
#define decimal cout<<fixed<<setprecision(15);
#define all(a) a.begin(),a.end()
#define rsort(a) {sort(all(a));reverse(all(a));}
#define dupli(a) {sort(all(a));a.erase(unique(all(a)),a.end());}
#define popcnt __builtin_popcountll
typedef long long ll;
typedef pair<ll,ll> P;
typedef tuple<ll,ll,ll> PP;
typedef tuple<ll,ll,ll,ll> PPP;
using vi=vector<ll>;
using vvi=vector<vi>;
using vvvi=vector<vvi>;
using vvvvi=vector<vvvi>;
using vp=vector<P>;
using vvp=vector<vp>;
using vb=vector<bool>;
using vvb=vector<vb>;
const ll inf=1001001001001001001;
const ll INF=1001001001;
const ll mod=1000000007;
const double eps=1e-10;
template<class T> bool chmin(T&a,T b){if(a>b){a=b;return true;}return false;}
template<class T> bool chmax(T&a,T b){if(a<b){a=b;return true;}return false;}
template<class T> void out(T a){cout<<a<<'\n';}
template<class T> void outp(T a){cout<<'('<<a.fi<<','<<a.se<<')'<<'\n';}
template<class T> void outvp(T v){rep(i,v.size())cout<<'('<<v[i].fi<<','<<v[i].se<<')';cout<<'\n';}
template<class T> void outvvp(T v){rep(i,v.size())outvp(v[i]);}
template<class T> void outv(T v){rep(i,v.size()){if(i)cout<<' ';cout<<v[i];}cout<<'\n';}
template<class T> void outvv(T v){rep(i,v.size())outv(v[i]);}
template<class T> bool isin(T x,T l,T r){return (l)<=(x)&&(x)<=(r);}
template<class T> void yesno(T b){if(b)out("yes");else out("no");}
template<class T> void YesNo(T b){if(b)out("Yes");else out("No");}
template<class T> void YESNO(T b){if(b)out("YES");else out("NO");}
template<class T> void outset(T s){auto itr=s.begin();while(itr!=s.end()){if(itr!=s.begin())cout<<' ';cout<<*itr;itr++;}cout<<'\n';}
void outs(ll a,ll b){if(a>=inf-100)out(b);else out(a);}
ll gcd(ll a,ll b){if(b==0)return a;return gcd(b,a%b);}
ll modpow(ll a,ll b){ll res=1;a%=mod;while(b){if(b&1)res=res*a%mod;a=a*a%mod;b>>=1;}return res;}

class BIT{
    vi bit;ll N;
    ll sum(ll i){
        chmin(i,N);
        ll res=0;
        for(ll j=i-1;j>=0;j=(j&(j+1))-1)res+=bit[j];
        return res;
    }
public:
    BIT(ll n):N(n){bit.resize(N);}
    void add(ll i,ll x){
        for(ll j=i;j<N;j|=j+1)bit[j]+=x;
    }
    ll get(ll a,ll b){
        return sum(b)-sum(a);
    }
};

ll count_inversion(vi v){
    ll res=0;
    vp a(v.size());
    rep(i,v.size())a[i]=P(v[i],i);
    sort(all(a));reverse(all(a));
    BIT bit(v.size());
    rep(i,v.size()){
        res+=bit.get(0,a[i].se);
        bit.add(a[i].se,1);
    }
    return res;
}
int main(){
    ll n,k;cin>>n>>k;
    vi v(n*k);rep(i,n*k)cin>>v[i];
    vvi al(k);vi cnt(n),tmp(n*k);
    rep(i,n*k){
        al[cnt[v[i]]].pb(v[i]);
        tmp[i]=cnt[v[i]];
        cnt[v[i]]++;
    }
    vi res(n);
    ll sum=count_inversion(tmp);
    rep(i,k){
        vi rv(n);rep(j,n)rv[al[i][j]]=j;
        vi ans(n);
        ans[0]=count_inversion(rv);
        rep(j,n-1){
            ans[j+1]=ans[j]+n-1-rv[j]-rv[j];
        }
        rep(j,n)res[j]+=ans[j];
    }
    ll ans=inf;
    rep(i,n)chmin(ans,res[i]);
    out(ans+sum);
}