#include<iostream>
#include<cstdio>
#include<vector>
#include<algorithm>
#define N 800
using namespace std;
int read()
{
    char c=0;
    int sum=0;
    while (c<'0'||c>'9') c=getchar();
    while ('0'<=c&&c<='9') sum=sum*10+c-'0',c=getchar();
    return sum;
}
struct reads
{
    int num,data;
};
struct node
{
    int x,y;
};
node st[N*N+1];
int n,m,leng,lengths,res,a[N*N+1],cnt[N+1],cur[N+1],tong[N+1],length;
vector<reads>E[N+1];
bool used[N*N+1];
void add(int x,int y)
{
    ++leng,E[x].push_back((reads){y,leng});
    return;
}
void adder(int x,int y)
{
    swap(a[x],a[y]),st[++lengths]=(node){x,y};
    return;
}
void dfs(int x)
{
    for (int &i=cur[x],t;i<E[x].size();++i)
	if (!used[E[x][i].data])
	    t=E[x][i].data,used[t]=1,dfs(E[x][i].num),tong[++length]=x;
    return;
}
int main()
{
    int x;
    vector<int>A;
    vector<int>B;
    n=read(),m=read();
    for (int i=1;i<=n*m;++i) a[i]=read();
    for (int i=1;i<=m;++i)
    {
	A.clear(),B.clear();
	for (int j=1;j<=n;++j) cnt[j]=0;
	for (int j=1;j<=n*m;++j)
	    if (a[j]==i)
		cnt[(j-1)/m+1]++;
	for (int j=1;j<=n;++j)
	{
	    if (cnt[j]>1)
	    {
		for (int k=1;k<=cnt[j]-1;++k) add(j,n+i);
	    }
	    else if (!cnt[j]) add(n+i,j);
	}
    }
    for (int i=1;i<=n;++i)
    {
	length=0,dfs(i);
	if (length)
	{
	    reverse(tong+1,tong+length+1);
	    for (int i=1;i<=length;i+=2)
	    {
		x=0;
		for (int j=(tong[i]-1)*m+1;j<=tong[i]*m;++j)
		    if (a[j]==tong[i+1]-n)
			x=j;
		adder(x,n*m+1);
	    }
	    for (int i=(tong[1]-1)*m+1;i<=tong[1]*m;++i)
		if (!a[i])
		    x=i;
	    adder(x,n*m+1);
	}
    }
    printf("%d\n",lengths);
    for (int i=1;i<=lengths;++i) printf("%d %d\n",st[i].x,st[i].y);
    return 0;
}